Estimating Optimal Treatment Rule for Major Depressive Disorder Using Penalized Regression Method.

Objectives: Major depressive disorder (MDD) stands as the primary contributor to disability worldwide. Identifying optimal treatment regimens for patients with MDD using advanced statistical techniques may help improve patient outcomes and reduce the number of hospitalizations. Methods: In a group of patients with MDD from north-western Iran, we compared treatments including work therapy (WT), WT plus electroconvulsive therapy (WT + ECT), WT plus family therapy (WT + FT), and other psychotherapeutic methods (PT). We also estimated the optimal treatment rule and identified essential variables in a loss-based framework using a penalized regression method. Results: The participants were 377 MDD patients of whom 198 (52.5%) received WT alone, 95 (25.2%) received WT + ECT, and 61 (16.2%) were given WT + FT. The remaining 23 (6.1%) patients were treated with PT. A comparison of the treatments revealed that a history of emotional problems was the important variable to consider when selecting WT + ECT, WT + FT, or PT, while patient education level and history of emotional problems were both important for WT + ECT. Applying the above optimal treatment rules is likely to reduce patients' hospital stay days. Conclusions: For patients with MDD, history of emotional problems and education level were the two most important variables for estimating the optimal treatment rules, including personalizing medications. Incorporating important variables into treatment regimens is likely to improve treatment outcomes and decrease the number of hospitalizations.