Competing risks occur in survival analysis when an individual is at risk of more than one type of event and one event's occurrence precludes another's. The cause-specific cumulative incidence function (CIF) is a measure of interest with competing-risks data. It gives the absolute (or crude) risk of having the event by time t, accounting for the fact that it is impossible to have the event if a competing event occurs first. The user-written command stcompet calculates nonparametric estimates of the cause-specific CIF, and the official Stata command stcrreg fits the Fine and Gray (1999, Journal of the American Statistical Association 94: 496–509) model for competing-risks data. Geskus (2011, Biometrics 67: 39–49) has recently shown that standard software can estimate some of the key measures in competing risks by restructuring the data and incorporating weights. This has a number of advantages because any tools developed for standard survival analysis can then be used to analyze competing-risks data. In this article, I describe the stcrprep command, which restructures the data and calculates the appropriate weights. After one uses stcrprep, a number of standard Stata survival analysis commands can then be used to analyze competing risks. For example, sts graph, failure will give a plot of the cause-specific CIF, and stcox will fit the Fine and Gray (1999) proportional subhazards model. Using stcrprep together with stcox is computationally much more efficient than using stcrreg. In addition, stcrprep opens up new opportunities for competing-risk models. I illustrate this by fitting flexible parametric survival models to the expanded data to directly model the cause-specific CIF.
The Estimation and Modeling of Cause-specific Cumulative Incidence Functions Using Time-dependent Weights
