Analysis of interval censored survival data in sequential multiple assignment randomized trials

This article delves into the analysis of interval-censored survival data within sequential multiple assignment randomized trials (SMARTs), a critical area for chronic disease management. The focus is on comparing adaptive treatment strategies, which involve sequential decision rules based on patient responses. The authors introduce a weighted spline-based sieve maximum likelihood method to handle the complexities of interval-censored data, a common issue in clinical trials where event times are only known to occur between two adjacent visits. The article also presents a simulation study to assess the performance of the proposed methodology and applies it to data from the STAR*D trial, comparing various treatment strategies for depression. The results highlight the effectiveness of the method in maintaining valid type I error rates and providing reliable estimates of treatment effects. This article offers a comprehensive approach to analyzing interval-censored data in SMARTs, making it an essential read for professionals seeking to improve the accuracy and reliability of their clinical trial analyses.