From d766f45980c967b7bf6d53af792a61b2ae5a748d Mon Sep 17 00:00:00 2001 From: NightlordTW Date: Sat, 9 Mar 2024 22:42:31 +0000 Subject: [PATCH] =?UTF-8?q?Deploying=20to=20gh-pages=20from=20@=20smartdat?= =?UTF-8?q?a-analysis-and-statistics/metamisc@d3aef8e71b15145c2e32b872dd11?= =?UTF-8?q?e406dfe26b8b=20=F0=9F=9A=80?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- articles/ma-pf.html | 2 +- index.html | 2 +- pkgdown.yml | 2 +- reference/metapred.html | 8 ++++---- reference/recalibrate.html | 4 ++-- search.json | 2 +- 6 files changed, 10 insertions(+), 10 deletions(-) diff --git a/articles/ma-pf.html b/articles/ma-pf.html index 473232f..d8f2943 100644 --- a/articles/ma-pf.html +++ b/articles/ma-pf.html @@ -574,7 +574,7 @@

Meta-analysis of the prog # Calculate the proportion of hazard ratios greater than 1.5 mean(HRsim > 1.5) -
## [1] 0.781016
+
## [1] 0.780348

Again, the probability that HER2 will yield a hazard ratio for overall survival of >1.5 in a new setting is 78%.

diff --git a/index.html b/index.html index 5297951..bf6ad9e 100644 --- a/index.html +++ b/index.html @@ -101,7 +101,7 @@

Funding
  • The Netherlands Organisation for Health Research and Development (grant 91617050).
  • The European Union’s Horizon 2020 research and innovation programme under ReCoDID grant agreement No 825746.
    • -
  • Smart Data Analysis and Statistics B.V., a limited company incorporated in the Netherlands and registered at the Dutch Chamber of Commerce under number 863595327.
    • +
  • Smart Data Analysis and Statistics B.V., a limited liability corporation registered at the Netherlands Chamber of Commerce under number 863595327.
    • diff --git a/pkgdown.yml b/pkgdown.yml index c286974..ca547cf 100644 --- a/pkgdown.yml +++ b/pkgdown.yml @@ -4,7 +4,7 @@ pkgdown_sha: ~ articles: ma-pf: ma-pf.html ma-pm: ma-pm.html -last_built: 2024-03-09T22:39Z +last_built: 2024-03-09T22:41Z urls: reference: https://smartdata-analysis-and-statistics.github.io/metamisc/reference article: https://smartdata-analysis-and-statistics.github.io/metamisc/articles diff --git a/reference/metapred.html b/reference/metapred.html index 9d871cc..562b519 100644 --- a/reference/metapred.html +++ b/reference/metapred.html @@ -245,7 +245,7 @@

    Examples#> #> Started with model: #> dvt ~ histdvt + ddimdich + sex + notraum -#> <environment: 0x55e86cb869d8> +#> <environment: 0x559b1011bad8> #> #> Generalizability: #> unchanged @@ -262,7 +262,7 @@

    Examples#> #> Started with model: #> dvt ~ histdvt + ddimdich + sex + notraum -#> <environment: 0x55e86cb869d8> +#> <environment: 0x559b1011bad8> #> #> Generalizability: #> unchanged @@ -274,7 +274,7 @@

    Examples#> #> Continued with model: #> dvt ~ histdvt + ddimdich + sex -#> <environment: 0x55e86cb869d8> +#> <environment: 0x559b1011bad8> #> #> Generalizability: #> ddimdich histdvt sex @@ -282,7 +282,7 @@

    Examples#> #> Continued with model: #> dvt ~ ddimdich + sex -#> <environment: 0x55e86cb869d8> +#> <environment: 0x559b1011bad8> #> #> Generalizability: #> ddimdich sex diff --git a/reference/recalibrate.html b/reference/recalibrate.html index fd2e98c..1c1c7ea 100644 --- a/reference/recalibrate.html +++ b/reference/recalibrate.html @@ -114,7 +114,7 @@

    Examples#> #> Started with model: #> dvt ~ vein + malign -#> <environment: 0x55e8796d8cf0> +#> <environment: 0x559b1b4a8590> #> #> Generalizability: #> unchanged @@ -126,7 +126,7 @@

    Examples#> #> Continued with model: #> dvt ~ malign -#> <environment: 0x55e8796d8cf0> +#> <environment: 0x559b1b4a8590> #> #> Generalizability: #> malign diff --git a/search.json b/search.json index 08df9e9..e416476 100644 --- a/search.json +++ b/search.json @@ -1 +1 @@ -[{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"introduction","dir":"Articles","previous_headings":"","what":"Introduction","title":"Meta-analysis of prognostic factors","text":"important task medical research identification assessment prognostic factors. prognostic factor measure , among people given health condition (, startpoint), associated subsequent clinical outcome (endpoint) (Riley 2013). Commonly investigated prognostic factors include simple measures age, sex, size tumor, can also include complex factors abnormal levels proteins catecholamines unusual genetic mutations (Sauerbrei Altman 2006). can useful modifiable targets interventions improve outcomes, building blocks prognostic models, predictors differential treatment response. past decades, numerous prognostic factor studies published medical literature. example, Riley Burchill (2003) identified 260 studies reporting associations 130 different tumour markers neuroblastoma. recently, Tzoulaki Ioannidis (2009) identified 79 studies reporting added value 86 different markers added Framingham risk score. Despite huge research effort, prognostic value traditional factors discussion uncertain usefulness many specific markers, prognostic indices, classification schemes still unproven (Sauerbrei Altman 2006). vignette aims illustrate results multiple prognostic factor studies can summarized sources -study heterogeneity can examined. Hereto, use R packages metamisc metafor. https://cran.r-project.org/package=metafor package provides comprehensive collection functions conducting meta-analyses R. https://cran.r-project.org/package=metamisc package provides additional functions facilitate meta-analysis prognosis studies. can load packages follows:","code":"library(metafor) library(metamisc)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"case-study","dir":"Articles","previous_headings":"","what":"Case Study","title":"Meta-analysis of prognostic factors","text":"Endometrial cancer (EC) fourth common malignancy women common gynecologic cancer. Overall, 5-year survival rates EC approximately 78–90% stage , 74% stage II, 36–57% stage III, 20% stage IV. poor outcomes raise urgent requirement accurate prognosis predictive markers applied EC guide therapy monitor disease progress individual patients. Several biological molecules proposed prognostic biomarkers EC. Among , hormone receptors estrogen receptors (ER), progesterone receptors (PR), human epidermal growth factor receptor 2 (HER2) attractive physiological functions. Recently, Zhang conducted systematic review evaluate overall risk hormone receptors endometrial cancer survival (Zhang Sun 2015). review included 16 studies recruiting 1764 patients HER2. study, estimates effect retrieved follows. simplest method consisted direct collection HR, odds ratio risk ratio, 95% CI original article, HR less 1 associated better outcome. available, total numbers observed deaths/cancer recurrences numbers patients group extracted calculate HR. data available Kaplan-Meier curves, data extracted graphical survival plots, estimation HR performed using described method. can load data 16 studies R follows: creates object Zhang contains summary data 14 studies reporting overall survival (OS) 6 studies reporting progression-free survival (PFS). total 14 studies assessed relation HER2 overall survival. corresponding hazard ratios (HR) given : Results suggest hormone receptor HER2 prognostic value survival prone substantial -study heterogeneity. example, hazard ratios appear much larger studies conducted USA. Possibly, variation treatment effect estimates caused differences baseline characteristics patients (age, tumor stage, race), differences cutoff value HER2, differences received treatments, differences duration follow-. Importantly, estimated hazard ratios adjusted patient-level covariates, particularly prone heterogeneity patient spectrum.","code":"data(Zhang)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"first-steps","dir":"Articles","previous_headings":"","what":"First steps","title":"Meta-analysis of prognostic factors","text":"facilitate quantitative analysis, information standard error different study effect sizes needed. Estimates standard error can obtained reported 95% confidence intervals (Altman Douglas G. Bland 2011). commonly assumed log hazard ratio follows Normal distribution, standard error (SE) log hazard ratio given : \\(\\mathrm{SE}=(\\log(u)-\\log(l))/(2*1.96)\\) upper lower limits 95% CI \\(u\\) \\(l\\) respectively. can implement calculation follows: often helpful display effect sizes studies forest plot. advantage forest plot allows visual inspection available evidence facilitates identification potential -study heterogeneity. forest plot overall survival can generated using forest function metamisc. requires provide information estimated hazard ratios (via argument theta), well bounds 95% confidence interval (via theta.ci.lb theta.ci.ub). can also generate forest plot using metafor:","code":"Zhang <- Zhang %>% mutate(logHR = log(HR), se.logHR = log(HR.975/HR.025)/(2 * qnorm(0.975))) library(dplyr) # Select the 14 studies investigating overall survival dat_os <- Zhang %>% filter(outcome == \"OS\") # Generate a forest plot of the log hazard ratio metamisc::forest(theta = dat_os$HR, theta.ci.lb = dat_os$HR.025, theta.ci.ub = dat_os$HR.975, theta.slab = dat_os$Study, xlab = \"Hazard ratio of HER2 versus OS\", refline = 1) metafor::forest(x = dat_os$HR, ci.lb = dat_os$HR.025, ci.ub = dat_os$HR.975, slab = dat_os$Study, xlab = \"Hazard ratio of HER2 versus OS\", refline = 1)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"assessment-of-publication-bias","dir":"Articles","previous_headings":"","what":"Assessment of publication bias","title":"Meta-analysis of prognostic factors","text":"presence small-study effects common threat systematic reviews meta-analyses. Small-study effects generic term phenomenon sometimes smaller studies show different, often stronger, effects large ones (Debray Riley 2018). One possible reason publication bias. Previously, D. G. Altman (2001) argued probable studies showing strong (often statistically significant) prognostic ability likely published. reason, important evaluate potential presence small-study effects, can verified visual inspection funnel plot. plot, estimate reported effect size plotted measure precision sample size included study meta-analysis. premise scatter plots reflect funnel shape, small-study effects exist (provided effect sizes substantially affected presence -study heterogeneity). However, small studies predominately one direction (usually direction larger effect sizes), asymmetry ensue. common approach construct funnel plot display individual observed effect sizes x-axis corresponding standard errors y-axis, use fixed effect summary estimate reference value. absence publication bias heterogeneity, one expect see points forming funnel shape, majority points falling inside pseudo-confidence region summary estimate. case study, study estimates fall within pseudo-confidence region, hence appears limited evidence publication bias. can formally test presence asymmetry funnel plot evaluating whether association estimated standard error estimated effect size. common use 10% level significance number studies meta-analysis usually low. case study, P-value 0.052, implies evidence funnel plot asymmetry . Funnel plot asymmetry tests can also performed using metamisc follows: yields , can construct funnel plot: caution warranted interpreting results funnel plot asymmetry tests (Debray Riley 2018). First, power detect asymmetry often limited meta-analyses usually include many studies. Second, many tests known yield inadequate type-error rates suffer low power. instance, demonstrated aforementioned test evaluate association estimated standard error effect size tends yield type-error rates high. Finally, funnel plot asymmetry may rather caused heterogeneity publication bias. therefore adjust aforementioned regression test use inverse total sample size (rather standard error) predictor. onwards, assume potential publication bias negligible, proceed standard meta-analysis methods.","code":"res <- rma(yi = logHR, sei = se.logHR, method = \"FE\", data = dat_os) funnel(res, xlab = \"Log Hazard Ratio\") regtest(x = dat_os$logHR, sei = dat_os$se.logHR, model = \"lm\", predictor = \"sei\") ## ## Regression Test for Funnel Plot Asymmetry ## ## Model: weighted regression with multiplicative dispersion ## Predictor: standard error ## ## Test for Funnel Plot Asymmetry: t = 2.1622, df = 12, p = 0.0515 ## Limit Estimate (as sei -> 0): b = 0.2590 (CI: -0.0760, 0.5939) regfit <- fat(b = dat_os$logHR, b.se = dat_os$se.logHR, method = \"E-FIV\") ## Call: fat(b = dat_os$logHR, b.se = dat_os$se.logHR, method = \"E-FIV\") ## ## Fixed effect summary estimate: 0.5193 ## ## test for funnel plot asymmetry: t =2.1622, df = 12, p = 0.0515 plot(regfit) regtest(x = dat_os$logHR, sei = dat_os$se.logHR, ni = dat_os$N, model = \"lm\", predictor = \"ninv\") ## ## Regression Test for Funnel Plot Asymmetry ## ## Model: weighted regression with multiplicative dispersion ## Predictor: inverse of the sample size ## ## Test for Funnel Plot Asymmetry: t = 0.1552, df = 12, p = 0.8793 ## Limit Estimate (as ni -> inf): b = 0.5088 (CI: 0.2226, 0.7950)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"meta-analysis-of-the-prognostic-value-of-her2","dir":"Articles","previous_headings":"","what":"Meta-analysis of the prognostic value of HER2","title":"Meta-analysis of prognostic factors","text":"Meta-analysis option identified studies considered sufficiently robust comparable, requires least two estimates statistic across studies. random effects approach often essential allow unexplained heterogeneity across studies due differences methods, time-scale, populations, cut-points, adjustment factors, treatments. standard random effects meta-analysis combines study estimates statistic interest (given log HR HER2) order estimate average effect (denoted \\(\\mu\\)) standard deviation (denoted \\(\\tau\\)) across studies. \\(\\hat b_i\\) \\(\\mathrm{var}(\\hat b_i)\\) denote estimate variance study \\(\\), general random effects meta-analysis model can specified : \\(\\hat b_i \\sim N(\\mu, \\mathrm{var}(\\hat b_i) + \\tau^2)\\) common first estimate heterogeneity parameter \\(\\tau\\) use resulting value estimate \\(\\mu\\). However, approach adequately reflect error associated parameter estimation, especially number studies small. reason, alternative estimators proposed simultaneously estimate \\(\\mu\\) \\(\\tau\\). , focus Restricted Maximum Likelihood Estimation (REML), implemented default metafor. pooled estimate log hazard ratio 0.667 standard error 0.135. -study standard deviation log hazard ratio 0.297. can extract key statistics follows: can use information derive summary estimate hazard ratio corresponding 95% confidence interval: summary HR HER2 statistically significant, indicating increased levels HER2 associated poorer survival. can also obtain summary estimate 95% CI HR HER2 simply using predict function: Although summary result (\\(\\hat \\mu\\)) usually main focus meta-analysis, reflects average across studies may hard translate clinical practice large -study heterogeneity. can quantify impact -study heterogeneity constructing \\(100(1-\\alpha/2)\\)% prediction interval, gives potential true prognostic effect new population conditional meta-analysis results (Riley Deeks 2011). approximate prediction interval (PI) given follows: \\(\\hat \\mu \\pm t_{\\alpha, N-2} \\sqrt{\\hat \\tau^2 + \\hat \\sigma^2}\\) \\(t_{\\alpha, N-2}\\) \\(100(1-\\alpha/2)\\)% percentile t-distribution \\(N-2\\) degrees freedom, \\(N\\) number studies, \\(\\hat \\sigma\\) estimated standard error \\(\\hat \\mu\\), \\(\\hat \\tau\\) estimated -study standard deviation. R, can calculate approximate 95% PI \\(\\hat \\mu\\) follows: 95% prediction interval ranges 0.956 3.969, suggests substantial heterogeneity prognostic value HER2. particular, although increased levels HER2 generally associated poorer survival, may also lead improved survival (HR < 1) certain settings. can add summary estimate prediction interval forest plot: possible approach enhance interpretation meta-analysis results calculate probability prognostic effect HER2 useful value (e.g. HR > 1.5 binary factor, indicates risk increased least 50%) new setting. can calculate probability follows: \\(Pr(\\mathrm{HR} > 1.5) = Pr(\\hat \\mu > \\log(1.5)) = 1 - Pr(\\hat \\mu \\leq \\log(1.5))\\) \\(Pr(\\hat \\mu \\leq \\log(1.5))\\) approximated using scaled Student-\\(t\\) distribution (similar calculation prediction interval): probability HER2 yield hazard ratio overall survival least 1.5 new setting 78%. means despite presence -study heterogeneity, likely HER2 provide substantial discriminative ability used single prognostic factor new setting. can also estimate probability means simulation: , probability HER2 yield hazard ratio overall survival >1.5 new setting 78%.","code":"resREML <- rma(yi = logHR, sei = se.logHR, method = \"REML\", slab = Study, data = dat_os) resREML ## ## Random-Effects Model (k = 14; tau^2 estimator: REML) ## ## tau^2 (estimated amount of total heterogeneity): 0.0883 (SE = 0.0854) ## tau (square root of estimated tau^2 value): 0.2972 ## I^2 (total heterogeneity / total variability): 49.17% ## H^2 (total variability / sampling variability): 1.97 ## ## Test for Heterogeneity: ## Q(df = 13) = 28.9214, p-val = 0.0067 ## ## Model Results: ## ## estimate se zval pval ci.lb ci.ub ## 0.6669 0.1354 4.9251 <.0001 0.4015 0.9324 *** ## ## --- ## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 # Summary estimate of the log hazard ratio for HER2 mu <- resREML$b # 95% confidence interval of the pooled log hazard ratio mu.ci <- c(resREML$ci.lb, resREML$ci.ub) # Between-study variance of the log hazard ratio tau2 <- resREML$tau2 # Error variance of the pooled log hazard ratio sigma2 <- as.numeric(vcov(resREML)) # Number of studies contributing to the meta-analyis numstudies <- resREML$k.all exp(mu) ## [,1] ## intrcpt 1.948268 exp(mu.ci) ## [1] 1.494104 2.540483 predict(resREML, transf = exp) ## ## pred ci.lb ci.ub pi.lb pi.ub ## 1.9483 1.4941 2.5405 1.0272 3.6954 level <- 0.05 crit <- qt(c(level/2, 1 - (level/2)), df = (numstudies - 2)) pi_lower <- exp(mu + crit[1] * sqrt(tau2 + sigma2)) pi_upper <- exp(mu + crit[2] * sqrt(tau2 + sigma2)) c(pi_lower, pi_upper) ## [1] 0.9563084 3.9691662 # Generate a forest plot of the log hazard ratio metamisc::forest(theta = dat_os$HR, theta.ci.lb = dat_os$HR.025, theta.ci.ub = dat_os$HR.975, theta.slab = dat_os$Study, theta.summary = exp(mu), theta.summary.ci.lb = exp(mu.ci[1]), theta.summary.ci.ub = exp(mu.ci[2]), theta.summary.pi.lb = pi_lower, theta.summary.pi.ub = pi_upper, xlab = \"Hazard ratio of HER2 versus OS\", refline = 1) probOS <- 1 - pt((log(1.5) - mu)/sqrt(tau2 + sigma2), df = (numstudies - 2)) probOS ## [,1] ## intrcpt 0.7805314 # Simulate 100000 new studies Nsim <- 1e+06 # Random draws from a Student T distribution rnd_t <- rt(Nsim, df = (numstudies - 2)) # Generate 1,000,000 hazard ratios HRsim <- exp(c(mu) + rnd_t * sqrt(tau2 + sigma2)) # Calculate the proportion of hazard ratios greater than 1.5 mean(HRsim > 1.5) ## [1] 0.781016"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"multivariate-meta-analysis","dir":"Articles","previous_headings":"","what":"Multivariate meta-analysis","title":"Meta-analysis of prognostic factors","text":"previous section, used 14 16 identified studies evaluate prognostic effect HER2 overall survival. Two studies excluded meta-analysis provide direct evidence overall survival. unwelcome, especially participants otherwise representative population, clinical settings, condition interest (Riley White 2017). reason, discuss multivariate meta-analysis methods can used borrow strength studies investigate primary outcome interest. Briefly, multivariate meta-analysis methods simultaneously summarize effect size across multiple outcomes whilst accounting correlation. example, six studies review Zhang Sun (2015) assessed hazard ratio HER2 progression-free survival, four also assessed overall survival. Hence, conducting multivariate meta-analysis can borrow strength two additional studies estimating hazard ratio overall survival. hazard ratios progression free survival depicted : first conduct univariate meta-analysis six studies investigating progression-free survival: Results indicate hormone receptor HER2 also prognostic value progression-free survival. Furthermore, reported HRs appear much homogeneous across studies, since -study standard deviation 0.17 progression-free survival whereas 0.30 overall survival. Note univariate meta-analysis progression-free survival based merely 6 studies, univariate meta-analysis overall survival based 14 studies. can now employ multivariate meta-analysis borrow information 4 studies report prognostic effects endpoints. , turn, allows studies contribute summary effect HER2 outcomes. first need define within-study covariance matrix estimated log hazard ratios progression-free survival overall survival. assume estimates hazard ratio independent within studies construct block diagonal matrix considers error variance estimate: multivariate random-effects model can now used simultaneously meta-analyze hazard ratios overall progression-free survival: summary estimate log hazard ratio overall survival 0.670 (multivariate meta-analysis) versus 0.667 (univariate meta-analysis) SE 0.132 , respectively, 0.135. Hence, gained precision including evidence 2 additional studies evaluated progression-free survival. Note estimation -study heterogeneity difficult progression-free survival due limited number studies. particular, found \\(\\tau^2\\)= 0.028 SE 0.145. multivariate meta-analysis, estimated -study variance PFS much larger (\\(\\tau^2\\)=0.077), based 16 rather merely 6 studies. summary, multivariate meta-analysis approach often helpful reduces need exclude relevant studies meta-analysis, thereby decreasing risk bias (e.g. due selective outcome reporting) potentially improving precision. indicated Riley White (2017), multivariate meta-analysis multiple outcomes beneficial outcomes highly correlated percentage studies missing outcomes large.","code":"dat_pfs <- Zhang %>% filter(outcome == \"PFS\") resPFS <- rma(yi = logHR, sei = se.logHR, method = \"REML\", slab = Study, data = dat_pfs) V <- diag(Zhang$se.logHR^2) res.MV <- rma.mv(yi = logHR, V = V, mods = ~outcome - 1, random = ~outcome | Study, struct = \"UN\", data = Zhang, method = \"REML\") res.MV ## ## Multivariate Meta-Analysis Model (k = 20; method: REML) ## ## Variance Components: ## ## outer factor: Study (nlvls = 16) ## inner factor: outcome (nlvls = 2) ## ## estim sqrt k.lvl fixed level ## tau^2.1 0.0865 0.2942 14 no OS ## tau^2.2 0.0770 0.2775 6 no PFS ## ## rho.OS rho.PFS OS PFS ## OS 1 - 4 ## PFS 1.0000 1 no - ## ## Test for Residual Heterogeneity: ## QE(df = 18) = 33.7664, p-val = 0.0135 ## ## Test of Moderators (coefficients 1:2): ## QM(df = 2) = 35.6315, p-val < .0001 ## ## Model Results: ## ## estimate se zval pval ci.lb ci.ub ## outcomeOS 0.6704 0.1318 5.0868 <.0001 0.4121 0.9287 *** ## outcomePFS 0.8734 0.2151 4.0606 <.0001 0.4518 1.2950 *** ## ## --- ## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1"},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pm.html","id":"case-study","dir":"Articles","previous_headings":"","what":"Case Study","title":"Meta-analysis of prediction model performance","text":"EuroSCORE II commonly used scoring rule estimating risk -hospital mortality patients undergoing major cardiac surgery. developed using data 16,828 adult patients 43 countries. Predictors include patient characteristics (e.g. age, gender), cardiac related factors (e.g. recent MI) surgery related factors (e.g. Surgery thoracic aorta). 2014, systematic review undertaken Guida et al. (2014) search articles assessing performance EuroSCORE II perioperative mortality cardiac surgery. systematic review identified 24 eligible validation studies, 22 studies included main analysis. case study, summarize results 22 studies, well results split-sample validation contained within original development article EuroSCORE II. use metamisc package derive summary estimates discrimination calibration performance EuroSCORE II, evaluate presence -study heterogeneity, identify potential sources -study heterogeneity. step--step tutorial provided Debray et al. (2017). can load data 23 validation studies follows:","code":"library(metamisc) data(EuroSCORE)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pm.html","id":"meta-analysis-of-calibration-performance","dir":"Articles","previous_headings":"","what":"Meta-analysis of calibration performance","title":"Meta-analysis of prediction model performance","text":"Calibration refers model’s accuracy predicted risk probabilities, indicates extent expected outcomes (predicted model) observed outcomes agree. Summarising estimates calibration performance challenging calibration plots often presented, studies tend report different types summary statistics calibration. example, case study, calibration assessed using Hosmer-Lemeshow test, calibration plots comparing observed mortality predicted EuroSCORE II (either overall groups patients). Within validation study, can compare total number observed events (O) total number expected (predicted) events deriving ratio O:E. total O:E ratio provides rough indication overall model calibration (across entire range predicted risks). describes whether (O:E > 1) fewer (O:E < 1) events occurred expected based model. Whilst O:E ratio explicitly reported studies, can calculated reported information: O:E ratio can also derived observed predicted mortality risk Po , respectively, Pe: recommended first transform extracted O:E ratios log (natural logarithm) scale applying meta-analysis.","code":"EuroSCORE <- EuroSCORE %>% mutate(oe = n.events/e.events) EuroSCORE %>% select(Po, Pe) %>% mutate(oe = Po/Pe) EuroSCORE <- EuroSCORE %>% mutate(logoe = log(oe))"},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/authors.html","id":null,"dir":"","previous_headings":"","what":"Authors","title":"Authors and Citation","text":"Thomas Debray. Author, maintainer. Valentijn de Jong. Author.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/authors.html","id":"citation","dir":"","previous_headings":"","what":"Citation","title":"Authors and Citation","text":"Debray T, de Jong V (2024). metamisc: Meta-Analysis Diagnosis Prognosis Research Studies. R package version 0.4.0.9000, https://smartdata-analysis--statistics.github.io/metamisc/, https://github.com/smartdata-analysis--statistics/metamisc.","code":"@Manual{, title = {metamisc: Meta-Analysis of Diagnosis and Prognosis Research Studies}, author = {Thomas Debray and Valentijn {de Jong}}, year = {2024}, note = {R package version 0.4.0.9000, https://smartdata-analysis-and-statistics.github.io/metamisc/}, url = {https://github.com/smartdata-analysis-and-statistics/metamisc}, }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"metamisc","dir":"","previous_headings":"","what":"Meta-Analysis of Diagnosis and Prognosis Research Studies","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"official repository R package metamisc, developed facilitate meta-analysis diagnosis prognosis research studies. package includes functions following tasks: develop validate multivariable prediction models datasets clustering (de Jong et al., 2021) summarize multiple estimates prediction model discrimination calibration performance (Debray et al., 2019) evaluate funnel plot asymmetry (Debray et al., 2018)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"installation","dir":"","previous_headings":"","what":"Installation","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"metamisc package can installed CRAN follows: can install development version metamisc GitHub :","code":"install.packages(\"metamisc\") # install.packages(\"devtools\") devtools::install_github(\"smartdata-analysis-and-statistics/metamisc\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"jasp","dir":"","previous_headings":"","what":"JASP","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"visual interface software implemented JASP https://jasp-stats.org/","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"funding","dir":"","previous_headings":"","what":"Funding","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"development R package funded following organisations: Netherlands Organisation Health Research Development (grant 91617050). European Union’s Horizon 2020 research innovation programme ReCoDID grant agreement 825746. Smart Data Analysis Statistics B.V., limited company incorporated Netherlands registered Dutch Chamber Commerce number 863595327.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"references","dir":"","previous_headings":"","what":"References","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"de Jong VMT, Moons KGM, Eijkemans MJC, Riley RD, Debray TPA. Developing generalizable prediction models pooled studies large clustered data sets. Stat Med. 2021 May 5;40(15):3533–59. Debray TPA, Moons KGM, Riley RD. Detecting small-study effects funnel plot asymmetry meta-analysis survival data: comparison new existing tests. Res Syn Meth. 2018;9(1):41–50. Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2019 Sep;28(9):2768–86.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":null,"dir":"Reference","previous_headings":"","what":"Collins data — Collins","title":"Collins data — Collins","text":"meta-analysis nine clinical trials investigating effect taking diuretics pregnancy risk pre-eclampsia.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Collins data — Collins","text":"","code":"data(Collins)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Collins data — Collins","text":"data frame 9 observations following 2 variables. logOR numeric vector treatment effect sizes (log odds ratio) SE numeric vector standard error treatment effect sizes","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Collins data — Collins","text":"Collins, R., Yusuf, S., Peto, R. Overview randomised trials diuretics pregnancy. British Medical Journal 1985, 290, 17--23. Hardy, R.J. Thompson, S.G. likelihood approach meta-analysis random effects. Statistics Medicine 1996; 15:619--629.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Collins data — Collins","text":"","code":"data(Collins)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":null,"dir":"Reference","previous_headings":"","what":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"hypothetical dataset 500 subjects suspected deep vein thrombosis (DVT).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"","code":"data(DVTipd)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"data frame 500 observations 16 variables. sex gender (0=female, 1=male) malign active malignancy (0=active malignancy, 1=active malignancy) par paresis (0=paresis, 1=paresis) surg recent surgery bedridden tend tenderness venous system oachst oral contraceptives hst leg entire leg swollen notraum absence leg trauma calfdif3 calf difference >= 3 cm pit pitting edema vein vein distension altdiagn alternative diagnosis present histdvt history previous DVT ddimdich dichotimized D-dimer value dvt final diagnosis DVT study study indicator","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"Hypothetical dataset derived Individual Participant Data Meta-Analysis Geersing et al (2014). dataset consists consecutive outpatients suspected deep vein thrombosis, documented information presence absence proximal deep vein thrombosis (dvt) acceptable reference test. Acceptable tests either compression ultrasonography venography initial presentation, , venous imaging performed, uneventful follow-least three months.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"Geersing GJ, Zuithoff NPA, Kearon C, Anderson DR, Ten Cate-Hoek AJ, Elf JL, et al. Exclusion deep vein thrombosis using Wells rule clinically important subgroups: individual patient data meta-analysis. BMJ. 2014;348:g1340.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"","code":"data(DVTipd) str(DVTipd) #> 'data.frame':\t500 obs. of 16 variables: #> $ sex : num 0 1 0 1 0 0 1 0 1 0 ... #> $ malign : num 0 0 0 0 0 0 0 0 0 0 ... #> $ par : num 0 0 1 0 0 0 0 0 0 0 ... #> $ surg : num 0 0 0 0 0 0 0 0 1 0 ... #> $ tend : num 1 1 0 1 1 0 0 1 1 1 ... #> $ oachst : num 0 0 0 0 0 0 0 0 0 0 ... #> $ leg : num 1 0 0 0 0 1 1 0 0 0 ... #> $ notraum : num 1 1 1 1 1 0 0 1 0 1 ... #> $ calfdif3: num 0 0 0 0 0 0 0 0 0 0 ... #> $ pit : num 0 0 0 0 0 1 0 1 1 1 ... #> $ vein : num 0 0 0 0 1 0 0 0 0 1 ... #> $ altdiagn: num 1 0 1 1 1 0 1 1 1 1 ... #> $ histdvt : num 0 1 0 0 0 0 1 0 0 0 ... #> $ ddimdich: num 1 0 0 0 0 1 1 0 1 1 ... #> $ dvt : num 0 0 0 0 0 0 0 0 0 0 ... #> $ study : Factor w/ 4 levels \"a\",\"b\",\"c\",\"d\": 1 4 1 4 1 4 4 4 4 2 ... summary(apply(DVTipd,2,as.factor)) #> sex malign par surg #> Length:500 Length:500 Length:500 Length:500 #> Class :character Class :character Class :character Class :character #> Mode :character Mode :character Mode :character Mode :character #> tend oachst leg notraum #> Length:500 Length:500 Length:500 Length:500 #> Class :character Class :character Class :character Class :character #> Mode :character Mode :character Mode :character Mode :character #> calfdif3 pit vein altdiagn #> Length:500 Length:500 Length:500 Length:500 #> Class :character Class :character Class :character Class :character #> Mode :character Mode :character Mode :character Mode :character #> histdvt ddimdich dvt study #> Length:500 Length:500 Length:500 Length:500 #> Class :character Class :character Class :character Class :character #> Mode :character Mode :character Mode :character Mode :character ## Develop a prediction model to predict presence of DVT model.dvt <- glm(\"dvt~sex+oachst+malign+surg+notraum+vein+calfdif3+ddimdich\", family=binomial, data=DVTipd) summary(model.dvt) #> #> Call: #> glm(formula = \"dvt~sex+oachst+malign+surg+notraum+vein+calfdif3+ddimdich\", #> family = binomial, data = DVTipd) #> #> Coefficients: #> Estimate Std. Error z value Pr(>|z|) #> (Intercept) -5.1664 0.6365 -8.117 4.76e-16 *** #> sex 0.8146 0.2825 2.883 0.00393 ** #> oachst 0.4324 0.6227 0.694 0.48739 #> malign 0.5679 0.4025 1.411 0.15826 #> surg 0.1002 0.4111 0.244 0.80734 #> notraum 0.3351 0.3700 0.906 0.36513 #> vein 0.4831 0.3186 1.516 0.12939 #> calfdif3 1.1841 0.2819 4.200 2.67e-05 *** #> ddimdich 2.6081 0.5310 4.911 9.04e-07 *** #> --- #> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 #> #> (Dispersion parameter for binomial family taken to be 1) #> #> Null deviance: 446.24 on 499 degrees of freedom #> Residual deviance: 345.98 on 491 degrees of freedom #> AIC: 363.98 #> #> Number of Fisher Scoring iterations: 6 #>"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":null,"dir":"Reference","previous_headings":"","what":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"Previously published prediction models predicting presence DVT.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"","code":"data(DVTmodels)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"object class litmodels following information literature model: study-level descriptives (\"descriptives\"), regression coefficient weight predictor (\"weights\") error variance regression coefficient weight (\"weights.var\").","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"Previously, several models (Gagne, Oudega) score charts (Wells, modified Wells, Hamilton) published evaluating presence DVT suspected patients. models combine information mulitple predictors weighted sum, can subsequently used obtain estimates absolute risk. See DVTipd information predictors.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"Wells PS, Anderson DR, Bormanis J, Guy F, Mitchell M, Gray L, Clement C, Robinson KS, Lewandowski B. Value assessment pretest probability deep-vein thrombosis clinical management. Lancet 1997; 350(9094):1795--1798. DOI: 10.1016/S0140-6736(97)08140-3. Wells PS, Anderson DR, Rodger M, Forgie M, Kearon C, Dreyer J, Kovacs G, Mitchell M, Lewandowski B, Kovacs MJ. Evaluation D-dimer diagnosis suspected deep-vein thrombosis. New England Journal Medicine 2003; 349(13):1227--1235. DOI: 10.1056/NEJMoa023153. Gagne P, Simon L, Le Pape F, Bressollette L, Mottier D, Le Gal G. Clinical prediction rule diagnosing deep vein thrombosis primary care. La Presse Medicale 2009; 38(4):525--533. DOI: 10.1016/j.lpm.2008.09.022. Subramaniam RM, Snyder B, Heath R, Tawse F, Sleigh J. Diagnosis lower limb deep venous thrombosis emergency department patients: performance Hamilton modified Wells scores. Annals Emergency Medicine 2006; 48(6):678--685. DOI: 10.1016/j.annemergmed.2006.04.010. Oudega R, Moons KGM, Hoes AW. Ruling deep venous thrombosis primary care. simple diagnostic algorithm including D-dimer testing. Thrombosis Haemostasis 2005; 94(1):200--205. DOI: 10.1160/TH04-12-0829.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"Debray TPA, Koffijberg H, Nieboer D, Vergouwe Y, Steyerberg EW, Moons KGM. Meta-analysis aggregation multiple published prediction models. Stat Med. 2014 Jun 30;33(14):2341--62. Debray TPA, Koffijberg H, Vergouwe Y, Moons KGM, Steyerberg EW. Aggregating published prediction models individual participant data: comparison different approaches. Stat Med. 2012 Oct 15;31(23):2697--712.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"","code":"data(DVTmodels)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":null,"dir":"Reference","previous_headings":"","what":"Daniels and Hughes data — Daniels","title":"Daniels and Hughes data — Daniels","text":"Data frame treatment differences CD4 cell count.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Daniels and Hughes data — Daniels","text":"","code":"data(\"Daniels\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Daniels and Hughes data — Daniels","text":"data frame 15 observations following 2 variables. Y1 Treatment differences log hazard ratio development AIDS death 2 years. vars1 Error variances Y1. Y2 Difference mean change CD4 cell count baseline 6 month studies AIDS Clinical Trial Group vars2 Error variances Y2.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Daniels and Hughes data — Daniels","text":"Daniels data comprises 15 phase II/III randomized clinical trials HIV Disease Section Adult AIDS Clinical Trials Group National Institutes Health, data available May 1996, least six months follow-patients least one patient developed AIDS died. data previously used Daniels Hughes (1997) assess whether change CD4 cell count surrogate time either development AIDS death drug trials patients HIV.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Daniels and Hughes data — Daniels","text":"Daniels MJ, Hughes MD. Meta-analysis evaluation potential surrogate markers. Statistics Medicine 1997; 16: 1965--1982.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":null,"dir":"Reference","previous_headings":"","what":"Predictive performance of EuroSCORE II — EuroSCORE","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"data set contains estimates predictive performance European system cardiac operative risk evaluation (EuroSCORE II) patients undergoing cardiac surgery. Results based original development study 22 validations identified Guida et al.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"","code":"data(\"EuroSCORE\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"data frame 23 observations following 13 variables. Study vector first author validation study n numeric vector total number patients performance estimates based n.events numeric vector total number observed events c.index numeric vector estimated concordance statistic validation se.c.index numeric vector standard error concordance statistics c.index.95CIl numeric vector lower bound 95% confidence interval estimated concordance statistics c.index.95CIu numeric vector upper bound 95% confidence interval estimated concordance statistics Po numeric vector overall observed event probability validation Pe numeric vector overall expected event probability validation SD.Pe numeric vector standard error Pe e.events numeric vector total number expected events validation multicentre logical vector describing whether study multicentre study mean.age numeric vector describing mean age patients sd.age numeric vector spread age patients pts..2010 logical vector describing whether studies included patients 2010 (.e., EuroSCORE II developed)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"Published 2012, EuroSCORE II developed using logistic regression dataset comprising 16,828 adult patients undergoing major cardiac surgery 154 hospitals 43 countries 12-week period (May-July) 2010. EuroSCORE II developed predict -hospital mortality patients undergoing type cardiac surgery. 2014, systematic review published evidence performance value euroSCORE II undertaken Guida et al. Twenty-two validations, including 145,592 patients 21 external validation articles (one study included two validations) split-sample validation contained within original development article included review; 23 validation studies total.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"Guida P, Mastro F, Scrascia G, Whitlock R, Paparella D. Performance European System Cardiac Operative Risk Evaluation II: meta-analysis 22 studies involving 145,592 cardiac surgery procedures. J Thorac Cardiovasc Surg. 2014; 148(6):3049--3057.e1. Nashef SAM, Roques F, Sharples LD, Nilsson J, Smith C, Goldstone AR, et al. EuroSCORE II. Eur J Cardiothorac Surg. 2012; 41(4):734-744; discussion 744-745.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"","code":"data(EuroSCORE)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":null,"dir":"Reference","previous_headings":"","what":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"Fibrinogen data set meta-analysis 31 studies association plasma fibrinogen concentration risk coronary heath disease (CHD) estimated.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"","code":"data(\"Fibrinogen\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"data frame 5 variables: N.total numeric vector describing total number patients study N.events numeric vector describing number observed events within study HR numeric vector describing estimated hazard ratio study HR.025 numeric vector describing lower boundary 95% confidence interval HR HR.975 numeric vector describing upper boundary 95% confidence interval HR","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"Fibrinogen Studies Collaboration. Collaborative meta-analysis prospective studies plasma fibrinogen cardiovascular disease. Eur J Cardiovasc Prev Rehabil. 2004 Feb;11(1):9-17. Thompson S, Kaptoge S, White , Wood , Perry P, Danesh J, et al. Statistical methods time--event analysis individual participant data multiple epidemiological studies. Int J Epidemiol. 2010 Oct;39(5):1345-59.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"","code":"data(Fibrinogen) ## maybe str(Fibrinogen) ; plot(Fibrinogen) ..."},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":null,"dir":"Reference","previous_headings":"","what":"Predictive performance of the Framingham Risk Score in male populations — Framingham","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"data set contains estimates performance Framingham model predicting coronary heart disease male populations (Wilson 1998). Results based original development study 20 validations identified Damen et al (BMC Med, 2017).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"","code":"data(\"Framingham\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"data frame 24 observations following 19 variables. AuthorYear vector describing study authors n numeric vector total number patients performance estimates based n.events numeric vector total number observed events c.index numeric vector estimated concordance statistic validation se.c.index numeric vector standard error concordance statistics c.index.95CIl numeric vector lower bound 95% confidence interval estimated concordance statistics c.index.95CIu numeric vector upper bound 95% confidence interval estimated concordance statistics Po numeric vector overall observed event probability validation Pe numeric vector overall expected event probability validation t.val numeric vector describing time period predictive performance assessed validation mean_age numeric vector describing mean age patients sd_age numeric vector spread age patients mean_SBP numeric vector mean systolic blood pressure validation studies (mm Hg) sd_SBP numeric vector spread systolic blood pressure validation studies mean_total_cholesterol numeric vector mean total cholesterol validation studies (mg/dL) sd_total_cholesterol numeric vector spread total cholesterol validation studies mean_hdl_cholesterol numeric vector mean high-density lipoprotein cholesterol validation studies (mg/dL) sd_hdl_cholesterol numeric vector spread high-density lipoprotein cholesterol validation studies pct_smoker numeric vector percentage smokers validation studies","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"Framingham Risk Score allows physicians predict 10-year coronary heart disease (CHD) risk patients without overt CHD. developed 1998 middle-aged white population sample, subsequently validated across different populations. current dataset contains original (internal validation) results, well 23 external validations identified systematic review. review, studies eligible inclusion described validation original Framingham model assessed performance fatal nonfatal CHD males general population setting.