Frailty models for understanding breast cancer survival heterogeneity

In this paper, we study frailty models as a means to understand heterogeneity in survival outcomes among breast cancer patients. We analyzed data of 200 patients treated at the Cancer Institute (WIA) in Chennai over an eight-year period, focusing on the variability in survival times. Utilizing both Cox Proportional Hazards (PH) and Accelerated Failure Time (AFT) regression models, we estimated parameters that differentiate between cured and uncured populations while accounting for shared frailty.Our findings indicate that the Lognormal AFT frailty model provides the best fit for examining the impact of covariates on survival times, while the Weibull distribution was identified as the most appropriate for the Cox PH model. This research highlights the importance of incorporating frailty to better understand the complexities of breast cancer survival and the influence of various factors on patient outcomes. Comprehensive analyses of these results are detailed in this paper.