sampleSWCRTLarge

Simulated cross-sectional individual-level SW-CRT data with 12 clusters and 5 periods. The cluster-period size is uniformly distributed between 20 and 30.
 The correlated binary and continuous outcomes are used for analysis as examples.

Population-averaged models have been increasingly used in the design and analysis of
cluster randomized trials (CRTs). To facilitate the applications of population-averaged
models in CRTs, the package implements the generalized estimating equations (GEE) and
matrix-adjusted estimating equations (MAEE) approaches to jointly estimate the marginal
mean models correlation models both for general CRTs and stepped wedge CRTs. Despite the
general GEE/MAEE approach, the package also implements a fast cluster-period GEE method by
Li et al. (2022) <doi:10.1093/biostatistics/kxaa056>
specifically for stepped wedge CRTs with large and variable cluster-period sizes and gives
a simple and efficient estimating equations approach based on the cluster-period means to
estimate the intervention effects as well as correlation parameters. In addition, the package
also provides functions for generating correlated binary data with specific mean vector and
correlation matrix based on the multivariate probit method in Emrich and Piedmonte (1991) <doi:10.1080/00031305.1991.10475828> or
the conditional linear family method in Qaqish (2003) <doi:10.1093/biomet/90.2.455>.

Hengshi Yu

geeCRT

Bias-Corrected GEE for Cluster Randomized Trials

Fan Li

Paul Rathouz

Elizabeth L. Turner

John Preisser

sampleSWCRTLarge function

A data frame with 1508 rows and 10 variables:<dl>
 <dt>period1</dt>
<dd>indicator of being at period 1</dd> <dt>period2</dt>
<dd>indicator of being at period 2</dd> <dt>period3</dt>
<dd>indicator of being at period 3</dd> <dt>period4</dt>
<dd>indicator of being at period 4</dd> <dt>period5</dt>
<dd>indicator of being at period 5</dd> <dt>treatment</dt>
<dd>indicator of being treated</dd> <dt>id</dt>
<dd>cluster identification number</dd> <dt>period</dt>
<dd>period order number</dd> <dt>y_bin</dt>
<dd>binary outcome variable</dd> <dt>y_con</dt>
<dd>continous outcome variable</dd>
</dl>

Format

simulated large SW-CRT data — sampleSWCRTLarge

A data frame with 1508 rows and 10 variables:<dl>
 <dt>period1</dt>
<dd>indicator of being at period 1</dd>

 <dt>period2</dt>
<dd>indicator of being at period 2</dd>

 <dt>period3</dt>
<dd>indicator of being at period 3</dd>

 <dt>period4</dt>
<dd>indicator of being at period 4</dd>

 <dt>period5</dt>
<dd>indicator of being at period 5</dd>

 <dt>treatment</dt>
<dd>indicator of being treated</dd>

 <dt>id</dt>
<dd>cluster identification number</dd>

 <dt>period</dt>
<dd>period order number</dd>

 <dt>y_bin</dt>
<dd>binary outcome variable</dd>

 <dt>y_con</dt>
<dd>continous outcome variable</dd>


</dl>

sampleSWCRTLarge: simulated large SW-CRT data

Description

Arguments

Format