Online Program Home
My Program

Abstract Details

Activity Number: 116 - Recent Advances in Cure Rate Models for Long-Term Survivors
Type: Topic Contributed
Date/Time: Monday, July 29, 2019 : 8:30 AM to 10:20 AM
Sponsor: Biometrics Section
Abstract #301696 Presentation
Title: Applying Cure Models with Competing Risks and Complex Censoring Patterns
Author(s): Jeremy Taylor* and Lauren Beesley
Companies: University of Michigan and University of Michigan
Keywords: cure models; competing risks; multistate models; head and neck cancer; Monte Carlo EM

Mixture cure models can be used to analyze time-to-event data when there is a non-susceptible group of patients who are not at risk for the event. The model is particularly applicable for cancers where it is believed that some patients can be cured and will not experience a recurrence of the cancer. Competing risks, such as death from other causes, are common and may prevent observation of the recurrence. The multistate cure model, which allows for competing risks, is a natural extension of the cure model. The multistate cure model specifies a model for the probability of cure and three intensity processes for the transitions from initial state to recurrence, from initial state to death, and from recurrence to death. The model can include baseline covariates that can affect the probability of cure and the intensity processes in different ways, which can enhance interpretation. Monte Carlo EM and Bayesian methods can be used for estimation. A common occurrence is unequal censoring times for recurrence and death. For this, we propose to impute recurrence events for those who are censored for recurrence before death. The methods will be applied to head and neck cancer data.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program