ASSESSING THE EFFECT OF TREATMENT REGIMES ON LONGITUDINAL OUTCOME DATA: APPLICATION TO REVAMP STUDY OF DEPRESSION

Abstract

Depression studies frequently adopt two-stage designs to examine the efficacy of augmenting
pharmacotherapy with psychotherapy. Initially subjects receive one of the several treatments;
if they respond, they continue the same treatment; however, if they fail to respond,
they move to the next stage and are randomized to other treatment options. Outcomes such
as 24-item Hamilton Rating Scale of Depression (HRSD24) scores are then collected repeatedly
to monitor the progress of the subject. The goal is to assess the effect of treatment
regimes (consisting of initial treatment, initial response and the second stage treatment
combinations) on HRSD24 profile. Statistical inference for assessing treatment regimes
using a summary outcome measure such as mean response has been well-studied in the
literature. Statistical methods for assessing the effect of treatment strategies on repeated
measures data focused mainly on estimating equations. In this article, we propose two
methods based on mixed models and multiple imputations to assess the effect of treatment
regimes on the longitudinal HRSD24 scores. Methods are compared through simulation
studies and through an application to a depression study. The simulation studies showed
that the estimates from both methods are approximately unbiased, and provide good coverage
rates for 95% confidence intervals.