737 Prediction of drop out in an adjunctive light treatment trial in patients with non-seasonal depression and evening chronotype

Abstract

Abstract Introduction Drop out during treatment hampers therapeutic effect of interventions. The current study examines the possible predictors of drop out during the five-week light treatment in patients with unipolar non-seasonal depression and evening-chronotype. Methods Baseline characteristics including demographics, sleep diary parameters, light treatment prescribed, and early clinical outcomes changes were compared between the Drop out and Non drop out group. Logistic regression analysis was used to examine predictors for drop out. All data were analyzed in a modified intention to treat analysis with last observation carried forward approach. Results A total of 91 subjects (Female 79%, 46.3 ± 11.8 years old) were included in the analysis. There was no significant difference in the baseline demographic and clinical characteristics between the Drop out and Non drop out group. There was also no significant difference in the improvement of clinical parameters over the first week among the two groups. However, treatment non-adherence (in terms of compliance of less than 80% of prescribed duration) over the first treatment week predicts a five-fold increase in risk of drop out during light therapy. (OR: 5.85, CI: 1.414–24.205, p=0.015) after controlling for potential confounders including age, gender, treatment group, patient expectation, and treatment-emergent adverse events. Conclusion This study found that baseline clinical characteristics including depression severity and improvement of depressive symptoms in the initial week did not differ between the Drop out and Non drop out group. The drop out was also not affected by the type of light (dim red versus bright red light), indirectly supporting dim red light as a valid placebo in bright light therapy trial. Treatment adherence is the early phase of light treatment is an important predictor of drop out. Support (if any):