-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathmeeting1.Rmd
95 lines (63 loc) · 3.48 KB
/
meeting1.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
---
title: "Meeting1"
author: "Kuan Liu"
date: '`r paste(substring(date(),5,10), substring(date(),21,24))`'
output:
pdf_document:
fig_caption: yes
keep_tex: yes
number_sections: yes
urlcolor: blue
---
# Chapter 1 Statistical approaches for clinicla trials
## Bayesian versus frequentist appraoches in clincial research
- Key differences
- parameters
- inference input data and prior
- **flexibility**, sequential learning
- probability statements, predictive probabilities
- decision theoretic
- comment on role of randomization
Additional reading
- [Beyond subjective and objective in statistics by Andrew Gelman and Christian Henning](http://www.stat.columbia.edu/~gelman/research/published/gelman_hennig_full_discussion.pdf)
- Bayesian is no more subjective than frequentist
- Prior can be used meaningfully both computationally and biologically
## Adaptivity
Bayesian is more commonly used for Phase I and Phase II trials. Some appealing adoption including non-fixed sample-size, early stopping and adoptive randomization.
### The book argue "full bayesian" with utility function is awkward.
- the expected utility function given poster distribution of the parameter
- an example utility gain (or loss): Y(Z=1, treated)-Y(Z=0, control)
$$ \int u(\theta, x) p(\theta |x)d\theta $$
### Bayes as a frequentist tool
- flexibility, sequential learning
- posterior predictive probability
- use of prior, borrow info from other studies
- adaptive randomization
- seamless phase II and III design
# Chapter 2 Basics of Bayesian
Additional reading on choice of prior
- https://github.com/stan-dev/stan/wiki/Prior-Choice-Recommendations
- http://www.stat.columbia.edu/~gelman/research/unpublished/prior_context_2.pdf
Additional reading on model selection (AIC, DIC, WAIC etc):
- [Understanding predictive information criteria for Bayesian models, Gelman, Hwang & Vehtari](http://www.stat.columbia.edu/~gelman/research/published/waic_understand3.pdf)
## Principles of Bayesian clincial trial design
Methods which can be used for clinical decision making
1. Posterior predictive methods
2. Bayesian indifference zone methods
{width=75%}
3. Use of priors
4. Operating characteristics - Bayesian trial sample size
Additional reading on Bayesian trial sample size
- [Bayesian techniques for sample size determination in clinical trials: a short review](https://doi.org/10.1191/0962280203sm345oa)
- [Using historical data for Bayesian sample size
determination]( https://doi.org/10.1111/j.1467-985X.2006.00438.)
- [Clinical Trials and Sample Size
Considerations: Another Perspective](https://projecteuclid.org/journals/statistical-science/volume-15/issue-2/Clinical-Trials-and-Sample-Size-Considerations-AnotherPerspective/10.1214/ss/1009212752.full)
Some thing interesting to read:
Bayesian and type I error, https://www.fharrell.com/post/pvalprobs/
5. Online Bayesian Phase I and II design apps, https://trialdesign.org/#newsSection
# Discussions
1. Example Bayesian trial, Advanced reperfusion strategies for patients with out-of-hospital cardiac arrest and refractory ventricular fibrillation (ARREST): a phase 2, single centre, open-label, randomised controlled trial, https://doi.org/10.1016/S0140-6736(20)32338-2
- The use of potentially optimistic treatment effect target belief
- Cost of using ECMO versus control
2. P-value versus Bayes factor, confidence interval versus credible region. (http://www.utstat.toronto.edu/mikevans/jeffrosenthal/chap7.pdf)