chapter_10_part1.Rmd

---
title: "Dealing with irregular and informative visits"
author: "Thomas Debray"
date: "This report was generated on `r format(Sys.time(), '%d %B, %Y')`"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
knitr::opts_chunk$set(message = FALSE)
source("Chapter 10 - Dealing with irregular and informative visits/sim.r")
source("Chapter 10 - Dealing with irregular and informative visits/fig_functions.r")

```

# Introduction

## Data generation

We simulate a randomized clinical trial with no treatment effect.

```{r}
sim_data <- sim_data_EDSS(tx_alloc_FUN = treatment_alloc_randomized, delta_xt = 0)
```

The data frame `sim_data` contains information on the treatment allocation `x` and two baseline covariates include `age` and `edss` (EDSS at treatment start). The outcome `y` represents the Expanded Disability Status Scale (EDSS) score after `time` months of treatment exposure. Briefly, the EDSS is a semi-continuous measure that varies from 0 (no disability) to 10 (death).

```{r}
plot_distribution_edss(sim_data)
```

Finally, we have a binary outcome `progression` representing disease progression which is defined as follows:

-   An increase in 1.5 points from `edss` \>= 0 and \< 1
-   An increase in 1.0 points from `edss` \>= 1 and \< 6
-   An increase of 0.5 points from `edss` \>= 6

```{r}
plot_distribution_progression(sim_data)
```

Simulate a randomized clinical trial with a (linear) treatment effect

```{r}
sim_data <- sim_data_EDSS(tx_alloc_FUN = treatment_alloc_randomized, delta_xt = -0.007)
plot_distribution_edss(sim_data)
plot_distribution_progression(sim_data)
```

Simulate a dataset with unequal treatment allocation

```{r}
sim_data <- sim_data_EDSS(tx_alloc_FUN = treatment_alloc_confounding, delta_xt = -0.007)
plot_distribution_edss(sim_data)
```

## Introduce missing visits

The following censoring functions are available to introduce informative patient visits:

-   `censor_visits_a1` : Patient visits are missing according to center only
-   `censor_visits_a2` : Patient visits are missing according to center and received treatment
-   `censor_visits_a3` : Visit schedules are regular but differ between treatment groups
-   `censor_visits_a4` : Patient visits are missing according to their received treatment and current EDSS score

```{r}
sim_data_censored <- censor_visits_a2(data = sim_data)
plot_prob_visit(sim_data_censored)
sim_data_censored <- censor_visits_a3(data = sim_data)
plot_prob_visit(sim_data_censored)
```

## Impute missing data