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"Damen JAAG, Hooft L, Schuit E, Debray TPA, Collins GS, Tzoulaki , et al. Prediction models cardiovascular disease risk general population: systematic review. BMJ. 2016;i2416. Damen JAAG, Pajouheshnia R, Heus P, Moons KGM, Reitsma JB, Scholten RJPM, et al. Performance Framingham risk models Pooled Cohort Equations: systematic review meta-analysis. BMC Med. 2017;17(1):109. Wilson PW, D'Agostino RB, Levy D, Belanger , Silbershatz H, Kannel WB. Prediction coronary heart disease using risk factor categories. Circulation. 1998; 97(18):1837--47.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"","code":"data(Framingham)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":null,"dir":"Reference","previous_headings":"","what":"Kertai data — Kertai","title":"Kertai data — Kertai","text":"Data frame diagnostic accuracy data exercise electrocardiography.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Kertai data — Kertai","text":"","code":"data(\"Kertai\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Kertai data — Kertai","text":"One data frame 4 variables. TP integer. number true positives FN integer. number false negatives FP integer. number false positives TN integer. number true negatives","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Kertai data — Kertai","text":"Kertai data set meta-analysis prognostic test studies comprises 7 studies diagnostic test accuracy exercise electrocardiography predicting cardiac events patients undergoing major vascular surgery measured.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Kertai data — Kertai","text":"Kertai MD, Boersma E, Bax JJ, Heijenbrok-Kal MH, Hunink MGM, L'talien GJ, Roelandt JRTC, van Urk H, Poldermans D. meta-analysis comparing prognostic accuracy six diagnostic tests predicting perioperative cardiac risk patients undergoing major vascular surgery. Heart 2003; 89: 1327--1334. Jackson D, Riley RD, & White IW. Multivariate meta-analysis: Potential promise. Statistics Medicine 2010; 30: 2481--2498.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":null,"dir":"Reference","previous_headings":"","what":"Roberts data — Roberts","title":"Roberts data — Roberts","text":"Data frame summary data 14 comparative studies.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Roberts data — Roberts","text":"","code":"data(\"Roberts\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Roberts data — Roberts","text":"One data frame 2 variables. SDM Effect sizes (standardized differences means) SE Standard error effect sizes","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Roberts data — Roberts","text":"Roberts data set meta-analysis 14 studies comparing 'set shifting' ability (ability move back forth different tasks) people eating disorders healthy controls.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Roberts data — Roberts","text":"Roberts , Tchanturia K, Stahl D, Southgate L, Treasure J. systematic review meta-analysis set-shifting ability eating disorders. Psychological Medicine 2007, 37: 1075--1084. Higgins JPT, Thompson SG, Spiegelhalter DJ. re-evaluation random-effects meta-analysis. Journal Royal Statistical Society. Series (Statistics Society) 2009, 172: 137--159.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":null,"dir":"Reference","previous_headings":"","what":"Diagnostic accuracy data — Scheidler","title":"Diagnostic accuracy data — Scheidler","text":"Data frame diagnostic accuracy data three imaging techniques diagnosis lymph node metastasis women cervical cancer.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Diagnostic accuracy data — Scheidler","text":"","code":"data(\"Scheidler\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Diagnostic accuracy data — Scheidler","text":"One data frame 6 variables. author string . author article modality integer . type test (1=CT, 2=LAG, 3=MRI) TP integer. number true positives FN integer. number false negatives FP integer. number false positives TN integer. number true negatives","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Diagnostic accuracy data — Scheidler","text":"Scheidler data comprises results meta-analysis three imaging techniques diagnosis lymph node metastasis women cervical cancer compared. Forty-four studies total included: 17 studies evaluated lymphangiography, another 17 studies examined computed tomography remaining 10 studies focused magnetic resonance imaging. Diagnosis metastatic disease lymphangiography (LAG) based presence nodal-filling defects, whereas computed tomography (CT) magnetic resonance imaging (MRI) rely nodal enlargement.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Diagnostic accuracy data — Scheidler","text":"Scheidler J, Hricak H, Yu KK, Subak L, Segal MR. Radiological evaluation lymph node metastases patients cervical cancer. meta-analysis. Journal American Medical Association 1997; 278: 1096--1101. Reitsma J, Glas , Rutjes , Scholten R, Bossuyt P, Zwinderman . Bivariate analysis sensitivity specificity produces informative summary measures diagnostic reviews. Journal Clinical Epidemiology 2005; 58: 982--990.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":null,"dir":"Reference","previous_headings":"","what":"The incremental value of cardiovascular risk factors — Tzoulaki","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"Tzoulaki et al. (2009) reviewed studies evaluated various candidate prognostic factors ability improve prediction coronary heart disease (CHD) outcomes beyond Framingham risk score (FRS) can achieve.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"","code":"data(\"Tzoulaki\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"data frame containing data 27 studies following 2 variables. PubmedID character vector Pubmed ID study N numeric vector describing study size N.events numeric vector describing observed number events FRS.orig.refitted boolean vector describing whether coefficients original Framingham Risk Score (FRS) re-estimated FRS.modif.refitted boolean vector describing whether coefficients modified Framingham Risk Score re-estimated predictors character vector indicating new risk factor(s) included modified FRS outcome character vector indicating primary outcome predicted AUC.orig numeric vector describing Area ROC curve (AUC) original FRS model AUC.orig.CIl numeric vector describing lower boundary 95% confidence interval AUC original FRS model AUC.orig.CIu numeric vector describing upper boundary 95% confidence interval AUC original FRS model AUC.modif numeric vector describing Area ROC curve (AUC) modified FRS model includes one new risk factors AUC.modif.CIl numeric vector describing lower boundary 95% confidence interval AUC modified FRS model AUC.modif.CIu numeric vector describing upper boundary 95% confidence interval AUC modified FRS model pval.AUCdiff numeric vector p-value difference AUC.orig AUC.modif sign.AUCdiff boolean vector indicating whether difference AUC.orig AUC.modif 0.05","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"Tzoulaki , Liberopoulos G, Ioannidis JPA. Assessment claims improved prediction beyond Framingham risk score. JAMA. 2009 Dec 2;302(21):2345–52.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"","code":"data(Tzoulaki) ## maybe str(Tzoulaki) ; plot(Tzoulaki) ..."},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":null,"dir":"Reference","previous_headings":"","what":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"dataset comprises results 16 studies assessing prognostic role human epidermal growth factor receptor 2 (HER2) endometrial cancer. studies previously identified systematic review Zhang et al. evaluate overall risk several hormone receptors endometrial cancer survival.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"","code":"data(\"Zhang\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"data frame 20 observations following 10 variables. Study factor 16 levels indicate study PrimaryAuthor factor indicating first author's last name year numeric vector indicating publication year Country factor indicating source country study data Disease factor indicating studied disease. Possible levels EC (endometrial cancer), EEC (endometrioid endometrial cancer) UPSC (uterine papillary serous carcinoma) N numeric vector describing total sample size study HR numeric vector describing estimated hazard ratio study HR.025 numeric vector describing lower boundary 95% confidence interval HR HR.975 numeric vector describing upper boundary 95% confidence interval HR outcome factor indicating studied outcome. Possible levels OS (overall survival) PFS (progression-free survival)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"Eligible studies identified searching PubMed EMBASE databases publications 1979 May 2014. Data collected studies comparing overall survival progression-free survival patients elevated levels human epidermal growth factor receptor 2 patients lower levels.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"Zhang Y, Zhao D, Gong C, Zhang F, J, Zhang W, et al. Prognostic role hormone receptors endometrial cancer: systematic review meta-analysis. World J Surg Oncol. 2015 Jun 25;13:208.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"Riley RD, Jackson D, Salanti G, Burke DL, Price M, Kirkham J, et al. Multivariate network meta-analysis multiple outcomes multiple treatments: rationale, concepts, examples. BMJ. 2017 13;358:j3932.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"","code":"data(Zhang) # Display the hazard ratios for overall survival in a forest plot ds <- subset(Zhang, outcome==\"OS\") with(ds, forest(theta = HR, theta.ci.lb = HR.025, theta.ci.ub = HR.975, theta.slab = Study, xlab = \"Hazard ratio of HER2 versus OS\", refline = 1))"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"Function display autocorrelation fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"","code":"acplot(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"generic function.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"Function display autocorrelation fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"","code":"# S3 method for mcmc.list acplot(x, P, nLags = 50, greek = FALSE, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"x object class \"mcmc.list\" P Optional dataframe describing parameters plot respective names nLags Integer indicating number lags autocorrelation plot. greek Logical value indicating whether parameter labels parsed get Greek letters. Defaults FALSE. ... Additional arguments passed ggs_autocorrelation","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"Function display autocorrelation fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"","code":"# S3 method for uvmeta acplot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"x object class \"valmeta\" ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"Results displayed estimated mean (mu) standard-deviation (tau) meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"","code":"if (FALSE) { data(Roberts) fit <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts), method=\"BAYES\")) acplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"Function display autocorrelation fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"","code":"# S3 method for valmeta acplot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"x object class \"valmeta\" ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"Results displayed estimated mean (mu) standard-deviation (tau) meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"","code":"if (FALSE) { data(EuroSCORE) fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) acplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate the concordance statistic — ccalc","title":"Calculate the concordance statistic — ccalc","text":"function calculates (transformed versions ) concordance (c-) statistic corresponding sampling variance.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate the concordance statistic — ccalc","text":"","code":"ccalc( cstat, cstat.se, cstat.cilb, cstat.ciub, cstat.cilv, sd.LP, N, O, Po, data, slab, subset, g = NULL, level = 0.95, approx.se.method = 4, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate the concordance statistic — ccalc","text":"cstat vector specify estimated c-statistics. cstat.se Optional vector specify corresponding standard errors. cstat.cilb Optional vector specify lower limits confidence interval. cstat.ciub Optional vector specify upper limits confidence interval. cstat.cilv Optional vector specify levels aformentioned confidence interval limits. (default: 0.95, corresponds 95% confidence interval). sd.LP Optional vector specify standard deviations linear predictor (prognostic index). N Optional vector specify sample/group sizes. O Optional vector specify total number observed events. Po Optional vector specify observed event probabilities. data Optional data frame containing variables given arguments . slab Optional vector labels studies. subset Optional vector indicating subset studies used. can logical vector numeric vector indicating indices studies include. g quoted string function transform estimates c-statistic; see details . level Optional numeric specify level confidence interval, default 0.95. approx.se.method integer specifying method used estimating standard error c-statistic (Newcombe, 2006). far, method 2 method 4 (default) implemented. ... Additional arguments.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate the concordance statistic — ccalc","text":"object class c(\"mm_perf\",\"data.frame\") following columns: \"theta\" (transformed) c-statistics. \"theta.se\" Standard errors (transformed) c-statistics. \"theta.cilb\" Lower confidence interval (transformed) c-statistics. level specified level. Intervals calculated scale theta assuming Normal distribution. \"theta.ciub\" Upper confidence interval (transformed) c-statistics. level specified level. Intervals calculated scale theta assuming Normal distribution. \"theta.source\" Method used calculating (transformed) c-statistic. \"theta.se.source\" Method used calculating standard error (transformed) c-statistic.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Calculate the concordance statistic — ccalc","text":"c-statistic measure discrimination, indicates ability prediction model distinguish patients developing developing outcome. c-statistic typically ranges 0.5 (discriminative ability) 1 (perfect discriminative ability). default, function ccalc derive c-statistic study, together corresponding standard error 95% confidence interval. However, also possible calculate transformed versions c-statistic. Appropriate standard errors derived using Delta method. instance, logit transformation can applied specifying g=\"log(cstat/(1-cstat))\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"restoring-the-c-statistic","dir":"Reference","previous_headings":"","what":"Restoring the c-statistic","title":"Calculate the concordance statistic — ccalc","text":"studies c-statistic missing, estimated standard deviation linear predictor (theta.source=\"std.dev(LP)\"). corresponding method described White et al. (2015).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"restoring-the-standard-error-of-the-c-statistic","dir":"Reference","previous_headings":"","what":"Restoring the standard error of the c-statistic","title":"Calculate the concordance statistic — ccalc","text":"missing, standard error c-statistic can estimated confidence interval. Alternatively, standard error can approximated combination reported c-statistic, total sample size total number events (Newcombe, 2006). can achieved adopting (modification ) method proposed Hanley McNeil, specified approx.se.method.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Calculate the concordance statistic — ccalc","text":"Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2018; press. Hanley JA, McNeil BJ. meaning use area receiver operating characteristic (ROC) curve. Radiology. 1982; 143(1):29--36. Newcombe RG. Confidence intervals effect size measure based Mann-Whitney statistic. Part 2: asymptotic methods evaluation. Stat Med. 2006; 25(4):559--73. Snell KI, Ensor J, Debray TP, Moons KG, Riley RD. Meta-analysis prediction model performance across multiple studies: scale helps ensure -study normality C -statistic calibration measures? Statistical Methods Medical Research. 2017. White IR, Rapsomaniki E, Emerging Risk Factors Collaboration. Covariate-adjusted measures discrimination survival data. Biom J. 2015;57(4):592--613.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Calculate the concordance statistic — ccalc","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Calculate the concordance statistic — ccalc","text":"","code":"######### Validation of prediction models with a binary outcome ######### data(EuroSCORE) # Calculate the c-statistic and its standard error est1 <- ccalc(cstat = c.index, cstat.se = se.c.index, cstat.cilb = c.index.95CIl, cstat.ciub = c.index.95CIu, N = n, O = n.events, data = EuroSCORE, slab = Study) est1 #> theta theta.se theta.cilb theta.ciub theta.source #> Nashef 0.8095 0.01377576 0.7820000 0.8360000 c-statistic #> Biancari 0.8670 0.03520473 0.7980000 0.9360000 c-statistic #> Di Dedda 0.8100 0.03571494 0.7400000 0.8800000 c-statistic #> Chalmers 0.7900 0.01000000 0.7704004 0.8095996 c-statistic #> Grant 0.8080 0.00800000 0.7923203 0.8236797 c-statistic #> Carneo 0.8500 0.01000000 0.8304004 0.8695996 c-statistic #> Kunt 0.7200 0.05100000 0.6200418 0.8199582 c-statistic #> Kirmani 0.8180 0.00700000 0.8042803 0.8317197 c-statistic #> Howell 0.6700 0.02972796 0.6117343 0.7282657 c-statistic #> Wang 0.7200 0.01500000 0.6906005 0.7493995 c-statistic #> Borde 0.7200 0.09044574 0.5427296 0.8972704 c-statistic #> Qadir 0.8400 0.02338189 0.7941723 0.8858277 c-statistic #> Spiliopoulos 0.7700 0.06700000 0.6386824 0.9013176 c-statistic #> Wendt 0.7200 0.03400000 0.6533612 0.7866388 c-statistic #> Laurent 0.7700 0.06100000 0.6504422 0.8895578 c-statistic #> Wang.1 0.6420 0.07100000 0.5028426 0.7811574 c-statistic #> Nishida 0.7697 0.04247895 0.6864428 0.8529572 c-statistic #> Barilli 0.8000 0.01500000 0.7706005 0.8293995 c-statistic #> Barilli.1 0.8200 0.02000000 0.7808007 0.8591993 c-statistic #> Paparella 0.8300 0.01200000 0.8064804 0.8535196 c-statistic #> Carosella 0.7600 0.05600000 0.6502420 0.8697580 c-statistic #> Borracci 0.8560 0.03300000 0.7913212 0.9206788 c-statistic #> Osnabrugge 0.7700 0.01000000 0.7504004 0.7895996 c-statistic #> theta.se.source #> Nashef Confidence Interval #> Biancari Confidence Interval #> Di Dedda Confidence Interval #> Chalmers Standard Error #> Grant Standard Error #> Carneo Standard Error #> Kunt Standard Error #> Kirmani Standard Error #> Howell Newcombe (Method 4) #> Wang Standard Error #> Borde Newcombe (Method 4) #> Qadir Newcombe (Method 4) #> Spiliopoulos Standard Error #> Wendt Standard Error #> Laurent Standard Error #> Wang.1 Standard Error #> Nishida Newcombe (Method 4) #> Barilli Standard Error #> Barilli.1 Standard Error #> Paparella Standard Error #> Carosella Standard Error #> Borracci Standard Error #> Osnabrugge Standard Error # Calculate the logit c-statistic and its standard error est2 <- ccalc(cstat = c.index, cstat.se = se.c.index, cstat.cilb = c.index.95CIl, cstat.ciub = c.index.95CIu, N = n, O = n.events, data = EuroSCORE, slab = Study, g = \"log(cstat/(1-cstat))\") est2 #> theta theta.se theta.cilb theta.ciub theta.source #> Nashef 1.4467646 0.08964514 1.27735968 1.6287622 c-statistic #> Biancari 1.8746898 0.33390703 1.37384090 2.6827324 c-statistic #> Di Dedda 1.4500102 0.24144872 1.04596856 1.9924302 c-statistic #> Chalmers 1.3249254 0.06027728 1.20678413 1.4430667 c-statistic #> Grant 1.4370667 0.05156766 1.33599594 1.5381374 c-statistic #> Carneo 1.7346011 0.07843137 1.58087839 1.8883237 c-statistic #> Kunt 0.9444616 0.25297619 0.44863739 1.4402858 c-statistic #> Kirmani 1.5028556 0.04701900 1.41070011 1.5950112 c-statistic #> Howell 0.7081851 0.13445481 0.44465847 0.9717116 c-statistic #> Wang 0.9444616 0.07440476 0.79863096 1.0902923 c-statistic #> Borde 0.9444616 0.44863958 0.06514418 1.8237790 c-statistic #> Qadir 1.6582281 0.17397241 1.31724841 1.9992077 c-statistic #> Spiliopoulos 1.2083112 0.37831733 0.46682285 1.9497996 c-statistic #> Wendt 0.9444616 0.16865079 0.61391213 1.2750111 c-statistic #> Laurent 1.2083112 0.34443817 0.53322480 1.8833976 c-statistic #> Wang.1 0.5840553 0.30891592 -0.02140877 1.1895194 c-statistic #> Nishida 1.2066180 0.23963947 0.73693330 1.6763027 c-statistic #> Barilli 1.3862944 0.09375000 1.20254774 1.5700410 c-statistic #> Barilli.1 1.5163475 0.13550136 1.25076971 1.7819253 c-statistic #> Paparella 1.5856273 0.08504607 1.41894004 1.7523145 c-statistic #> Carosella 1.1526795 0.30701754 0.55093618 1.7544228 c-statistic #> Borracci 1.7824571 0.26771807 1.25773930 2.3071748 c-statistic #> Osnabrugge 1.2083112 0.05646527 1.09764130 1.3189811 c-statistic #> theta.se.source #> Nashef Confidence Interval #> Biancari Confidence Interval #> Di Dedda Confidence Interval #> Chalmers Standard Error #> Grant Standard Error #> Carneo Standard Error #> Kunt Standard Error #> Kirmani Standard Error #> Howell Newcombe (Method 4) #> Wang Standard Error #> Borde Newcombe (Method 4) #> Qadir Newcombe (Method 4) #> Spiliopoulos Standard Error #> Wendt Standard Error #> Laurent Standard Error #> Wang.1 Standard Error #> Nishida Newcombe (Method 4) #> Barilli Standard Error #> Barilli.1 Standard Error #> Paparella Standard Error #> Carosella Standard Error #> Borracci Standard Error #> Osnabrugge Standard Error # Display the results of all studies in a forest plot plot(est1)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":null,"dir":"Reference","previous_headings":"","what":"Convert a correlation matrix into a covariance matrix — cor2cov","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"Convert correlation matrix covariance matrix","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"","code":"cor2cov(sigma, cormat)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"sigma vector standard deviations. order standard deviations correspond column order 'cormat'. cormat symmetric numeric correlation matrix","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"covariance matrix","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":null,"dir":"Reference","previous_headings":"","what":"Posterior distribution of estimated model parameters — dplot","title":"Posterior distribution of estimated model parameters — dplot","text":"Generate plot posterior distribution","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Posterior distribution of estimated model parameters — dplot","text":"","code":"dplot(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Posterior distribution of estimated model parameters — dplot","text":"... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Posterior distribution of estimated model parameters — dplot","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Posterior distribution of estimated model parameters — dplot","text":"generic function.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Posterior distribution of estimated model parameters — dplot","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":null,"dir":"Reference","previous_headings":"","what":"Posterior distribution of estimated model parameters — dplot.mcmc.list","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"Generate plot posterior distribution","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"","code":"# S3 method for mcmc.list dplot(x, P, plot_type = \"dens\", ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"x object class \"mcmc.list\" P Optional dataframe describing parameters plot respective names plot_type Optional character string specify whether density plot (\"dens\") histogram (\"hist\") displayed. ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"Function generate plots prior posterior distribution Bayesian meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"","code":"# S3 method for uvmeta dplot(x, par, distr_type, plot_type = \"dens\", ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"x object class \"uvmeta\" par Character string specify parameter plot generated. Options \"mu\" (mean random effects model) \"tau\" (standard deviation random effects model). distr_type Character string specify whether prior distribution (\"prior\") posterior distribution (\"posterior\") displayed. plot_type Character string specify whether density plot (\"dens\") histogram (\"hist\") displayed. ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"","code":"if (FALSE) { data(Roberts) fit <- with(Roberts, uvmeta(r=SDM, r.se=SE, method=\"BAYES\")) dplot(fit) dplot(fit, distr_type = \"posterior\") dplot(fit, par = \"tau\", distr_type = \"prior\") dplot(fit, plot_type = \"hist\") }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"Function generate plots prior posterior distribution Bayesian meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"","code":"# S3 method for valmeta dplot(x, par, distr_type, plot_type = \"dens\", ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"x object class \"valmeta\" par Character string specify parameter plot generated. Options \"mu\" (mean random effects model) \"tau\" (standard deviation random effects model). distr_type Character string specify whether prior distribution (\"prior\") posterior distribution (\"posterior\") displayed. plot_type Character string specify whether density plot (\"dens\") histogram (\"hist\") displayed. ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"","code":"if (FALSE) { data(EuroSCORE) # Meta-analysis of the concordance statistic fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) dplot(fit) dplot(fit, distr_type = \"posterior\") dplot(fit, par = \"tau\", distr_type = \"prior\") # Meta-analysis of the O:E ratio EuroSCORE.new <- EuroSCORE EuroSCORE.new$n[c(1, 2, 5, 10, 20)] <- NA pars <- list(hp.tau.dist=\"dhalft\", # Prior for the between-study standard deviation hp.tau.sigma=1.5, # Standard deviation for 'hp.tau.dist' hp.tau.df=3, # Degrees of freedom for 'hp.tau.dist' hp.tau.max=10) # Maximum value for the between-study standard deviation fit2 <- valmeta(measure=\"OE\", O=n.events, E=e.events, N=n, data=EuroSCORE.new, method=\"BAYES\", slab=Study, pars=pars) dplot(fit2, plot_type = \"hist\") }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":null,"dir":"Reference","previous_headings":"","what":"Regression tests for detecting funnel plot asymmetry — fat","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"presence small-study effects common threat systematic reviews meta-analyses, especially due publication bias, occurs small primary studies likely reported (published) findings positive. presence small-study effects can verified visual inspection funnel plot, included study meta-analysis, estimate reported effect size depicted measure precision sample size. premise scatter plots reflect funnel shape, small-study effects exist. However, small studies predominately one direction (usually direction larger effect sizes), asymmetry ensue. fat function implements several tests detecting funnel plot asymmetry, can used presence -study heterogeneity treatment effect relatively low.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"","code":"fat(b, b.se, n.total, d.total, d1, d2, method = \"E-FIV\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"b Vector effect size study. Examples log odds ratio, log hazards ratio, log relative risk. b.se Optional vector standard error effect size study n.total Optional vector total sample size study d.total Optional vector total number observed events study d1 Optional vector total number observed events exposed groups d2 Optional vector total number observed events unexposed groups method Method testing funnel plot asymmetry, defaults \"E-FIV\" (Egger's test multiplicative dispersion). options E-UW, M-FIV, M-FPV, D-FIV D-FAV. info \"Details\"","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"list containing following entries: \"pval\" two-sided P-value indicating statistical significance funnel plot asymettry test. Values significance level (usually defined 10%) support presence funnel plot asymmetry, thus small-study effects. \"model\" fitted glm object, representing estimated regression model used testing funnel plot asymmetry.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"egger-regression-method","dir":"Reference","previous_headings":"","what":"Egger regression method","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"common approach test presence small-study effects estimate regression model standardized effect estimate (effect/SE) regressed measure precision (1/SE), (method=\"E-UW\", Egger 1997). possible allow -study heterogeneity adopting multiplicative overdispersion parameter variance study multiplied (method=\"E-FIV\", Sterne 2000). Unfortunately, demonstrated aforementioned two tests biased : () independent variable subject sampling variability; (ii) standardized treatment effect correlated estimated precision; (iii) binary data, independent regression variable biased estimate true precision, larger bias smaller sample sizes (Macaskill et al. 2001).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"macaskill-regression-method","dir":"Reference","previous_headings":"","what":"Macaskill regression method","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"standard approach estimates regression model effect size function study size (method=\"M-FIV\", Macaskill et al. 2001). study weighted precision treatment effect estimate allow possible heteroscedasticity. alternative approach weight study pooled' estimate outcome proportion (method=\"M-FPV\") studies zero events, continuity correction applied adding 0.5 cell counts.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"peters-regression-method","dir":"Reference","previous_headings":"","what":"Peters regression method","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"approach (method=\"P-FPV\") estimates regression model treatment effect function inverse total sample size (Peters et al. 2006). studies zero events, continuity correction applied adding 0.5 cell counts.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"debray-regression-method","dir":"Reference","previous_headings":"","what":"Debray regression method","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"approach proposed survival data, uses total number events independent variable weighted regression model (Debray et al. 2017). study weights based inverse variance (method=\"D-FIV\") approximation thereof (method=\"D-FAV\").","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"Debray TPA, Moons KGM, Riley RD. Detecting small-study effects funnel plot asymmetry meta-analysis survival data: comparison new existing tests. Res Syn Meth. 2018;9(1):41--50. Egger M, Davey Smith G, Schneider M, Minder C. Bias meta-analysis detected simple, graphical test. BMJ. 1997;315(7109):629--34. Macaskill P, Walter SD, Irwig L. comparison methods detect publication bias meta-analysis. Stat Med. 2001;20(4):641--54. Peters JL, Sutton AJ, Jones DR, Abrams KR, Rushton L. Comparison two methods detect publication bias meta-analysis. JAMA. 2006 Feb 8;295(6):676--80. Sterne JA, Gavaghan D, Egger M. Publication related bias meta-analysis: power statistical tests prevalence literature. J Clin Epidemiol. 2000;53(11):1119--29.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"","code":"data(Fibrinogen) b <- log(Fibrinogen$HR) b.se <- ((log(Fibrinogen$HR.975) - log(Fibrinogen$HR.025))/(2*qnorm(0.975))) n.total <- Fibrinogen$N.total d.total <- Fibrinogen$N.events fat(b=b, b.se=b.se) #> Call: fat(b = b, b.se = b.se) #> #> Fixed effect summary estimate: 0.4186 #> #> test for funnel plot asymmetry: t =1.9021, df = 29, p = 0.0671 fat(b=b, b.se=b.se, d.total=d.total, method=\"D-FIV\") #> Call: fat(b = b, b.se = b.se, d.total = d.total, method = \"D-FIV\") #> #> Fixed effect summary estimate: 0.4186 #> #> test for funnel plot asymmetry: t =1.6847, df = 29, p = 0.1028 # Note that many tests are also available via metafor require(metafor) #> Loading required package: metafor #> Loading required package: Matrix #> Loading required package: metadat #> Loading required package: numDeriv #> #> Loading the 'metafor' package (version 4.4-0). For an #> introduction to the package please type: help(metafor) #> #> Attaching package: ‘metafor’ #> The following object is masked from ‘package:metamisc’: #> #> forest fat(b=b, b.se=b.se, n.total=n.total, method=\"M-FIV\") #> Call: fat(b = b, b.se = b.se, n.total = n.total, method = \"M-FIV\") #> #> Fixed effect summary estimate: 0.4186 #> #> test for funnel plot asymmetry: t =-1.4275, df = 29, p = 0.1641 regtest(x=b, sei=b.se, ni=n.total, model=\"lm\", predictor=\"ni\") #> #> Regression Test for Funnel Plot Asymmetry #> #> Model: weighted regression with multiplicative dispersion #> Predictor: sample size #> #> Test for Funnel Plot Asymmetry: t = -1.4275, df = 29, p = 0.1641 #>"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract Model Fitted Values — fitted.metapred","title":"Extract Model Fitted Values — fitted.metapred","text":"Extract fitted values metapred object. default returns fitted values model cross-validation procedure.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract Model Fitted Values — fitted.metapred","text":"","code":"# S3 method for metapred fitted( object, select = \"cv\", step = NULL, model = NULL, as.stratified = TRUE, type = \"response\", ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract Model Fitted Values — fitted.metapred","text":"object object class metapred select character. Select fitted values \"cv\" (default) \"global\" model. step character numeric. Name number step select select = \"cv\". Defaults best step. model character numeric. Name number model select select = \"cv\". Defaults best model. .stratified logical. select = \"cv\" determines whether returned predictions stratified list (TRUE, default) original order (FALSE). type character. Type fitted value. ... compatibility .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Extract Model Fitted Values — fitted.metapred","text":"Function still development, use caution. returns type = \"response\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Extract Model Fitted Values — fitted.metapred","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest plot — forest.default","title":"Forest plot — forest.default","text":"Generate forest plot specifying various effect sizes, confidence intervals summary estimate.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest plot — forest.default","text":"","code":"# S3 method for default forest( theta, theta.ci.lb, theta.ci.ub, theta.slab, theta.summary, theta.summary.ci.lb, theta.summary.ci.ub, theta.summary.pi.lb, theta.summary.pi.ub, title, sort = \"asc\", theme = theme_bw(), predint.linetype = 1, xlim, xlab = \"\", refline = 0, label.summary = \"Summary Estimate\", label.predint = \"Prediction Interval\", nrows.before.summary = 1, study.digits = 2, study.shape = 15, col.diamond = \"white\", col.predint = \"black\", size.study = 0.5, size.predint = 1, lty.ref = \"dotted\", ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest plot — forest.default","text":"theta Numeric vector effect size study theta.ci.lb Numeric vector specifying lower bound confidence interval effect sizes theta.ci.ub Numeric vector specifying upper bound confidence interval effect sizes theta.slab Character vector specifying study labels theta.summary Meta-analysis summary estimate effect sizes theta.summary.ci.lb Lower bound confidence (credibility) interval summary estimate theta.summary.ci.ub Upper bound confidence (credibility) interval summary estimate theta.summary.pi.lb Lower bound (approximate) prediction interval summary estimate. theta.summary.pi.ub Upper bound (approximate) prediction interval summary estimate. title Title forest plot sort default, studies sorted ascending effect size (sort=\"asc\"). Set \"desc\" sorting reverse order, value ignore sorting. theme Theme generate forest plot. default, classic dark--light ggplot2 theme used. See ggtheme information. predint.linetype linetype prediction interval xlim x limits (x1, x2) forest plot xlab Optional character string specifying X label refline Optional numeric specifying reference line label.summary Optional character string specifying label summary estimate label.predint Optional character string specifying label (approximate) prediction interval nrows..summary many empty rows introduced study results summary estimates study.digits many significant digits used print stuy results study.shape Plotting symbol use study results. default, filled square used. col.diamond filling color diamond representing summary estimate. E.g. \"red\", \"blue\", hex color code (\"#2e8aff\") col.predint Line color prediction interval. E.g. \"red\", \"blue\", hex color code (\"#2e8aff\") size.study Line width study results mm size.predint Line width prediction interval mm lty.ref Line type reference line ... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Forest plot — forest.default","text":"object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest plot — forest.default","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest plot — forest","title":"Forest plot — forest","text":"Generate forest plot specifying various effect sizes, confidence intervals summary estimate.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest plot — forest","text":"","code":"forest(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest plot — forest","text":"... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Forest plot — forest","text":"generic function. See forest.default making forest plots summary statistics, forest.metapred plotting metapred objects, forest.mp.cv.val plotting mp.cv.val objects.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest plot — forest","text":"Thomas Debray Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest plot of a metapred fit — forest.metapred","title":"Forest plot of a metapred fit — forest.metapred","text":"Draw forest plot performance internally-externally cross-validated model. default final model shown.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest plot of a metapred fit — forest.metapred","text":"","code":"# S3 method for metapred forest(x, perfFUN = 1, step = NULL, method = \"REML\", model = NULL, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest plot of a metapred fit — forest.metapred","text":"x metapred fit object perfFUN Numeric character. performance statistic plotted? Defaults first. step step plotted? Defaults best step. numeric converted name step: 0 unchanged model, 1 first change... method character string specifying whether fixed- random-effects model used summarize prediction model performance. fixed-effects model fitted using method=\"FE\". Random-effects models fitted setting method equal one following: \"DL\", \"\", \"SJ\", \"ML\", \"REML\", \"EB\", \"HS\", \"GENQ\". Default \"REML\". model model change plotted? NULL (default, best change) character name variable (integer) index model change. ... arguments passed plotting internals. E.g. title. See forest.default details.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest plot of a metapred fit — forest.metapred","text":"Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Forest plot of a metapred fit — forest.metapred","text":"","code":"data(DVTipd) # Internal-external cross-validation of a pre-specified model 'f' f <- dvt ~ histdvt + ddimdich + sex + notraum fit <- metapred(DVTipd, strata = \"study\", formula = f, scope = f, family = binomial) # Display the model's external performance (expressed as mean squared error by default) # for each study forest(fit) #> Error in forest.default(fit): Must specify either 'vi', 'sei', or ('ci.lb', 'ci.ub') pairs."},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.mp.cv.val.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest plot of a validation object. — forest.mp.cv.val","title":"Forest plot of a validation object. — forest.mp.cv.val","text":"Draw forest plot performance internally-externally cross-validated model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.mp.cv.val.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest plot of a validation object. — forest.mp.cv.val","text":"","code":"# S3 method for mp.cv.val forest(x, perfFUN = 1, method = \"REML\", xlab = NULL, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.mp.cv.val.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest plot of a validation object. — forest.mp.cv.val","text":"x mp.cv.val perf object. perfFUN Numeric character. performance statistic plotted? Defaults first. method character string specifying whether fixed- random-effects model used summarize prediction model performance. fixed-effects model fitted using method=\"FE\". Random-effects models fitted setting method equal one following: \"DL\", \"\", \"SJ\", \"ML\", \"REML\", \"EB\", \"HS\", \"GENQ\". Default \"REML\". xlab Label x-axis. Defaults name performance function. ... arguments passed plotting internals. E.g. title. See forest.default details.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.mp.cv.val.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest plot of a validation object. — forest.mp.cv.val","text":"Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":null,"dir":"Reference","previous_headings":"","what":"Gelman-Rubin-Brooks plot — gelmanplot","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"plot shows evolution Gelman Rubin's shrink factor number iterations increases. code adapted R package coda.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"","code":"gelmanplot(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":null,"dir":"Reference","previous_headings":"","what":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"plot shows evolution Gelman Rubin's shrink factor number iterations increases. code adapted R package coda.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"","code":"# S3 method for mcmc.list gelmanplot( x, P, confidence = 0.95, max.bins = 50, autoburnin = TRUE, greek = FALSE, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"x mcmc object P Optional dataframe describing parameters plot respective names confidence coverage probability confidence interval potential scale reduction factor max.bins Maximum number bins, excluding last one. autoburnin Logical flag indicating whether second half series used computation. set TRUE (default) start(x) less end(x)/2 start series adjusted second half series used. greek Logical value indicating whether parameter labels parsed get Greek letters. Defaults false. ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"plot shows evolution Gelman Rubin's shrink factor number iterations increases. code adapted R package coda.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"","code":"# S3 method for uvmeta gelmanplot(x, confidence = 0.95, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"x mcmc object confidence coverage probability confidence interval potential scale reduction factor ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"plot shows evolution Gelman Rubin's shrink factor number iterations increases. code adapted R package coda.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"","code":"# S3 method for valmeta gelmanplot(x, confidence = 0.95, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"x mcmc object confidence coverage probability confidence interval potential scale reduction factor ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"","code":"if (FALSE) { data(EuroSCORE) # Meta-analysis of the concordance statistic fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) gelmanplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":null,"dir":"Reference","previous_headings":"","what":"Generalizability estimates — gen","title":"Generalizability estimates — gen","text":"Obtain generalizability estimates model fit.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Generalizability estimates — gen","text":"","code":"gen(object, ...) generalizability(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Generalizability estimates — gen","text":"object model fit object, either metapred subset(metapred) object. ... default, final model selected. parameter allows arguments passed subset.metapred generalizability estimates steps/models may returned..","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Generalizability estimates — gen","text":"named values indices parameter genFUN one estimates generalizability can selected. Use genFUN = 0 select .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Generalizability estimates — gen","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":null,"dir":"Reference","previous_headings":"","what":"IMPACT data — impact","title":"IMPACT data — impact","text":"IMPACT dataset comprises 15 studies patients suffering traumatic brain injury, including individual patient data 11 randomized controlled trials four observational studies.