Proportional Hazards Cure Model for the Analysis of Time to Event with Frequently Unidentifiable Causes

• Published in: Biometrics Vol. 63; no. 4; pp. 1237 - 1244
• Main Authors: Dahlberg, Suzanne E, Wang, Molin
• Format: Journal Article
• Language: English
• Published: Malden, USA Blackwell Publishing Inc 01.12.2007; International Biometric Society; Blackwell Publishing Ltd
• Subjects: Aesthetic proportionality; Algorithms; Biometrics; Breast cancer; Breast Neoplasms - epidemiology; Breast Neoplasms - therapy; Causality; Censorship; Competing risks; Cure rate; Estimation methods; Female; Humans; Logistic regression; Masked cause; Measurement techniques; Menopause; Modeling; Outcome Assessment (Health Care) - methods; Parametric models; Proportional hazards model; Proportional Hazards Models; Reliability functions; Research methodology; Risk Assessment - methods; Risk Factors; Sensitivity and Specificity; Time Factors; Treatment Outcome
• ISSN: 0006-341X; 1541-0420
• DOI: 10.1111/j.1541-0420.2007.00811.x

We propose a semiparametric method for the analysis of masked-cause failure data that are also subject to a cure. We present estimators for the failure time distribution, the cure rate, and the covariate effect on each of these, assuming a proportional hazards cure model for the time to event of interest and we use the expectation-maximization algorithm to conduct the likelihood maximization. The method is applied to data from a breast cancer clinical trial.