codmi

codmi is an R algorithm which allows for the inclusion of oncological COVID-19 deaths in traditional survival analysis.
For a given sample of time-to-event data, typically oncological deaths with censored observations, where covid deaths are also observed, codmi algorithm includes covid-death cases in the standard data by mean-imputation. Data completion is done through an expectation-maximization algorithm based on the Kaplan-Meier estimator. The completed data are equipped with the corresponding Kaplan-Meier survival function estimates and the appropriate confidence intervals. These are based on an extended Greenwood’s formula where the variance is adjusted for the mean-imputation.

This branch of the repository contains the basic functions of the algorithm and a suite of examples. A self-contained installable package is in the works.

usage

• codmi needs a working installation of the R programming language to run;
• install the survival and the stats third party packages;
• download the repository and execute the code in examples.R