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"IMPACT data — impact","text":"","code":"data(\"impact\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"IMPACT data — impact","text":"data frame 11022 observations following 11 variables. name Name study type Type study, RCT: randomized controlled trial,OBS: observational cohort age Age patient motor_score Glasgow Coma Scale motor score pupil Pupillary reactivity ct Marshall Computerized Tomography classification hypox Hypoxia (0=, 1=yes) hypots Hypotension (0=, 1=yes) tsah Traumatic subarachnoid hemorrhage (0=, 1=yes) edh Epidural hematoma (0=, 1=yes) mort 6-month mortality (0=alive, 1=dead)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"IMPACT data — impact","text":"included studies part IMPACT project, total 25 prognostic factors considered prediction 6-month mortality. Missing values imputed using study fixed effect imputation model (Steyerberg et al, 2008).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"IMPACT data — impact","text":"Steyerberg EW, Nieboer D, Debray TPA, Van Houwelingen JC. Assessment heterogeneity individual participant data meta-analysis prediction models: overview illustration. Stat Med. 2019;38(22):4290--309.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"IMPACT data — impact","text":"Murray GD, Butcher , McHugh GS, et al. Multivariable prognostic analysis traumatic brain injury: results IMPACT study. J Neurotrauma. 2007;24(2):329--337. Steyerberg EW, Mushkudiani N, Perel P, et al. Predicting outcome traumatic brain injury: development international validation prognostic scores based admission characteristics. PLOS Med. 2008;5(8):e165.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"IMPACT data — impact","text":"","code":"data(impact) by(impact, impact$name, summary) #> impact$name: APOE #> name type age motor_score pupil ct #> APOE :756 OBS:756 Min. :14.00 1/2: 7 Both:638 I/II:377 #> CSTAT : 0 RCT: 0 1st Qu.:25.00 3 :331 None: 79 III :151 #> EBIC : 0 Median :37.00 4 : 18 One : 39 IV/V:228 #> HIT I : 0 Mean :41.14 5/6:400 #> HIT II : 0 3rd Qu.:56.00 #> NABIS : 0 Max. :93.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.000 Min. :0.00000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.000 1st Qu.:0.00000 #> Median :0.0000 Median :0.0000 Median :0.000 Median :0.00000 #> Mean :0.2712 Mean :0.1071 Mean :0.254 Mean :0.06614 #> 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:1.000 3rd Qu.:0.00000 #> Max. :1.0000 Max. :1.0000 Max. :1.000 Max. :1.00000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.1548 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: CSTAT #> name type age motor_score pupil ct #> CSTAT :517 OBS: 0 Min. :15.00 1/2:128 Both:368 I/II:212 #> APOE : 0 RCT:517 1st Qu.:20.00 3 : 86 None: 63 III :105 #> EBIC : 0 Median :28.00 4 :180 One : 86 IV/V:200 #> HIT I : 0 Mean :31.79 5/6:123 #> HIT II : 0 3rd Qu.:41.00 #> NABIS : 0 Max. :68.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.1335 Mean :0.1663 Mean :0.4778 Mean :0.1161 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2224 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: EBIC #> name type age motor_score pupil ct #> EBIC :822 OBS:822 Min. :14.00 1/2:230 Both:531 I/II:333 #> APOE : 0 RCT: 0 1st Qu.:24.00 3 :198 None:209 III : 82 #> CSTAT : 0 Median :37.50 4 :113 One : 82 IV/V:407 #> HIT I : 0 Mean :41.79 5/6:281 #> HIT II : 0 3rd Qu.:59.00 #> NABIS : 0 Max. :92.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.00000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.00000 #> Mean :0.2871 Mean :0.2445 Mean :0.4136 Mean :0.09246 #> 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.00000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.00000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.3418 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: HIT I #> name type age motor_score pupil ct #> HIT I :350 OBS: 0 Min. :14.00 1/2:163 Both:232 I/II:127 #> APOE : 0 RCT:350 1st Qu.:21.00 3 : 54 None: 67 III : 79 #> CSTAT : 0 Median :34.00 4 : 56 One : 51 IV/V:144 #> EBIC : 0 Mean :35.49 5/6: 77 #> HIT II : 0 3rd Qu.:47.00 #> NABIS : 0 Max. :71.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.00 Min. :0.00000 Min. :0.0000 Min. :0.00 #> 1st Qu.:0.00 1st Qu.:0.00000 1st Qu.:0.0000 1st Qu.:0.00 #> Median :0.00 Median :0.00000 Median :0.0000 Median :0.00 #> Mean :0.18 Mean :0.05429 Mean :0.3257 Mean :0.18 #> 3rd Qu.:0.00 3rd Qu.:0.00000 3rd Qu.:1.0000 3rd Qu.:0.00 #> Max. :1.00 Max. :1.00000 Max. :1.0000 Max. :1.00 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2829 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: HIT II #> name type age motor_score pupil ct #> HIT II :819 OBS: 0 Min. :15.0 1/2:280 Both:583 I/II:345 #> APOE : 0 RCT:819 1st Qu.:22.0 3 :151 None:138 III : 90 #> CSTAT : 0 Median :33.0 4 :181 One : 98 IV/V:384 #> EBIC : 0 Mean :36.3 5/6:207 #> HIT I : 0 3rd Qu.:49.0 #> NABIS : 0 Max. :74.0 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.199 Mean :0.1001 Mean :0.3272 Mean :0.1636 #> 3rd Qu.:0.000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2295 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: NABIS #> name type age motor_score pupil ct #> NABIS :385 OBS: 0 Min. :14.00 1/2:144 Both:236 I/II: 45 #> APOE : 0 RCT:385 1st Qu.:21.00 3 : 65 None: 99 III :204 #> CSTAT : 0 Median :30.00 4 : 76 One : 50 IV/V:136 #> EBIC : 0 Mean :31.82 5/6:100 #> HIT I : 0 3rd Qu.:40.00 #> HIT II : 0 Max. :68.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.3299 Mean :0.1558 Mean :0.4779 Mean :0.1325 #> 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2623 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: PEG #> name type age motor_score pupil ct #> PEG :1510 OBS: 0 Min. :15.00 1/2:655 Both:787 I/II:576 #> APOE : 0 RCT:1510 1st Qu.:20.00 3 :165 None:564 III :345 #> CSTAT : 0 Median :27.00 4 :334 One :159 IV/V:589 #> EBIC : 0 Mean :30.44 5/6:356 #> HIT I : 0 3rd Qu.:38.00 #> HIT II : 0 Max. :86.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.00000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.00000 #> Mean :0.2199 Mean :0.1695 Mean :0.4113 Mean :0.09801 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.00000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.00000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2397 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: PHARMOS #> name type age motor_score pupil ct #> PHARMOS:856 OBS: 0 Min. :16.00 1/2:134 Both:667 I/II:429 #> APOE : 0 RCT:856 1st Qu.:23.00 3 :262 None: 37 III :205 #> CSTAT : 0 Median :33.00 4 :225 One :152 IV/V:222 #> EBIC : 0 Mean :35.01 5/6:235 #> HIT I : 0 3rd Qu.:45.25 #> HIT II : 0 Max. :66.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :1.000 Median :0.0000 #> Mean :0.2477 Mean :0.1542 Mean :0.597 Mean :0.1939 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.1694 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: SAP #> name type age motor_score pupil ct #> SAP :919 OBS: 0 Min. :15.00 1/2:264 Both:639 I/II:400 #> APOE : 0 RCT:919 1st Qu.:23.00 3 :146 None:155 III :151 #> CSTAT : 0 Median :32.00 4 :223 One :125 IV/V:368 #> EBIC : 0 Mean :35.63 5/6:286 #> HIT I : 0 3rd Qu.:47.00 #> HIT II : 0 Max. :71.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.1371 Mean :0.1523 Mean :0.4418 Mean :0.2057 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2307 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: SKB #> name type age motor_score pupil ct #> SKB :126 OBS: 0 Min. :16.00 1/2:56 Both:81 I/II:54 #> APOE : 0 RCT:126 1st Qu.:20.00 3 :31 None:28 III :40 #> CSTAT : 0 Median :27.00 4 :16 One :17 IV/V:32 #> EBIC : 0 Mean :30.61 5/6:23 #> HIT I : 0 3rd Qu.:39.00 #> HIT II : 0 Max. :70.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:1.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :1.0000 Median :0.0000 #> Mean :0.2937 Mean :0.1905 Mean :0.7857 Mean :0.1111 #> 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2698 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: SLIN #> name type age motor_score pupil ct #> SLIN :409 OBS: 0 Min. :15.00 1/2: 55 Both:316 I/II:154 #> APOE : 0 RCT:409 1st Qu.:21.00 3 : 91 None: 16 III : 94 #> CSTAT : 0 Median :28.00 4 :127 One : 77 IV/V:161 #> EBIC : 0 Mean :32.35 5/6:136 #> HIT I : 0 3rd Qu.:43.00 #> HIT II : 0 Max. :79.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.00000 Min. :0.0000 Min. :0.00 Min. :0.0000 #> 1st Qu.:0.00000 1st Qu.:0.0000 1st Qu.:1.00 1st Qu.:0.0000 #> Median :0.00000 Median :0.0000 Median :1.00 Median :0.0000 #> Mean :0.05868 Mean :0.1369 Mean :0.78 Mean :0.1516 #> 3rd Qu.:0.00000 3rd Qu.:0.0000 3rd Qu.:1.00 3rd Qu.:0.0000 #> Max. :1.00000 Max. :1.0000 Max. :1.00 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2298 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: TCDB #> name type age motor_score pupil ct #> TCDB :603 OBS:603 Min. :16.00 1/2:243 Both:299 I/II:219 #> APOE : 0 RCT: 0 1st Qu.:21.00 3 :105 None:249 III :119 #> CSTAT : 0 Median :26.00 4 :121 One : 55 IV/V:265 #> EBIC : 0 Mean :32.97 5/6:134 #> HIT I : 0 3rd Qu.:40.00 #> HIT II : 0 Max. :93.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.1808 Mean :0.2371 Mean :0.4428 Mean :0.1045 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.4378 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: TINT #> name type age motor_score pupil ct #> TINT :1118 OBS: 0 Min. :14.00 1/2:141 Both:807 I/II:474 #> APOE : 0 RCT:1118 1st Qu.:21.00 3 :237 None:138 III :221 #> CSTAT : 0 Median :30.00 4 :327 One :173 IV/V:423 #> EBIC : 0 Mean :33.61 5/6:413 #> HIT I : 0 3rd Qu.:45.00 #> HIT II : 0 Max. :79.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :1.0000 Median :0.0000 #> Mean :0.1592 Mean :0.1404 Mean :0.5259 Mean :0.1655 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2487 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: TIUS #> name type age motor_score pupil ct #> TIUS :1041 OBS: 0 Min. :14.00 1/2:152 Both:708 I/II:463 #> APOE : 0 RCT:1041 1st Qu.:23.00 3 :132 None:211 III :198 #> CSTAT : 0 Median :30.00 4 :300 One :122 IV/V:380 #> EBIC : 0 Mean :32.78 5/6:457 #> HIT I : 0 3rd Qu.:41.00 #> HIT II : 0 Max. :77.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.00 Min. :0.0000 Min. :0.00000 #> 1st Qu.:0.0000 1st Qu.:0.00 1st Qu.:0.0000 1st Qu.:0.00000 #> Median :0.0000 Median :0.00 Median :0.0000 Median :0.00000 #> Mean :0.2795 Mean :0.22 Mean :0.4284 Mean :0.08646 #> 3rd Qu.:1.0000 3rd Qu.:0.00 3rd Qu.:1.0000 3rd Qu.:0.00000 #> Max. :1.0000 Max. :1.00 Max. :1.0000 Max. :1.00000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2161 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: UK4 #> name type age motor_score pupil ct #> UK4 :791 OBS:791 Min. :14.00 1/2:198 Both:433 I/II:271 #> APOE : 0 RCT: 0 1st Qu.:22.00 3 :231 None:243 III :155 #> CSTAT : 0 Median :36.00 4 :141 One :115 IV/V:365 #> EBIC : 0 Mean :39.64 5/6:221 #> HIT I : 0 3rd Qu.:55.00 #> HIT II : 0 Max. :87.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :1.0000 Median :0.0000 #> Mean :0.2566 Mean :0.2642 Mean :0.5183 Mean :0.1429 #> 3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.4539 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> # Plot the distribution of age by study library(ggplot2) e <- ggplot(impact, aes(x = name, y = age)) e + geom_violin(aes(fill = type), trim = FALSE) + theme(axis.text.x = element_text(angle = 45)) + xlab(\"Study\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":null,"dir":"Reference","previous_headings":"","what":"Impute missing values by their conditional mean — impute_conditional_mean","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"function imputes missing values conditional mean","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"","code":"impute_conditional_mean(x, mu, Sigma)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"x vector observations, may missing (indicated NA) mu vector population means 'x'. missing values allowed . Sigma matrix describing population covariance 'x'","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"vector missing values 'x' replaced conditional mean","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"","code":"# Define the population means mu <- c(0, 1, 2) # Define the covariance of the population Sigma <- diag(1,3) Sigma[1,2] <- Sigma[2,1] <- 0.3 Sigma[2,3] <- Sigma[3,2] <- 0.1 Sigma[1,3] <- Sigma[3,1] <- -0.2 # Generate a 'random' sample from the population that is partially observed x <- c(NA, 2, 4) # Impute the missing values impute_conditional_mean (x=x, mu=mu, Sigma=Sigma) #> [1] -0.1414141 2.0000000 4.0000000"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":null,"dir":"Reference","previous_headings":"","what":"Apply the inverse logit tranformation — inv.logit","title":"Apply the inverse logit tranformation — inv.logit","text":"Transforms linear predictor probability.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Apply the inverse logit tranformation — inv.logit","text":"","code":"inv.logit(x)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Apply the inverse logit tranformation — inv.logit","text":"x vector numerics (-Inf Inf)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Apply the inverse logit tranformation — inv.logit","text":"vector numerics 0 1.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Apply the inverse logit tranformation — inv.logit","text":"Thomas Debray ","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Print the log-likelihood — logLik.riley","title":"Print the log-likelihood — logLik.riley","text":"function provides (restricted) log-likelihood fitted model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Print the log-likelihood — logLik.riley","text":"","code":"# S3 method for riley logLik(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Print the log-likelihood — logLik.riley","text":"object riley object, representing fitted alternative model bivariate random-effects meta-analysis within-study correlations unknown. ... Additional arguments passed functions, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Print the log-likelihood — logLik.riley","text":"Returns object class logLik. (restricted) log-likelihood model represented object evaluated estimated coefficients. contains least one attribute, \"df\" (degrees freedom), giving number (estimated) parameters model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Print the log-likelihood — logLik.riley","text":"Riley RD, Thompson JR, Abrams KR. alternative model bivariate random-effects meta-analysis within-study correlations unknown. Biostatistics 2008; 9: 172--186.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Print the log-likelihood — logLik.riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Print the log-likelihood — logLik.riley","text":"","code":"data(Daniels) fit <- riley(Daniels,control=list(maxit=10000)) logLik(fit) #> 'log Lik.' -48.85119 (df=5)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":null,"dir":"Reference","previous_headings":"","what":"Apply logit tranformation — logit","title":"Apply logit tranformation — logit","text":"Transforms values 0 1 values -Inf Inf.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Apply logit tranformation — logit","text":"","code":"logit(x)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Apply logit tranformation — logit","text":"x vector numerics (0 1)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Apply logit tranformation — logit","text":"vector numerics (-Inf Inf).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Apply logit tranformation — logit","text":"Thomas Debray ","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":null,"dir":"Reference","previous_headings":"","what":"Random effects meta-analysis — ma","title":"Random effects meta-analysis — ma","text":"Meta-analysis performance coefficients metapred object. Caution: still development.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Random effects meta-analysis — ma","text":"","code":"ma(object, method, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Random effects meta-analysis — ma","text":"object model fit object, metapred object. method Character, method meta-analysis passed valmeta uvmeta. Defaults \"REML\". ... arguments passed metapred, valmeta uvmeta.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Random effects meta-analysis — ma","text":"Produces different object types depending input.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Random effects meta-analysis — ma","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metamisc-package.html","id":null,"dir":"Reference","previous_headings":"","what":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","text":"Facilitate frequentist Bayesian meta-analysis diagnosis prognosis research studies. includes functions summarize multiple estimates prediction model discrimination calibration performance (Debray et al., 2019) . also includes functions evaluate funnel plot asymmetry (Debray et al., 2018) . Finally, package provides functions developing multivariable prediction models datasets clustering (de Jong et al., 2021) .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metamisc-package.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","text":"following functionality currently implemented: univariate meta-analysis summary data (uvmeta), bivariate meta-analysis correlated outcomes (riley), meta-analysis prediction model performance (valmeta), evaluation funnel plot asymmetry (fat). metamisc package also provides comprehensive framework developing prediction models patient-level data multiple studies settings available (metapred).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metamisc-package.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","text":"Thomas Debray , Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metamisc-package.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","text":"de Jong VMT, Moons KGM, Eijkemans MJC, Riley RD, Debray TPA. Developing generalizable prediction models pooled studies large clustered data sets. Stat Med. 2021;40(15):3533--59. Debray TPA, Moons KGM, Ahmed , Koffijberg H, Riley RD. framework developing, implementing, evaluating clinical prediction models individual participant data meta-analysis. Stat Med. 2013;32(18):3158--80. Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2019 Sep;28(9):2768--86. Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. Debray TPA, Moons KGM, Riley RD. Detecting small-study effects funnel plot asymmetry meta-analysis survival data: comparison new existing tests. Res Syn Meth. 2018;9(1):41--50. Riley RD, Moons K, Snell KIE, Ensor J, Hooft L, Altman D, et al. guide systematic review meta-analysis prognostic factor studies. BMJ. 2019;364:k4597. Riley RD, Tierney JF, Stewart LA. Individual participant data meta-analysis: handbook healthcare research. Hoboken, NJ: Wiley; 2021. ISBN: 978-1-119-33372-2. Schmid CH, Stijnen T, White IR. Handbook meta-analysis. First edition. Boca Raton: Taylor Francis; 2020. ISBN: 978-1-315-11940-3. Steyerberg EW, Nieboer D, Debray TPA, Van Houwelingen JC. Assessment heterogeneity individual participant data meta-analysis prediction models: overview illustration. Stat Med. 2019;38(22):4290--309.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":null,"dir":"Reference","previous_headings":"","what":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"Generalized stepwise regression obtaining prediction model validated (stepwise) internal-external cross-validation, obtain adequate performance across data sets. Requires data individuals multiple studies.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"","code":"metapred( data, strata, formula, estFUN = \"glm\", scope = NULL, retest = FALSE, max.steps = 1000, center = FALSE, recal.int = FALSE, cvFUN = NULL, cv.k = NULL, metaFUN = NULL, meta.method = NULL, predFUN = NULL, perfFUN = NULL, genFUN = NULL, selFUN = \"which.min\", gen.of.perf = \"first\", ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"data data.frame containing data. Note metapred removes observations missing data listwise variables formula scope, ensure data used model step. outcome variable numeric coercible numeric .numeric(). strata Character specify name strata (e.g. studies clusters) variable formula formula first model evaluated. metapred start formula update using terms scope. Defaults full main effects model, first column data assumed outcome remaining columns (except strata) predictors. See formula formulas general. estFUN Function estimating model first stage. Currently \"lm\", \"glm\" \"logistfirth\" supported. scope formula. difference formula scope defines range models examined stepwise search. Defaults NULL, leads intercept-model. scope nested formula, implies backwards selection applied (default). scope nested formula, implies forward selection applied. equal, stepwise selection applied. retest Logical. added (removed) terms retested removal (addition)? TRUE implies bi-directional stepwise search. max.steps Integer. Maximum number steps (additions removals terms) take. Defaults 1000, essentially many takes. 0 implies stepwise selection. center logical. numeric predictors centered around cluster mean? recal.int Logical. intercept recalibrated validation? cvFUN Cross-validation method, study (.e. cluster stratum) level. \"l1o\" leave-one-cross-validation (default). \"bootstrap\" bootstrap. \"fixed\", one data sets used validation. user written function may supplied well. cv.k Parameter cvFUN. cvFUN=\"bootstrap\", number bootstraps. cvFUN=\"fixed\", vector indices (sorted) data sets. used cvFUN=\"l1o\". metaFUN Function computing meta-analytic coefficient estimates two-stage MA. default, rma.uni, metafor package used. Default settings univariate random effects, estimated \"REML\". Method can passed trough meta.method argument. meta.method Name method meta-analysis. Default \"REML\". options see rma.uni. predFUN Function predicting new values. Defaults predicted probability outcome, using link function glm() lm(). perfFUN Function computing performance prediction models. Default: mean squared error (perfFUN=\"mse\", aka Brier score binomial outcomes).options \"var.e\" (variance prediction error), \"auc\" (area curve), \"cal_int\" (calibration intercept), \"cal_slope\" (multiplicative calibration slope) \"cal_add_slope\" (additive calibration slope), list , first used model selection. genFUN Function list named functions computing generalizability performance. Default: rema, summary statistic random effects meta-analysis. Choose \"rema_tau\" heterogeneity estimate random effects meta-analysis, genFUN=\"abs_mean\" (absolute) mean, coefficient_of_variation coefficient variation. list containing , first used model selection. selFUN Function selecting best method. Default: lowest value genFUN. set \".max\" high values genFUN indicate good model. gen..perf performance measures generalizability measures computed? \"first\" (default) first. \"respective\" matching generalizability measure performance measure location list. \"factorial\" applying generalizability measures performance estimates. ... pass arguments estFUN (e.g. family = \"binomial\"), FUNctions.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"list class metapred, containing final model global.model, stepwise tree estimates coefficients, performance measures, generalizability measures stepwise.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"Use subset.metapred obtain individual prediction model metapred object. Note formula.changes currently unordered; represent order changes stepwise procedure. metapred still development, use care.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"Debray TPA, Moons KGM, Ahmed , Koffijberg H, Riley RD. framework developing, implementing, evaluating clinical prediction models individual participant data meta-analysis. Stat Med. 2013;32(18):3158-80. de Jong VMT, Moons KGM, Eijkemans MJC, Riley RD, Debray TPA. Developing generalizable prediction models pooled studies large clustered data sets. Stat Med. 2021;40(15):3533--59. Riley RD, Tierney JF, Stewart LA. Individual participant data meta-analysis: handbook healthcare research. Hoboken, NJ: Wiley; 2021. ISBN: 978-1-119-33372-2. Schmid CH, Stijnen T, White IR. Handbook meta-analysis. First edition. Boca Raton: Taylor Francis; 2020. ISBN: 978-1-315-11940-3.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"","code":"data(DVTipd) if (FALSE) { # Explore heterogeneity in intercept and assocation of 'ddimdich' glmer(dvt ~ 0 + cluster + (ddimdich|study), family = binomial(), data = DVTipd) } # Scope f <- dvt ~ histdvt + ddimdich + sex + notraum # Internal-external cross-validation of a pre-specified model 'f' fit <- metapred(DVTipd, strata = \"study\", formula = f, scope = f, family = binomial) fit #> Call: metapred(data = DVTipd, strata = \"study\", formula = f, scope = f, #> family = binomial) #> #> Started with model: #> dvt ~ histdvt + ddimdich + sex + notraum #> #> #> Generalizability: #> unchanged #> 1 0.1484983 #> #> Cross-validation stopped after 0 steps, as no changes were requested. Final model: #> Meta-analytic model of prediction models estimated in 4 strata. Coefficients: #> (Intercept) histdvt ddimdich sex notraum #> -4.1180636 0.6174010 1.6962441 0.9647970 0.3761707 # Let's try to simplify model 'f' in order to improve its external validity metapred(DVTipd, strata = \"study\", formula = f, family = binomial) #> Call: metapred(data = DVTipd, strata = \"study\", formula = f, family = binomial) #> #> Started with model: #> dvt ~ histdvt + ddimdich + sex + notraum #> #> #> Generalizability: #> unchanged #> 1 0.1484983 #> #> Generalizability: #> ddimdich histdvt notraum sex #> 1 0.136086 0.1375105 0.12977 0.141173 #> #> Continued with model: #> dvt ~ histdvt + ddimdich + sex #> #> #> Generalizability: #> ddimdich histdvt sex #> 1 0.1366828 0.1279623 0.1319755 #> #> Continued with model: #> dvt ~ ddimdich + sex #> #> #> Generalizability: #> ddimdich sex #> 1 0.1355548 0.1303254 #> #> Cross-validation stopped after 3 steps, as no improvement was possible. Final model: #> Meta-analytic model of prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich sex #> -3.6187987 1.7130967 0.8784071 # We can also try to build a generalizable model from scratch if (FALSE) { # Some additional examples: metapred(DVTipd, strata = \"study\", formula = dvt ~ 1, scope = f, family = binomial) # Forwards metapred(DVTipd, strata = \"study\", formula = f, scope = f, family = binomial) # no selection metapred(DVTipd, strata = \"study\", formula = f, max.steps = 0, family = binomial) # no selection metapred(DVTipd, strata = \"study\", formula = f, recal.int = TRUE, family = binomial) metapred(DVTipd, strata = \"study\", formula = f, meta.method = \"REML\", family = binomial) } # By default, metapred assumes the first column is the outcome. newdat <- data.frame(dvt=0, histdvt=0, ddimdich=0, sex=1, notraum=0) fitted <- predict(fit, newdata = newdat)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate the total O:E ratio — oecalc","title":"Calculate the total O:E ratio — oecalc","text":"function calculates (transformed versions ) ratio total number observed versus expected events corresponding sampling variance.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate the total O:E ratio — oecalc","text":"","code":"oecalc( OE, OE.se, OE.cilb, OE.ciub, OE.cilv, EO, EO.se, citl, citl.se, N, O, E, Po, Po.se, Pe, data, slab, add = 1/2, g = NULL, level = 0.95, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate the total O:E ratio — oecalc","text":"OE vector estimated ratio total observed versus total expected events OE.se Optional vector standard errors estimated O:E ratios. OE.cilb Optional vector specify lower limits confidence interval OE. OE.ciub Optional vector specify upper limits confidence interval OE. OE.cilv Optional vector specify levels aformentioned confidence interval limits. (default: 0.95, corresponds 95% confidence interval). EO Optional vector estimated ratio total expected versus total observed events EO.se Optional vector standard errors estimated E:O ratios citl Optional vector estimated calibration---large statistics citl.se Optional vector standard error calibration---large statistics N Optional vector specify sample/group sizes. O Optional vector specify total number observed events. E Optional vector specify total number expected events Po Optional vector specify (cumulative) observed event probabilities. Po.se Optional vector standard errors Po. time--event data, also SE observed survival probabilities (e.g. obtained Kaplan-Meier analysis) Pe Optional vector specify (cumulative) expected event probabilites (specified, time t.val) data Optional data frame containing variables given arguments . slab Optional vector labels studies. add non-negative number indicating amount add zero counts. See `Details' g quoted string function transform estimates total O:E ratio; see details . level level confidence interval, default 0.95. ... Additional arguments.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate the total O:E ratio — oecalc","text":"object class c(\"mm_perf\",\"data.frame\") following columns: \"theta\" (transformed) O:E ratio. \"theta.se\" Standard errors (transformed) O:E ratio. \"theta.cilb\" Lower confidence interval (transformed) O:E ratios. level specified level. Intervals calculated scale theta assuming Normal distribution. \"theta.ciub\" Upper confidence interval (transformed) c-statistics. level specified level. Intervals calculated scale theta assuming Normal distribution. \"theta.source\" Method used calculating (transformed) O:E ratio. \"theta.se.source\" Method used calculating standard error (transformed) O:E ratio.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Calculate the total O:E ratio — oecalc","text":"Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2019 Sep;28(9):2768--86. Snell KI, Ensor J, Debray TP, Moons KG, Riley RD. Meta-analysis prediction model performance across multiple studies: scale helps ensure -study normality C -statistic calibration measures? Stat Methods Med Res. 2017.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Calculate the total O:E ratio — oecalc","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Calculate the total O:E ratio — oecalc","text":"","code":"######### Validation of prediction models with a binary outcome ######### data(EuroSCORE) # Calculate the total O:E ratio and its standard error est1 <- oecalc(O = n.events, E = e.events, N = n, data = EuroSCORE, slab = Study) est1 #> theta theta.se theta.cilb theta.ciub theta.source O #> Nashef 1.0450450 0.06716203 0.9134099 1.1766802 O, E and N 232 #> Biancari 0.6086957 0.11345371 0.3863305 0.8310608 O, E and N 28 #> Di Dedda 1.2058824 0.18475130 0.8437765 1.5679882 O, E and N 41 #> Chalmers 0.7318008 0.05203645 0.6298112 0.8337903 O, E and N 191 #> Grant 0.9214541 0.03320253 0.8563783 0.9865298 O, E and N 746 #> Carneo 1.2573099 0.08328543 1.0940735 1.4205464 O, E and N 215 #> Kunt 4.8571429 0.79922240 3.2906957 6.4235900 O, E and N 34 #> Kirmani 1.4134367 0.05935803 1.2970971 1.5297763 O, E and N 547 #> Howell 0.8571429 0.08588255 0.6888162 1.0254696 O, E and N 90 #> Wang 0.7793103 0.05131185 0.6787410 0.8798797 O, E and N 226 #> Borde 0.8000000 0.28056169 0.2501092 1.3498908 O, E and N 8 #> Qadir 1.0270270 0.11555260 0.8005481 1.2535060 O, E and N 76 #> Spiliopoulos 1.5555556 0.40204098 0.7675697 2.3435414 O, E and N 14 #> Wendt 1.3235294 0.19309081 0.9450784 1.7019804 O, E and N 45 #> Laurent 2.5714286 0.58846311 1.4180621 3.7247951 O, E and N 18 #> Wang.1 0.6190476 0.17032315 0.2852204 0.9528749 O, E and N 13 #> Nishida 0.9705882 0.16279815 0.6515097 1.2896668 O, E and N 33 #> Barilli 0.6885246 0.04710205 0.5962063 0.7808429 O, E and N 210 #> Barilli.1 1.2019231 0.10340170 0.9992595 1.4045867 O, E and N 125 #> Paparella 1.1029412 0.06211634 0.9811954 1.2246870 O, E and N 300 #> Carosella 2.2500000 0.73637626 0.8067290 3.6932710 O, E and N 9 #> Borracci 1.3125000 0.28036848 0.7629879 1.8620121 O, E and N 21 #> Osnabrugge 0.6830357 0.02064915 0.6425641 0.7235073 O, E and N 1071 #> E N #> Nashef 222.00 5553 #> Biancari 46.00 1027 #> Di Dedda 34.00 1090 #> Chalmers 261.00 5576 #> Grant 809.59 23740 #> Carneo 171.00 3798 #> Kunt 7.00 428 #> Kirmani 387.00 15497 #> Howell 105.00 933 #> Wang 290.00 11170 #> Borde 10.00 498 #> Qadir 74.00 2004 #> Spiliopoulos 9.00 216 #> Wendt 34.00 1066 #> Laurent 7.00 314 #> Wang.1 21.00 818 #> Nishida 34.00 461 #> Barilli 305.00 12201 #> Barilli.1 104.00 1670 #> Paparella 272.00 6191 #> Carosella 4.00 250 #> Borracci 16.00 503 #> Osnabrugge 1568.00 50588 # Calculate the log of the total O:E ratio and its standard error est2 <- oecalc(O = n.events, E = e.events, N = n, data = EuroSCORE, slab = Study, g = \"log(OE)\") est2 #> theta theta.se theta.cilb theta.ciub theta.source O #> Nashef 0.04405999 0.06426711 -0.081901240 0.17002122 O, E and N 232 #> Biancari -0.49643689 0.18638824 -0.861751123 -0.13112265 O, E and N 28 #> Di Dedda 0.18721154 0.15320840 -0.113071397 0.48749448 O, E and N 41 #> Chalmers -0.31224698 0.07110740 -0.451614919 -0.17287904 O, E and N 191 #> Grant -0.08180235 0.03603276 -0.152425252 -0.01117944 O, E and N 746 #> Carneo 0.22897447 0.06624097 0.099144553 0.35880439 O, E and N 215 #> Kunt 1.58045038 0.16454579 1.257946559 1.90295419 O, E and N 34 #> Kirmani 0.34602411 0.04199553 0.263714374 0.42833385 O, E and N 547 #> Howell -0.15415068 0.10019631 -0.350531831 0.04223047 O, E and N 90 #> Wang -0.24934592 0.06584264 -0.378395127 -0.12029672 O, E and N 226 #> Borde -0.22314355 0.35070211 -0.910507050 0.46421995 O, E and N 8 #> Qadir 0.02666825 0.11251174 -0.193850721 0.24718721 O, E and N 76 #> Spiliopoulos 0.44183275 0.25845491 -0.064729568 0.94839507 O, E and N 14 #> Wendt 0.28030197 0.14589084 -0.005638818 0.56624275 O, E and N 45 #> Laurent 0.94446161 0.22884677 0.495930190 1.39299303 O, E and N 18 #> Wang.1 -0.47957308 0.27513739 -1.018832454 0.05968629 O, E and N 13 #> Nishida -0.02985296 0.16773143 -0.358600527 0.29889460 O, E and N 33 #> Barilli -0.37320425 0.06841012 -0.507285614 -0.23912288 O, E and N 210 #> Barilli.1 0.18392284 0.08603021 0.015306718 0.35253896 O, E and N 125 #> Paparella 0.09798041 0.05631881 -0.012402434 0.20836325 O, E and N 300 #> Carosella 0.81093022 0.32727834 0.169476459 1.45238397 O, E and N 9 #> Borracci 0.27193372 0.21361408 -0.146742192 0.69060962 O, E and N 21 #> Osnabrugge -0.38120813 0.03023143 -0.440460642 -0.32195562 O, E and N 1071 #> E N #> Nashef 222.00 5553 #> Biancari 46.00 1027 #> Di Dedda 34.00 1090 #> Chalmers 261.00 5576 #> Grant 809.59 23740 #> Carneo 171.00 3798 #> Kunt 7.00 428 #> Kirmani 387.00 15497 #> Howell 105.00 933 #> Wang 290.00 11170 #> Borde 10.00 498 #> Qadir 74.00 2004 #> Spiliopoulos 9.00 216 #> Wendt 34.00 1066 #> Laurent 7.00 314 #> Wang.1 21.00 818 #> Nishida 34.00 461 #> Barilli 305.00 12201 #> Barilli.1 104.00 1670 #> Paparella 272.00 6191 #> Carosella 4.00 250 #> Borracci 16.00 503 #> Osnabrugge 1568.00 50588 # Display the results of all studies in a forest plot plot(est1)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/perf.html","id":null,"dir":"Reference","previous_headings":"","what":"Performance estimates — perf","title":"Performance estimates — perf","text":"Obtain performance estimates model fit.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/perf.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Performance estimates — perf","text":"","code":"perf(object, ...) performance(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/perf.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Performance estimates — perf","text":"object model fit object, either metapred subset(metapred) object. ... default, final model selected. parameter allows arguments passed subset.metapred performance estimates steps/models may returned. Use perfFUN = 0 select .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/perf.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Performance estimates — perf","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":null,"dir":"Reference","previous_headings":"","what":"Display results from the funnel plot asymmetry test — plot.fat","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"Generates funnel plot fitted fat object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"","code":"# S3 method for fat plot( x, ref, confint = TRUE, confint.level = 0.1, confint.col = \"skyblue\", confint.alpha = 0.5, confint.density = NULL, xlab = \"Effect size\", add.pval = TRUE, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"x object class fat ref numeric value indicating fixed random effects summary estimate. value provided retrieved fixed effects meta-analysis (possible). confint logical indicator. TRUE, confidence interval displayed estimated regression model (based Student-T distribution) confint.level Significance level constructing confidence interval. confint.col color filling confidence interval. Choose NA leave polygons unfilled. confint.density specified positive value gives color shading lines. confint.alpha numeric value 0 1 indicating opacity confidence region. confint.density density shading lines, lines per inch. default value NULL means shading lines drawn. zero value density means shading filling whereas negative values NA suppress shading (allow color filling). xlab title x axis add.pval Logical indicate whether P-value added plot ... Additional arguments.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"Thomas Debray Frantisek Bartos ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"","code":"data(Fibrinogen) b <- log(Fibrinogen$HR) b.se <- ((log(Fibrinogen$HR.975) - log(Fibrinogen$HR.025))/(2*qnorm(0.975))) n.total <- Fibrinogen$N.total # A very simple funnel plot plot(fat(b=b, b.se=b.se), xlab = \"Log hazard ratio\") # Plot the funnel for an alternative test plot(fat(b=b, b.se=b.se, n.total=n.total, method=\"M-FIV\"), xlab = \"Log hazard ratio\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest Plots — plot.mm_perf","title":"Forest Plots — plot.mm_perf","text":"Function create forest plots objects class \"mm_perf\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest Plots — plot.mm_perf","text":"","code":"# S3 method for mm_perf plot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest Plots — plot.mm_perf","text":"x object class \"mm_perf\" ... Additional arguments passed forest.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Forest Plots — plot.mm_perf","text":"object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Forest Plots — plot.mm_perf","text":"forest plot shows performance estimates study corresponding confidence intervals.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Forest Plots — plot.mm_perf","text":"Lewis S, Clarke M. Forest plots: trying see wood trees. BMJ. 2001; 322(7300):1479--80.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest Plots — plot.mm_perf","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Forest Plots — plot.mm_perf","text":"","code":"data(EuroSCORE) # Calculate the c-statistic and its standard error est1 <- ccalc(cstat = c.index, cstat.se = se.c.index, cstat.cilb = c.index.95CIl, cstat.ciub = c.index.95CIu, N = n, O = n.events, data = EuroSCORE, slab = Study) plot(est1) # Calculate the total O:E ratio and its standard error est2 <- oecalc(O = n.events, E = e.events, N = n, data = EuroSCORE, slab = Study) plot(est2)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"Generates forest plot outcome bivariate meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"","code":"# S3 method for riley plot(x, title, sort = \"asc\", xlim, refline, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"x object class riley title Title forest plot sort default, studies ordered ascending effect size (sort=\"asc\"). study ordering descending effect size, choose sort=\"desc\". value, study ordering ignored. xlim x limits (x1, x2) forest plot refline Optional numeric specifying reference line ... Additional parameters generating forest plots","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"Riley RD, Thompson JR, Abrams KR. alternative model bivariate random-effects meta-analysis within-study correlations unknown. Biostatistics 2008; 9: 172--186.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"","code":"data(Scheidler) #Perform the analysis fit <- riley(Scheidler[which(Scheidler$modality==1),]) plot(fit) require(ggplot2) plot(fit, sort=\"none\", theme=theme_gray())"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest Plots — plot.uvmeta","title":"Forest Plots — plot.uvmeta","text":"Function create forest plots objects class \"uvmeta\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest Plots — plot.uvmeta","text":"","code":"# S3 method for uvmeta plot(x, sort = \"asc\", ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest Plots — plot.uvmeta","text":"x object class \"uvmeta\" sort default, studies ordered ascending effect size (sort=\"asc\"). study ordering descending effect size, choose sort=\"desc\". value, study ordering ignored. ... Additional arguments passed forest.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Forest Plots — plot.uvmeta","text":"forest plot shows performance estimates validation corresponding confidence intervals. polygon added bottom forest plot, showing summary estimate based model. 95% prediction interval added default random-effects models, dotted line indicates (approximate) bounds","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Forest Plots — plot.uvmeta","text":"Full lines indicate confidence intervals credibility intervals (case Bayesian meta-analysis). Dashed lines indicate prediction intervals. width intervals defined significance level chosen meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Forest Plots — plot.uvmeta","text":"Lewis S, Clarke M. Forest plots: trying see wood trees. BMJ. 2001; 322(7300):1479--80. Riley RD, Higgins JPT, Deeks JJ. Interpretation random effects meta-analyses. BMJ. 2011 342:d549--d549.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest Plots — plot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Forest Plots — plot.uvmeta","text":"","code":"data(Roberts) # Frequentist random-effects meta-analysis fit <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts))) plot(fit)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest Plots — plot.valmeta","title":"Forest Plots — plot.valmeta","text":"Function create forest plots objects class \"valmeta\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest Plots — plot.valmeta","text":"","code":"# S3 method for valmeta plot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest Plots — plot.valmeta","text":"x object class \"valmeta\" ... Additional arguments passed forest.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Forest Plots — plot.valmeta","text":"object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Forest Plots — plot.valmeta","text":"forest plot shows performance estimates validation corresponding confidence intervals. polygon added bottom forest plot, showing summary estimate based model. 95% prediction interval added default random-effects models, dotted line indicates (approximate) bounds.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Forest Plots — plot.valmeta","text":"Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. Lewis S, Clarke M. Forest plots: trying see wood trees. BMJ. 2001; 322(7300):1479--80. Riley RD, Higgins JPT, Deeks JJ. Interpretation random effects meta-analyses. BMJ. 2011 342:d549--d549.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest Plots — plot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Forest Plots — plot.valmeta","text":"","code":"data(EuroSCORE) fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE) plot(fit) library(ggplot2) plot(fit, theme=theme_grey())"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Prediction Interval — predict.riley","title":"Prediction Interval — predict.riley","text":"Calculates prediction interval summary parameters Riley's alternative model bivariate random-effects meta-analysis. interval predicts range future observations fall given already observed.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Prediction Interval — predict.riley","text":"","code":"# S3 method for riley predict(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Prediction Interval — predict.riley","text":"object riley object. ... Additional arguments (currently ignored)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Prediction Interval — predict.riley","text":"Data frame containing prediction intervals summary estimates beta1 beta2 (effect size data), mean sensitivity false positive rate (diagnostic test accuracy data).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Prediction Interval — predict.riley","text":"Prediction intervals based Student's t-distribution (numstudies - 5) degrees freedom. width interval specified significance level chosen meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Prediction Interval — predict.riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":null,"dir":"Reference","previous_headings":"","what":"Recalibrate a Prediction Model — recalibrate","title":"Recalibrate a Prediction Model — recalibrate","text":"recalibrate used recalibrate prediction model classes metapred, glm lm.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Recalibrate a Prediction Model — recalibrate","text":"","code":"recalibrate(object, newdata, f = ~1, estFUN = NULL, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Recalibrate a Prediction Model — recalibrate","text":"object model fit object recalibrated, class metapred, glm lm, . newdata data.frame containing new data set updating. f formula. coefficients model updated? Default: intercept . Left-hand side may left . See formula details. estFUN Function model estimation. left NULL, function automatically retrieved metapred objects. objects, function name corresponding first class object taken. E.g. glm() glm objects. ... Optional arguments pass estFUN.