Joint modeling of longitudinal outcomes and survival using latent growth modeling approach in a mesothelioma trial

Ping Wang,Mark Ernest Boye,Wei Shen

doi:10.1007/s10742-012-0092-z

Abstract

Joint modeling of longitudinal and survival data can provide more efficient and less biased estimates of treatment effects through accounting for the associations between these two data types. Sponsors of oncology clinical trials routinely and increasingly include patient-reported outcome (PRO) instruments to evaluate the effect of treatment on symptoms, functioning, and quality of life. Known publications of these trials typically do not include jointly modeled analyses and results. We formulated several joint models based on a latent growth model for longitudinal PRO data and a Cox proportional hazards model for survival data. The longitudinal and survival components were linked through either a latent growth trajectory or shared random effects. We applied these models to data from a randomized phase III oncology clinical trial in mesothelioma. We compared the results derived under different model specifications and showed that the use of joint modeling may result in improved estimates of the overall treatment effect.

Highlights

Clinical research often generates both longitudinal and survival data
We fitted the Cox proportional hazards model with treatment as the only covariate and the hazard ratio (HR) for treatment was HR0 = 0.73 (P-value = 0.001)
Our joint models produced different and seemingly more accurate results compared with models focused on patient-reported outcome (PRO) alone, or survival alone, or the naive model ignoring measurement errors in PROs by directly handling informative censoring and accounting for measurement errors

Summary

Introduction

Clinical research often generates both longitudinal and survival (time-to-event) data. Semiparametric methods such as Cox proportional hazards models and parametric methods such as Weibull models are commonly used. Separate analyses of each type of outcome may not be able to provide adequate answers to some important research questions. One such example is whether CD4 lymphocyte count could serve as a good surrogate marker for clinical progression in AIDS clinical trials (Tsiatis et al 1995).

Methods

Results

Conclusion