Cognitive-behavioral therapy vs. light therapy for preventing winter depression recurrence: study protocol for a randomized controlled trial

Background: Seasonal affective disorder (SAD) is a subtype of recurrent depression involving major depressive episodes during the fall and/or winter months that remit in the spring. The central public health challenge in the management of SAD is prevention of winter depression recurrence. Light therapy (LT) is the established and best available acute SAD treatment. However, long-term compliance with daily LT from first symptom through spontaneous springtime remission every fall/winter season is poor. Time-limited alternative treatments with effects that endure beyond the cessation of acute treatment are needed to prevent the annual recurrence of SAD. Methods/design: This is an NIMH-funded R01-level randomized clinical trial to test the efficacy of a novel, SAD-tailored cognitive-behavioral group therapy (CBT) against LT in a head-to-head comparison on next winter outcomes. This project is designed to test for a clinically meaningful difference between CBT and LT on depression recurrence in the next winter (the primary outcome). This is a concurrent two-arm study that will randomize 160 currently symptomatic community adults with major depression, recurrent with seasonal pattern, to CBT or LT. After 6 weeks of treatment in the initial winter, participants are followed in the subsequent summer, the next winter, and two winters later. Key methodological issues surround timing study procedures for a predictably recurrent and time-limited disorder with a focus on long-term outcomes. Discussion: The chosen design answers the primary question of whether prior exposure to CBT is associated with a substantially lower likelihood of depression recurrence the next winter than LT. This design does not test the relative contributions of the cognitive-behavioral treatment components vs. nonspecific factors to CBT's outcomes and is not adequately powered to test for differences or equivalence between cells at treatment endpoint. Alternative designs addressing these limitations would have required more patients, increased costs, and reduced power to detect a difference in the primary outcome.