On Share Frailty Cure Model: An Application On Cervical Cancer

Survival analyses are greatly used in medical research especially frailty models which are mostly used to account for heterogeneity in time-to-event. Over the years treatment of cancer has progressed with some patients being cured from different type of cancer. Survival analysis is more focused on subjects that are less at risk of recurrences, metastasis or death after the first treatment as these set of subjects are regarded as being cured. The general assumption of standard frailty model is that all subjects have the same frailty. These assumptions ignore the heterogeneity of such frailties and will lead to incorrect results and conclusions. To address the identified deficiencies in previous studies, this research will propose a shared frailty cure model. Shared frailty assumes that within a cluster the value of frailty term is improved with constant and common frailty to all subjects in the same group clusters by measuring the correlation between event times within the cluster, hence representing changes over time in clusters or population heterogeneity. These structures can be achieved by introducing covariates that are rank specific by the process Shared frailty model, addressing the weakness of the cure frailty model by considering the homogeneity in groups or clusters were failure can be similar by having the same frailty.