The estimation and modelling of cause-specific cumulative incidence functions using time-dependent weights

Area of Science:

• Biostatistics
• Survival Analysis
• Epidemiology

Background:

• Competing risks present challenges in survival analysis, where multiple event types exist and one event prevents others.
• The cause-specific cumulative incidence function (CIF) is crucial for estimating absolute risk in competing risks scenarios.
• Existing Stata commands like stcompet and stcrreg have limitations in flexibility and efficiency for competing risks.

Purpose of the Study:

• To introduce the stcrprep command for restructuring competing risks data in Stata.
• To enable the application of standard survival analysis tools to competing risks data.
• To offer a more computationally efficient alternative to existing methods for competing risks analysis.

Main Methods:

• Data restructuring and weighting using the novel stcrprep command.
• Application of standard Stata survival analysis commands (e.g., sts graph, stcox) to the prepared data.
• Fitting flexible parametric survival models to the expanded dataset for direct CIF modeling.

Main Results:

• The stcrprep command facilitates the use of standard survival analysis tools for competing risks.
• Using stcrprep with stcox is more computationally efficient than the stcrreg command.
• The approach allows for advanced modeling, including flexible parametric models for cause-specific CIFs.

Conclusions:

• The stcrprep command enhances the analysis of competing risks data in Stata.
• This method provides a flexible, efficient, and accessible approach to competing risks modeling.
• It opens new avenues for research by integrating competing risks analysis with standard survival tools.