Comparison of Methods to Analyze Time-to-Event Endpoints with Delayed Treatment Effect

Simulation study comparing statistical methods for analyzing time-to-event endpoints when treatment effects are delayed, violating proportional hazards assumptions. Evaluates type I error and power of weighted logrank tests, combinations, and regression-based alternatives.

Use Cases

• Method comparison for selecting appropriate tests when proportional hazards is violated
• Power analysis under non-proportional hazards scenarios
• Simulation-based evaluation of weighted logrank tests

Strengths

• Extensive simulation covering 40+ methods
• Focus on type I error control and power
• Robust results not sensitive to parameter misspecification

Limitations

• Interpretation complicated by absence of interpretable summary measures
• No column-level documentation; field semantics must be inferred
• Dataset scale unknown; may limit suitability assessment

Provenance

Source

figshare

Collection Method

Platform upload