Accelerated failure model with empirical analysis and application to colon cancer data: Testing and validation

This paper introduces and examines a novel accelerated failure time (AFT) model, a versatile multi-parameter regression model suitable for representing diverse time-to-event datasets. Serving as an alternative to the Cox model and other established AFT models, the proposed model’s parameters are estimated through maximum likelihood estimation. A Monte Carlo simulation study is conducted to evaluate the proposed model’s performance across various scenarios, demonstrating its superiority over competing AFT models. Application of the proposed model to a colon cancer dataset reveal its superior parametric fit compared to the Cox proportional hazards (Cox-PH) model and other competing AFT models, as indicated by information criteria and goodness-of-fit measures. This new AFT model contributes to the repertoire of tools/models for analyzing survival datasets and offers an advantageous hazard-based regression approach.