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Recalibrate a Prediction Model — recalibrate","text":"Recalibrated model fit object, class object. Generally, updated coefficients can retrieved coef().","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Recalibrate a Prediction Model — recalibrate","text":"Currently coefficients updated variances aspects left untouched. updating entire model statistics, see update.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Recalibrate a Prediction Model — recalibrate","text":"","code":"data(DVTipd) DVTipd$cluster <- 1:4 # Add a fictional clustering to the data set. # Suppose we estimated the model in three studies: DVTipd123 <- DVTipd[DVTipd$cluster <= 3, ] mp <- metamisc:::metapred(DVTipd123, strata = \"cluster\", f = dvt ~ vein + malign, family = binomial) # and now want to recalibrate it for the fourth: DVTipd4 <- DVTipd[DVTipd$cluster == 4, ] metamisc:::recalibrate(mp, newdata = DVTipd4) #> Call: metamisc:::recalibrate(object = mp, newdata = DVTipd4) #> #> Started with model: #> dvt ~ vein + malign #> #> #> Generalizability: #> unchanged #> 1 0.1332079 #> #> Generalizability: #> malign vein #> 1 0.1344432 0.1314326 #> #> Continued with model: #> dvt ~ malign #> #> #> Generalizability: #> malign #> 1 0.1342213 #> #> Cross-validation stopped after 2 steps, as no improvement was possible. Final model: #> Meta-analytic model of prediction models estimated in 3 strata. Coefficients: #> (Intercept) malign #> -1.775619 1.145096"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Fit the alternative model for bivariate random-effects meta-analysis — riley","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"function fits alternative model bivariate random-effects meta-analysis within-study correlations unknown. bivariate model proposed Riley et al. (2008) similar general bivariate random-effects model (van Houwelingen et al. 2002), includes overall correlation parameter rather separating (usually unknown) within- -study correlation. consequence, alternative model fully hierarchical, estimates additional variation beyond sampling error (psi) directly equivalent -study variation (tau) general model. model particularly useful large within-study variability, primary studies available general model estimates -study correlation 1 -1.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"","code":"riley(X, slab, optimization = \"Nelder-Mead\", control = list(), pars, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"X data frame containing integer variables Y1, vars1, Y2 vars2, columns Y1 Y2 represent effect sizes outcome 1 , respectively, outcome 2. columns vars1 vars2 represent error variances Y1 , respectively, Y2. Alternatively, considering meta-analysis diagnostic test accuracy data, columns TP, FN, FP TN may specified. Corresponding values represent number true positives, number false negatives, number false positives , respectively, number true negatives. slab Optional vector specifying label study optimization optimization method used minimizing negative (restricted) log-likelihood function. default method implementation Nelder Mead (1965), uses function values robust relatively slow. methods described optim. control list control parameters pass optim. pars List additional arguments. width confidence, credibility prediction intervals defined level (defaults 0.95). ... Arguments passed functions. See \"Details\" information.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"object class riley many standard methods available. warning message casted Hessian matrix contains negative eigenvalues, implies identified solution saddle point thus optimal.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"Parameters estimated iteratively maximizing restriced log-likelihood using Newton-Raphson procedure. results univariate random-effects meta-analysis method--moments estimator used starting values beta1, beta2, psi1 psi2 optim command. Standard errors parameters obtained inverse Hessian matrix.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"meta-analysis-of-effect-sizes","dir":"Reference","previous_headings":"","what":"Meta-analysis of effect sizes","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"following parameters estimated iteratively maximizing restriced log-likelihood using Newton-Raphson procedure: pooled effect size outcome 1 (beta1), pooled effect size outcome 2 (beta2), additional variation beta1 beyond sampling error (psi1), additional variation beta2 beyond sampling error (psi2) correlation rho psi1 psi2.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"meta-analysis-of-diagnostic-test-accuracy","dir":"Reference","previous_headings":"","what":"Meta-analysis of diagnostic test accuracy","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"Although model can also used diagnostic test accuracy data substantial within-study correlations expected, assuming zero within-study correlations (.e. applying Reitsma's approach) usually justified (Reitsma et al. 2005, Daniels Hughes 1997, Korn et al. 2005, Thompson et al. 2005, Van Houwelingen et al. 2002). logit transformation applied sensitivities ans false positive rates study, order meet normality assumptions. zero cell counts occur, continuity corrections may required. correction value can specified using correction (defaults 0.5). , argument correction.control set \"\" (default) continuity correction added whole data one cell one study zero. correction.control=\"single\" correction applied rows data zero. following parameters estimated: logit sensitivity (beta1), logit false positive rate (beta2), additional variation beta1 beyond sampling error (psi1), additional variation beta2 beyond sampling error (psi2) correlation (rho) psi1 psi2.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"overall correlation parameter rho transformed estimation ensure corresponding values remain -1 1. transformed correlation rhoT given logit((rho+1)/2). optimization, starting value rhoT set 0. standard error rho derived rhoT using Delta method. Similarly, Delta methods used derive standard error sensitivity false positive rate beta1 , respectively, beta2. Algorithms dealing missing data currently implemented, Bayesian approaches become available later versions.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"Korn EL, Albert PS, McShane LM. Assessing surrogates trial endpoints using mixed models. Statistics Medicine 2005; 24: 163--182. Nelder JA, Mead R. simplex algorithm function minimization. Computer Journal (1965); 7: 308--313. Reitsma J, Glas , Rutjes , Scholten R, Bossuyt P, Zwinderman . Bivariate analysis sensitivity specificity produces informative summary measures diagnostic reviews. Journal Clinical Epidemiology 2005; 58: 982--990. Riley RD, Thompson JR, Abrams KR. alternative model bivariate random-effects meta-analysis within-study correlations unknown. Biostatistics 2008; 9: 172--186. Thompson JR, Minelli C, Abrams KR, Tobin MD, Riley RD. Meta-analysis genetic studies using mendelian randomization--multivariate approach. Statistics Medicine 2005; 24: 2241--2254. van Houwelingen HC, Arends LR, Stijnen T. Advanced methods meta-analysis: multivariate approach meta-regression. Statistics Medicine 2002; 21: 589--624.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"","code":"data(Scheidler) data(Daniels) data(Kertai) # Meta-analysis of potential surrogate markers data # The results obtained by Riley (2008) were as follows: # beta1 = -0.042 (SE = 0.063), # beta2 = 14.072 (SE = 4.871) # rho = -0.759 if (FALSE) { fit1 <- riley(Daniels) #maxit reached, try again with more iterations } fit1 <- riley(Daniels, control=list(maxit=10000)) summary(fit1) #> $call #> riley.default(X = Daniels, control = list(maxit = 10000)) #> #> $confints #> Estimate SE 2.5 % 97.5 % #> beta1 0.005298983 0.06479973 -0.12170616 0.1323041 #> beta2 13.505678310 4.99256719 3.72042644 23.2909302 #> psi1 0.134785102 0.09190903 -0.04535329 0.3149235 #> psi2 18.076027226 4.00992798 10.21671280 25.9353417 #> rho -0.748689375 0.15266774 -0.89430728 -0.4596724 #> #> attr(,\"class\") #> [1] \"summary.riley\" # Meta-analysis of prognostic test studies fit2 <- riley(Kertai) fit2 #> Call: #> riley.default(X = Kertai) #> #> Coefficients #> beta1 beta2 psi1 psi2 rho #> 0.8164679 -0.9715821 0.3499043 0.7692122 0.1537878 #> #> Degrees of Freedom: 9 Residual # Meta-analysis of computed tomography data ds <- Scheidler[which(Scheidler$modality==1),] fit3 <- riley(ds) fit3 #> Call: #> riley.default(X = ds) #> #> Coefficients #> beta1 beta2 psi1 psi2 rho #> -0.01731291 -2.32166611 0.71181410 0.38103153 0.70119871 #> #> Degrees of Freedom: 29 Residual"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the running means of a Bayesian meta-analysis model — rmplot","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"Function display running means fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"","code":"rmplot(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"generic function.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"Function display running means fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"","code":"# S3 method for mcmc.list rmplot(x, P, greek = FALSE, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"x object class \"mcmc.list\" P Optional dataframe describing parameters plot respective names greek Logical value indicating whether parameter labels parsed get Greek letters. Defaults FALSE. ... Additional arguments passed ggs_running","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"Function display running means fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"","code":"# S3 method for uvmeta rmplot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"x object class \"valmeta\" ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"Results displayed estimated mean (mu) standard-deviation (tau) meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"","code":"if (FALSE) { data(Roberts) fit <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts), method=\"BAYES\")) rmplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"Function display running means fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"","code":"# S3 method for valmeta rmplot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"x object class \"valmeta\" ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"Results displayed estimated mean (mu) standard-deviation (tau) meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"","code":"if (FALSE) { data(EuroSCORE) fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) rmplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":null,"dir":"Reference","previous_headings":"","what":"Standard errors and variances — se","title":"Standard errors and variances — se","text":"Obtain standard errors variances model fit","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Standard errors and variances — se","text":"","code":"se(object, ...) variances(object, ...) tau(object, ...) tau2(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Standard errors and variances — se","text":"object model fit object ... arguments","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Standard errors and variances — se","text":"se standard errors object, variances variances. tau heterogeneity coefficients.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Standard errors and variances — se","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/stackedglm.html","id":null,"dir":"Reference","previous_headings":"","what":"Stacked Regression — stackedglm","title":"Stacked Regression — stackedglm","text":"function combines one existing prediction models /called meta-model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/stackedglm.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Stacked Regression — stackedglm","text":"","code":"stackedglm(models, family = binomial, data)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/stackedglm.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Stacked Regression — stackedglm","text":"models list containing historical prediction models, can defined several ways. instance, historical regression models can specified using named vector containing regression coefficients individual predictors (need include intercept term). List items may also represent object function predict() exists. family description error distribution link function used meta-model. can character string naming family function, family function result call family function. (See family details family functions.) data optional data frame, list environment (object coercible .data.frame data frame) containing variables model. found data, variables taken environment(formula), typically environment stackedglm called.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/stackedglm.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Stacked Regression — stackedglm","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":null,"dir":"Reference","previous_headings":"","what":"Subsetting metapred fits — subset.metapred","title":"Subsetting metapred fits — subset.metapred","text":"Return model cross-validation procedure final 'global' model. Caution: function still development.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Subsetting metapred fits — subset.metapred","text":"","code":"# S3 method for metapred subset( x, select = \"cv\", step = NULL, model = NULL, stratum = NULL, add = TRUE, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Subsetting metapred fits — subset.metapred","text":"x metapred object select type model select: \"cv\" (default), \"global\", (experimental) \"stratified\", \"stratum\". step step selected? Defaults best step. numeric converted name step: 0 unchanged model, 1 first change... model model change selected? NULL (default, best change) character name variable (integer) index model change. stratum Experimental. Stratum return select = \"stratum\". add Logical. Add data, options functions resulting object? Defaults TRUE. Experimental. ... compatibility .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Subsetting metapred fits — subset.metapred","text":"object class mp.cv select = \"cv\" object class mp.global select = \"global\". cases, additional data added resulting object, thereby making suitable methods.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Subsetting metapred fits — subset.metapred","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Subsetting metapred fits — subset.metapred","text":"","code":"data(DVTipd) DVTipd$cluster <- letters[1:4] # Add a fictional clustering to the data. mp <- metapred(DVTipd, strata = \"cluster\", formula = dvt ~ histdvt + ddimdich, family = binomial) subset(mp) # best cross-validated model #> Prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich #> a -3.198673 2.135779 #> b -3.555348 2.251292 #> c -19.566069 18.540216 #> d -3.891820 2.947359 #> #> Meta-analytic models, estimated in 4 fold combinations. Coefficients: #> (Intercept) ddimdich #> b, c, d -3.724268 2.600774 #> a, c, d -3.432711 2.421694 #> a, b, d -3.463478 2.377572 #> a, b, c -3.318460 2.176257 #> #> Cross-validation at stratum level yields the following performance: #> val.strata estimate se var ci.lb ci.ub #> b, c, d a 0.1285223 0.01986149 0.0003944788 0.08921088 0.1678338 #> a, c, d b 0.1293544 0.01729276 0.0002990394 0.09512723 0.1635816 #> a, b, d c 0.1123561 0.01717570 0.0002950048 0.07836058 0.1463516 #> a, b, c d 0.1297595 0.01948263 0.0003795729 0.09119796 0.1683211 #> measure #> b, c, d mse #> a, c, d mse #> a, b, d mse #> a, b, c mse #> #> Generalizability: #> 1 #> 0.1244262 #> subset(mp, select = \"global\") # Final model fitted on all strata. #> Meta-analytic model of prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich #> -3.463480 2.377574 subset(mp, step = 1) # The best model of step 1 #> Prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich #> a -3.198673 2.135779 #> b -3.555348 2.251292 #> c -19.566069 18.540216 #> d -3.891820 2.947359 #> #> Meta-analytic models, estimated in 4 fold combinations. Coefficients: #> (Intercept) ddimdich #> b, c, d -3.724268 2.600774 #> a, c, d -3.432711 2.421694 #> a, b, d -3.463478 2.377572 #> a, b, c -3.318460 2.176257 #> #> Cross-validation at stratum level yields the following performance: #> val.strata estimate se var ci.lb ci.ub #> b, c, d a 0.1285223 0.01986149 0.0003944788 0.08921088 0.1678338 #> a, c, d b 0.1293544 0.01729276 0.0002990394 0.09512723 0.1635816 #> a, b, d c 0.1123561 0.01717570 0.0002950048 0.07836058 0.1463516 #> a, b, c d 0.1297595 0.01948263 0.0003795729 0.09119796 0.1683211 #> measure #> b, c, d mse #> a, c, d mse #> a, b, d mse #> a, b, c mse #> #> Generalizability: #> 1 #> 0.1244262 #> subset(mp, step = 1, model = \"histdvt\") # The model in which histdvt was removed, in step 1. #> Prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich #> a -3.198673 2.135779 #> b -3.555348 2.251292 #> c -19.566069 18.540216 #> d -3.891820 2.947359 #> #> Meta-analytic models, estimated in 4 fold combinations. Coefficients: #> (Intercept) ddimdich #> b, c, d -3.724268 2.600774 #> a, c, d -3.432711 2.421694 #> a, b, d -3.463478 2.377572 #> a, b, c -3.318460 2.176257 #> #> Cross-validation at stratum level yields the following performance: #> val.strata estimate se var ci.lb ci.ub #> b, c, d a 0.1285223 0.01986149 0.0003944788 0.08921088 0.1678338 #> a, c, d b 0.1293544 0.01729276 0.0002990394 0.09512723 0.1635816 #> a, b, d c 0.1123561 0.01717570 0.0002950048 0.07836058 0.1463516 #> a, b, c d 0.1297595 0.01948263 0.0003795729 0.09119796 0.1683211 #> measure #> b, c, d mse #> a, c, d mse #> a, b, d mse #> a, b, c mse #> #> Generalizability: #> 1 #> 0.1244262 #>"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"Parameter summaries Provides summary estimates alternative model bivariate random-effects meta-analysis Riley et al. (2008) corresponding standard errors (derived inverse Hessian). confidence intervals, asymptotic normality assumed.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"","code":"# S3 method for riley summary(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"object riley object ... Arguments passed functions (currently ignored)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"array confidence intervals estimated model parameters. diagnostic test accuracy data, resulting summary sensitivity false positive rate included.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"meta-analysis diagnostic test accuracy data, beta1 equals logit sensitivity (Sens) beta2 equals logit false positive rate (FPR).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"overall correlation (rho) confidence intervals derived using transformation logit((rho+1)/2). Similarly, logit transformation used derive confidence intervals summary sensitivity false positive rate.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"Riley RD, Thompson JR, Abrams KR. alternative model bivariate random-effects meta-analysis within-study correlations unknown. Biostatistics 2008; 9: 172--186.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"function provides summary estimates fitted univariate meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"","code":"# S3 method for uvmeta summary(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"object object class \"uvmeta\" ... Optional arguments passed functions","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. basic introduction fixed-effect random-effects models meta-analysis. Research Synthesis Methods 2010; 1: 97--111. DerSimonian R, Laird N. Meta-analysis clinical trials. Controlled Clinical Trials 1986; 7: 177--188. Riley RD, Higgins JPT, Deeks JJ. Interpretation random effects meta-analyses. British Medical Journal 2011; 342: d549.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":null,"dir":"Reference","previous_headings":"","what":"Class ","title":"Class ","text":"class encapsulates results univariate meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":"objects-from-the-class","dir":"Reference","previous_headings":"","what":"Objects from the Class","title":"Class ","text":"Objects can created calls form uvmeta.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":"slots","dir":"Reference","previous_headings":"","what":"Slots","title":"Class ","text":"call: (language) call uvmeta. data: (data frame) data used meta-analysis. results: (data frame) Contains pooled effect size (mu), -study variability (tausq), Cochran's Q statistic (Q) Higgins' Thompson's square statistic (Isq). estimate, error variances provided predefined confidence (method=\"MOM\") credibility (method=\"bayes\") intervals. model: (character) meta-analysis model used. method: (character) estimator used. na.action: (character) Information action applied object NAs handled specially, NULL. df: (numeric) Degrees freedom. numstudies: (numeric) amount studies used meta-analysis. pred.int: (data frame) prediction interval, predicting range future effect sizes fall given already observed (based Student's t-distribution, cfr. Riley 2011) formula: (character) formula specified, character vector giving formula parameter specifications.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":"methods","dir":"Reference","previous_headings":"","what":"Methods","title":"Class ","text":"print signature(object = \"uvmeta\"): Print object summary. forest signature(object = \"uvmeta\"): Plot forest plot summary estimate. summary signature(object = \"uvmeta\"): Generate object summary.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Class ","text":"","code":"data(Collins) #Extract effect size and error variance r <- Collins$logOR vars <- Collins$SE**2 #Frequentist random-effects meta-analysis fit1 <- uvmeta(r,vars) #Extract results fit1$results #> NULL"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Univariate meta-analysis — uvmeta","title":"Univariate meta-analysis — uvmeta","text":"function summarizes multiple estimates single parameter assuming fixed (.e. common) effect random effects across studies. summary estimate obtained calculating weighted mean accounts sample size (case random effects assumed) -study heterogeneity.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Univariate meta-analysis — uvmeta","text":"","code":"uvmeta( r, r.se, r.vi, method = \"REML\", test = \"knha\", labels, na.action, n.chains = 4, pars, verbose = FALSE, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Univariate meta-analysis — uvmeta","text":"r Vector numerics containing effect size study r.se Vector numerics containing standard error effect sizes r.vi Vector numerics containing sampling variance effect sizes method Character string specifying whether fixed-effect random-effects model fitted. fixed-effect model fitted using method=\"FE\". Random-effects models fitted setting method equal one following: \"REML\" (Default), \"DL\", \"\", \"SJ\", \"ML\", \"EB\", \"HS\", \"GENQ\" \"BAYES\". See 'Details'. test Optional character string method!=\"BAYES\" specify test statistics confidence intervals fixed effects computed. default (test=\"knha\"), method Knapp Hartung (2003) used adjusting test statistics confidence intervals. Type '?rma' details. labels Optional vector characters containing labels studies na.action function indicates happen data contain NAs. Defaults \"na.fail\", options \"na.omit\", \"na.exclude\" \"na.pass\". n.chains Optional numeric specifying number chains use Gibbs sampler (method=\"BAYES\"). chains improve sensitivity convergence diagnostic, cause simulation run slowly. default number chains 4. pars Optional list additional arguments. width confidence, credibility prediction intervals defined level (defaults 0.95). following parameters configure MCMC sampling procedure: hp.mu.mean (mean prior distribution random effects model, defaults 0), hp.mu.var (variance prior distribution random effects model, defaults 1E6), hp.tau.min (minimum value -study standard deviation, defaults 0), hp.tau.max (maximum value -study standard deviation, defaults 2), hp.tau.sigma (standard deviation prior distribution -study standard-deviation), hp.tau.dist (prior distribution -study standard-deviation. Defaults \"dunif\"), hp.tau.df (degrees freedom prior distribution -study standard-deviation. Defaults 3). verbose TRUE messages generated fitting process displayed. ... Additional arguments passed rma runjags (method=\"BAYES\").","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Univariate meta-analysis — uvmeta","text":"object class uvmeta many standard methods available. \"data\" array (transformed) data used meta-analysis, method(s) used restoring missing information. \"method\" character string specifying meta-analysis method. \"est\" estimated performance statistic model. Bayesian meta-analysis, posterior median returned. \"se\" standard error (posterior standard deviation) summary estimate. \"tau2\" estimated amount (residual) heterogeneity. Always 0 method=\"FE\". Bayesian meta-analysis, posterior median returned. \"se.tau2\" estimated standard error (posterior standard deviation) -study variation. \"ci.lb\" lower bound confidence (credibility) interval summary estimate \"ci.ub\" upper bound confidence (credibility) interval summary estimate \"pi.lb\" lower bound (approximate) prediction interval summary estimate \"pi.ub\" upper bound (approximate) prediction interval summary estimate \"fit\" full results fitted model \"slab\" vector specifying label study.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Univariate meta-analysis — uvmeta","text":"Unless specified otherwise, meta-analysis models assume random effects fitted using restricted maximum likelihood estimation metafor package (Viechtbauer 2010). , confidence intervals average performance based Hartung-Knapp-Sidik-Jonkman method, better account uncertainty estimated -study heterogeneity (Debray 2016). Bayesian meta-analysis can performed specifying method=\"BAYES\". case, R packages runjags rjags must installed.] random-effects models, prediction interval pooled effect size displayed. interval predicts range future effect sizes fall given already observed (Higgins 2009, Riley 2011).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"bayesian-meta-analysis-models","dir":"Reference","previous_headings":"","what":"Bayesian meta-analysis models","title":"Univariate meta-analysis — uvmeta","text":"Bayesian meta-analysis models involve Gibbs sampler (method=\"BAYES\"), R packages runjags rjags must installed. Bayesian approach uses uninformative Normal prior mean uniform prior -study variance pooled effect size (Higgins 2009). default, Normal prior mean 0 variance 1000. hyperparameters can, however, altered variables hp.mu.mean hp.mu.var argument pars. prior distribution -study standard deviation given uniform distribution, default bounded 0 100.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Univariate meta-analysis — uvmeta","text":"Biggerstaff BJ, Tweedie RL. Incorporating variability estimates heterogeneity random effects model meta-analysis. Statistics Medicine 1997; 16: 753--768. Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. basic introduction fixed-effect random-effects models meta-analysis. Research Synthesis Methods 2010; 1: 97--111. doi:10.1002/jrsm.12 DerSimonian R, Laird N. Meta-analysis clinical trials. Controlled Clinical Trials 1986; 7: 177--188. Graham PL, Moran JL. Robust meta-analytic conclusions mandate provision prediction intervals meta-analysis summaries. Journal Clinical Epidemiology 2012; 65: 503--510. Higgins JPT, Thompson SG. Quantifying heterogeneity meta-analysis. Statistics Medicine 2002; 21: 1539--1558. Higgins JPT, Thompson SG, Spiegelhalter DJ. re-evaluation random-effects meta-analysis. J R Stat Soc Ser Stat Soc. 2009;172:137--59. doi:10.1111/j.1467-985X.2008.00552.x Riley RD, Higgins JPT, Deeks JJ. Interpretation random effects meta-analyses. British Medical Journal 2011; 342: d549. doi:10.1136/bmj.d549 Viechtbauer W. Conducting Meta-Analyses R metafor Package. Journal Statistical Software. 2010; 36. doi:10.18637/jss.v036.i03","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Univariate meta-analysis — uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Univariate meta-analysis — uvmeta","text":"","code":"data(Roberts) # Frequentist random-effects meta-analysis fit1 <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts))) summary(fit1) #> Call: #> uvmeta.default(r = SDM, r.se = SE, labels = rownames(Roberts)) #> #> Random effects summary:\t 0.36195 (SE: 0.0859) #> Tau squared: \t\t 0.01322 (SE: 0.03431) plot(fit1) #show a forest plot fit1 #> Summary estimate with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 0.36194806 0.17636909 0.54752703 0.04923206 0.67466405 if (FALSE) { # Bayesian random effects meta-analysis fit2 <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts), method=\"BAYES\")) plot(fit2) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Meta-analysis of prediction model performance — valmeta","title":"Meta-analysis of prediction model performance — valmeta","text":"function provides summary estimates concordance statistic, total observed-expected ratio calibration slope. appropriate, data transformations applied missing information derived available quantities. Unless specified otherwise, meta-analysis models assume random effects fitted using restricted maximum likelihood estimation metafor package (Viechtbauer 2010). , confidence intervals average performance based Hartung-Knapp-Sidik-Jonkman method. conducting Bayesian meta-analysis, R packages runjags rjags must installed.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Meta-analysis of prediction model performance — valmeta","text":"","code":"valmeta( measure = \"cstat\", cstat, cstat.se, cstat.cilb, cstat.ciub, cstat.cilv, sd.LP, OE, OE.se, OE.cilb, OE.ciub, OE.cilv, citl, citl.se, N, O, E, Po, Po.se, Pe, data, method = \"REML\", test = \"knha\", verbose = FALSE, slab, n.chains = 4, pars, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Meta-analysis of prediction model performance — valmeta","text":"measure character string indicating summary performance measure calculated. Options \"cstat\" (meta-analysis concordance statistic) \"OE\" (meta-analysis total observed-expected ratio). See `Details' information. cstat Optional vector estimated c-statistic valiation cstat.se Optional vector standard error estimated c-statistics cstat.cilb Optional vector specify lower limits confidence interval. cstat.ciub Optional vector specify upper limits confidence interval. cstat.cilv Optional vector specify levels aformentioned confidence interval limits. (default: 0.95, corresponds 95% confidence interval). sd.LP Optional vector standard deviation linear predictor (prognostic index) OE Optional vector estimated ratio total observed versus total expected events OE.se Optional vector standard errors estimated O:E ratios OE.cilb Optional vector specify lower limits confidence interval OE. OE.ciub Optional vector specify upper limits confidence interval OE. OE.cilv Optional vector specify levels aformentioned confidence interval limits. (default: 0.95, corresponds 95% confidence interval). citl Optional vector estimated calibration---large valiation citl.se Optional vector standard error estimated calibration---large statistics N Optional vector total number participants valiation O Optional vector total number observed events valiation (specified, time t.val) E Optional vector total number expected events valiation (specified, time t.val) Po Optional vector (cumulative) observed event probability valiation (specified, time t.val) Po.se Optional vector standard errors Po. Pe Optional vector (cumulative) expected event probability validation (specified, time t.val) data optional data frame containing variables given arguments . method Character string specifying whether fixed- random-effects model fitted. fixed-effects model fitted using method=\"FE\". Random-effects models fitted setting method equal one following: \"REML\" (Default), \"DL\", \"\", \"SJ\", \"ML\", \"EB\", \"HS\", \"GENQ\" \"BAYES\". See 'Details'. test Optional character string specifying test statistics confidence intervals fixed effects computed. default (test=\"knha\"), method Knapp Hartung (2003) used adjusting test statistics confidence intervals. Type '?rma' details. verbose TRUE messages generated fitting process displayed. slab Optional vector specifying label study n.chains Optional numeric specifying number chains use Gibbs sampler (method=\"BAYES\"). chains improve sensitivity convergence diagnostic, cause simulation run slowly. default number chains 4. pars list additional arguments. See 'Details' information. following parameters configure MCMC sampling procedure: hp.mu.mean (mean prior distribution random effects model, defaults 0), hp.mu.var (variance prior distribution random effects model, defaults 1000), hp.tau.min (minimum value -study standard deviation, defaults 0), hp.tau.max (maximum value -study standard deviation, defaults 2), hp.tau.sigma (standard deviation prior distribution -study standard-deviation), hp.tau.dist (prior distribution -study standard-deviation. Defaults \"dunif\"), hp.tau.df (degrees freedom prior distribution -study standard-deviation. Defaults 3). arguments method.restore.c.se (method restoring missing estimates standard error c-statistic. See ccalc information), model.cstat (likelihood/link modeling c-statistic; see \"Details\"), model.oe (likelihood/link modeling O:E ratio; see \"Details\"), seed (integer indicate random seed). ... Additional arguments passed rma runjags (method=\"BAYES\").","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Meta-analysis of prediction model performance — valmeta","text":"object class valmeta following elements: \"data\" array (transformed) data used meta-analysis, method(s) used restoring missing information. \"measure\" character string specifying performance measure meta-analysed. \"method\" character string specifying meta-analysis method. \"model\" character string specifying meta-analysis model (link function). \"est\" summary estimate performance statistic. Bayesian meta-analysis, posterior median returned. \"ci.lb\" lower bound confidence (credibility) interval summary performance estimate. \"ci.ub\" upper bound confidence (credibility) interval summary performance estimate. \"pi.lb\" lower bound (approximate) prediction interval summary performance estimate. \"pi.ub\" upper bound (approximate) prediction interval summary performance estimate. \"fit\" full results fitted model. \"slab\" vector specifying label study.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"meta-analysis-of-the-concordance-statistic","dir":"Reference","previous_headings":"","what":"Meta-analysis of the concordance statistic","title":"Meta-analysis of prediction model performance — valmeta","text":"summary estimate concorcance (c-) statistic can obtained specifying measure=\"cstat\". c-statistic measure discrimination, indicates ability prediction model distinguish patients developing developing outcome. c-statistic typically ranges 0.5 (discriminative ability) 1 (perfect discriminative ability). missing, c-statistic /standard error derived reported information. See ccalc information. default, assumed logit c-statistic Normally distributed within across studies (pars$model.cstat = \"normal/logit\"). Alternatively, possible assume raw c-statistic Normally distributed across studies pars$model.cstat = \"normal/identity\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"meta-analysis-of-the-total-observed-versus-expected-ratio","dir":"Reference","previous_headings":"","what":"Meta-analysis of the total observed versus expected ratio","title":"Meta-analysis of prediction model performance — valmeta","text":"summary estimate total observed versus expected (O:E) ratio can obtained specifying measure=\"OE\". total O:E ratio provides rough indication overall model calibration (across entire range predicted risks). missing, total O:E ratio /standard error derived reported information. See oecalc information. frequentist meta-analysis, within-study variation can either modeled using Normal (model.oe = \"normal/log\" model.oe = \"normal/identity\") Poisson distribution (model.oe = \"poisson/log\"). performing Bayesian meta-analysis, data modeled using one-stage random effects (hierarchical related regression) model. particular, binomial distribution (O, E N known), Poisson distribution (O E known) Normal distribution (OE OE.se OE.95CI known) selected separately study.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"bayesian-meta-analysis","dir":"Reference","previous_headings":"","what":"Bayesian meta-analysis","title":"Meta-analysis of prediction model performance — valmeta","text":"Bayesian meta-analysis models assume presence random effects. Summary estimates based posterior mean. Credibility prediction intervals directly obtained corresponding posterior quantiles. prior distribution (transformed) performance estimate modeled using Normal distribution, mean hp.mu.mean (defaults 0) variance hp.mu.var (defaults 1000). meta-analysis total O:E ratio, maximum value hp.mu.var 100. default, prior distribution -study standard deviation modeled using uniform distribution (hp.tau.dist=\"dunif\"), boundaries hp.tau.min hp.tau.max. Alternative choices truncated Student-t distribution (hp.tau.dist=\"dhalft\") mean hp.tau.mean, standard deviation hp.tau.sigma hp.tau.df degrees freedom. distribution restricted range hp.tau.min hp.tau.max.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Meta-analysis of prediction model performance — valmeta","text":"width calculated confidence, credibility prediction intervals can specified using level pars argument (defaults 0.95).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Meta-analysis of prediction model performance — valmeta","text":"Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. doi:10.1136/bmj.i6460 Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2019;28:2768--86. doi:10.1177/0962280218785504 Riley RD, Tierney JF, Stewart LA. Individual participant data meta-analysis: handbook healthcare research. Hoboken, NJ: Wiley; 2021. ISBN: 978-1-119-33372-2. Steyerberg EW, Nieboer D, Debray TPA, van Houwelingen HC. Assessment heterogeneity individual participant data meta-analysis prediction models: overview illustration. Stat Med. 2019; 38:4290--309. doi:10.1002/sim.8296 Viechtbauer W. Conducting Meta-Analyses R metafor Package. Journal Statistical Software. 2010; 36. doi:10.18637/jss.v036.i03","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Meta-analysis of prediction model performance — valmeta","text":"","code":"######### Validation of prediction models with a binary outcome ######### data(EuroSCORE) # Meta-analysis of the c-statistic (random effects) fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, cstat.cilv=0.95, N=n, O=n.events, slab=Study, data=EuroSCORE) plot(fit) # Nearly identical results when we need to estimate the SE valmeta(cstat=c.index, N=n, O=n.events, slab=Study, data=EuroSCORE) #> Summary c-statistic with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 0.7889020 0.7650864 0.8109000 0.6818676 0.8669518 #> #> Number of studies included: 23 # Two-stage meta-analysis of the total O:E ratio (random effects) valmeta(measure=\"OE\", O=n.events, E=e.events, N=n, slab=Study, data=EuroSCORE) #> Summary Total O:E ratio with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 1.1075973 0.8998973 1.3632352 0.4295250 2.8561122 #> #> Number of studies included: 23 valmeta(measure=\"OE\", O=n.events, E=e.events, data=EuroSCORE) #> Summary Total O:E ratio with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 1.1059784 0.8990028 1.3606056 0.4316383 2.8338269 #> #> Number of studies included: 23 valmeta(measure=\"OE\", Po=Po, Pe=Pe, N=n, data=EuroSCORE) #> Summary Total O:E ratio with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 1.1230955 0.9212978 1.3690944 0.4549877 2.7722586 #> #> Number of studies included: 23 if (FALSE) { # One-stage meta-analysis of the total O:E ratio (random effects) valmeta(measure=\"OE\", O=n.events, E=e.events, data=EuroSCORE, method=\"ML\", pars=list(model.oe=\"poisson/log\")) # Bayesian random effects meta-analysis of the c-statistic fit2 <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, cstat.cilv=0.95, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) # Bayesian one-stage random effects meta-analysis of the total O:E ratio # Consider that some (but not all) studies do not provide information on N # A Poisson distribution will be used for studies 1, 2, 5, 10 and 20 # A Binomial distribution will be used for the remaining studies EuroSCORE.new <- EuroSCORE EuroSCORE.new$n[c(1, 2, 5, 10, 20)] <- NA pars <- list(hp.tau.dist=\"dhalft\", # Prior for the between-study standard deviation hp.tau.sigma=1.5, # Standard deviation for 'hp.tau.dist' hp.tau.df=3, # Degrees of freedom for 'hp.tau.dist' hp.tau.max=10, # Maximum value for the between-study standard deviation seed=5) # Set random seed for the simulations fit3 <- valmeta(measure=\"OE\", O=n.events, E=e.events, N=n, data=EuroSCORE.new, method=\"BAYES\", slab=Study, pars=pars) plot(fit3) print(fit3$fit$model) # Inspect the JAGS model print(fit3$fit$data) # Inspect the JAGS data } ######### Validation of prediction models with a time-to-event outcome ######### data(Framingham) # Meta-analysis of total O:E ratio after 10 years of follow-up valmeta(measure=\"OE\", Po=Po, Pe=Pe, N=n, data=Framingham) #> Warning: 8 studies with NAs omitted from model fitting. #> Summary Total O:E ratio with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 0.5781061 0.4400900 0.7594053 0.1935434 1.7267794 #> #> Number of studies included: 16"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"Returns variance-covariance matrix main parameters fitted model object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"","code":"# S3 method for riley vcov(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"object riley object. ... arguments passed functions","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"variance-covariance matrix obtained inverse Hessian provided optim.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"matrix estimated covariances parameter estimates Riley model: logit sensitivity (mu1), logit false positive rate (mu2), additional variation mu1 beyond sampling error (psi1), additional variation mu2 beyond sampling error (psi2) transformation correlation psi1 psi2 (rhoT). original correlation given inv.logit(rhoT)*2-1.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"warning message casted Hessian matrix contains negative eigenvalues. implies identified minimum (restricted) negative log-likelihood saddle point, solution therefore optimal.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"Riley, RD., Thompson, JR., & Abrams, KR. (2008). “alternative model bivariate random-effects meta-analysis within-study correlations unknown.” Biostatistics, 9, 172--186.