The goal of multipleOutcomes is to fit statistical models for multiple outcomes simultaneously. It computes estimates of parameters across fitted models and returns the matrix of asymptotic covariance. Various applications of this package, including PATED (Prognostic Variables Assisted Treatment Effect Detection), multiple comparison adjustment, conditional power are illustrated.

Installation

Install the development version from GitHub:

if(!requireNamespace("remotes", quietly = TRUE)) {
  install.packages("remotes")
}
remotes::install_github(
  "zhangh12/multipleOutcomes",
  dependencies = TRUE
)

Example

Let’s get started by analyzing a real randomized trial data. The data was from a randomized clinical trial of indomethacin to prevent post-ERCP pancreatitis. The binary endpoint is post-ERCP pancreatitis (yes/no). It was collected, reformatted and shared by Dr. Peter Higgins in his R package medicaldata. For more information, please refer to the data set indo_rct in medicaldata, and the manuscript.

Here we use the PATED method to estimate marginal treatment effect on post-ERCP pancreatitis, adjusting for a set of risk factors for PEP as potential prognostic covariates. I am not an expert of this disease, so I simply follow the description information of the covariates in medicaldata.

Here we compare PATED with unconditional logistic regression model in estimating the treatment effect. For each of the covariates, a regression model covar ~ rx is fitted to compute the slope coefficient