Data maturity and follow-up in time-to-event analyses

Area of Science:

• Biostatistics
• Clinical Trials
• Survival Analysis

Background:

• Kaplan-Meier survival plots are crucial for time-to-event outcomes.
• Determining the meaningful duration of these plots is essential for accurate interpretation.
• Current methods may lack clear guidelines for plot curtailment.

Purpose of the Study:

• To propose methods for determining the minimum number of subjects at risk for meaningful Kaplan-Meier survival estimates.
• To provide guidance on when to curtail survival plots to avoid unreliable conclusions.
• To offer practical approaches for assessing data maturity in survival analyses.

Main Methods:

• Investigating the decrease in survival estimates S(t) if an additional event occurs.
• Implementing an absolute decrease threshold for S(t).
• Calculating a minimum acceptable number of subjects at risk using a confidence interval approach.

Main Results:

• Two distinct methods are proposed to calculate limits for the number of subjects still at risk.
• The methods allow investigators to make informed decisions based on study context.
• The approaches are illustrated with published studies, addressing sample size and censoring variations.

Conclusions:

• The proposed methods provide a framework for determining the appropriate curtailment of Kaplan-Meier survival plots.
• These methods enhance the reliability of survival estimates by defining meaningful endpoints for analysis.
• The simple calculation using standard statistical package outputs facilitates widespread adoption.