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"Thomas Debray ","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/news/index.html","id":"metamisc-0409000","dir":"Changelog","previous_headings":"","what":"metamisc 0.4.0.9000","title":"metamisc 0.4.0.9000","text":"Added NEWS.md file track changes package.","code":""}] +[{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"introduction","dir":"Articles","previous_headings":"","what":"Introduction","title":"Meta-analysis of prognostic factors","text":"important task medical research identification assessment prognostic factors. prognostic factor measure , among people given health condition (, startpoint), associated subsequent clinical outcome (endpoint) (Riley 2013). Commonly investigated prognostic factors include simple measures age, sex, size tumor, can also include complex factors abnormal levels proteins catecholamines unusual genetic mutations (Sauerbrei Altman 2006). can useful modifiable targets interventions improve outcomes, building blocks prognostic models, predictors differential treatment response. past decades, numerous prognostic factor studies published medical literature. example, Riley Burchill (2003) identified 260 studies reporting associations 130 different tumour markers neuroblastoma. recently, Tzoulaki Ioannidis (2009) identified 79 studies reporting added value 86 different markers added Framingham risk score. Despite huge research effort, prognostic value traditional factors discussion uncertain usefulness many specific markers, prognostic indices, classification schemes still unproven (Sauerbrei Altman 2006). vignette aims illustrate results multiple prognostic factor studies can summarized sources -study heterogeneity can examined. Hereto, use R packages metamisc metafor. https://cran.r-project.org/package=metafor package provides comprehensive collection functions conducting meta-analyses R. https://cran.r-project.org/package=metamisc package provides additional functions facilitate meta-analysis prognosis studies. can load packages follows:","code":"library(metafor) library(metamisc)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"case-study","dir":"Articles","previous_headings":"","what":"Case Study","title":"Meta-analysis of prognostic factors","text":"Endometrial cancer (EC) fourth common malignancy women common gynecologic cancer. Overall, 5-year survival rates EC approximately 78–90% stage , 74% stage II, 36–57% stage III, 20% stage IV. poor outcomes raise urgent requirement accurate prognosis predictive markers applied EC guide therapy monitor disease progress individual patients. Several biological molecules proposed prognostic biomarkers EC. Among , hormone receptors estrogen receptors (ER), progesterone receptors (PR), human epidermal growth factor receptor 2 (HER2) attractive physiological functions. Recently, Zhang conducted systematic review evaluate overall risk hormone receptors endometrial cancer survival (Zhang Sun 2015). review included 16 studies recruiting 1764 patients HER2. study, estimates effect retrieved follows. simplest method consisted direct collection HR, odds ratio risk ratio, 95% CI original article, HR less 1 associated better outcome. available, total numbers observed deaths/cancer recurrences numbers patients group extracted calculate HR. data available Kaplan-Meier curves, data extracted graphical survival plots, estimation HR performed using described method. can load data 16 studies R follows: creates object Zhang contains summary data 14 studies reporting overall survival (OS) 6 studies reporting progression-free survival (PFS). total 14 studies assessed relation HER2 overall survival. corresponding hazard ratios (HR) given : Results suggest hormone receptor HER2 prognostic value survival prone substantial -study heterogeneity. example, hazard ratios appear much larger studies conducted USA. Possibly, variation treatment effect estimates caused differences baseline characteristics patients (age, tumor stage, race), differences cutoff value HER2, differences received treatments, differences duration follow-. Importantly, estimated hazard ratios adjusted patient-level covariates, particularly prone heterogeneity patient spectrum.","code":"data(Zhang)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"first-steps","dir":"Articles","previous_headings":"","what":"First steps","title":"Meta-analysis of prognostic factors","text":"facilitate quantitative analysis, information standard error different study effect sizes needed. Estimates standard error can obtained reported 95% confidence intervals (Altman Douglas G. Bland 2011). commonly assumed log hazard ratio follows Normal distribution, standard error (SE) log hazard ratio given : \\(\\mathrm{SE}=(\\log(u)-\\log(l))/(2*1.96)\\) upper lower limits 95% CI \\(u\\) \\(l\\) respectively. can implement calculation follows: often helpful display effect sizes studies forest plot. advantage forest plot allows visual inspection available evidence facilitates identification potential -study heterogeneity. forest plot overall survival can generated using forest function metamisc. requires provide information estimated hazard ratios (via argument theta), well bounds 95% confidence interval (via theta.ci.lb theta.ci.ub). can also generate forest plot using metafor:","code":"Zhang <- Zhang %>% mutate(logHR = log(HR), se.logHR = log(HR.975/HR.025)/(2 * qnorm(0.975))) library(dplyr) # Select the 14 studies investigating overall survival dat_os <- Zhang %>% filter(outcome == \"OS\") # Generate a forest plot of the log hazard ratio metamisc::forest(theta = dat_os$HR, theta.ci.lb = dat_os$HR.025, theta.ci.ub = dat_os$HR.975, theta.slab = dat_os$Study, xlab = \"Hazard ratio of HER2 versus OS\", refline = 1) metafor::forest(x = dat_os$HR, ci.lb = dat_os$HR.025, ci.ub = dat_os$HR.975, slab = dat_os$Study, xlab = \"Hazard ratio of HER2 versus OS\", refline = 1)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"assessment-of-publication-bias","dir":"Articles","previous_headings":"","what":"Assessment of publication bias","title":"Meta-analysis of prognostic factors","text":"presence small-study effects common threat systematic reviews meta-analyses. Small-study effects generic term phenomenon sometimes smaller studies show different, often stronger, effects large ones (Debray Riley 2018). One possible reason publication bias. Previously, D. G. Altman (2001) argued probable studies showing strong (often statistically significant) prognostic ability likely published. reason, important evaluate potential presence small-study effects, can verified visual inspection funnel plot. plot, estimate reported effect size plotted measure precision sample size included study meta-analysis. premise scatter plots reflect funnel shape, small-study effects exist (provided effect sizes substantially affected presence -study heterogeneity). However, small studies predominately one direction (usually direction larger effect sizes), asymmetry ensue. common approach construct funnel plot display individual observed effect sizes x-axis corresponding standard errors y-axis, use fixed effect summary estimate reference value. absence publication bias heterogeneity, one expect see points forming funnel shape, majority points falling inside pseudo-confidence region summary estimate. case study, study estimates fall within pseudo-confidence region, hence appears limited evidence publication bias. can formally test presence asymmetry funnel plot evaluating whether association estimated standard error estimated effect size. common use 10% level significance number studies meta-analysis usually low. case study, P-value 0.052, implies evidence funnel plot asymmetry . Funnel plot asymmetry tests can also performed using metamisc follows: yields , can construct funnel plot: caution warranted interpreting results funnel plot asymmetry tests (Debray Riley 2018). First, power detect asymmetry often limited meta-analyses usually include many studies. Second, many tests known yield inadequate type-error rates suffer low power. instance, demonstrated aforementioned test evaluate association estimated standard error effect size tends yield type-error rates high. Finally, funnel plot asymmetry may rather caused heterogeneity publication bias. therefore adjust aforementioned regression test use inverse total sample size (rather standard error) predictor. onwards, assume potential publication bias negligible, proceed standard meta-analysis methods.","code":"res <- rma(yi = logHR, sei = se.logHR, method = \"FE\", data = dat_os) funnel(res, xlab = \"Log Hazard Ratio\") regtest(x = dat_os$logHR, sei = dat_os$se.logHR, model = \"lm\", predictor = \"sei\") ## ## Regression Test for Funnel Plot Asymmetry ## ## Model: weighted regression with multiplicative dispersion ## Predictor: standard error ## ## Test for Funnel Plot Asymmetry: t = 2.1622, df = 12, p = 0.0515 ## Limit Estimate (as sei -> 0): b = 0.2590 (CI: -0.0760, 0.5939) regfit <- fat(b = dat_os$logHR, b.se = dat_os$se.logHR, method = \"E-FIV\") ## Call: fat(b = dat_os$logHR, b.se = dat_os$se.logHR, method = \"E-FIV\") ## ## Fixed effect summary estimate: 0.5193 ## ## test for funnel plot asymmetry: t =2.1622, df = 12, p = 0.0515 plot(regfit) regtest(x = dat_os$logHR, sei = dat_os$se.logHR, ni = dat_os$N, model = \"lm\", predictor = \"ninv\") ## ## Regression Test for Funnel Plot Asymmetry ## ## Model: weighted regression with multiplicative dispersion ## Predictor: inverse of the sample size ## ## Test for Funnel Plot Asymmetry: t = 0.1552, df = 12, p = 0.8793 ## Limit Estimate (as ni -> inf): b = 0.5088 (CI: 0.2226, 0.7950)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"meta-analysis-of-the-prognostic-value-of-her2","dir":"Articles","previous_headings":"","what":"Meta-analysis of the prognostic value of HER2","title":"Meta-analysis of prognostic factors","text":"Meta-analysis option identified studies considered sufficiently robust comparable, requires least two estimates statistic across studies. random effects approach often essential allow unexplained heterogeneity across studies due differences methods, time-scale, populations, cut-points, adjustment factors, treatments. standard random effects meta-analysis combines study estimates statistic interest (given log HR HER2) order estimate average effect (denoted \\(\\mu\\)) standard deviation (denoted \\(\\tau\\)) across studies. \\(\\hat b_i\\) \\(\\mathrm{var}(\\hat b_i)\\) denote estimate variance study \\(\\), general random effects meta-analysis model can specified : \\(\\hat b_i \\sim N(\\mu, \\mathrm{var}(\\hat b_i) + \\tau^2)\\) common first estimate heterogeneity parameter \\(\\tau\\) use resulting value estimate \\(\\mu\\). However, approach adequately reflect error associated parameter estimation, especially number studies small. reason, alternative estimators proposed simultaneously estimate \\(\\mu\\) \\(\\tau\\). , focus Restricted Maximum Likelihood Estimation (REML), implemented default metafor. pooled estimate log hazard ratio 0.667 standard error 0.135. -study standard deviation log hazard ratio 0.297. can extract key statistics follows: can use information derive summary estimate hazard ratio corresponding 95% confidence interval: summary HR HER2 statistically significant, indicating increased levels HER2 associated poorer survival. can also obtain summary estimate 95% CI HR HER2 simply using predict function: Although summary result (\\(\\hat \\mu\\)) usually main focus meta-analysis, reflects average across studies may hard translate clinical practice large -study heterogeneity. can quantify impact -study heterogeneity constructing \\(100(1-\\alpha/2)\\)% prediction interval, gives potential true prognostic effect new population conditional meta-analysis results (Riley Deeks 2011). approximate prediction interval (PI) given follows: \\(\\hat \\mu \\pm t_{\\alpha, N-2} \\sqrt{\\hat \\tau^2 + \\hat \\sigma^2}\\) \\(t_{\\alpha, N-2}\\) \\(100(1-\\alpha/2)\\)% percentile t-distribution \\(N-2\\) degrees freedom, \\(N\\) number studies, \\(\\hat \\sigma\\) estimated standard error \\(\\hat \\mu\\), \\(\\hat \\tau\\) estimated -study standard deviation. R, can calculate approximate 95% PI \\(\\hat \\mu\\) follows: 95% prediction interval ranges 0.956 3.969, suggests substantial heterogeneity prognostic value HER2. particular, although increased levels HER2 generally associated poorer survival, may also lead improved survival (HR < 1) certain settings. can add summary estimate prediction interval forest plot: possible approach enhance interpretation meta-analysis results calculate probability prognostic effect HER2 useful value (e.g. HR > 1.5 binary factor, indicates risk increased least 50%) new setting. can calculate probability follows: \\(Pr(\\mathrm{HR} > 1.5) = Pr(\\hat \\mu > \\log(1.5)) = 1 - Pr(\\hat \\mu \\leq \\log(1.5))\\) \\(Pr(\\hat \\mu \\leq \\log(1.5))\\) approximated using scaled Student-\\(t\\) distribution (similar calculation prediction interval): probability HER2 yield hazard ratio overall survival least 1.5 new setting 78%. means despite presence -study heterogeneity, likely HER2 provide substantial discriminative ability used single prognostic factor new setting. can also estimate probability means simulation: , probability HER2 yield hazard ratio overall survival >1.5 new setting 78%.","code":"resREML <- rma(yi = logHR, sei = se.logHR, method = \"REML\", slab = Study, data = dat_os) resREML ## ## Random-Effects Model (k = 14; tau^2 estimator: REML) ## ## tau^2 (estimated amount of total heterogeneity): 0.0883 (SE = 0.0854) ## tau (square root of estimated tau^2 value): 0.2972 ## I^2 (total heterogeneity / total variability): 49.17% ## H^2 (total variability / sampling variability): 1.97 ## ## Test for Heterogeneity: ## Q(df = 13) = 28.9214, p-val = 0.0067 ## ## Model Results: ## ## estimate se zval pval ci.lb ci.ub ## 0.6669 0.1354 4.9251 <.0001 0.4015 0.9324 *** ## ## --- ## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 # Summary estimate of the log hazard ratio for HER2 mu <- resREML$b # 95% confidence interval of the pooled log hazard ratio mu.ci <- c(resREML$ci.lb, resREML$ci.ub) # Between-study variance of the log hazard ratio tau2 <- resREML$tau2 # Error variance of the pooled log hazard ratio sigma2 <- as.numeric(vcov(resREML)) # Number of studies contributing to the meta-analyis numstudies <- resREML$k.all exp(mu) ## [,1] ## intrcpt 1.948268 exp(mu.ci) ## [1] 1.494104 2.540483 predict(resREML, transf = exp) ## ## pred ci.lb ci.ub pi.lb pi.ub ## 1.9483 1.4941 2.5405 1.0272 3.6954 level <- 0.05 crit <- qt(c(level/2, 1 - (level/2)), df = (numstudies - 2)) pi_lower <- exp(mu + crit[1] * sqrt(tau2 + sigma2)) pi_upper <- exp(mu + crit[2] * sqrt(tau2 + sigma2)) c(pi_lower, pi_upper) ## [1] 0.9563084 3.9691662 # Generate a forest plot of the log hazard ratio metamisc::forest(theta = dat_os$HR, theta.ci.lb = dat_os$HR.025, theta.ci.ub = dat_os$HR.975, theta.slab = dat_os$Study, theta.summary = exp(mu), theta.summary.ci.lb = exp(mu.ci[1]), theta.summary.ci.ub = exp(mu.ci[2]), theta.summary.pi.lb = pi_lower, theta.summary.pi.ub = pi_upper, xlab = \"Hazard ratio of HER2 versus OS\", refline = 1) probOS <- 1 - pt((log(1.5) - mu)/sqrt(tau2 + sigma2), df = (numstudies - 2)) probOS ## [,1] ## intrcpt 0.7805314 # Simulate 100000 new studies Nsim <- 1e+06 # Random draws from a Student T distribution rnd_t <- rt(Nsim, df = (numstudies - 2)) # Generate 1,000,000 hazard ratios HRsim <- exp(c(mu) + rnd_t * sqrt(tau2 + sigma2)) # Calculate the proportion of hazard ratios greater than 1.5 mean(HRsim > 1.5) ## [1] 0.780348"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pf.html","id":"multivariate-meta-analysis","dir":"Articles","previous_headings":"","what":"Multivariate meta-analysis","title":"Meta-analysis of prognostic factors","text":"previous section, used 14 16 identified studies evaluate prognostic effect HER2 overall survival. Two studies excluded meta-analysis provide direct evidence overall survival. unwelcome, especially participants otherwise representative population, clinical settings, condition interest (Riley White 2017). reason, discuss multivariate meta-analysis methods can used borrow strength studies investigate primary outcome interest. Briefly, multivariate meta-analysis methods simultaneously summarize effect size across multiple outcomes whilst accounting correlation. example, six studies review Zhang Sun (2015) assessed hazard ratio HER2 progression-free survival, four also assessed overall survival. Hence, conducting multivariate meta-analysis can borrow strength two additional studies estimating hazard ratio overall survival. hazard ratios progression free survival depicted : first conduct univariate meta-analysis six studies investigating progression-free survival: Results indicate hormone receptor HER2 also prognostic value progression-free survival. Furthermore, reported HRs appear much homogeneous across studies, since -study standard deviation 0.17 progression-free survival whereas 0.30 overall survival. Note univariate meta-analysis progression-free survival based merely 6 studies, univariate meta-analysis overall survival based 14 studies. can now employ multivariate meta-analysis borrow information 4 studies report prognostic effects endpoints. , turn, allows studies contribute summary effect HER2 outcomes. first need define within-study covariance matrix estimated log hazard ratios progression-free survival overall survival. assume estimates hazard ratio independent within studies construct block diagonal matrix considers error variance estimate: multivariate random-effects model can now used simultaneously meta-analyze hazard ratios overall progression-free survival: summary estimate log hazard ratio overall survival 0.670 (multivariate meta-analysis) versus 0.667 (univariate meta-analysis) SE 0.132 , respectively, 0.135. Hence, gained precision including evidence 2 additional studies evaluated progression-free survival. Note estimation -study heterogeneity difficult progression-free survival due limited number studies. particular, found \\(\\tau^2\\)= 0.028 SE 0.145. multivariate meta-analysis, estimated -study variance PFS much larger (\\(\\tau^2\\)=0.077), based 16 rather merely 6 studies. summary, multivariate meta-analysis approach often helpful reduces need exclude relevant studies meta-analysis, thereby decreasing risk bias (e.g. due selective outcome reporting) potentially improving precision. indicated Riley White (2017), multivariate meta-analysis multiple outcomes beneficial outcomes highly correlated percentage studies missing outcomes large.","code":"dat_pfs <- Zhang %>% filter(outcome == \"PFS\") resPFS <- rma(yi = logHR, sei = se.logHR, method = \"REML\", slab = Study, data = dat_pfs) V <- diag(Zhang$se.logHR^2) res.MV <- rma.mv(yi = logHR, V = V, mods = ~outcome - 1, random = ~outcome | Study, struct = \"UN\", data = Zhang, method = \"REML\") res.MV ## ## Multivariate Meta-Analysis Model (k = 20; method: REML) ## ## Variance Components: ## ## outer factor: Study (nlvls = 16) ## inner factor: outcome (nlvls = 2) ## ## estim sqrt k.lvl fixed level ## tau^2.1 0.0865 0.2942 14 no OS ## tau^2.2 0.0770 0.2775 6 no PFS ## ## rho.OS rho.PFS OS PFS ## OS 1 - 4 ## PFS 1.0000 1 no - ## ## Test for Residual Heterogeneity: ## QE(df = 18) = 33.7664, p-val = 0.0135 ## ## Test of Moderators (coefficients 1:2): ## QM(df = 2) = 35.6315, p-val < .0001 ## ## Model Results: ## ## estimate se zval pval ci.lb ci.ub ## outcomeOS 0.6704 0.1318 5.0868 <.0001 0.4121 0.9287 *** ## outcomePFS 0.8734 0.2151 4.0606 <.0001 0.4518 1.2950 *** ## ## --- ## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1"},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pm.html","id":"case-study","dir":"Articles","previous_headings":"","what":"Case Study","title":"Meta-analysis of prediction model performance","text":"EuroSCORE II commonly used scoring rule estimating risk -hospital mortality patients undergoing major cardiac surgery. developed using data 16,828 adult patients 43 countries. Predictors include patient characteristics (e.g. age, gender), cardiac related factors (e.g. recent MI) surgery related factors (e.g. Surgery thoracic aorta). 2014, systematic review undertaken Guida et al. (2014) search articles assessing performance EuroSCORE II perioperative mortality cardiac surgery. systematic review identified 24 eligible validation studies, 22 studies included main analysis. case study, summarize results 22 studies, well results split-sample validation contained within original development article EuroSCORE II. use metamisc package derive summary estimates discrimination calibration performance EuroSCORE II, evaluate presence -study heterogeneity, identify potential sources -study heterogeneity. step--step tutorial provided Debray et al. (2017). can load data 23 validation studies follows:","code":"library(metamisc) data(EuroSCORE)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/articles/ma-pm.html","id":"meta-analysis-of-calibration-performance","dir":"Articles","previous_headings":"","what":"Meta-analysis of calibration performance","title":"Meta-analysis of prediction model performance","text":"Calibration refers model’s accuracy predicted risk probabilities, indicates extent expected outcomes (predicted model) observed outcomes agree. Summarising estimates calibration performance challenging calibration plots often presented, studies tend report different types summary statistics calibration. example, case study, calibration assessed using Hosmer-Lemeshow test, calibration plots comparing observed mortality predicted EuroSCORE II (either overall groups patients). Within validation study, can compare total number observed events (O) total number expected (predicted) events deriving ratio O:E. total O:E ratio provides rough indication overall model calibration (across entire range predicted risks). describes whether (O:E > 1) fewer (O:E < 1) events occurred expected based model. Whilst O:E ratio explicitly reported studies, can calculated reported information: O:E ratio can also derived observed predicted mortality risk Po , respectively, Pe: recommended first transform extracted O:E ratios log (natural logarithm) scale applying meta-analysis.","code":"EuroSCORE <- EuroSCORE %>% mutate(oe = n.events/e.events) EuroSCORE %>% select(Po, Pe) %>% mutate(oe = Po/Pe) EuroSCORE <- EuroSCORE %>% mutate(logoe = log(oe))"},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/authors.html","id":null,"dir":"","previous_headings":"","what":"Authors","title":"Authors and Citation","text":"Thomas Debray. Author, maintainer. Valentijn de Jong. Author.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/authors.html","id":"citation","dir":"","previous_headings":"","what":"Citation","title":"Authors and Citation","text":"Debray T, de Jong V (2024). metamisc: Meta-Analysis Diagnosis Prognosis Research Studies. R package version 0.4.0.9000, https://smartdata-analysis--statistics.github.io/metamisc/, https://github.com/smartdata-analysis--statistics/metamisc.","code":"@Manual{, title = {metamisc: Meta-Analysis of Diagnosis and Prognosis Research Studies}, author = {Thomas Debray and Valentijn {de Jong}}, year = {2024}, note = {R package version 0.4.0.9000, https://smartdata-analysis-and-statistics.github.io/metamisc/}, url = {https://github.com/smartdata-analysis-and-statistics/metamisc}, }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"metamisc","dir":"","previous_headings":"","what":"Meta-Analysis of Diagnosis and Prognosis Research Studies","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"official repository R package metamisc, developed facilitate meta-analysis diagnosis prognosis research studies. package includes functions following tasks: develop validate multivariable prediction models datasets clustering (de Jong et al., 2021) summarize multiple estimates prediction model discrimination calibration performance (Debray et al., 2019) evaluate funnel plot asymmetry (Debray et al., 2018)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"installation","dir":"","previous_headings":"","what":"Installation","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"metamisc package can installed CRAN follows: can install development version metamisc GitHub :","code":"install.packages(\"metamisc\") # install.packages(\"devtools\") devtools::install_github(\"smartdata-analysis-and-statistics/metamisc\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"jasp","dir":"","previous_headings":"","what":"JASP","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"visual interface software implemented JASP https://jasp-stats.org/","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"funding","dir":"","previous_headings":"","what":"Funding","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"development R package funded following organisations: Netherlands Organisation Health Research Development (grant 91617050). European Union’s Horizon 2020 research innovation programme ReCoDID grant agreement 825746. Smart Data Analysis Statistics B.V., limited liability corporation registered Netherlands Chamber Commerce number 863595327.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/index.html","id":"references","dir":"","previous_headings":"","what":"References","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies","text":"de Jong VMT, Moons KGM, Eijkemans MJC, Riley RD, Debray TPA. Developing generalizable prediction models pooled studies large clustered data sets. Stat Med. 2021 May 5;40(15):3533–59. Debray TPA, Moons KGM, Riley RD. Detecting small-study effects funnel plot asymmetry meta-analysis survival data: comparison new existing tests. Res Syn Meth. 2018;9(1):41–50. Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2019 Sep;28(9):2768–86.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":null,"dir":"Reference","previous_headings":"","what":"Collins data — Collins","title":"Collins data — Collins","text":"meta-analysis nine clinical trials investigating effect taking diuretics pregnancy risk pre-eclampsia.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Collins data — Collins","text":"","code":"data(Collins)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Collins data — Collins","text":"data frame 9 observations following 2 variables. logOR numeric vector treatment effect sizes (log odds ratio) SE numeric vector standard error treatment effect sizes","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Collins data — Collins","text":"Collins, R., Yusuf, S., Peto, R. Overview randomised trials diuretics pregnancy. British Medical Journal 1985, 290, 17--23. Hardy, R.J. Thompson, S.G. likelihood approach meta-analysis random effects. Statistics Medicine 1996; 15:619--629.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Collins.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Collins data — Collins","text":"","code":"data(Collins)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":null,"dir":"Reference","previous_headings":"","what":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"hypothetical dataset 500 subjects suspected deep vein thrombosis (DVT).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"","code":"data(DVTipd)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"data frame 500 observations 16 variables. sex gender (0=female, 1=male) malign active malignancy (0=active malignancy, 1=active malignancy) par paresis (0=paresis, 1=paresis) surg recent surgery bedridden tend tenderness venous system oachst oral contraceptives hst leg entire leg swollen notraum absence leg trauma calfdif3 calf difference >= 3 cm pit pitting edema vein vein distension altdiagn alternative diagnosis present histdvt history previous DVT ddimdich dichotimized D-dimer value dvt final diagnosis DVT study study indicator","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"Hypothetical dataset derived Individual Participant Data Meta-Analysis Geersing et al (2014). dataset consists consecutive outpatients suspected deep vein thrombosis, documented information presence absence proximal deep vein thrombosis (dvt) acceptable reference test. Acceptable tests either compression ultrasonography venography initial presentation, , venous imaging performed, uneventful follow-least three months.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"Geersing GJ, Zuithoff NPA, Kearon C, Anderson DR, Ten Cate-Hoek AJ, Elf JL, et al. Exclusion deep vein thrombosis using Wells rule clinically important subgroups: individual patient data meta-analysis. BMJ. 2014;348:g1340.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTipd.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Hypothetical dataset for diagnosis of Deep Vein Thrombosis (DVT) — DVTipd","text":"","code":"data(DVTipd) str(DVTipd) #> 'data.frame':\t500 obs. of 16 variables: #> $ sex : num 0 1 0 1 0 0 1 0 1 0 ... #> $ malign : num 0 0 0 0 0 0 0 0 0 0 ... #> $ par : num 0 0 1 0 0 0 0 0 0 0 ... #> $ surg : num 0 0 0 0 0 0 0 0 1 0 ... #> $ tend : num 1 1 0 1 1 0 0 1 1 1 ... #> $ oachst : num 0 0 0 0 0 0 0 0 0 0 ... #> $ leg : num 1 0 0 0 0 1 1 0 0 0 ... #> $ notraum : num 1 1 1 1 1 0 0 1 0 1 ... #> $ calfdif3: num 0 0 0 0 0 0 0 0 0 0 ... #> $ pit : num 0 0 0 0 0 1 0 1 1 1 ... #> $ vein : num 0 0 0 0 1 0 0 0 0 1 ... #> $ altdiagn: num 1 0 1 1 1 0 1 1 1 1 ... #> $ histdvt : num 0 1 0 0 0 0 1 0 0 0 ... #> $ ddimdich: num 1 0 0 0 0 1 1 0 1 1 ... #> $ dvt : num 0 0 0 0 0 0 0 0 0 0 ... #> $ study : Factor w/ 4 levels \"a\",\"b\",\"c\",\"d\": 1 4 1 4 1 4 4 4 4 2 ... summary(apply(DVTipd,2,as.factor)) #> sex malign par surg #> Length:500 Length:500 Length:500 Length:500 #> Class :character Class :character Class :character Class :character #> Mode :character Mode :character Mode :character Mode :character #> tend oachst leg notraum #> Length:500 Length:500 Length:500 Length:500 #> Class :character Class :character Class :character Class :character #> Mode :character Mode :character Mode :character Mode :character #> calfdif3 pit vein altdiagn #> Length:500 Length:500 Length:500 Length:500 #> Class :character Class :character Class :character Class :character #> Mode :character Mode :character Mode :character Mode :character #> histdvt ddimdich dvt study #> Length:500 Length:500 Length:500 Length:500 #> Class :character Class :character Class :character Class :character #> Mode :character Mode :character Mode :character Mode :character ## Develop a prediction model to predict presence of DVT model.dvt <- glm(\"dvt~sex+oachst+malign+surg+notraum+vein+calfdif3+ddimdich\", family=binomial, data=DVTipd) summary(model.dvt) #> #> Call: #> glm(formula = \"dvt~sex+oachst+malign+surg+notraum+vein+calfdif3+ddimdich\", #> family = binomial, data = DVTipd) #> #> Coefficients: #> Estimate Std. Error z value Pr(>|z|) #> (Intercept) -5.1664 0.6365 -8.117 4.76e-16 *** #> sex 0.8146 0.2825 2.883 0.00393 ** #> oachst 0.4324 0.6227 0.694 0.48739 #> malign 0.5679 0.4025 1.411 0.15826 #> surg 0.1002 0.4111 0.244 0.80734 #> notraum 0.3351 0.3700 0.906 0.36513 #> vein 0.4831 0.3186 1.516 0.12939 #> calfdif3 1.1841 0.2819 4.200 2.67e-05 *** #> ddimdich 2.6081 0.5310 4.911 9.04e-07 *** #> --- #> Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 #> #> (Dispersion parameter for binomial family taken to be 1) #> #> Null deviance: 446.24 on 499 degrees of freedom #> Residual deviance: 345.98 on 491 degrees of freedom #> AIC: 363.98 #> #> Number of Fisher Scoring iterations: 6 #>"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":null,"dir":"Reference","previous_headings":"","what":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"Previously published prediction models predicting presence DVT.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"","code":"data(DVTmodels)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"object class litmodels following information literature model: study-level descriptives (\"descriptives\"), regression coefficient weight predictor (\"weights\") error variance regression coefficient weight (\"weights.var\").","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"Previously, several models (Gagne, Oudega) score charts (Wells, modified Wells, Hamilton) published evaluating presence DVT suspected patients. models combine information mulitple predictors weighted sum, can subsequently used obtain estimates absolute risk. See DVTipd information predictors.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"Wells PS, Anderson DR, Bormanis J, Guy F, Mitchell M, Gray L, Clement C, Robinson KS, Lewandowski B. Value assessment pretest probability deep-vein thrombosis clinical management. Lancet 1997; 350(9094):1795--1798. DOI: 10.1016/S0140-6736(97)08140-3. Wells PS, Anderson DR, Rodger M, Forgie M, Kearon C, Dreyer J, Kovacs G, Mitchell M, Lewandowski B, Kovacs MJ. Evaluation D-dimer diagnosis suspected deep-vein thrombosis. New England Journal Medicine 2003; 349(13):1227--1235. DOI: 10.1056/NEJMoa023153. Gagne P, Simon L, Le Pape F, Bressollette L, Mottier D, Le Gal G. Clinical prediction rule diagnosing deep vein thrombosis primary care. La Presse Medicale 2009; 38(4):525--533. DOI: 10.1016/j.lpm.2008.09.022. Subramaniam RM, Snyder B, Heath R, Tawse F, Sleigh J. Diagnosis lower limb deep venous thrombosis emergency department patients: performance Hamilton modified Wells scores. Annals Emergency Medicine 2006; 48(6):678--685. DOI: 10.1016/j.annemergmed.2006.04.010. Oudega R, Moons KGM, Hoes AW. Ruling deep venous thrombosis primary care. simple diagnostic algorithm including D-dimer testing. Thrombosis Haemostasis 2005; 94(1):200--205. DOI: 10.1160/TH04-12-0829.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"Debray TPA, Koffijberg H, Nieboer D, Vergouwe Y, Steyerberg EW, Moons KGM. Meta-analysis aggregation multiple published prediction models. Stat Med. 2014 Jun 30;33(14):2341--62. Debray TPA, Koffijberg H, Vergouwe Y, Moons KGM, Steyerberg EW. Aggregating published prediction models individual participant data: comparison different approaches. Stat Med. 2012 Oct 15;31(23):2697--712.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/DVTmodels.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Risk prediction models for diagnosing Deep Venous Thrombosis (DVT) — DVTmodels","text":"","code":"data(DVTmodels)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":null,"dir":"Reference","previous_headings":"","what":"Daniels and Hughes data — Daniels","title":"Daniels and Hughes data — Daniels","text":"Data frame treatment differences CD4 cell count.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Daniels and Hughes data — Daniels","text":"","code":"data(\"Daniels\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Daniels and Hughes data — Daniels","text":"data frame 15 observations following 2 variables. Y1 Treatment differences log hazard ratio development AIDS death 2 years. vars1 Error variances Y1. Y2 Difference mean change CD4 cell count baseline 6 month studies AIDS Clinical Trial Group vars2 Error variances Y2.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Daniels and Hughes data — Daniels","text":"Daniels data comprises 15 phase II/III randomized clinical trials HIV Disease Section Adult AIDS Clinical Trials Group National Institutes Health, data available May 1996, least six months follow-patients least one patient developed AIDS died. data previously used Daniels Hughes (1997) assess whether change CD4 cell count surrogate time either development AIDS death drug trials patients HIV.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Daniels.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Daniels and Hughes data — Daniels","text":"Daniels MJ, Hughes MD. Meta-analysis evaluation potential surrogate markers. Statistics Medicine 1997; 16: 1965--1982.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":null,"dir":"Reference","previous_headings":"","what":"Predictive performance of EuroSCORE II — EuroSCORE","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"data set contains estimates predictive performance European system cardiac operative risk evaluation (EuroSCORE II) patients undergoing cardiac surgery. Results based original development study 22 validations identified Guida et al.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"","code":"data(\"EuroSCORE\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"data frame 23 observations following 13 variables. Study vector first author validation study n numeric vector total number patients performance estimates based n.events numeric vector total number observed events c.index numeric vector estimated concordance statistic validation se.c.index numeric vector standard error concordance statistics c.index.95CIl numeric vector lower bound 95% confidence interval estimated concordance statistics c.index.95CIu numeric vector upper bound 95% confidence interval estimated concordance statistics Po numeric vector overall observed event probability validation Pe numeric vector overall expected event probability validation SD.Pe numeric vector standard error Pe e.events numeric vector total number expected events validation multicentre logical vector describing whether study multicentre study mean.age numeric vector describing mean age patients sd.age numeric vector spread age patients pts..2010 logical vector describing whether studies included patients 2010 (.e., EuroSCORE II developed)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"Published 2012, EuroSCORE II developed using logistic regression dataset comprising 16,828 adult patients undergoing major cardiac surgery 154 hospitals 43 countries 12-week period (May-July) 2010. EuroSCORE II developed predict -hospital mortality patients undergoing type cardiac surgery. 2014, systematic review published evidence performance value euroSCORE II undertaken Guida et al. Twenty-two validations, including 145,592 patients 21 external validation articles (one study included two validations) split-sample validation contained within original development article included review; 23 validation studies total.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"Guida P, Mastro F, Scrascia G, Whitlock R, Paparella D. Performance European System Cardiac Operative Risk Evaluation II: meta-analysis 22 studies involving 145,592 cardiac surgery procedures. J Thorac Cardiovasc Surg. 2014; 148(6):3049--3057.e1. Nashef SAM, Roques F, Sharples LD, Nilsson J, Smith C, Goldstone AR, et al. EuroSCORE II. Eur J Cardiothorac Surg. 2012; 41(4):734-744; discussion 744-745.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/EuroSCORE.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Predictive performance of EuroSCORE II — EuroSCORE","text":"","code":"data(EuroSCORE)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":null,"dir":"Reference","previous_headings":"","what":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"Fibrinogen data set meta-analysis 31 studies association plasma fibrinogen concentration risk coronary heath disease (CHD) estimated.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"","code":"data(\"Fibrinogen\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"data frame 5 variables: N.total numeric vector describing total number patients study N.events numeric vector describing number observed events within study HR numeric vector describing estimated hazard ratio study HR.025 numeric vector describing lower boundary 95% confidence interval HR HR.975 numeric vector describing upper boundary 95% confidence interval HR","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"Fibrinogen Studies Collaboration. Collaborative meta-analysis prospective studies plasma fibrinogen cardiovascular disease. Eur J Cardiovasc Prev Rehabil. 2004 Feb;11(1):9-17. Thompson S, Kaptoge S, White , Wood , Perry P, Danesh J, et al. Statistical methods time--event analysis individual participant data multiple epidemiological studies. Int J Epidemiol. 2010 Oct;39(5):1345-59.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Fibrinogen.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Meta-analysis of the association between plasma fibrinogen concentration and the risk of coronary heath disease — Fibrinogen","text":"","code":"data(Fibrinogen) ## maybe str(Fibrinogen) ; plot(Fibrinogen) ..."},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":null,"dir":"Reference","previous_headings":"","what":"Predictive performance of the Framingham Risk Score in male populations — Framingham","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"data set contains estimates performance Framingham model predicting coronary heart disease male populations (Wilson 1998). Results based original development study 20 validations identified Damen et al (BMC Med, 2017).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"","code":"data(\"Framingham\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"data frame 24 observations following 19 variables. AuthorYear vector describing study authors n numeric vector total number patients performance estimates based n.events numeric vector total number observed events c.index numeric vector estimated concordance statistic validation se.c.index numeric vector standard error concordance statistics c.index.95CIl numeric vector lower bound 95% confidence interval estimated concordance statistics c.index.95CIu numeric vector upper bound 95% confidence interval estimated concordance statistics Po numeric vector overall observed event probability validation Pe numeric vector overall expected event probability validation t.val numeric vector describing time period predictive performance assessed validation mean_age numeric vector describing mean age patients sd_age numeric vector spread age patients mean_SBP numeric vector mean systolic blood pressure validation studies (mm Hg) sd_SBP numeric vector spread systolic blood pressure validation studies mean_total_cholesterol numeric vector mean total cholesterol validation studies (mg/dL) sd_total_cholesterol numeric vector spread total cholesterol validation studies mean_hdl_cholesterol numeric vector mean high-density lipoprotein cholesterol validation studies (mg/dL) sd_hdl_cholesterol numeric vector spread high-density lipoprotein cholesterol validation studies pct_smoker numeric vector percentage smokers validation studies","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"Framingham Risk Score allows physicians predict 10-year coronary heart disease (CHD) risk patients without overt CHD. developed 1998 middle-aged white population sample, subsequently validated across different populations. current dataset contains original (internal validation) results, well 23 external validations identified systematic review. review, studies eligible inclusion described validation original Framingham model assessed performance fatal nonfatal CHD males general population setting.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"Damen JAAG, Hooft L, Schuit E, Debray TPA, Collins GS, Tzoulaki , et al. Prediction models cardiovascular disease risk general population: systematic review. BMJ. 2016;i2416. Damen JAAG, Pajouheshnia R, Heus P, Moons KGM, Reitsma JB, Scholten RJPM, et al. Performance Framingham risk models Pooled Cohort Equations: systematic review meta-analysis. BMC Med. 2017;17(1):109. Wilson PW, D'Agostino RB, Levy D, Belanger , Silbershatz H, Kannel WB. Prediction coronary heart disease using risk factor categories. Circulation. 1998; 97(18):1837--47.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Framingham.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Predictive performance of the Framingham Risk Score in male populations — Framingham","text":"","code":"data(Framingham)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":null,"dir":"Reference","previous_headings":"","what":"Kertai data — Kertai","title":"Kertai data — Kertai","text":"Data frame diagnostic accuracy data exercise electrocardiography.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Kertai data — Kertai","text":"","code":"data(\"Kertai\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Kertai data — Kertai","text":"One data frame 4 variables. TP integer. number true positives FN integer. number false negatives FP integer. number false positives TN integer. number true negatives","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Kertai data — Kertai","text":"Kertai data set meta-analysis prognostic test studies comprises 7 studies diagnostic test accuracy exercise electrocardiography predicting cardiac events patients undergoing major vascular surgery measured.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Kertai.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Kertai data — Kertai","text":"Kertai MD, Boersma E, Bax JJ, Heijenbrok-Kal MH, Hunink MGM, L'talien GJ, Roelandt JRTC, van Urk H, Poldermans D. meta-analysis comparing prognostic accuracy six diagnostic tests predicting perioperative cardiac risk patients undergoing major vascular surgery. Heart 2003; 89: 1327--1334. Jackson D, Riley RD, & White IW. Multivariate meta-analysis: Potential promise. Statistics Medicine 2010; 30: 2481--2498.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":null,"dir":"Reference","previous_headings":"","what":"Roberts data — Roberts","title":"Roberts data — Roberts","text":"Data frame summary data 14 comparative studies.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Roberts data — Roberts","text":"","code":"data(\"Roberts\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Roberts data — Roberts","text":"One data frame 2 variables. SDM Effect sizes (standardized differences means) SE Standard error effect sizes","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Roberts data — Roberts","text":"Roberts data set meta-analysis 14 studies comparing 'set shifting' ability (ability move back forth different tasks) people eating disorders healthy controls.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Roberts.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Roberts data — Roberts","text":"Roberts , Tchanturia K, Stahl D, Southgate L, Treasure J. systematic review meta-analysis set-shifting ability eating disorders. Psychological Medicine 2007, 37: 1075--1084. Higgins JPT, Thompson SG, Spiegelhalter DJ. re-evaluation random-effects meta-analysis. Journal Royal Statistical Society. Series (Statistics Society) 2009, 172: 137--159.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":null,"dir":"Reference","previous_headings":"","what":"Diagnostic accuracy data — Scheidler","title":"Diagnostic accuracy data — Scheidler","text":"Data frame diagnostic accuracy data three imaging techniques diagnosis lymph node metastasis women cervical cancer.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Diagnostic accuracy data — Scheidler","text":"","code":"data(\"Scheidler\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Diagnostic accuracy data — Scheidler","text":"One data frame 6 variables. author string . author article modality integer . type test (1=CT, 2=LAG, 3=MRI) TP integer. number true positives FN integer. number false negatives FP integer. number false positives TN integer. number true negatives","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Diagnostic accuracy data — Scheidler","text":"Scheidler data comprises results meta-analysis three imaging techniques diagnosis lymph node metastasis women cervical cancer compared. Forty-four studies total included: 17 studies evaluated lymphangiography, another 17 studies examined computed tomography remaining 10 studies focused magnetic resonance imaging. Diagnosis metastatic disease lymphangiography (LAG) based presence nodal-filling defects, whereas computed tomography (CT) magnetic resonance imaging (MRI) rely nodal enlargement.