On the catch-up time method for analyzing cancer screening trials

Area of Science:

• Biostatistics
• Epidemiology
• Clinical Trial Design

Background:

• Cancer screening trials face challenges with time-varying mortality ratios due to lag and dilution effects.
• Conventional analysis tools like the logrank test may lose power in these scenarios.

Purpose of the Study:

• To evaluate the bias and power properties of the "catch-up time" method proposed by Aron and Prorok.
• To investigate alternative adaptive tests for cancer screening trial analysis.

Main Methods:

• Simulated data resembling the Health Insurance Plan of Greater New York study cohort were used.
• Models for postdiagnosis survival were considered under both null (no screening effect) and alternative (screening effect) hypotheses.
• The bias and power of the catch-up time method were thoroughly evaluated.

Main Results:

• The catch-up time method can introduce significant bias into cancer screening trial analyses.
• The standard logrank test may be less powerful due to time-dependent effects.
• Adaptive tests were identified as potentially more effective.

Conclusions:

• The catch-up time method is not always reliable for analyzing cancer screening trial data.
• Adaptive statistical tests offer a promising alternative for maintaining test size and improving power in the presence of complex disease dynamics.