Destructive cure models with proportional hazards lifetimes and associated likelihood inference

TL;DR

This paper develops and evaluates destructive cure models with proportional hazards lifetimes, using flexible distributions and likelihood inference, applied to survival data including melanoma cases.

Contribution

It introduces a new class of destructive cure models with flexible initial cause distributions and proportional hazards, along with EM and profile likelihood estimation methods.

Findings

Simulation studies show accurate and robust parameter estimation.

Model misspecification impacts estimates significantly.

Real data analysis demonstrates practical applicability.

Abstract

In survival analysis, cure models have gained much importance due to rapid advancements in medical sciences. More recently, a subset of cure models, called destructive cure models, have been studied extensively under competing risks scenario wherein initial competing risks undergo a destructive process, such as under a chemotherapy. In this article, we study destructive cure models by assuming a flexible weighted Poisson distribution (exponentially weighted Poisson, length biased Poisson and negative binomial distributions) for the initial number of competing causes and with lifetimes of the susceptible individuals following proportional hazards. The expectation-maximization (EM) algorithm and profile likelihood approach are made use of for estimating the model parameters. An extensive simulation study is carried out under various parameter settings to examine the properties of the models, and the accuracy and robustness of the proposed estimation technique. Effects of model misspecification on the parameter estimates are also discussed in detail. Finally, for the illustration of the proposed methodology, a real-life cutaneous melanoma data set is analyzed.