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Scheidler.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Diagnostic accuracy data — Scheidler","text":"Scheidler J, Hricak H, Yu KK, Subak L, Segal MR. Radiological evaluation lymph node metastases patients cervical cancer. meta-analysis. Journal American Medical Association 1997; 278: 1096--1101. Reitsma J, Glas , Rutjes , Scholten R, Bossuyt P, Zwinderman . Bivariate analysis sensitivity specificity produces informative summary measures diagnostic reviews. Journal Clinical Epidemiology 2005; 58: 982--990.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":null,"dir":"Reference","previous_headings":"","what":"The incremental value of cardiovascular risk factors — Tzoulaki","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"Tzoulaki et al. (2009) reviewed studies evaluated various candidate prognostic factors ability improve prediction coronary heart disease (CHD) outcomes beyond Framingham risk score (FRS) can achieve.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"","code":"data(\"Tzoulaki\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"data frame containing data 27 studies following 2 variables. PubmedID character vector Pubmed ID study N numeric vector describing study size N.events numeric vector describing observed number events FRS.orig.refitted boolean vector describing whether coefficients original Framingham Risk Score (FRS) re-estimated FRS.modif.refitted boolean vector describing whether coefficients modified Framingham Risk Score re-estimated predictors character vector indicating new risk factor(s) included modified FRS outcome character vector indicating primary outcome predicted AUC.orig numeric vector describing Area ROC curve (AUC) original FRS model AUC.orig.CIl numeric vector describing lower boundary 95% confidence interval AUC original FRS model AUC.orig.CIu numeric vector describing upper boundary 95% confidence interval AUC original FRS model AUC.modif numeric vector describing Area ROC curve (AUC) modified FRS model includes one new risk factors AUC.modif.CIl numeric vector describing lower boundary 95% confidence interval AUC modified FRS model AUC.modif.CIu numeric vector describing upper boundary 95% confidence interval AUC modified FRS model pval.AUCdiff numeric vector p-value difference AUC.orig AUC.modif sign.AUCdiff boolean vector indicating whether difference AUC.orig AUC.modif 0.05","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"Tzoulaki , Liberopoulos G, Ioannidis JPA. Assessment claims improved prediction beyond Framingham risk score. JAMA. 2009 Dec 2;302(21):2345–52.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Tzoulaki.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"The incremental value of cardiovascular risk factors — Tzoulaki","text":"","code":"data(Tzoulaki) ## maybe str(Tzoulaki) ; plot(Tzoulaki) ..."},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":null,"dir":"Reference","previous_headings":"","what":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"dataset comprises results 16 studies assessing prognostic role human epidermal growth factor receptor 2 (HER2) endometrial cancer. studies previously identified systematic review Zhang et al. evaluate overall risk several hormone receptors endometrial cancer survival.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"","code":"data(\"Zhang\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"data frame 20 observations following 10 variables. Study factor 16 levels indicate study PrimaryAuthor factor indicating first author's last name year numeric vector indicating publication year Country factor indicating source country study data Disease factor indicating studied disease. Possible levels EC (endometrial cancer), EEC (endometrioid endometrial cancer) UPSC (uterine papillary serous carcinoma) N numeric vector describing total sample size study HR numeric vector describing estimated hazard ratio study HR.025 numeric vector describing lower boundary 95% confidence interval HR HR.975 numeric vector describing upper boundary 95% confidence interval HR outcome factor indicating studied outcome. Possible levels OS (overall survival) PFS (progression-free survival)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"Eligible studies identified searching PubMed EMBASE databases publications 1979 May 2014. Data collected studies comparing overall survival progression-free survival patients elevated levels human epidermal growth factor receptor 2 patients lower levels.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"Zhang Y, Zhao D, Gong C, Zhang F, J, Zhang W, et al. Prognostic role hormone receptors endometrial cancer: systematic review meta-analysis. World J Surg Oncol. 2015 Jun 25;13:208.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"Riley RD, Jackson D, Salanti G, Burke DL, Price M, Kirkham J, et al. Multivariate network meta-analysis multiple outcomes multiple treatments: rationale, concepts, examples. BMJ. 2017 13;358:j3932.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/Zhang.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Meta-analysis of the prognostic role of hormone receptors in endometrial cancer — Zhang","text":"","code":"data(Zhang) # Display the hazard ratios for overall survival in a forest plot ds <- subset(Zhang, outcome==\"OS\") with(ds, forest(theta = HR, theta.ci.lb = HR.025, theta.ci.ub = HR.975, theta.slab = Study, xlab = \"Hazard ratio of HER2 versus OS\", refline = 1))"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"Function display autocorrelation fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"","code":"acplot(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"generic function.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"Function display autocorrelation fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"","code":"# S3 method for mcmc.list acplot(x, P, nLags = 50, greek = FALSE, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"x object class \"mcmc.list\" P Optional dataframe describing parameters plot respective names nLags Integer indicating number lags autocorrelation plot. greek Logical value indicating whether parameter labels parsed get Greek letters. Defaults FALSE. ... Additional arguments passed ggs_autocorrelation","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.mcmc.list.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.mcmc.list","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"Function display autocorrelation fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"","code":"# S3 method for uvmeta acplot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"x object class \"valmeta\" ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"Results displayed estimated mean (mu) standard-deviation (tau) meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.uvmeta","text":"","code":"if (FALSE) { data(Roberts) fit <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts), method=\"BAYES\")) acplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"Function display autocorrelation fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"","code":"# S3 method for valmeta acplot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"x object class \"valmeta\" ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"Results displayed estimated mean (mu) standard-deviation (tau) meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/acplot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the autocorrelation of a Bayesian meta-analysis model — acplot.valmeta","text":"","code":"if (FALSE) { data(EuroSCORE) fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) acplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate the concordance statistic — ccalc","title":"Calculate the concordance statistic — ccalc","text":"function calculates (transformed versions ) concordance (c-) statistic corresponding sampling variance.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate the concordance statistic — ccalc","text":"","code":"ccalc( cstat, cstat.se, cstat.cilb, cstat.ciub, cstat.cilv, sd.LP, N, O, Po, data, slab, subset, g = NULL, level = 0.95, approx.se.method = 4, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate the concordance statistic — ccalc","text":"cstat vector specify estimated c-statistics. cstat.se Optional vector specify corresponding standard errors. cstat.cilb Optional vector specify lower limits confidence interval. cstat.ciub Optional vector specify upper limits confidence interval. cstat.cilv Optional vector specify levels aformentioned confidence interval limits. (default: 0.95, corresponds 95% confidence interval). sd.LP Optional vector specify standard deviations linear predictor (prognostic index). N Optional vector specify sample/group sizes. O Optional vector specify total number observed events. Po Optional vector specify observed event probabilities. data Optional data frame containing variables given arguments . slab Optional vector labels studies. subset Optional vector indicating subset studies used. can logical vector numeric vector indicating indices studies include. g quoted string function transform estimates c-statistic; see details . level Optional numeric specify level confidence interval, default 0.95. approx.se.method integer specifying method used estimating standard error c-statistic (Newcombe, 2006). far, method 2 method 4 (default) implemented. ... Additional arguments.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate the concordance statistic — ccalc","text":"object class c(\"mm_perf\",\"data.frame\") following columns: \"theta\" (transformed) c-statistics. \"theta.se\" Standard errors (transformed) c-statistics. \"theta.cilb\" Lower confidence interval (transformed) c-statistics. level specified level. Intervals calculated scale theta assuming Normal distribution. \"theta.ciub\" Upper confidence interval (transformed) c-statistics. level specified level. Intervals calculated scale theta assuming Normal distribution. \"theta.source\" Method used calculating (transformed) c-statistic. \"theta.se.source\" Method used calculating standard error (transformed) c-statistic.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Calculate the concordance statistic — ccalc","text":"c-statistic measure discrimination, indicates ability prediction model distinguish patients developing developing outcome. c-statistic typically ranges 0.5 (discriminative ability) 1 (perfect discriminative ability). default, function ccalc derive c-statistic study, together corresponding standard error 95% confidence interval. However, also possible calculate transformed versions c-statistic. Appropriate standard errors derived using Delta method. instance, logit transformation can applied specifying g=\"log(cstat/(1-cstat))\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"restoring-the-c-statistic","dir":"Reference","previous_headings":"","what":"Restoring the c-statistic","title":"Calculate the concordance statistic — ccalc","text":"studies c-statistic missing, estimated standard deviation linear predictor (theta.source=\"std.dev(LP)\"). corresponding method described White et al. (2015).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"restoring-the-standard-error-of-the-c-statistic","dir":"Reference","previous_headings":"","what":"Restoring the standard error of the c-statistic","title":"Calculate the concordance statistic — ccalc","text":"missing, standard error c-statistic can estimated confidence interval. Alternatively, standard error can approximated combination reported c-statistic, total sample size total number events (Newcombe, 2006). can achieved adopting (modification ) method proposed Hanley McNeil, specified approx.se.method.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Calculate the concordance statistic — ccalc","text":"Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2018; press. Hanley JA, McNeil BJ. meaning use area receiver operating characteristic (ROC) curve. Radiology. 1982; 143(1):29--36. Newcombe RG. Confidence intervals effect size measure based Mann-Whitney statistic. Part 2: asymptotic methods evaluation. Stat Med. 2006; 25(4):559--73. Snell KI, Ensor J, Debray TP, Moons KG, Riley RD. Meta-analysis prediction model performance across multiple studies: scale helps ensure -study normality C -statistic calibration measures? Statistical Methods Medical Research. 2017. White IR, Rapsomaniki E, Emerging Risk Factors Collaboration. Covariate-adjusted measures discrimination survival data. Biom J. 2015;57(4):592--613.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Calculate the concordance statistic — ccalc","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ccalc.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Calculate the concordance statistic — ccalc","text":"","code":"######### Validation of prediction models with a binary outcome ######### data(EuroSCORE) # Calculate the c-statistic and its standard error est1 <- ccalc(cstat = c.index, cstat.se = se.c.index, cstat.cilb = c.index.95CIl, cstat.ciub = c.index.95CIu, N = n, O = n.events, data = EuroSCORE, slab = Study) est1 #> theta theta.se theta.cilb theta.ciub theta.source #> Nashef 0.8095 0.01377576 0.7820000 0.8360000 c-statistic #> Biancari 0.8670 0.03520473 0.7980000 0.9360000 c-statistic #> Di Dedda 0.8100 0.03571494 0.7400000 0.8800000 c-statistic #> Chalmers 0.7900 0.01000000 0.7704004 0.8095996 c-statistic #> Grant 0.8080 0.00800000 0.7923203 0.8236797 c-statistic #> Carneo 0.8500 0.01000000 0.8304004 0.8695996 c-statistic #> Kunt 0.7200 0.05100000 0.6200418 0.8199582 c-statistic #> Kirmani 0.8180 0.00700000 0.8042803 0.8317197 c-statistic #> Howell 0.6700 0.02972796 0.6117343 0.7282657 c-statistic #> Wang 0.7200 0.01500000 0.6906005 0.7493995 c-statistic #> Borde 0.7200 0.09044574 0.5427296 0.8972704 c-statistic #> Qadir 0.8400 0.02338189 0.7941723 0.8858277 c-statistic #> Spiliopoulos 0.7700 0.06700000 0.6386824 0.9013176 c-statistic #> Wendt 0.7200 0.03400000 0.6533612 0.7866388 c-statistic #> Laurent 0.7700 0.06100000 0.6504422 0.8895578 c-statistic #> Wang.1 0.6420 0.07100000 0.5028426 0.7811574 c-statistic #> Nishida 0.7697 0.04247895 0.6864428 0.8529572 c-statistic #> Barilli 0.8000 0.01500000 0.7706005 0.8293995 c-statistic #> Barilli.1 0.8200 0.02000000 0.7808007 0.8591993 c-statistic #> Paparella 0.8300 0.01200000 0.8064804 0.8535196 c-statistic #> Carosella 0.7600 0.05600000 0.6502420 0.8697580 c-statistic #> Borracci 0.8560 0.03300000 0.7913212 0.9206788 c-statistic #> Osnabrugge 0.7700 0.01000000 0.7504004 0.7895996 c-statistic #> theta.se.source #> Nashef Confidence Interval #> Biancari Confidence Interval #> Di Dedda Confidence Interval #> Chalmers Standard Error #> Grant Standard Error #> Carneo Standard Error #> Kunt Standard Error #> Kirmani Standard Error #> Howell Newcombe (Method 4) #> Wang Standard Error #> Borde Newcombe (Method 4) #> Qadir Newcombe (Method 4) #> Spiliopoulos Standard Error #> Wendt Standard Error #> Laurent Standard Error #> Wang.1 Standard Error #> Nishida Newcombe (Method 4) #> Barilli Standard Error #> Barilli.1 Standard Error #> Paparella Standard Error #> Carosella Standard Error #> Borracci Standard Error #> Osnabrugge Standard Error # Calculate the logit c-statistic and its standard error est2 <- ccalc(cstat = c.index, cstat.se = se.c.index, cstat.cilb = c.index.95CIl, cstat.ciub = c.index.95CIu, N = n, O = n.events, data = EuroSCORE, slab = Study, g = \"log(cstat/(1-cstat))\") est2 #> theta theta.se theta.cilb theta.ciub theta.source #> Nashef 1.4467646 0.08964514 1.27735968 1.6287622 c-statistic #> Biancari 1.8746898 0.33390703 1.37384090 2.6827324 c-statistic #> Di Dedda 1.4500102 0.24144872 1.04596856 1.9924302 c-statistic #> Chalmers 1.3249254 0.06027728 1.20678413 1.4430667 c-statistic #> Grant 1.4370667 0.05156766 1.33599594 1.5381374 c-statistic #> Carneo 1.7346011 0.07843137 1.58087839 1.8883237 c-statistic #> Kunt 0.9444616 0.25297619 0.44863739 1.4402858 c-statistic #> Kirmani 1.5028556 0.04701900 1.41070011 1.5950112 c-statistic #> Howell 0.7081851 0.13445481 0.44465847 0.9717116 c-statistic #> Wang 0.9444616 0.07440476 0.79863096 1.0902923 c-statistic #> Borde 0.9444616 0.44863958 0.06514418 1.8237790 c-statistic #> Qadir 1.6582281 0.17397241 1.31724841 1.9992077 c-statistic #> Spiliopoulos 1.2083112 0.37831733 0.46682285 1.9497996 c-statistic #> Wendt 0.9444616 0.16865079 0.61391213 1.2750111 c-statistic #> Laurent 1.2083112 0.34443817 0.53322480 1.8833976 c-statistic #> Wang.1 0.5840553 0.30891592 -0.02140877 1.1895194 c-statistic #> Nishida 1.2066180 0.23963947 0.73693330 1.6763027 c-statistic #> Barilli 1.3862944 0.09375000 1.20254774 1.5700410 c-statistic #> Barilli.1 1.5163475 0.13550136 1.25076971 1.7819253 c-statistic #> Paparella 1.5856273 0.08504607 1.41894004 1.7523145 c-statistic #> Carosella 1.1526795 0.30701754 0.55093618 1.7544228 c-statistic #> Borracci 1.7824571 0.26771807 1.25773930 2.3071748 c-statistic #> Osnabrugge 1.2083112 0.05646527 1.09764130 1.3189811 c-statistic #> theta.se.source #> Nashef Confidence Interval #> Biancari Confidence Interval #> Di Dedda Confidence Interval #> Chalmers Standard Error #> Grant Standard Error #> Carneo Standard Error #> Kunt Standard Error #> Kirmani Standard Error #> Howell Newcombe (Method 4) #> Wang Standard Error #> Borde Newcombe (Method 4) #> Qadir Newcombe (Method 4) #> Spiliopoulos Standard Error #> Wendt Standard Error #> Laurent Standard Error #> Wang.1 Standard Error #> Nishida Newcombe (Method 4) #> Barilli Standard Error #> Barilli.1 Standard Error #> Paparella Standard Error #> Carosella Standard Error #> Borracci Standard Error #> Osnabrugge Standard Error # Display the results of all studies in a forest plot plot(est1)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":null,"dir":"Reference","previous_headings":"","what":"Convert a correlation matrix into a covariance matrix — cor2cov","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"Convert correlation matrix covariance matrix","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"","code":"cor2cov(sigma, cormat)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"sigma vector standard deviations. order standard deviations correspond column order 'cormat'. cormat symmetric numeric correlation matrix","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"covariance matrix","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/cor2cov.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Convert a correlation matrix into a covariance matrix — cor2cov","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":null,"dir":"Reference","previous_headings":"","what":"Posterior distribution of estimated model parameters — dplot","title":"Posterior distribution of estimated model parameters — dplot","text":"Generate plot posterior distribution","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Posterior distribution of estimated model parameters — dplot","text":"","code":"dplot(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Posterior distribution of estimated model parameters — dplot","text":"... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Posterior distribution of estimated model parameters — dplot","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Posterior distribution of estimated model parameters — dplot","text":"generic function.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Posterior distribution of estimated model parameters — dplot","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":null,"dir":"Reference","previous_headings":"","what":"Posterior distribution of estimated model parameters — dplot.mcmc.list","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"Generate plot posterior distribution","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"","code":"# S3 method for mcmc.list dplot(x, P, plot_type = \"dens\", ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"x object class \"mcmc.list\" P Optional dataframe describing parameters plot respective names plot_type Optional character string specify whether density plot (\"dens\") histogram (\"hist\") displayed. ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.mcmc.list.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Posterior distribution of estimated model parameters — dplot.mcmc.list","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"Function generate plots prior posterior distribution Bayesian meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"","code":"# S3 method for uvmeta dplot(x, par, distr_type, plot_type = \"dens\", ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"x object class \"uvmeta\" par Character string specify parameter plot generated. Options \"mu\" (mean random effects model) \"tau\" (standard deviation random effects model). distr_type Character string specify whether prior distribution (\"prior\") posterior distribution (\"posterior\") displayed. plot_type Character string specify whether density plot (\"dens\") histogram (\"hist\") displayed. ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.uvmeta","text":"","code":"if (FALSE) { data(Roberts) fit <- with(Roberts, uvmeta(r=SDM, r.se=SE, method=\"BAYES\")) dplot(fit) dplot(fit, distr_type = \"posterior\") dplot(fit, par = \"tau\", distr_type = \"prior\") dplot(fit, plot_type = \"hist\") }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"Function generate plots prior posterior distribution Bayesian meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"","code":"# S3 method for valmeta dplot(x, par, distr_type, plot_type = \"dens\", ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"x object class \"valmeta\" par Character string specify parameter plot generated. Options \"mu\" (mean random effects model) \"tau\" (standard deviation random effects model). distr_type Character string specify whether prior distribution (\"prior\") posterior distribution (\"posterior\") displayed. plot_type Character string specify whether density plot (\"dens\") histogram (\"hist\") displayed. ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/dplot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the prior and posterior distribution of a meta-analysis model — dplot.valmeta","text":"","code":"if (FALSE) { data(EuroSCORE) # Meta-analysis of the concordance statistic fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) dplot(fit) dplot(fit, distr_type = \"posterior\") dplot(fit, par = \"tau\", distr_type = \"prior\") # Meta-analysis of the O:E ratio EuroSCORE.new <- EuroSCORE EuroSCORE.new$n[c(1, 2, 5, 10, 20)] <- NA pars <- list(hp.tau.dist=\"dhalft\", # Prior for the between-study standard deviation hp.tau.sigma=1.5, # Standard deviation for 'hp.tau.dist' hp.tau.df=3, # Degrees of freedom for 'hp.tau.dist' hp.tau.max=10) # Maximum value for the between-study standard deviation fit2 <- valmeta(measure=\"OE\", O=n.events, E=e.events, N=n, data=EuroSCORE.new, method=\"BAYES\", slab=Study, pars=pars) dplot(fit2, plot_type = \"hist\") }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":null,"dir":"Reference","previous_headings":"","what":"Regression tests for detecting funnel plot asymmetry — fat","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"presence small-study effects common threat systematic reviews meta-analyses, especially due publication bias, occurs small primary studies likely reported (published) findings positive. presence small-study effects can verified visual inspection funnel plot, included study meta-analysis, estimate reported effect size depicted measure precision sample size. premise scatter plots reflect funnel shape, small-study effects exist. However, small studies predominately one direction (usually direction larger effect sizes), asymmetry ensue. fat function implements several tests detecting funnel plot asymmetry, can used presence -study heterogeneity treatment effect relatively low.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"","code":"fat(b, b.se, n.total, d.total, d1, d2, method = \"E-FIV\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"b Vector effect size study. Examples log odds ratio, log hazards ratio, log relative risk. b.se Optional vector standard error effect size study n.total Optional vector total sample size study d.total Optional vector total number observed events study d1 Optional vector total number observed events exposed groups d2 Optional vector total number observed events unexposed groups method Method testing funnel plot asymmetry, defaults \"E-FIV\" (Egger's test multiplicative dispersion). options E-UW, M-FIV, M-FPV, D-FIV D-FAV. info \"Details\"","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"list containing following entries: \"pval\" two-sided P-value indicating statistical significance funnel plot asymettry test. Values significance level (usually defined 10%) support presence funnel plot asymmetry, thus small-study effects. \"model\" fitted glm object, representing estimated regression model used testing funnel plot asymmetry.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"egger-regression-method","dir":"Reference","previous_headings":"","what":"Egger regression method","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"common approach test presence small-study effects estimate regression model standardized effect estimate (effect/SE) regressed measure precision (1/SE), (method=\"E-UW\", Egger 1997). possible allow -study heterogeneity adopting multiplicative overdispersion parameter variance study multiplied (method=\"E-FIV\", Sterne 2000). Unfortunately, demonstrated aforementioned two tests biased : () independent variable subject sampling variability; (ii) standardized treatment effect correlated estimated precision; (iii) binary data, independent regression variable biased estimate true precision, larger bias smaller sample sizes (Macaskill et al. 2001).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"macaskill-regression-method","dir":"Reference","previous_headings":"","what":"Macaskill regression method","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"standard approach estimates regression model effect size function study size (method=\"M-FIV\", Macaskill et al. 2001). study weighted precision treatment effect estimate allow possible heteroscedasticity. alternative approach weight study pooled' estimate outcome proportion (method=\"M-FPV\") studies zero events, continuity correction applied adding 0.5 cell counts.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"peters-regression-method","dir":"Reference","previous_headings":"","what":"Peters regression method","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"approach (method=\"P-FPV\") estimates regression model treatment effect function inverse total sample size (Peters et al. 2006). studies zero events, continuity correction applied adding 0.5 cell counts.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"debray-regression-method","dir":"Reference","previous_headings":"","what":"Debray regression method","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"approach proposed survival data, uses total number events independent variable weighted regression model (Debray et al. 2017). study weights based inverse variance (method=\"D-FIV\") approximation thereof (method=\"D-FAV\").","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"Debray TPA, Moons KGM, Riley RD. Detecting small-study effects funnel plot asymmetry meta-analysis survival data: comparison new existing tests. Res Syn Meth. 2018;9(1):41--50. Egger M, Davey Smith G, Schneider M, Minder C. Bias meta-analysis detected simple, graphical test. BMJ. 1997;315(7109):629--34. Macaskill P, Walter SD, Irwig L. comparison methods detect publication bias meta-analysis. Stat Med. 2001;20(4):641--54. Peters JL, Sutton AJ, Jones DR, Abrams KR, Rushton L. Comparison two methods detect publication bias meta-analysis. JAMA. 2006 Feb 8;295(6):676--80. Sterne JA, Gavaghan D, Egger M. Publication related bias meta-analysis: power statistical tests prevalence literature. J Clin Epidemiol. 2000;53(11):1119--29.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fat.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Regression tests for detecting funnel plot asymmetry — fat","text":"","code":"data(Fibrinogen) b <- log(Fibrinogen$HR) b.se <- ((log(Fibrinogen$HR.975) - log(Fibrinogen$HR.025))/(2*qnorm(0.975))) n.total <- Fibrinogen$N.total d.total <- Fibrinogen$N.events fat(b=b, b.se=b.se) #> Call: fat(b = b, b.se = b.se) #> #> Fixed effect summary estimate: 0.4186 #> #> test for funnel plot asymmetry: t =1.9021, df = 29, p = 0.0671 fat(b=b, b.se=b.se, d.total=d.total, method=\"D-FIV\") #> Call: fat(b = b, b.se = b.se, d.total = d.total, method = \"D-FIV\") #> #> Fixed effect summary estimate: 0.4186 #> #> test for funnel plot asymmetry: t =1.6847, df = 29, p = 0.1028 # Note that many tests are also available via metafor require(metafor) #> Loading required package: metafor #> Loading required package: Matrix #> Loading required package: metadat #> Loading required package: numDeriv #> #> Loading the 'metafor' package (version 4.4-0). For an #> introduction to the package please type: help(metafor) #> #> Attaching package: ‘metafor’ #> The following object is masked from ‘package:metamisc’: #> #> forest fat(b=b, b.se=b.se, n.total=n.total, method=\"M-FIV\") #> Call: fat(b = b, b.se = b.se, n.total = n.total, method = \"M-FIV\") #> #> Fixed effect summary estimate: 0.4186 #> #> test for funnel plot asymmetry: t =-1.4275, df = 29, p = 0.1641 regtest(x=b, sei=b.se, ni=n.total, model=\"lm\", predictor=\"ni\") #> #> Regression Test for Funnel Plot Asymmetry #> #> Model: weighted regression with multiplicative dispersion #> Predictor: sample size #> #> Test for Funnel Plot Asymmetry: t = -1.4275, df = 29, p = 0.1641 #>"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract Model Fitted Values — fitted.metapred","title":"Extract Model Fitted Values — fitted.metapred","text":"Extract fitted values metapred object. default returns fitted values model cross-validation procedure.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract Model Fitted Values — fitted.metapred","text":"","code":"# S3 method for metapred fitted( object, select = \"cv\", step = NULL, model = NULL, as.stratified = TRUE, type = \"response\", ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract Model Fitted Values — fitted.metapred","text":"object object class metapred select character. Select fitted values \"cv\" (default) \"global\" model. step character numeric. Name number step select select = \"cv\". Defaults best step. model character numeric. Name number model select select = \"cv\". Defaults best model. .stratified logical. select = \"cv\" determines whether returned predictions stratified list (TRUE, default) original order (FALSE). type character. Type fitted value. ... compatibility .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Extract Model Fitted Values — fitted.metapred","text":"Function still development, use caution. returns type = \"response\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/fitted.metapred.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Extract Model Fitted Values — fitted.metapred","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest plot — forest.default","title":"Forest plot — forest.default","text":"Generate forest plot specifying various effect sizes, confidence intervals summary estimate.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest plot — forest.default","text":"","code":"# S3 method for default forest( theta, theta.ci.lb, theta.ci.ub, theta.slab, theta.summary, theta.summary.ci.lb, theta.summary.ci.ub, theta.summary.pi.lb, theta.summary.pi.ub, title, sort = \"asc\", theme = theme_bw(), predint.linetype = 1, xlim, xlab = \"\", refline = 0, label.summary = \"Summary Estimate\", label.predint = \"Prediction Interval\", nrows.before.summary = 1, study.digits = 2, study.shape = 15, col.diamond = \"white\", col.predint = \"black\", size.study = 0.5, size.predint = 1, lty.ref = \"dotted\", ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest plot — forest.default","text":"theta Numeric vector effect size study theta.ci.lb Numeric vector specifying lower bound confidence interval effect sizes theta.ci.ub Numeric vector specifying upper bound confidence interval effect sizes theta.slab Character vector specifying study labels theta.summary Meta-analysis summary estimate effect sizes theta.summary.ci.lb Lower bound confidence (credibility) interval summary estimate theta.summary.ci.ub Upper bound confidence (credibility) interval summary estimate theta.summary.pi.lb Lower bound (approximate) prediction interval summary estimate. theta.summary.pi.ub Upper bound (approximate) prediction interval summary estimate. title Title forest plot sort default, studies sorted ascending effect size (sort=\"asc\"). Set \"desc\" sorting reverse order, value ignore sorting. theme Theme generate forest plot. default, classic dark--light ggplot2 theme used. See ggtheme information. predint.linetype linetype prediction interval xlim x limits (x1, x2) forest plot xlab Optional character string specifying X label refline Optional numeric specifying reference line label.summary Optional character string specifying label summary estimate label.predint Optional character string specifying label (approximate) prediction interval nrows..summary many empty rows introduced study results summary estimates study.digits many significant digits used print stuy results study.shape Plotting symbol use study results. default, filled square used. col.diamond filling color diamond representing summary estimate. E.g. \"red\", \"blue\", hex color code (\"#2e8aff\") col.predint Line color prediction interval. E.g. \"red\", \"blue\", hex color code (\"#2e8aff\") size.study Line width study results mm size.predint Line width prediction interval mm lty.ref Line type reference line ... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Forest plot — forest.default","text":"object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.default.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest plot — forest.default","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest plot — forest","title":"Forest plot — forest","text":"Generate forest plot specifying various effect sizes, confidence intervals summary estimate.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest plot — forest","text":"","code":"forest(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest plot — forest","text":"... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Forest plot — forest","text":"generic function. See forest.default making forest plots summary statistics, forest.metapred plotting metapred objects, forest.mp.cv.val plotting mp.cv.val objects.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest plot — forest","text":"Thomas Debray Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest plot of a metapred fit — forest.metapred","title":"Forest plot of a metapred fit — forest.metapred","text":"Draw forest plot performance internally-externally cross-validated model. default final model shown.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest plot of a metapred fit — forest.metapred","text":"","code":"# S3 method for metapred forest(x, perfFUN = 1, step = NULL, method = \"REML\", model = NULL, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest plot of a metapred fit — forest.metapred","text":"x metapred fit object perfFUN Numeric character. performance statistic plotted? Defaults first. step step plotted? Defaults best step. numeric converted name step: 0 unchanged model, 1 first change... method character string specifying whether fixed- random-effects model used summarize prediction model performance. fixed-effects model fitted using method=\"FE\". Random-effects models fitted setting method equal one following: \"DL\", \"\", \"SJ\", \"ML\", \"REML\", \"EB\", \"HS\", \"GENQ\". Default \"REML\". model model change plotted? NULL (default, best change) character name variable (integer) index model change. ... arguments passed plotting internals. E.g. title. See forest.default details.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest plot of a metapred fit — forest.metapred","text":"Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.metapred.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Forest plot of a metapred fit — forest.metapred","text":"","code":"data(DVTipd) # Internal-external cross-validation of a pre-specified model 'f' f <- dvt ~ histdvt + ddimdich + sex + notraum fit <- metapred(DVTipd, strata = \"study\", formula = f, scope = f, family = binomial) # Display the model's external performance (expressed as mean squared error by default) # for each study forest(fit) #> Error in forest.default(fit): Must specify either 'vi', 'sei', or ('ci.lb', 'ci.ub') pairs."},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.mp.cv.val.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest plot of a validation object. — forest.mp.cv.val","title":"Forest plot of a validation object. — forest.mp.cv.val","text":"Draw forest plot performance internally-externally cross-validated model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.mp.cv.val.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest plot of a validation object. — forest.mp.cv.val","text":"","code":"# S3 method for mp.cv.val forest(x, perfFUN = 1, method = \"REML\", xlab = NULL, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.mp.cv.val.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest plot of a validation object. — forest.mp.cv.val","text":"x mp.cv.val perf object. perfFUN Numeric character. performance statistic plotted? Defaults first. method character string specifying whether fixed- random-effects model used summarize prediction model performance. fixed-effects model fitted using method=\"FE\". Random-effects models fitted setting method equal one following: \"DL\", \"\", \"SJ\", \"ML\", \"REML\", \"EB\", \"HS\", \"GENQ\". Default \"REML\". xlab Label x-axis. Defaults name performance function. ... arguments passed plotting internals. E.g. title. See forest.default details.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/forest.mp.cv.val.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest plot of a validation object. — forest.mp.cv.val","text":"Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":null,"dir":"Reference","previous_headings":"","what":"Gelman-Rubin-Brooks plot — gelmanplot","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"plot shows evolution Gelman Rubin's shrink factor number iterations increases. code adapted R package coda.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"","code":"gelmanplot(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Gelman-Rubin-Brooks plot — gelmanplot","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":null,"dir":"Reference","previous_headings":"","what":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"plot shows evolution Gelman Rubin's shrink factor number iterations increases. code adapted R package coda.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"","code":"# S3 method for mcmc.list gelmanplot( x, P, confidence = 0.95, max.bins = 50, autoburnin = TRUE, greek = FALSE, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"x mcmc object P Optional dataframe describing parameters plot respective names confidence coverage probability confidence interval potential scale reduction factor max.bins Maximum number bins, excluding last one. autoburnin Logical flag indicating whether second half series used computation. set TRUE (default) start(x) less end(x)/2 start series adjusted second half series used. greek Logical value indicating whether parameter labels parsed get Greek letters. Defaults false. ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.mcmc.list.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Gelman-Rubin-Brooks plot — gelmanplot.mcmc.list","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"plot shows evolution Gelman Rubin's shrink factor number iterations increases. code adapted R package coda.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"","code":"# S3 method for uvmeta gelmanplot(x, confidence = 0.95, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"x mcmc object confidence coverage probability confidence interval potential scale reduction factor ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Gelman-Rubin-Brooks plot — gelmanplot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"plot shows evolution Gelman Rubin's shrink factor number iterations increases. code adapted R package coda.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"","code":"# S3 method for valmeta gelmanplot(x, confidence = 0.95, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"x mcmc object confidence coverage probability confidence interval potential scale reduction factor ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gelmanplot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Gelman-Rubin-Brooks plot — gelmanplot.valmeta","text":"","code":"if (FALSE) { data(EuroSCORE) # Meta-analysis of the concordance statistic fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) gelmanplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":null,"dir":"Reference","previous_headings":"","what":"Generalizability estimates — gen","title":"Generalizability estimates — gen","text":"Obtain generalizability estimates model fit.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Generalizability estimates — gen","text":"","code":"gen(object, ...) generalizability(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Generalizability estimates — gen","text":"object model fit object, either metapred subset(metapred) object. ... default, final model selected. parameter allows arguments passed subset.metapred generalizability estimates steps/models may returned..","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Generalizability estimates — gen","text":"named values indices parameter genFUN one estimates generalizability can selected. Use genFUN = 0 select .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/gen.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Generalizability estimates — gen","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":null,"dir":"Reference","previous_headings":"","what":"IMPACT data — impact","title":"IMPACT data — impact","text":"IMPACT dataset comprises 15 studies patients suffering traumatic brain injury, including individual patient data 11 randomized controlled trials four observational studies.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"IMPACT data — impact","text":"","code":"data(\"impact\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"IMPACT data — impact","text":"data frame 11022 observations following 11 variables. name Name study type Type study, RCT: randomized controlled trial,OBS: observational cohort age Age patient motor_score Glasgow Coma Scale motor score pupil Pupillary reactivity ct Marshall Computerized Tomography classification hypox Hypoxia (0=, 1=yes) hypots Hypotension (0=, 1=yes) tsah Traumatic subarachnoid hemorrhage (0=, 1=yes) edh Epidural hematoma (0=, 1=yes) mort 6-month mortality (0=alive, 1=dead)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"IMPACT data — impact","text":"included studies part IMPACT project, total 25 prognostic factors considered prediction 6-month mortality. Missing values imputed using study fixed effect imputation model (Steyerberg et al, 2008).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"IMPACT data — impact","text":"Steyerberg EW, Nieboer D, Debray TPA, Van Houwelingen JC. Assessment heterogeneity individual participant data meta-analysis prediction models: overview illustration. Stat Med. 2019;38(22):4290--309.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"IMPACT data — impact","text":"Murray GD, Butcher , McHugh GS, et al. Multivariable prognostic analysis traumatic brain injury: results IMPACT study. J Neurotrauma. 2007;24(2):329--337. Steyerberg EW, Mushkudiani N, Perel P, et al. Predicting outcome traumatic brain injury: development international validation prognostic scores based admission characteristics. PLOS Med. 2008;5(8):e165.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impact.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"IMPACT data — impact","text":"","code":"data(impact) by(impact, impact$name, summary) #> impact$name: APOE #> name type age motor_score pupil ct #> APOE :756 OBS:756 Min. :14.00 1/2: 7 Both:638 I/II:377 #> CSTAT : 0 RCT: 0 1st Qu.:25.00 3 :331 None: 79 III :151 #> EBIC : 0 Median :37.00 4 : 18 One : 39 IV/V:228 #> HIT I : 0 Mean :41.14 5/6:400 #> HIT II : 0 3rd Qu.:56.00 #> NABIS : 0 Max. :93.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.000 Min. :0.00000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.000 1st Qu.:0.00000 #> Median :0.0000 Median :0.0000 Median :0.000 Median :0.00000 #> Mean :0.2712 Mean :0.1071 Mean :0.254 Mean :0.06614 #> 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:1.000 3rd Qu.:0.00000 #> Max. :1.0000 Max. :1.0000 Max. :1.000 Max. :1.00000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.1548 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: CSTAT #> name type age motor_score pupil ct #> CSTAT :517 OBS: 0 Min. :15.00 1/2:128 Both:368 I/II:212 #> APOE : 0 RCT:517 1st Qu.:20.00 3 : 86 None: 63 III :105 #> EBIC : 0 Median :28.00 4 :180 One : 86 IV/V:200 #> HIT I : 0 Mean :31.79 5/6:123 #> HIT II : 0 3rd Qu.:41.00 #> NABIS : 0 Max. :68.