Modelling survival data with a cured fraction using frailty models.

Area of Science:

• Biostatistics
• Survival Analysis
• Medical Statistics

Background:

• Cure models are traditionally used for time-to-event data with a cured fraction.
• These models analyze patient outcomes, considering a portion of the population that will not experience the event of interest.

Purpose of the Study:

• To evaluate frailty models as an alternative to standard cure models for time-to-event data.
• To incorporate heterogeneity in risk and a cured component in the analysis.
• To analyze leukaemia recurrence data in patients undergoing autologous transplantation.

Main Methods:

• Utilized frailty models, specifically the gamma frailty mixture model and the compound Poisson model.
• Employed maximum likelihood techniques for model fitting.
• Applied models to real-world data on leukaemia recurrence post-transplantation.

Main Results:

• Frailty models demonstrated an improved fit compared to the standard cure model for leukaemia recurrence data.
• The gamma frailty mixture and compound Poisson models provided a better analysis of the time-to-event data.
• Heterogeneity in risk and the cured fraction were effectively modeled.

Conclusions:

• Frailty models are a valuable alternative for analyzing time-to-event data with cured fractions.
• These models offer enhanced capabilities in handling risk heterogeneity.
• The study highlights the utility of frailty models in cancer recurrence studies, such as leukaemia.