A generalized log-rank-type test for comparing survivals with doubly interval-censored data

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Methods

Background:

• Doubly interval-censored data presents challenges in survival analysis as both event times are not precisely observed.
• Existing methods, like Sun's nonparametric test, rely on estimating marginal survival functions.

Purpose of the Study:

• To develop a generalized log-rank-type test for comparing survival functions in multiple groups with doubly interval-censored data.
• To introduce a novel method that avoids estimating marginal survival functions.

Main Methods:

• A new nonparametric test utilizing uniform weights based on risk set sizes.
• The method simplifies to the standard log-rank test for right-censored data.
• Comparison with existing methods through simulation studies.

Main Results:

• The proposed test demonstrates good performance in terms of statistical size and power.
• Simulation results indicate the test is effective for analyzing doubly interval-censored data.
• The method was successfully applied to AIDS cohort study data.

Conclusions:

• The developed uniform-weighted test is a valuable tool for survival analysis with doubly interval-censored data.
• This approach offers a more direct and potentially more powerful alternative to existing methods.
• The test's ability to reduce to the log-rank test enhances its applicability.