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.1335 Mean :0.1663 Mean :0.4778 Mean :0.1161 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2224 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: EBIC #> name type age motor_score pupil ct #> EBIC :822 OBS:822 Min. :14.00 1/2:230 Both:531 I/II:333 #> APOE : 0 RCT: 0 1st Qu.:24.00 3 :198 None:209 III : 82 #> CSTAT : 0 Median :37.50 4 :113 One : 82 IV/V:407 #> HIT I : 0 Mean :41.79 5/6:281 #> HIT II : 0 3rd Qu.:59.00 #> NABIS : 0 Max. :92.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.00000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.00000 #> Mean :0.2871 Mean :0.2445 Mean :0.4136 Mean :0.09246 #> 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.00000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.00000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.3418 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: HIT I #> name type age motor_score pupil ct #> HIT I :350 OBS: 0 Min. :14.00 1/2:163 Both:232 I/II:127 #> APOE : 0 RCT:350 1st Qu.:21.00 3 : 54 None: 67 III : 79 #> CSTAT : 0 Median :34.00 4 : 56 One : 51 IV/V:144 #> EBIC : 0 Mean :35.49 5/6: 77 #> HIT II : 0 3rd Qu.:47.00 #> NABIS : 0 Max. :71.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.00 Min. :0.00000 Min. :0.0000 Min. :0.00 #> 1st Qu.:0.00 1st Qu.:0.00000 1st Qu.:0.0000 1st Qu.:0.00 #> Median :0.00 Median :0.00000 Median :0.0000 Median :0.00 #> Mean :0.18 Mean :0.05429 Mean :0.3257 Mean :0.18 #> 3rd Qu.:0.00 3rd Qu.:0.00000 3rd Qu.:1.0000 3rd Qu.:0.00 #> Max. :1.00 Max. :1.00000 Max. :1.0000 Max. :1.00 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2829 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: HIT II #> name type age motor_score pupil ct #> HIT II :819 OBS: 0 Min. :15.0 1/2:280 Both:583 I/II:345 #> APOE : 0 RCT:819 1st Qu.:22.0 3 :151 None:138 III : 90 #> CSTAT : 0 Median :33.0 4 :181 One : 98 IV/V:384 #> EBIC : 0 Mean :36.3 5/6:207 #> HIT I : 0 3rd Qu.:49.0 #> NABIS : 0 Max. :74.0 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.199 Mean :0.1001 Mean :0.3272 Mean :0.1636 #> 3rd Qu.:0.000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2295 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: NABIS #> name type age motor_score pupil ct #> NABIS :385 OBS: 0 Min. :14.00 1/2:144 Both:236 I/II: 45 #> APOE : 0 RCT:385 1st Qu.:21.00 3 : 65 None: 99 III :204 #> CSTAT : 0 Median :30.00 4 : 76 One : 50 IV/V:136 #> EBIC : 0 Mean :31.82 5/6:100 #> HIT I : 0 3rd Qu.:40.00 #> HIT II : 0 Max. :68.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.3299 Mean :0.1558 Mean :0.4779 Mean :0.1325 #> 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2623 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: PEG #> name type age motor_score pupil ct #> PEG :1510 OBS: 0 Min. :15.00 1/2:655 Both:787 I/II:576 #> APOE : 0 RCT:1510 1st Qu.:20.00 3 :165 None:564 III :345 #> CSTAT : 0 Median :27.00 4 :334 One :159 IV/V:589 #> EBIC : 0 Mean :30.44 5/6:356 #> HIT I : 0 3rd Qu.:38.00 #> HIT II : 0 Max. :86.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.00000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.00000 #> Mean :0.2199 Mean :0.1695 Mean :0.4113 Mean :0.09801 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.00000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.00000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2397 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: PHARMOS #> name type age motor_score pupil ct #> PHARMOS:856 OBS: 0 Min. :16.00 1/2:134 Both:667 I/II:429 #> APOE : 0 RCT:856 1st Qu.:23.00 3 :262 None: 37 III :205 #> CSTAT : 0 Median :33.00 4 :225 One :152 IV/V:222 #> EBIC : 0 Mean :35.01 5/6:235 #> HIT I : 0 3rd Qu.:45.25 #> HIT II : 0 Max. :66.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :1.000 Median :0.0000 #> Mean :0.2477 Mean :0.1542 Mean :0.597 Mean :0.1939 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.1694 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: SAP #> name type age motor_score pupil ct #> SAP :919 OBS: 0 Min. :15.00 1/2:264 Both:639 I/II:400 #> APOE : 0 RCT:919 1st Qu.:23.00 3 :146 None:155 III :151 #> CSTAT : 0 Median :32.00 4 :223 One :125 IV/V:368 #> EBIC : 0 Mean :35.63 5/6:286 #> HIT I : 0 3rd Qu.:47.00 #> HIT II : 0 Max. :71.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.1371 Mean :0.1523 Mean :0.4418 Mean :0.2057 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2307 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: SKB #> name type age motor_score pupil ct #> SKB :126 OBS: 0 Min. :16.00 1/2:56 Both:81 I/II:54 #> APOE : 0 RCT:126 1st Qu.:20.00 3 :31 None:28 III :40 #> CSTAT : 0 Median :27.00 4 :16 One :17 IV/V:32 #> EBIC : 0 Mean :30.61 5/6:23 #> HIT I : 0 3rd Qu.:39.00 #> HIT II : 0 Max. :70.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:1.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :1.0000 Median :0.0000 #> Mean :0.2937 Mean :0.1905 Mean :0.7857 Mean :0.1111 #> 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2698 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: SLIN #> name type age motor_score pupil ct #> SLIN :409 OBS: 0 Min. :15.00 1/2: 55 Both:316 I/II:154 #> APOE : 0 RCT:409 1st Qu.:21.00 3 : 91 None: 16 III : 94 #> CSTAT : 0 Median :28.00 4 :127 One : 77 IV/V:161 #> EBIC : 0 Mean :32.35 5/6:136 #> HIT I : 0 3rd Qu.:43.00 #> HIT II : 0 Max. :79.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.00000 Min. :0.0000 Min. :0.00 Min. :0.0000 #> 1st Qu.:0.00000 1st Qu.:0.0000 1st Qu.:1.00 1st Qu.:0.0000 #> Median :0.00000 Median :0.0000 Median :1.00 Median :0.0000 #> Mean :0.05868 Mean :0.1369 Mean :0.78 Mean :0.1516 #> 3rd Qu.:0.00000 3rd Qu.:0.0000 3rd Qu.:1.00 3rd Qu.:0.0000 #> Max. :1.00000 Max. :1.0000 Max. :1.00 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2298 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: TCDB #> name type age motor_score pupil ct #> TCDB :603 OBS:603 Min. :16.00 1/2:243 Both:299 I/II:219 #> APOE : 0 RCT: 0 1st Qu.:21.00 3 :105 None:249 III :119 #> CSTAT : 0 Median :26.00 4 :121 One : 55 IV/V:265 #> EBIC : 0 Mean :32.97 5/6:134 #> HIT I : 0 3rd Qu.:40.00 #> HIT II : 0 Max. :93.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :0.0000 Median :0.0000 #> Mean :0.1808 Mean :0.2371 Mean :0.4428 Mean :0.1045 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.4378 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: TINT #> name type age motor_score pupil ct #> TINT :1118 OBS: 0 Min. :14.00 1/2:141 Both:807 I/II:474 #> APOE : 0 RCT:1118 1st Qu.:21.00 3 :237 None:138 III :221 #> CSTAT : 0 Median :30.00 4 :327 One :173 IV/V:423 #> EBIC : 0 Mean :33.61 5/6:413 #> HIT I : 0 3rd Qu.:45.00 #> HIT II : 0 Max. :79.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :1.0000 Median :0.0000 #> Mean :0.1592 Mean :0.1404 Mean :0.5259 Mean :0.1655 #> 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2487 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: TIUS #> name type age motor_score pupil ct #> TIUS :1041 OBS: 0 Min. :14.00 1/2:152 Both:708 I/II:463 #> APOE : 0 RCT:1041 1st Qu.:23.00 3 :132 None:211 III :198 #> CSTAT : 0 Median :30.00 4 :300 One :122 IV/V:380 #> EBIC : 0 Mean :32.78 5/6:457 #> HIT I : 0 3rd Qu.:41.00 #> HIT II : 0 Max. :77.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.00 Min. :0.0000 Min. :0.00000 #> 1st Qu.:0.0000 1st Qu.:0.00 1st Qu.:0.0000 1st Qu.:0.00000 #> Median :0.0000 Median :0.00 Median :0.0000 Median :0.00000 #> Mean :0.2795 Mean :0.22 Mean :0.4284 Mean :0.08646 #> 3rd Qu.:1.0000 3rd Qu.:0.00 3rd Qu.:1.0000 3rd Qu.:0.00000 #> Max. :1.0000 Max. :1.00 Max. :1.0000 Max. :1.00000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.2161 #> 3rd Qu.:0.0000 #> Max. :1.0000 #> #> ------------------------------------------------------------ #> impact$name: UK4 #> name type age motor_score pupil ct #> UK4 :791 OBS:791 Min. :14.00 1/2:198 Both:433 I/II:271 #> APOE : 0 RCT: 0 1st Qu.:22.00 3 :231 None:243 III :155 #> CSTAT : 0 Median :36.00 4 :141 One :115 IV/V:365 #> EBIC : 0 Mean :39.64 5/6:221 #> HIT I : 0 3rd Qu.:55.00 #> HIT II : 0 Max. :87.00 #> (Other): 0 #> hypox hypots tsah edh #> Min. :0.0000 Min. :0.0000 Min. :0.0000 Min. :0.0000 #> 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000 #> Median :0.0000 Median :0.0000 Median :1.0000 Median :0.0000 #> Mean :0.2566 Mean :0.2642 Mean :0.5183 Mean :0.1429 #> 3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:0.0000 #> Max. :1.0000 Max. :1.0000 Max. :1.0000 Max. :1.0000 #> #> mort #> Min. :0.0000 #> 1st Qu.:0.0000 #> Median :0.0000 #> Mean :0.4539 #> 3rd Qu.:1.0000 #> Max. :1.0000 #> # Plot the distribution of age by study library(ggplot2) e <- ggplot(impact, aes(x = name, y = age)) e + geom_violin(aes(fill = type), trim = FALSE) + theme(axis.text.x = element_text(angle = 45)) + xlab(\"Study\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":null,"dir":"Reference","previous_headings":"","what":"Impute missing values by their conditional mean — impute_conditional_mean","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"function imputes missing values conditional mean","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"","code":"impute_conditional_mean(x, mu, Sigma)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"x vector observations, may missing (indicated NA) mu vector population means 'x'. missing values allowed . Sigma matrix describing population covariance 'x'","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"vector missing values 'x' replaced conditional mean","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/impute_conditional_mean.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Impute missing values by their conditional mean — impute_conditional_mean","text":"","code":"# Define the population means mu <- c(0, 1, 2) # Define the covariance of the population Sigma <- diag(1,3) Sigma[1,2] <- Sigma[2,1] <- 0.3 Sigma[2,3] <- Sigma[3,2] <- 0.1 Sigma[1,3] <- Sigma[3,1] <- -0.2 # Generate a 'random' sample from the population that is partially observed x <- c(NA, 2, 4) # Impute the missing values impute_conditional_mean (x=x, mu=mu, Sigma=Sigma) #> [1] -0.1414141 2.0000000 4.0000000"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":null,"dir":"Reference","previous_headings":"","what":"Apply the inverse logit tranformation — inv.logit","title":"Apply the inverse logit tranformation — inv.logit","text":"Transforms linear predictor probability.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Apply the inverse logit tranformation — inv.logit","text":"","code":"inv.logit(x)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Apply the inverse logit tranformation — inv.logit","text":"x vector numerics (-Inf Inf)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Apply the inverse logit tranformation — inv.logit","text":"vector numerics 0 1.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/inv.logit.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Apply the inverse logit tranformation — inv.logit","text":"Thomas Debray ","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Print the log-likelihood — logLik.riley","title":"Print the log-likelihood — logLik.riley","text":"function provides (restricted) log-likelihood fitted model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Print the log-likelihood — logLik.riley","text":"","code":"# S3 method for riley logLik(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Print the log-likelihood — logLik.riley","text":"object riley object, representing fitted alternative model bivariate random-effects meta-analysis within-study correlations unknown. ... Additional arguments passed functions, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Print the log-likelihood — logLik.riley","text":"Returns object class logLik. (restricted) log-likelihood model represented object evaluated estimated coefficients. contains least one attribute, \"df\" (degrees freedom), giving number (estimated) parameters model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Print the log-likelihood — logLik.riley","text":"Riley RD, Thompson JR, Abrams KR. alternative model bivariate random-effects meta-analysis within-study correlations unknown. Biostatistics 2008; 9: 172--186.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Print the log-likelihood — logLik.riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logLik.riley.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Print the log-likelihood — logLik.riley","text":"","code":"data(Daniels) fit <- riley(Daniels,control=list(maxit=10000)) logLik(fit) #> 'log Lik.' -48.85119 (df=5)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":null,"dir":"Reference","previous_headings":"","what":"Apply logit tranformation — logit","title":"Apply logit tranformation — logit","text":"Transforms values 0 1 values -Inf Inf.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Apply logit tranformation — logit","text":"","code":"logit(x)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Apply logit tranformation — logit","text":"x vector numerics (0 1)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Apply logit tranformation — logit","text":"vector numerics (-Inf Inf).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/logit.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Apply logit tranformation — logit","text":"Thomas Debray ","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":null,"dir":"Reference","previous_headings":"","what":"Random effects meta-analysis — ma","title":"Random effects meta-analysis — ma","text":"Meta-analysis performance coefficients metapred object. Caution: still development.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Random effects meta-analysis — ma","text":"","code":"ma(object, method, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Random effects meta-analysis — ma","text":"object model fit object, metapred object. method Character, method meta-analysis passed valmeta uvmeta. Defaults \"REML\". ... arguments passed metapred, valmeta uvmeta.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Random effects meta-analysis — ma","text":"Produces different object types depending input.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/ma.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Random effects meta-analysis — ma","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metamisc-package.html","id":null,"dir":"Reference","previous_headings":"","what":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","text":"Facilitate frequentist Bayesian meta-analysis diagnosis prognosis research studies. includes functions summarize multiple estimates prediction model discrimination calibration performance (Debray et al., 2019) . also includes functions evaluate funnel plot asymmetry (Debray et al., 2018) . Finally, package provides functions developing multivariable prediction models datasets clustering (de Jong et al., 2021) .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metamisc-package.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","text":"following functionality currently implemented: univariate meta-analysis summary data (uvmeta), bivariate meta-analysis correlated outcomes (riley), meta-analysis prediction model performance (valmeta), evaluation funnel plot asymmetry (fat). metamisc package also provides comprehensive framework developing prediction models patient-level data multiple studies settings available (metapred).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metamisc-package.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","text":"Thomas Debray , Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metamisc-package.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Meta-Analysis of Diagnosis and Prognosis Research Studies — metamisc-package","text":"de Jong VMT, Moons KGM, Eijkemans MJC, Riley RD, Debray TPA. Developing generalizable prediction models pooled studies large clustered data sets. Stat Med. 2021;40(15):3533--59. Debray TPA, Moons KGM, Ahmed , Koffijberg H, Riley RD. framework developing, implementing, evaluating clinical prediction models individual participant data meta-analysis. Stat Med. 2013;32(18):3158--80. Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2019 Sep;28(9):2768--86. Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. Debray TPA, Moons KGM, Riley RD. Detecting small-study effects funnel plot asymmetry meta-analysis survival data: comparison new existing tests. Res Syn Meth. 2018;9(1):41--50. Riley RD, Moons K, Snell KIE, Ensor J, Hooft L, Altman D, et al. guide systematic review meta-analysis prognostic factor studies. BMJ. 2019;364:k4597. Riley RD, Tierney JF, Stewart LA. Individual participant data meta-analysis: handbook healthcare research. Hoboken, NJ: Wiley; 2021. ISBN: 978-1-119-33372-2. Schmid CH, Stijnen T, White IR. Handbook meta-analysis. First edition. Boca Raton: Taylor Francis; 2020. ISBN: 978-1-315-11940-3. Steyerberg EW, Nieboer D, Debray TPA, Van Houwelingen JC. Assessment heterogeneity individual participant data meta-analysis prediction models: overview illustration. Stat Med. 2019;38(22):4290--309.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":null,"dir":"Reference","previous_headings":"","what":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"Generalized stepwise regression obtaining prediction model validated (stepwise) internal-external cross-validation, obtain adequate performance across data sets. Requires data individuals multiple studies.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"","code":"metapred( data, strata, formula, estFUN = \"glm\", scope = NULL, retest = FALSE, max.steps = 1000, center = FALSE, recal.int = FALSE, cvFUN = NULL, cv.k = NULL, metaFUN = NULL, meta.method = NULL, predFUN = NULL, perfFUN = NULL, genFUN = NULL, selFUN = \"which.min\", gen.of.perf = \"first\", ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"data data.frame containing data. Note metapred removes observations missing data listwise variables formula scope, ensure data used model step. outcome variable numeric coercible numeric .numeric(). strata Character specify name strata (e.g. studies clusters) variable formula formula first model evaluated. metapred start formula update using terms scope. Defaults full main effects model, first column data assumed outcome remaining columns (except strata) predictors. See formula formulas general. estFUN Function estimating model first stage. Currently \"lm\", \"glm\" \"logistfirth\" supported. scope formula. difference formula scope defines range models examined stepwise search. Defaults NULL, leads intercept-model. scope nested formula, implies backwards selection applied (default). scope nested formula, implies forward selection applied. equal, stepwise selection applied. retest Logical. added (removed) terms retested removal (addition)? TRUE implies bi-directional stepwise search. max.steps Integer. Maximum number steps (additions removals terms) take. Defaults 1000, essentially many takes. 0 implies stepwise selection. center logical. numeric predictors centered around cluster mean? recal.int Logical. intercept recalibrated validation? cvFUN Cross-validation method, study (.e. cluster stratum) level. \"l1o\" leave-one-cross-validation (default). \"bootstrap\" bootstrap. \"fixed\", one data sets used validation. user written function may supplied well. cv.k Parameter cvFUN. cvFUN=\"bootstrap\", number bootstraps. cvFUN=\"fixed\", vector indices (sorted) data sets. used cvFUN=\"l1o\". metaFUN Function computing meta-analytic coefficient estimates two-stage MA. default, rma.uni, metafor package used. Default settings univariate random effects, estimated \"REML\". Method can passed trough meta.method argument. meta.method Name method meta-analysis. Default \"REML\". options see rma.uni. predFUN Function predicting new values. Defaults predicted probability outcome, using link function glm() lm(). perfFUN Function computing performance prediction models. Default: mean squared error (perfFUN=\"mse\", aka Brier score binomial outcomes).options \"var.e\" (variance prediction error), \"auc\" (area curve), \"cal_int\" (calibration intercept), \"cal_slope\" (multiplicative calibration slope) \"cal_add_slope\" (additive calibration slope), list , first used model selection. genFUN Function list named functions computing generalizability performance. Default: rema, summary statistic random effects meta-analysis. Choose \"rema_tau\" heterogeneity estimate random effects meta-analysis, genFUN=\"abs_mean\" (absolute) mean, coefficient_of_variation coefficient variation. list containing , first used model selection. selFUN Function selecting best method. Default: lowest value genFUN. set \".max\" high values genFUN indicate good model. gen..perf performance measures generalizability measures computed? \"first\" (default) first. \"respective\" matching generalizability measure performance measure location list. \"factorial\" applying generalizability measures performance estimates. ... pass arguments estFUN (e.g. family = \"binomial\"), FUNctions.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"list class metapred, containing final model global.model, stepwise tree estimates coefficients, performance measures, generalizability measures stepwise.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"Use subset.metapred obtain individual prediction model metapred object. Note formula.changes currently unordered; represent order changes stepwise procedure. metapred still development, use care.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"Debray TPA, Moons KGM, Ahmed , Koffijberg H, Riley RD. framework developing, implementing, evaluating clinical prediction models individual participant data meta-analysis. Stat Med. 2013;32(18):3158-80. de Jong VMT, Moons KGM, Eijkemans MJC, Riley RD, Debray TPA. Developing generalizable prediction models pooled studies large clustered data sets. Stat Med. 2021;40(15):3533--59. Riley RD, Tierney JF, Stewart LA. Individual participant data meta-analysis: handbook healthcare research. Hoboken, NJ: Wiley; 2021. ISBN: 978-1-119-33372-2. Schmid CH, Stijnen T, White IR. Handbook meta-analysis. First edition. Boca Raton: Taylor Francis; 2020. ISBN: 978-1-315-11940-3.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"Valentijn de Jong ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/metapred.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Generalized Stepwise Regression for Prediction Models in Clustered Data — metapred","text":"","code":"data(DVTipd) if (FALSE) { # Explore heterogeneity in intercept and assocation of 'ddimdich' glmer(dvt ~ 0 + cluster + (ddimdich|study), family = binomial(), data = DVTipd) } # Scope f <- dvt ~ histdvt + ddimdich + sex + notraum # Internal-external cross-validation of a pre-specified model 'f' fit <- metapred(DVTipd, strata = \"study\", formula = f, scope = f, family = binomial) fit #> Call: metapred(data = DVTipd, strata = \"study\", formula = f, scope = f, #> family = binomial) #> #> Started with model: #> dvt ~ histdvt + ddimdich + sex + notraum #> #> #> Generalizability: #> unchanged #> 1 0.1484983 #> #> Cross-validation stopped after 0 steps, as no changes were requested. Final model: #> Meta-analytic model of prediction models estimated in 4 strata. Coefficients: #> (Intercept) histdvt ddimdich sex notraum #> -4.1180636 0.6174010 1.6962441 0.9647970 0.3761707 # Let's try to simplify model 'f' in order to improve its external validity metapred(DVTipd, strata = \"study\", formula = f, family = binomial) #> Call: metapred(data = DVTipd, strata = \"study\", formula = f, family = binomial) #> #> Started with model: #> dvt ~ histdvt + ddimdich + sex + notraum #> #> #> Generalizability: #> unchanged #> 1 0.1484983 #> #> Generalizability: #> ddimdich histdvt notraum sex #> 1 0.136086 0.1375105 0.12977 0.141173 #> #> Continued with model: #> dvt ~ histdvt + ddimdich + sex #> #> #> Generalizability: #> ddimdich histdvt sex #> 1 0.1366828 0.1279623 0.1319755 #> #> Continued with model: #> dvt ~ ddimdich + sex #> #> #> Generalizability: #> ddimdich sex #> 1 0.1355548 0.1303254 #> #> Cross-validation stopped after 3 steps, as no improvement was possible. Final model: #> Meta-analytic model of prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich sex #> -3.6187987 1.7130967 0.8784071 # We can also try to build a generalizable model from scratch if (FALSE) { # Some additional examples: metapred(DVTipd, strata = \"study\", formula = dvt ~ 1, scope = f, family = binomial) # Forwards metapred(DVTipd, strata = \"study\", formula = f, scope = f, family = binomial) # no selection metapred(DVTipd, strata = \"study\", formula = f, max.steps = 0, family = binomial) # no selection metapred(DVTipd, strata = \"study\", formula = f, recal.int = TRUE, family = binomial) metapred(DVTipd, strata = \"study\", formula = f, meta.method = \"REML\", family = binomial) } # By default, metapred assumes the first column is the outcome. newdat <- data.frame(dvt=0, histdvt=0, ddimdich=0, sex=1, notraum=0) fitted <- predict(fit, newdata = newdat)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate the total O:E ratio — oecalc","title":"Calculate the total O:E ratio — oecalc","text":"function calculates (transformed versions ) ratio total number observed versus expected events corresponding sampling variance.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate the total O:E ratio — oecalc","text":"","code":"oecalc( OE, OE.se, OE.cilb, OE.ciub, OE.cilv, EO, EO.se, citl, citl.se, N, O, E, Po, Po.se, Pe, data, slab, add = 1/2, g = NULL, level = 0.95, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate the total O:E ratio — oecalc","text":"OE vector estimated ratio total observed versus total expected events OE.se Optional vector standard errors estimated O:E ratios. OE.cilb Optional vector specify lower limits confidence interval OE. OE.ciub Optional vector specify upper limits confidence interval OE. OE.cilv Optional vector specify levels aformentioned confidence interval limits. (default: 0.95, corresponds 95% confidence interval). EO Optional vector estimated ratio total expected versus total observed events EO.se Optional vector standard errors estimated E:O ratios citl Optional vector estimated calibration---large statistics citl.se Optional vector standard error calibration---large statistics N Optional vector specify sample/group sizes. O Optional vector specify total number observed events. E Optional vector specify total number expected events Po Optional vector specify (cumulative) observed event probabilities. Po.se Optional vector standard errors Po. time--event data, also SE observed survival probabilities (e.g. obtained Kaplan-Meier analysis) Pe Optional vector specify (cumulative) expected event probabilites (specified, time t.val) data Optional data frame containing variables given arguments . slab Optional vector labels studies. add non-negative number indicating amount add zero counts. See `Details' g quoted string function transform estimates total O:E ratio; see details . level level confidence interval, default 0.95. ... Additional arguments.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate the total O:E ratio — oecalc","text":"object class c(\"mm_perf\",\"data.frame\") following columns: \"theta\" (transformed) O:E ratio. \"theta.se\" Standard errors (transformed) O:E ratio. \"theta.cilb\" Lower confidence interval (transformed) O:E ratios. level specified level. Intervals calculated scale theta assuming Normal distribution. \"theta.ciub\" Upper confidence interval (transformed) c-statistics. level specified level. Intervals calculated scale theta assuming Normal distribution. \"theta.source\" Method used calculating (transformed) O:E ratio. \"theta.se.source\" Method used calculating standard error (transformed) O:E ratio.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Calculate the total O:E ratio — oecalc","text":"Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2019 Sep;28(9):2768--86. Snell KI, Ensor J, Debray TP, Moons KG, Riley RD. Meta-analysis prediction model performance across multiple studies: scale helps ensure -study normality C -statistic calibration measures? Stat Methods Med Res. 2017.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Calculate the total O:E ratio — oecalc","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/oecalc.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Calculate the total O:E ratio — oecalc","text":"","code":"######### Validation of prediction models with a binary outcome ######### data(EuroSCORE) # Calculate the total O:E ratio and its standard error est1 <- oecalc(O = n.events, E = e.events, N = n, data = EuroSCORE, slab = Study) est1 #> theta theta.se theta.cilb theta.ciub theta.source O #> Nashef 1.0450450 0.06716203 0.9134099 1.1766802 O, E and N 232 #> Biancari 0.6086957 0.11345371 0.3863305 0.8310608 O, E and N 28 #> Di Dedda 1.2058824 0.18475130 0.8437765 1.5679882 O, E and N 41 #> Chalmers 0.7318008 0.05203645 0.6298112 0.8337903 O, E and N 191 #> Grant 0.9214541 0.03320253 0.8563783 0.9865298 O, E and N 746 #> Carneo 1.2573099 0.08328543 1.0940735 1.4205464 O, E and N 215 #> Kunt 4.8571429 0.79922240 3.2906957 6.4235900 O, E and N 34 #> Kirmani 1.4134367 0.05935803 1.2970971 1.5297763 O, E and N 547 #> Howell 0.8571429 0.08588255 0.6888162 1.0254696 O, E and N 90 #> Wang 0.7793103 0.05131185 0.6787410 0.8798797 O, E and N 226 #> Borde 0.8000000 0.28056169 0.2501092 1.3498908 O, E and N 8 #> Qadir 1.0270270 0.11555260 0.8005481 1.2535060 O, E and N 76 #> Spiliopoulos 1.5555556 0.40204098 0.7675697 2.3435414 O, E and N 14 #> Wendt 1.3235294 0.19309081 0.9450784 1.7019804 O, E and N 45 #> Laurent 2.5714286 0.58846311 1.4180621 3.7247951 O, E and N 18 #> Wang.1 0.6190476 0.17032315 0.2852204 0.9528749 O, E and N 13 #> Nishida 0.9705882 0.16279815 0.6515097 1.2896668 O, E and N 33 #> Barilli 0.6885246 0.04710205 0.5962063 0.7808429 O, E and N 210 #> Barilli.1 1.2019231 0.10340170 0.9992595 1.4045867 O, E and N 125 #> Paparella 1.1029412 0.06211634 0.9811954 1.2246870 O, E and N 300 #> Carosella 2.2500000 0.73637626 0.8067290 3.6932710 O, E and N 9 #> Borracci 1.3125000 0.28036848 0.7629879 1.8620121 O, E and N 21 #> Osnabrugge 0.6830357 0.02064915 0.6425641 0.7235073 O, E and N 1071 #> E N #> Nashef 222.00 5553 #> Biancari 46.00 1027 #> Di Dedda 34.00 1090 #> Chalmers 261.00 5576 #> Grant 809.59 23740 #> Carneo 171.00 3798 #> Kunt 7.00 428 #> Kirmani 387.00 15497 #> Howell 105.00 933 #> Wang 290.00 11170 #> Borde 10.00 498 #> Qadir 74.00 2004 #> Spiliopoulos 9.00 216 #> Wendt 34.00 1066 #> Laurent 7.00 314 #> Wang.1 21.00 818 #> Nishida 34.00 461 #> Barilli 305.00 12201 #> Barilli.1 104.00 1670 #> Paparella 272.00 6191 #> Carosella 4.00 250 #> Borracci 16.00 503 #> Osnabrugge 1568.00 50588 # Calculate the log of the total O:E ratio and its standard error est2 <- oecalc(O = n.events, E = e.events, N = n, data = EuroSCORE, slab = Study, g = \"log(OE)\") est2 #> theta theta.se theta.cilb theta.ciub theta.source O #> Nashef 0.04405999 0.06426711 -0.081901240 0.17002122 O, E and N 232 #> Biancari -0.49643689 0.18638824 -0.861751123 -0.13112265 O, E and N 28 #> Di Dedda 0.18721154 0.15320840 -0.113071397 0.48749448 O, E and N 41 #> Chalmers -0.31224698 0.07110740 -0.451614919 -0.17287904 O, E and N 191 #> Grant -0.08180235 0.03603276 -0.152425252 -0.01117944 O, E and N 746 #> Carneo 0.22897447 0.06624097 0.099144553 0.35880439 O, E and N 215 #> Kunt 1.58045038 0.16454579 1.257946559 1.90295419 O, E and N 34 #> Kirmani 0.34602411 0.04199553 0.263714374 0.42833385 O, E and N 547 #> Howell -0.15415068 0.10019631 -0.350531831 0.04223047 O, E and N 90 #> Wang -0.24934592 0.06584264 -0.378395127 -0.12029672 O, E and N 226 #> Borde -0.22314355 0.35070211 -0.910507050 0.46421995 O, E and N 8 #> Qadir 0.02666825 0.11251174 -0.193850721 0.24718721 O, E and N 76 #> Spiliopoulos 0.44183275 0.25845491 -0.064729568 0.94839507 O, E and N 14 #> Wendt 0.28030197 0.14589084 -0.005638818 0.56624275 O, E and N 45 #> Laurent 0.94446161 0.22884677 0.495930190 1.39299303 O, E and N 18 #> Wang.1 -0.47957308 0.27513739 -1.018832454 0.05968629 O, E and N 13 #> Nishida -0.02985296 0.16773143 -0.358600527 0.29889460 O, E and N 33 #> Barilli -0.37320425 0.06841012 -0.507285614 -0.23912288 O, E and N 210 #> Barilli.1 0.18392284 0.08603021 0.015306718 0.35253896 O, E and N 125 #> Paparella 0.09798041 0.05631881 -0.012402434 0.20836325 O, E and N 300 #> Carosella 0.81093022 0.32727834 0.169476459 1.45238397 O, E and N 9 #> Borracci 0.27193372 0.21361408 -0.146742192 0.69060962 O, E and N 21 #> Osnabrugge -0.38120813 0.03023143 -0.440460642 -0.32195562 O, E and N 1071 #> E N #> Nashef 222.00 5553 #> Biancari 46.00 1027 #> Di Dedda 34.00 1090 #> Chalmers 261.00 5576 #> Grant 809.59 23740 #> Carneo 171.00 3798 #> Kunt 7.00 428 #> Kirmani 387.00 15497 #> Howell 105.00 933 #> Wang 290.00 11170 #> Borde 10.00 498 #> Qadir 74.00 2004 #> Spiliopoulos 9.00 216 #> Wendt 34.00 1066 #> Laurent 7.00 314 #> Wang.1 21.00 818 #> Nishida 34.00 461 #> Barilli 305.00 12201 #> Barilli.1 104.00 1670 #> Paparella 272.00 6191 #> Carosella 4.00 250 #> Borracci 16.00 503 #> Osnabrugge 1568.00 50588 # Display the results of all studies in a forest plot plot(est1)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/perf.html","id":null,"dir":"Reference","previous_headings":"","what":"Performance estimates — perf","title":"Performance estimates — perf","text":"Obtain performance estimates model fit.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/perf.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Performance estimates — perf","text":"","code":"perf(object, ...) performance(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/perf.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Performance estimates — perf","text":"object model fit object, either metapred subset(metapred) object. ... default, final model selected. parameter allows arguments passed subset.metapred performance estimates steps/models may returned. Use perfFUN = 0 select .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/perf.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Performance estimates — perf","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":null,"dir":"Reference","previous_headings":"","what":"Display results from the funnel plot asymmetry test — plot.fat","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"Generates funnel plot fitted fat object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"","code":"# S3 method for fat plot( x, ref, confint = TRUE, confint.level = 0.1, confint.col = \"skyblue\", confint.alpha = 0.5, confint.density = NULL, xlab = \"Effect size\", add.pval = TRUE, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"x object class fat ref numeric value indicating fixed random effects summary estimate. value provided retrieved fixed effects meta-analysis (possible). confint logical indicator. TRUE, confidence interval displayed estimated regression model (based Student-T distribution) confint.level Significance level constructing confidence interval. confint.col color filling confidence interval. Choose NA leave polygons unfilled. confint.density specified positive value gives color shading lines. confint.alpha numeric value 0 1 indicating opacity confidence region. confint.density density shading lines, lines per inch. default value NULL means shading lines drawn. zero value density means shading filling whereas negative values NA suppress shading (allow color filling). xlab title x axis add.pval Logical indicate whether P-value added plot ... Additional arguments.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"Thomas Debray Frantisek Bartos ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.fat.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Display results from the funnel plot asymmetry test — plot.fat","text":"","code":"data(Fibrinogen) b <- log(Fibrinogen$HR) b.se <- ((log(Fibrinogen$HR.975) - log(Fibrinogen$HR.025))/(2*qnorm(0.975))) n.total <- Fibrinogen$N.total # A very simple funnel plot plot(fat(b=b, b.se=b.se), xlab = \"Log hazard ratio\") # Plot the funnel for an alternative test plot(fat(b=b, b.se=b.se, n.total=n.total, method=\"M-FIV\"), xlab = \"Log hazard ratio\")"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest Plots — plot.mm_perf","title":"Forest Plots — plot.mm_perf","text":"Function create forest plots objects class \"mm_perf\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest Plots — plot.mm_perf","text":"","code":"# S3 method for mm_perf plot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest Plots — plot.mm_perf","text":"x object class \"mm_perf\" ... Additional arguments passed forest.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Forest Plots — plot.mm_perf","text":"object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Forest Plots — plot.mm_perf","text":"forest plot shows performance estimates study corresponding confidence intervals.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Forest Plots — plot.mm_perf","text":"Lewis S, Clarke M. Forest plots: trying see wood trees. BMJ. 2001; 322(7300):1479--80.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest Plots — plot.mm_perf","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.mm_perf.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Forest Plots — plot.mm_perf","text":"","code":"data(EuroSCORE) # Calculate the c-statistic and its standard error est1 <- ccalc(cstat = c.index, cstat.se = se.c.index, cstat.cilb = c.index.95CIl, cstat.ciub = c.index.95CIu, N = n, O = n.events, data = EuroSCORE, slab = Study) plot(est1) # Calculate the total O:E ratio and its standard error est2 <- oecalc(O = n.events, E = e.events, N = n, data = EuroSCORE, slab = Study) plot(est2)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"Generates forest plot outcome bivariate meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"","code":"# S3 method for riley plot(x, title, sort = \"asc\", xlim, refline, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"x object class riley title Title forest plot sort default, studies ordered ascending effect size (sort=\"asc\"). study ordering descending effect size, choose sort=\"desc\". value, study ordering ignored. xlim x limits (x1, x2) forest plot refline Optional numeric specifying reference line ... Additional parameters generating forest plots","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"Riley RD, Thompson JR, Abrams KR. alternative model bivariate random-effects meta-analysis within-study correlations unknown. Biostatistics 2008; 9: 172--186.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.riley.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the summary of the bivariate model from Riley et al. (2008). — plot.riley","text":"","code":"data(Scheidler) #Perform the analysis fit <- riley(Scheidler[which(Scheidler$modality==1),]) plot(fit) require(ggplot2) plot(fit, sort=\"none\", theme=theme_gray())"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest Plots — plot.uvmeta","title":"Forest Plots — plot.uvmeta","text":"Function create forest plots objects class \"uvmeta\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest Plots — plot.uvmeta","text":"","code":"# S3 method for uvmeta plot(x, sort = \"asc\", ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest Plots — plot.uvmeta","text":"x object class \"uvmeta\" sort default, studies ordered ascending effect size (sort=\"asc\"). study ordering descending effect size, choose sort=\"desc\". value, study ordering ignored. ... Additional arguments passed forest.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Forest Plots — plot.uvmeta","text":"forest plot shows performance estimates validation corresponding confidence intervals. polygon added bottom forest plot, showing summary estimate based model. 95% prediction interval added default random-effects models, dotted line indicates (approximate) bounds","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Forest Plots — plot.uvmeta","text":"Full lines indicate confidence intervals credibility intervals (case Bayesian meta-analysis). Dashed lines indicate prediction intervals. width intervals defined significance level chosen meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Forest Plots — plot.uvmeta","text":"Lewis S, Clarke M. Forest plots: trying see wood trees. BMJ. 2001; 322(7300):1479--80. Riley RD, Higgins JPT, Deeks JJ. Interpretation random effects meta-analyses. BMJ. 2011 342:d549--d549.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest Plots — plot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Forest Plots — plot.uvmeta","text":"","code":"data(Roberts) # Frequentist random-effects meta-analysis fit <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts))) plot(fit)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Forest Plots — plot.valmeta","title":"Forest Plots — plot.valmeta","text":"Function create forest plots objects class \"valmeta\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Forest Plots — plot.valmeta","text":"","code":"# S3 method for valmeta plot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Forest Plots — plot.valmeta","text":"x object class \"valmeta\" ... Additional arguments passed forest.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Forest Plots — plot.valmeta","text":"object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Forest Plots — plot.valmeta","text":"forest plot shows performance estimates validation corresponding confidence intervals. polygon added bottom forest plot, showing summary estimate based model. 95% prediction interval added default random-effects models, dotted line indicates (approximate) bounds.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Forest Plots — plot.valmeta","text":"Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. Lewis S, Clarke M. Forest plots: trying see wood trees. BMJ. 2001; 322(7300):1479--80. Riley RD, Higgins JPT, Deeks JJ. Interpretation random effects meta-analyses. BMJ. 2011 342:d549--d549.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Forest Plots — plot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/plot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Forest Plots — plot.valmeta","text":"","code":"data(EuroSCORE) fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE) plot(fit) library(ggplot2) plot(fit, theme=theme_grey())"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Prediction Interval — predict.riley","title":"Prediction Interval — predict.riley","text":"Calculates prediction interval summary parameters Riley's alternative model bivariate random-effects meta-analysis. interval predicts range future observations fall given already observed.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Prediction Interval — predict.riley","text":"","code":"# S3 method for riley predict(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Prediction Interval — predict.riley","text":"object riley object. ... Additional arguments (currently ignored)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Prediction Interval — predict.riley","text":"Data frame containing prediction intervals summary estimates beta1 beta2 (effect size data), mean sensitivity false positive rate (diagnostic test accuracy data).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Prediction Interval — predict.riley","text":"Prediction intervals based Student's t-distribution (numstudies - 5) degrees freedom. width interval specified significance level chosen meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/predict.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Prediction Interval — predict.riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":null,"dir":"Reference","previous_headings":"","what":"Recalibrate a Prediction Model — recalibrate","title":"Recalibrate a Prediction Model — recalibrate","text":"recalibrate used recalibrate prediction model classes metapred, glm lm.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Recalibrate a Prediction Model — recalibrate","text":"","code":"recalibrate(object, newdata, f = ~1, estFUN = NULL, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Recalibrate a Prediction Model — recalibrate","text":"object model fit object recalibrated, class metapred, glm lm, . newdata data.frame containing new data set updating. f formula. coefficients model updated? Default: intercept . Left-hand side may left . See formula details. estFUN Function model estimation. left NULL, function automatically retrieved metapred objects. objects, function name corresponding first class object taken. E.g. glm() glm objects. ... Optional arguments pass estFUN.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Recalibrate a Prediction Model — recalibrate","text":"Recalibrated model fit object, class object. Generally, updated coefficients can retrieved coef().","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Recalibrate a Prediction Model — recalibrate","text":"Currently coefficients updated variances aspects left untouched. updating entire model statistics, see update.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/recalibrate.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Recalibrate a Prediction Model — recalibrate","text":"","code":"data(DVTipd) DVTipd$cluster <- 1:4 # Add a fictional clustering to the data set. # Suppose we estimated the model in three studies: DVTipd123 <- DVTipd[DVTipd$cluster <= 3, ] mp <- metamisc:::metapred(DVTipd123, strata = \"cluster\", f = dvt ~ vein + malign, family = binomial) # and now want to recalibrate it for the fourth: DVTipd4 <- DVTipd[DVTipd$cluster == 4, ] metamisc:::recalibrate(mp, newdata = DVTipd4) #> Call: metamisc:::recalibrate(object = mp, newdata = DVTipd4) #> #> Started with model: #> dvt ~ vein + malign #> #> #> Generalizability: #> unchanged #> 1 0.1332079 #> #> Generalizability: #> malign vein #> 1 0.1344432 0.1314326 #> #> Continued with model: #> dvt ~ malign #> #> #> Generalizability: #> malign #> 1 0.1342213 #> #> Cross-validation stopped after 2 steps, as no improvement was possible. Final model: #> Meta-analytic model of prediction models estimated in 3 strata. Coefficients: #> (Intercept) malign #> -1.775619 1.145096"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Fit the alternative model for bivariate random-effects meta-analysis — riley","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"function fits alternative model bivariate random-effects meta-analysis within-study correlations unknown. bivariate model proposed Riley et al. (2008) similar general bivariate random-effects model (van Houwelingen et al. 2002), includes overall correlation parameter rather separating (usually unknown) within- -study correlation. consequence, alternative model fully hierarchical, estimates additional variation beyond sampling error (psi) directly equivalent -study variation (tau) general model. model particularly useful large within-study variability, primary studies available general model estimates -study correlation 1 -1.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"","code":"riley(X, slab, optimization = \"Nelder-Mead\", control = list(), pars, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"X data frame containing integer variables Y1, vars1, Y2 vars2, columns Y1 Y2 represent effect sizes outcome 1 , respectively, outcome 2. columns vars1 vars2 represent error variances Y1 , respectively, Y2. Alternatively, considering meta-analysis diagnostic test accuracy data, columns TP, FN, FP TN may specified. Corresponding values represent number true positives, number false negatives, number false positives , respectively, number true negatives. slab Optional vector specifying label study optimization optimization method used minimizing negative (restricted) log-likelihood function. default method implementation Nelder Mead (1965), uses function values robust relatively slow. methods described optim. control list control parameters pass optim. pars List additional arguments. width confidence, credibility prediction intervals defined level (defaults 0.95). ... Arguments passed functions. See \"Details\" information.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"object class riley many standard methods available. warning message casted Hessian matrix contains negative eigenvalues, implies identified solution saddle point thus optimal.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"Parameters estimated iteratively maximizing restriced log-likelihood using Newton-Raphson procedure. results univariate random-effects meta-analysis method--moments estimator used starting values beta1, beta2, psi1 psi2 optim command. Standard errors parameters obtained inverse Hessian matrix.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"meta-analysis-of-effect-sizes","dir":"Reference","previous_headings":"","what":"Meta-analysis of effect sizes","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"following parameters estimated iteratively maximizing restriced log-likelihood using Newton-Raphson procedure: pooled effect size outcome 1 (beta1), pooled effect size outcome 2 (beta2), additional variation beta1 beyond sampling error (psi1), additional variation beta2 beyond sampling error (psi2) correlation rho psi1 psi2.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"meta-analysis-of-diagnostic-test-accuracy","dir":"Reference","previous_headings":"","what":"Meta-analysis of diagnostic test accuracy","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"Although model can also used diagnostic test accuracy data substantial within-study correlations expected, assuming zero within-study correlations (.e. applying Reitsma's approach) usually justified (Reitsma et al. 2005, Daniels Hughes 1997, Korn et al. 2005, Thompson et al. 2005, Van Houwelingen et al. 2002). logit transformation applied sensitivities ans false positive rates study, order meet normality assumptions. zero cell counts occur, continuity corrections may required. correction value can specified using correction (defaults 0.5). , argument correction.control set \"\" (default) continuity correction added whole data one cell one study zero. correction.control=\"single\" correction applied rows data zero. following parameters estimated: logit sensitivity (beta1), logit false positive rate (beta2), additional variation beta1 beyond sampling error (psi1), additional variation beta2 beyond sampling error (psi2) correlation (rho) psi1 psi2.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"overall correlation parameter rho transformed estimation ensure corresponding values remain -1 1. transformed correlation rhoT given logit((rho+1)/2). optimization, starting value rhoT set 0. standard error rho derived rhoT using Delta method. Similarly, Delta methods used derive standard error sensitivity false positive rate beta1 , respectively, beta2. Algorithms dealing missing data currently implemented, Bayesian approaches become available later versions.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"Korn EL, Albert PS, McShane LM. Assessing surrogates trial endpoints using mixed models. Statistics Medicine 2005; 24: 163--182. Nelder JA, Mead R. simplex algorithm function minimization. Computer Journal (1965); 7: 308--313. Reitsma J, Glas , Rutjes , Scholten R, Bossuyt P, Zwinderman . Bivariate analysis sensitivity specificity produces informative summary measures diagnostic reviews. Journal Clinical Epidemiology 2005; 58: 982--990. Riley RD, Thompson JR, Abrams KR. alternative model bivariate random-effects meta-analysis within-study correlations unknown. Biostatistics 2008; 9: 172--186. Thompson JR, Minelli C, Abrams KR, Tobin MD, Riley RD. Meta-analysis genetic studies using mendelian randomization--multivariate approach. Statistics Medicine 2005; 24: 2241--2254. van Houwelingen HC, Arends LR, Stijnen T. Advanced methods meta-analysis: multivariate approach meta-regression. Statistics Medicine 2002; 21: 589--624.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/riley.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Fit the alternative model for bivariate random-effects meta-analysis — riley","text":"","code":"data(Scheidler) data(Daniels) data(Kertai) # Meta-analysis of potential surrogate markers data # The results obtained by Riley (2008) were as follows: # beta1 = -0.042 (SE = 0.063), # beta2 = 14.072 (SE = 4.871) # rho = -0.759 if (FALSE) { fit1 <- riley(Daniels) #maxit reached, try again with more iterations } fit1 <- riley(Daniels, control=list(maxit=10000)) summary(fit1) #> $call #> riley.default(X = Daniels, control = list(maxit = 10000)) #> #> $confints #> Estimate SE 2.5 % 97.5 % #> beta1 0.005298983 0.06479973 -0.12170616 0.1323041 #> beta2 13.505678310 4.99256719 3.72042644 23.2909302 #> psi1 0.134785102 0.09190903 -0.04535329 0.3149235 #> psi2 18.076027226 4.00992798 10.21671280 25.9353417 #> rho -0.748689375 0.15266774 -0.89430728 -0.4596724 #> #> attr(,\"class\") #> [1] \"summary.riley\" # Meta-analysis of prognostic test studies fit2 <- riley(Kertai) fit2 #> Call: #> riley.default(X = Kertai) #> #> Coefficients #> beta1 beta2 psi1 psi2 rho #> 0.8164679 -0.9715821 0.3499043 0.7692122 0.1537878 #> #> Degrees of Freedom: 9 Residual # Meta-analysis of computed tomography data ds <- Scheidler[which(Scheidler$modality==1),] fit3 <- riley(ds) fit3 #> Call: #> riley.default(X = ds) #> #> Coefficients #> beta1 beta2 psi1 psi2 rho #> -0.01731291 -2.32166611 0.71181410 0.38103153 0.70119871 #> #> Degrees of Freedom: 29 Residual"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the running means of a Bayesian meta-analysis model — rmplot","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"Function display running means fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"","code":"rmplot(...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"... Additional arguments, currently ignored.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"generic function.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the running means of a Bayesian meta-analysis model — rmplot","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"Function display running means fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"","code":"# S3 method for mcmc.list rmplot(x, P, greek = FALSE, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"x object class \"mcmc.list\" P Optional dataframe describing parameters plot respective names greek Logical value indicating whether parameter labels parsed get Greek letters. Defaults FALSE. ... Additional arguments passed ggs_running","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"ggplot object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.mcmc.list.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.mcmc.list","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"Function display running means fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"","code":"# S3 method for uvmeta rmplot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"x object class \"valmeta\" ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"Results displayed estimated mean (mu) standard-deviation (tau) meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.uvmeta","text":"","code":"if (FALSE) { data(Roberts) fit <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts), method=\"BAYES\")) rmplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"Function display running means fitted Bayesian meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"","code":"# S3 method for valmeta rmplot(x, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"x object class \"valmeta\" ... Additional arguments currently used","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"ggplot object. object class ggplot","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"Results displayed estimated mean (mu) standard-deviation (tau) meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/rmplot.valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot the running means of a Bayesian meta-analysis model — rmplot.valmeta","text":"","code":"if (FALSE) { data(EuroSCORE) fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) rmplot(fit) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":null,"dir":"Reference","previous_headings":"","what":"Standard errors and variances — se","title":"Standard errors and variances — se","text":"Obtain standard errors variances model fit","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Standard errors and variances — se","text":"","code":"se(object, ...) variances(object, ...) tau(object, ...) tau2(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Standard errors and variances — se","text":"object model fit object ... arguments","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Standard errors and variances — se","text":"se standard errors object, variances variances. tau heterogeneity coefficients.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/se.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Standard errors and variances — se","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/stackedglm.html","id":null,"dir":"Reference","previous_headings":"","what":"Stacked Regression — stackedglm","title":"Stacked Regression — stackedglm","text":"function combines one existing prediction models /called meta-model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/stackedglm.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Stacked Regression — stackedglm","text":"","code":"stackedglm(models, family = binomial, data)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/stackedglm.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Stacked Regression — stackedglm","text":"models list containing historical prediction models, can defined several ways. instance, historical regression models can specified using named vector containing regression coefficients individual predictors (need include intercept term). List items may also represent object function predict() exists. family description error distribution link function used meta-model. can character string naming family function, family function result call family function. (See family details family functions.) data optional data frame, list environment (object coercible .data.frame data frame) containing variables model. found data, variables taken environment(formula), typically environment stackedglm called.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/stackedglm.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Stacked Regression — stackedglm","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":null,"dir":"Reference","previous_headings":"","what":"Subsetting metapred fits — subset.metapred","title":"Subsetting metapred fits — subset.metapred","text":"Return model cross-validation procedure final 'global' model. Caution: function still development.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Subsetting metapred fits — subset.metapred","text":"","code":"# S3 method for metapred subset( x, select = \"cv\", step = NULL, model = NULL, stratum = NULL, add = TRUE, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Subsetting metapred fits — subset.metapred","text":"x metapred object select type model select: \"cv\" (default), \"global\", (experimental) \"stratified\", \"stratum\". step step selected? Defaults best step. numeric converted name step: 0 unchanged model, 1 first change... model model change selected? NULL (default, best change) character name variable (integer) index model change. stratum Experimental. Stratum return select = \"stratum\". add Logical. Add data, options functions resulting object? Defaults TRUE. Experimental. ... compatibility .","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Subsetting metapred fits — subset.metapred","text":"object class mp.cv select = \"cv\" object class mp.global select = \"global\". cases, additional data added resulting object, thereby making suitable methods.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Subsetting metapred fits — subset.metapred","text":"Valentijn de Jong","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/subset.metapred.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Subsetting metapred fits — subset.metapred","text":"","code":"data(DVTipd) DVTipd$cluster <- letters[1:4] # Add a fictional clustering to the data. mp <- metapred(DVTipd, strata = \"cluster\", formula = dvt ~ histdvt + ddimdich, family = binomial) subset(mp) # best cross-validated model #> Prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich #> a -3.198673 2.135779 #> b -3.555348 2.251292 #> c -19.566069 18.540216 #> d -3.891820 2.947359 #> #> Meta-analytic models, estimated in 4 fold combinations. Coefficients: #> (Intercept) ddimdich #> b, c, d -3.724268 2.600774 #> a, c, d -3.432711 2.421694 #> a, b, d -3.463478 2.377572 #> a, b, c -3.318460 2.176257 #> #> Cross-validation at stratum level yields the following performance: #> val.strata estimate se var ci.lb ci.ub #> b, c, d a 0.1285223 0.01986149 0.0003944788 0.08921088 0.1678338 #> a, c, d b 0.1293544 0.01729276 0.0002990394 0.09512723 0.1635816 #> a, b, d c 0.1123561 0.01717570 0.0002950048 0.07836058 0.1463516 #> a, b, c d 0.1297595 0.01948263 0.0003795729 0.09119796 0.1683211 #> measure #> b, c, d mse #> a, c, d mse #> a, b, d mse #> a, b, c mse #> #> Generalizability: #> 1 #> 0.1244262 #> subset(mp, select = \"global\") # Final model fitted on all strata. #> Meta-analytic model of prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich #> -3.463480 2.377574 subset(mp, step = 1) # The best model of step 1 #> Prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich #> a -3.198673 2.135779 #> b -3.555348 2.251292 #> c -19.566069 18.540216 #> d -3.891820 2.947359 #> #> Meta-analytic models, estimated in 4 fold combinations. Coefficients: #> (Intercept) ddimdich #> b, c, d -3.724268 2.600774 #> a, c, d -3.432711 2.421694 #> a, b, d -3.463478 2.377572 #> a, b, c -3.318460 2.176257 #> #> Cross-validation at stratum level yields the following performance: #> val.strata estimate se var ci.lb ci.ub #> b, c, d a 0.1285223 0.01986149 0.0003944788 0.08921088 0.1678338 #> a, c, d b 0.1293544 0.01729276 0.0002990394 0.09512723 0.1635816 #> a, b, d c 0.1123561 0.01717570 0.0002950048 0.07836058 0.1463516 #> a, b, c d 0.1297595 0.01948263 0.0003795729 0.09119796 0.1683211 #> measure #> b, c, d mse #> a, c, d mse #> a, b, d mse #> a, b, c mse #> #> Generalizability: #> 1 #> 0.1244262 #> subset(mp, step = 1, model = \"histdvt\") # The model in which histdvt was removed, in step 1. #> Prediction models estimated in 4 strata. Coefficients: #> (Intercept) ddimdich #> a -3.198673 2.135779 #> b -3.555348 2.251292 #> c -19.566069 18.540216 #> d -3.891820 2.947359 #> #> Meta-analytic models, estimated in 4 fold combinations. Coefficients: #> (Intercept) ddimdich #> b, c, d -3.724268 2.600774 #> a, c, d -3.432711 2.421694 #> a, b, d -3.463478 2.377572 #> a, b, c -3.318460 2.176257 #> #> Cross-validation at stratum level yields the following performance: #> val.strata estimate se var ci.lb ci.ub #> b, c, d a 0.1285223 0.01986149 0.0003944788 0.08921088 0.1678338 #> a, c, d b 0.1293544 0.01729276 0.0002990394 0.09512723 0.1635816 #> a, b, d c 0.1123561 0.01717570 0.0002950048 0.07836058 0.1463516 #> a, b, c d 0.1297595 0.01948263 0.0003795729 0.09119796 0.1683211 #> measure #> b, c, d mse #> a, c, d mse #> a, b, d mse #> a, b, c mse #> #> Generalizability: #> 1 #> 0.1244262 #>"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"Parameter summaries Provides summary estimates alternative model bivariate random-effects meta-analysis Riley et al. (2008) corresponding standard errors (derived inverse Hessian). confidence intervals, asymptotic normality assumed.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"","code":"# S3 method for riley summary(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"object riley object ... Arguments passed functions (currently ignored)","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"array confidence intervals estimated model parameters. diagnostic test accuracy data, resulting summary sensitivity false positive rate included.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"meta-analysis diagnostic test accuracy data, beta1 equals logit sensitivity (Sens) beta2 equals logit false positive rate (FPR).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"overall correlation (rho) confidence intervals derived using transformation logit((rho+1)/2). Similarly, logit transformation used derive confidence intervals summary sensitivity false positive rate.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"Riley RD, Thompson JR, Abrams KR. alternative model bivariate random-effects meta-analysis within-study correlations unknown. Biostatistics 2008; 9: 172--186.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Parameter summaries\nProvides the summary estimates of the alternative model for bivariate random-effects meta-analysis by Riley et al. \n(2008) with their corresponding standard errors (derived from the inverse Hessian). For confidence intervals,\nasymptotic normality is assumed. — summary.riley","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"function provides summary estimates fitted univariate meta-analysis model.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"","code":"# S3 method for uvmeta summary(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"object object class \"uvmeta\" ... Optional arguments passed functions","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. basic introduction fixed-effect random-effects models meta-analysis. Research Synthesis Methods 2010; 1: 97--111. DerSimonian R, Laird N. Meta-analysis clinical trials. Controlled Clinical Trials 1986; 7: 177--188. Riley RD, Higgins JPT, Deeks JJ. Interpretation random effects meta-analyses. British Medical Journal 2011; 342: d549.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/summary.uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Summarizing Univariate Meta-Analysis Models — summary.uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":null,"dir":"Reference","previous_headings":"","what":"Class ","title":"Class ","text":"class encapsulates results univariate meta-analysis.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":"objects-from-the-class","dir":"Reference","previous_headings":"","what":"Objects from the Class","title":"Class ","text":"Objects can created calls form uvmeta.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":"slots","dir":"Reference","previous_headings":"","what":"Slots","title":"Class ","text":"call: (language) call uvmeta. data: (data frame) data used meta-analysis. results: (data frame) Contains pooled effect size (mu), -study variability (tausq), Cochran's Q statistic (Q) Higgins' Thompson's square statistic (Isq). estimate, error variances provided predefined confidence (method=\"MOM\") credibility (method=\"bayes\") intervals. model: (character) meta-analysis model used. method: (character) estimator used. na.action: (character) Information action applied object NAs handled specially, NULL. df: (numeric) Degrees freedom. numstudies: (numeric) amount studies used meta-analysis. pred.int: (data frame) prediction interval, predicting range future effect sizes fall given already observed (based Student's t-distribution, cfr. Riley 2011) formula: (character) formula specified, character vector giving formula parameter specifications.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":"methods","dir":"Reference","previous_headings":"","what":"Methods","title":"Class ","text":"print signature(object = \"uvmeta\"): Print object summary. forest signature(object = \"uvmeta\"): Plot forest plot summary estimate. summary signature(object = \"uvmeta\"): Generate object summary.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta-class.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Class ","text":"","code":"data(Collins) #Extract effect size and error variance r <- Collins$logOR vars <- Collins$SE**2 #Frequentist random-effects meta-analysis fit1 <- uvmeta(r,vars) #Extract results fit1$results #> NULL"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Univariate meta-analysis — uvmeta","title":"Univariate meta-analysis — uvmeta","text":"function summarizes multiple estimates single parameter assuming fixed (.e. common) effect random effects across studies. summary estimate obtained calculating weighted mean accounts sample size (case random effects assumed) -study heterogeneity.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Univariate meta-analysis — uvmeta","text":"","code":"uvmeta( r, r.se, r.vi, method = \"REML\", test = \"knha\", labels, na.action, n.chains = 4, pars, verbose = FALSE, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Univariate meta-analysis — uvmeta","text":"r Vector numerics containing effect size study r.se Vector numerics containing standard error effect sizes r.vi Vector numerics containing sampling variance effect sizes method Character string specifying whether fixed-effect random-effects model fitted. fixed-effect model fitted using method=\"FE\". Random-effects models fitted setting method equal one following: \"REML\" (Default), \"DL\", \"\", \"SJ\", \"ML\", \"EB\", \"HS\", \"GENQ\" \"BAYES\". See 'Details'. test Optional character string method!=\"BAYES\" specify test statistics confidence intervals fixed effects computed. default (test=\"knha\"), method Knapp Hartung (2003) used adjusting test statistics confidence intervals. Type '?rma' details. labels Optional vector characters containing labels studies na.action function indicates happen data contain NAs. Defaults \"na.fail\", options \"na.omit\", \"na.exclude\" \"na.pass\". n.chains Optional numeric specifying number chains use Gibbs sampler (method=\"BAYES\"). chains improve sensitivity convergence diagnostic, cause simulation run slowly. default number chains 4. pars Optional list additional arguments. width confidence, credibility prediction intervals defined level (defaults 0.95). following parameters configure MCMC sampling procedure: hp.mu.mean (mean prior distribution random effects model, defaults 0), hp.mu.var (variance prior distribution random effects model, defaults 1E6), hp.tau.min (minimum value -study standard deviation, defaults 0), hp.tau.max (maximum value -study standard deviation, defaults 2), hp.tau.sigma (standard deviation prior distribution -study standard-deviation), hp.tau.dist (prior distribution -study standard-deviation. Defaults \"dunif\"), hp.tau.df (degrees freedom prior distribution -study standard-deviation. Defaults 3). verbose TRUE messages generated fitting process displayed. ... Additional arguments passed rma runjags (method=\"BAYES\").","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Univariate meta-analysis — uvmeta","text":"object class uvmeta many standard methods available. \"data\" array (transformed) data used meta-analysis, method(s) used restoring missing information. \"method\" character string specifying meta-analysis method. \"est\" estimated performance statistic model. Bayesian meta-analysis, posterior median returned. \"se\" standard error (posterior standard deviation) summary estimate. \"tau2\" estimated amount (residual) heterogeneity. Always 0 method=\"FE\". Bayesian meta-analysis, posterior median returned. \"se.tau2\" estimated standard error (posterior standard deviation) -study variation. \"ci.lb\" lower bound confidence (credibility) interval summary estimate \"ci.ub\" upper bound confidence (credibility) interval summary estimate \"pi.lb\" lower bound (approximate) prediction interval summary estimate \"pi.ub\" upper bound (approximate) prediction interval summary estimate \"fit\" full results fitted model \"slab\" vector specifying label study.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Univariate meta-analysis — uvmeta","text":"Unless specified otherwise, meta-analysis models assume random effects fitted using restricted maximum likelihood estimation metafor package (Viechtbauer 2010). , confidence intervals average performance based Hartung-Knapp-Sidik-Jonkman method, better account uncertainty estimated -study heterogeneity (Debray 2016). Bayesian meta-analysis can performed specifying method=\"BAYES\". case, R packages runjags rjags must installed.] random-effects models, prediction interval pooled effect size displayed. interval predicts range future effect sizes fall given already observed (Higgins 2009, Riley 2011).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"bayesian-meta-analysis-models","dir":"Reference","previous_headings":"","what":"Bayesian meta-analysis models","title":"Univariate meta-analysis — uvmeta","text":"Bayesian meta-analysis models involve Gibbs sampler (method=\"BAYES\"), R packages runjags rjags must installed. Bayesian approach uses uninformative Normal prior mean uniform prior -study variance pooled effect size (Higgins 2009). default, Normal prior mean 0 variance 1000. hyperparameters can, however, altered variables hp.mu.mean hp.mu.var argument pars. prior distribution -study standard deviation given uniform distribution, default bounded 0 100.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Univariate meta-analysis — uvmeta","text":"Biggerstaff BJ, Tweedie RL. Incorporating variability estimates heterogeneity random effects model meta-analysis. Statistics Medicine 1997; 16: 753--768. Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. basic introduction fixed-effect random-effects models meta-analysis. Research Synthesis Methods 2010; 1: 97--111. doi:10.1002/jrsm.12 DerSimonian R, Laird N. Meta-analysis clinical trials. Controlled Clinical Trials 1986; 7: 177--188. Graham PL, Moran JL. Robust meta-analytic conclusions mandate provision prediction intervals meta-analysis summaries. Journal Clinical Epidemiology 2012; 65: 503--510. Higgins JPT, Thompson SG. Quantifying heterogeneity meta-analysis. Statistics Medicine 2002; 21: 1539--1558. Higgins JPT, Thompson SG, Spiegelhalter DJ. re-evaluation random-effects meta-analysis. J R Stat Soc Ser Stat Soc. 2009;172:137--59. doi:10.1111/j.1467-985X.2008.00552.x Riley RD, Higgins JPT, Deeks JJ. Interpretation random effects meta-analyses. British Medical Journal 2011; 342: d549. doi:10.1136/bmj.d549 Viechtbauer W. Conducting Meta-Analyses R metafor Package. Journal Statistical Software. 2010; 36. doi:10.18637/jss.v036.i03","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Univariate meta-analysis — uvmeta","text":"Thomas Debray ","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/uvmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Univariate meta-analysis — uvmeta","text":"","code":"data(Roberts) # Frequentist random-effects meta-analysis fit1 <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts))) summary(fit1) #> Call: #> uvmeta.default(r = SDM, r.se = SE, labels = rownames(Roberts)) #> #> Random effects summary:\t 0.36195 (SE: 0.0859) #> Tau squared: \t\t 0.01322 (SE: 0.03431) plot(fit1) #show a forest plot fit1 #> Summary estimate with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 0.36194806 0.17636909 0.54752703 0.04923206 0.67466405 if (FALSE) { # Bayesian random effects meta-analysis fit2 <- with(Roberts, uvmeta(r=SDM, r.se=SE, labels=rownames(Roberts), method=\"BAYES\")) plot(fit2) }"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":null,"dir":"Reference","previous_headings":"","what":"Meta-analysis of prediction model performance — valmeta","title":"Meta-analysis of prediction model performance — valmeta","text":"function provides summary estimates concordance statistic, total observed-expected ratio calibration slope. appropriate, data transformations applied missing information derived available quantities. Unless specified otherwise, meta-analysis models assume random effects fitted using restricted maximum likelihood estimation metafor package (Viechtbauer 2010). , confidence intervals average performance based Hartung-Knapp-Sidik-Jonkman method. conducting Bayesian meta-analysis, R packages runjags rjags must installed.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Meta-analysis of prediction model performance — valmeta","text":"","code":"valmeta( measure = \"cstat\", cstat, cstat.se, cstat.cilb, cstat.ciub, cstat.cilv, sd.LP, OE, OE.se, OE.cilb, OE.ciub, OE.cilv, citl, citl.se, N, O, E, Po, Po.se, Pe, data, method = \"REML\", test = \"knha\", verbose = FALSE, slab, n.chains = 4, pars, ... )"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Meta-analysis of prediction model performance — valmeta","text":"measure character string indicating summary performance measure calculated. Options \"cstat\" (meta-analysis concordance statistic) \"OE\" (meta-analysis total observed-expected ratio). See `Details' information. cstat Optional vector estimated c-statistic valiation cstat.se Optional vector standard error estimated c-statistics cstat.cilb Optional vector specify lower limits confidence interval. cstat.ciub Optional vector specify upper limits confidence interval. cstat.cilv Optional vector specify levels aformentioned confidence interval limits. (default: 0.95, corresponds 95% confidence interval). sd.LP Optional vector standard deviation linear predictor (prognostic index) OE Optional vector estimated ratio total observed versus total expected events OE.se Optional vector standard errors estimated O:E ratios OE.cilb Optional vector specify lower limits confidence interval OE. OE.ciub Optional vector specify upper limits confidence interval OE. OE.cilv Optional vector specify levels aformentioned confidence interval limits. (default: 0.95, corresponds 95% confidence interval). citl Optional vector estimated calibration---large valiation citl.se Optional vector standard error estimated calibration---large statistics N Optional vector total number participants valiation O Optional vector total number observed events valiation (specified, time t.val) E Optional vector total number expected events valiation (specified, time t.val) Po Optional vector (cumulative) observed event probability valiation (specified, time t.val) Po.se Optional vector standard errors Po. Pe Optional vector (cumulative) expected event probability validation (specified, time t.val) data optional data frame containing variables given arguments . method Character string specifying whether fixed- random-effects model fitted. fixed-effects model fitted using method=\"FE\". Random-effects models fitted setting method equal one following: \"REML\" (Default), \"DL\", \"\", \"SJ\", \"ML\", \"EB\", \"HS\", \"GENQ\" \"BAYES\". See 'Details'. test Optional character string specifying test statistics confidence intervals fixed effects computed. default (test=\"knha\"), method Knapp Hartung (2003) used adjusting test statistics confidence intervals. Type '?rma' details. verbose TRUE messages generated fitting process displayed. slab Optional vector specifying label study n.chains Optional numeric specifying number chains use Gibbs sampler (method=\"BAYES\"). chains improve sensitivity convergence diagnostic, cause simulation run slowly. default number chains 4. pars list additional arguments. See 'Details' information. following parameters configure MCMC sampling procedure: hp.mu.mean (mean prior distribution random effects model, defaults 0), hp.mu.var (variance prior distribution random effects model, defaults 1000), hp.tau.min (minimum value -study standard deviation, defaults 0), hp.tau.max (maximum value -study standard deviation, defaults 2), hp.tau.sigma (standard deviation prior distribution -study standard-deviation), hp.tau.dist (prior distribution -study standard-deviation. Defaults \"dunif\"), hp.tau.df (degrees freedom prior distribution -study standard-deviation. Defaults 3). arguments method.restore.c.se (method restoring missing estimates standard error c-statistic. See ccalc information), model.cstat (likelihood/link modeling c-statistic; see \"Details\"), model.oe (likelihood/link modeling O:E ratio; see \"Details\"), seed (integer indicate random seed). ... Additional arguments passed rma runjags (method=\"BAYES\").","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Meta-analysis of prediction model performance — valmeta","text":"object class valmeta following elements: \"data\" array (transformed) data used meta-analysis, method(s) used restoring missing information. \"measure\" character string specifying performance measure meta-analysed. \"method\" character string specifying meta-analysis method. \"model\" character string specifying meta-analysis model (link function). \"est\" summary estimate performance statistic. Bayesian meta-analysis, posterior median returned. \"ci.lb\" lower bound confidence (credibility) interval summary performance estimate. \"ci.ub\" upper bound confidence (credibility) interval summary performance estimate. \"pi.lb\" lower bound (approximate) prediction interval summary performance estimate. \"pi.ub\" upper bound (approximate) prediction interval summary performance estimate. \"fit\" full results fitted model. \"slab\" vector specifying label study.","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"meta-analysis-of-the-concordance-statistic","dir":"Reference","previous_headings":"","what":"Meta-analysis of the concordance statistic","title":"Meta-analysis of prediction model performance — valmeta","text":"summary estimate concorcance (c-) statistic can obtained specifying measure=\"cstat\". c-statistic measure discrimination, indicates ability prediction model distinguish patients developing developing outcome. c-statistic typically ranges 0.5 (discriminative ability) 1 (perfect discriminative ability). missing, c-statistic /standard error derived reported information. See ccalc information. default, assumed logit c-statistic Normally distributed within across studies (pars$model.cstat = \"normal/logit\"). Alternatively, possible assume raw c-statistic Normally distributed across studies pars$model.cstat = \"normal/identity\".","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"meta-analysis-of-the-total-observed-versus-expected-ratio","dir":"Reference","previous_headings":"","what":"Meta-analysis of the total observed versus expected ratio","title":"Meta-analysis of prediction model performance — valmeta","text":"summary estimate total observed versus expected (O:E) ratio can obtained specifying measure=\"OE\". total O:E ratio provides rough indication overall model calibration (across entire range predicted risks). missing, total O:E ratio /standard error derived reported information. See oecalc information. frequentist meta-analysis, within-study variation can either modeled using Normal (model.oe = \"normal/log\" model.oe = \"normal/identity\") Poisson distribution (model.oe = \"poisson/log\"). performing Bayesian meta-analysis, data modeled using one-stage random effects (hierarchical related regression) model. particular, binomial distribution (O, E N known), Poisson distribution (O E known) Normal distribution (OE OE.se OE.95CI known) selected separately study.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"bayesian-meta-analysis","dir":"Reference","previous_headings":"","what":"Bayesian meta-analysis","title":"Meta-analysis of prediction model performance — valmeta","text":"Bayesian meta-analysis models assume presence random effects. Summary estimates based posterior mean. Credibility prediction intervals directly obtained corresponding posterior quantiles. prior distribution (transformed) performance estimate modeled using Normal distribution, mean hp.mu.mean (defaults 0) variance hp.mu.var (defaults 1000). meta-analysis total O:E ratio, maximum value hp.mu.var 100. default, prior distribution -study standard deviation modeled using uniform distribution (hp.tau.dist=\"dunif\"), boundaries hp.tau.min hp.tau.max. Alternative choices truncated Student-t distribution (hp.tau.dist=\"dhalft\") mean hp.tau.mean, standard deviation hp.tau.sigma hp.tau.df degrees freedom. distribution restricted range hp.tau.min hp.tau.max.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Meta-analysis of prediction model performance — valmeta","text":"width calculated confidence, credibility prediction intervals can specified using level pars argument (defaults 0.95).","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Meta-analysis of prediction model performance — valmeta","text":"Debray TPA, Damen JAAG, Snell KIE, Ensor J, Hooft L, Reitsma JB, et al. guide systematic review meta-analysis prediction model performance. BMJ. 2017;356:i6460. doi:10.1136/bmj.i6460 Debray TPA, Damen JAAG, Riley R, Snell KIE, Reitsma JB, Hooft L, et al. framework meta-analysis prediction model studies binary time--event outcomes. Stat Methods Med Res. 2019;28:2768--86. doi:10.1177/0962280218785504 Riley RD, Tierney JF, Stewart LA. Individual participant data meta-analysis: handbook healthcare research. Hoboken, NJ: Wiley; 2021. ISBN: 978-1-119-33372-2. Steyerberg EW, Nieboer D, Debray TPA, van Houwelingen HC. Assessment heterogeneity individual participant data meta-analysis prediction models: overview illustration. Stat Med. 2019; 38:4290--309. doi:10.1002/sim.8296 Viechtbauer W. Conducting Meta-Analyses R metafor Package. Journal Statistical Software. 2010; 36. doi:10.18637/jss.v036.i03","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/valmeta.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Meta-analysis of prediction model performance — valmeta","text":"","code":"######### Validation of prediction models with a binary outcome ######### data(EuroSCORE) # Meta-analysis of the c-statistic (random effects) fit <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, cstat.cilv=0.95, N=n, O=n.events, slab=Study, data=EuroSCORE) plot(fit) # Nearly identical results when we need to estimate the SE valmeta(cstat=c.index, N=n, O=n.events, slab=Study, data=EuroSCORE) #> Summary c-statistic with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 0.7889020 0.7650864 0.8109000 0.6818676 0.8669518 #> #> Number of studies included: 23 # Two-stage meta-analysis of the total O:E ratio (random effects) valmeta(measure=\"OE\", O=n.events, E=e.events, N=n, slab=Study, data=EuroSCORE) #> Summary Total O:E ratio with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 1.1075973 0.8998973 1.3632352 0.4295250 2.8561122 #> #> Number of studies included: 23 valmeta(measure=\"OE\", O=n.events, E=e.events, data=EuroSCORE) #> Summary Total O:E ratio with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 1.1059784 0.8990028 1.3606056 0.4316383 2.8338269 #> #> Number of studies included: 23 valmeta(measure=\"OE\", Po=Po, Pe=Pe, N=n, data=EuroSCORE) #> Summary Total O:E ratio with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 1.1230955 0.9212978 1.3690944 0.4549877 2.7722586 #> #> Number of studies included: 23 if (FALSE) { # One-stage meta-analysis of the total O:E ratio (random effects) valmeta(measure=\"OE\", O=n.events, E=e.events, data=EuroSCORE, method=\"ML\", pars=list(model.oe=\"poisson/log\")) # Bayesian random effects meta-analysis of the c-statistic fit2 <- valmeta(cstat=c.index, cstat.se=se.c.index, cstat.cilb=c.index.95CIl, cstat.ciub=c.index.95CIu, cstat.cilv=0.95, N=n, O=n.events, data=EuroSCORE, method=\"BAYES\", slab=Study) # Bayesian one-stage random effects meta-analysis of the total O:E ratio # Consider that some (but not all) studies do not provide information on N # A Poisson distribution will be used for studies 1, 2, 5, 10 and 20 # A Binomial distribution will be used for the remaining studies EuroSCORE.new <- EuroSCORE EuroSCORE.new$n[c(1, 2, 5, 10, 20)] <- NA pars <- list(hp.tau.dist=\"dhalft\", # Prior for the between-study standard deviation hp.tau.sigma=1.5, # Standard deviation for 'hp.tau.dist' hp.tau.df=3, # Degrees of freedom for 'hp.tau.dist' hp.tau.max=10, # Maximum value for the between-study standard deviation seed=5) # Set random seed for the simulations fit3 <- valmeta(measure=\"OE\", O=n.events, E=e.events, N=n, data=EuroSCORE.new, method=\"BAYES\", slab=Study, pars=pars) plot(fit3) print(fit3$fit$model) # Inspect the JAGS model print(fit3$fit$data) # Inspect the JAGS data } ######### Validation of prediction models with a time-to-event outcome ######### data(Framingham) # Meta-analysis of total O:E ratio after 10 years of follow-up valmeta(measure=\"OE\", Po=Po, Pe=Pe, N=n, data=Framingham) #> Warning: 8 studies with NAs omitted from model fitting. #> Summary Total O:E ratio with 95% confidence and (approximate) 95% prediction interval: #> #> Estimate CIl CIu PIl PIu #> 0.5781061 0.4400900 0.7594053 0.1935434 1.7267794 #> #> Number of studies included: 16"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"Returns variance-covariance matrix main parameters fitted model object.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"","code":"# S3 method for riley vcov(object, ...)"},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"object riley object. ... arguments passed functions","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"variance-covariance matrix obtained inverse Hessian provided optim.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"matrix estimated covariances parameter estimates Riley model: logit sensitivity (mu1), logit false positive rate (mu2), additional variation mu1 beyond sampling error (psi1), additional variation mu2 beyond sampling error (psi2) transformation correlation psi1 psi2 (rhoT). original correlation given inv.logit(rhoT)*2-1.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"warning message casted Hessian matrix contains negative eigenvalues. implies identified minimum (restricted) negative log-likelihood saddle point, solution therefore optimal.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"Riley, RD., Thompson, JR., & Abrams, KR. (2008). “alternative model bivariate random-effects meta-analysis within-study correlations unknown.” Biostatistics, 9, 172--186.","code":""},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/reference/vcov.riley.html","id":"author","dir":"Reference","previous_headings":"","what":"Author","title":"Calculate Variance-Covariance Matrix for a Fitted Riley Model Object — vcov.riley","text":"Thomas Debray ","code":""},{"path":[]},{"path":"https://smartdata-analysis-and-statistics.github.io/metamisc/news/index.html","id":"metamisc-0409000","dir":"Changelog","previous_headings":"","what":"metamisc 0.4.0.9000","title":"metamisc 0.4.0.9000","text":"Added NEWS.md file track changes package.","code":""}]