The potential of predictive analytics to provide clinical decision support in depression treatment planning

Abstract

Purpose of review: To review progress developing clinical decision support tools for personalized treatment of major depressive disorder (MDD).

Recent findings: Over the years, a variety of individual indicators ranging from biomarkers to clinical observations and self-report scales have been used to predict various aspects of differential MDD treatment response. Most of this work focused on predicting remission either with antidepressant medications versus psychotherapy, some antidepressant medications versus others, some psychotherapies versus others, and combination therapies versus monotherapies. However, to date, none of the individual predictors in these studies has been strong enough to guide optimal treatment selection for most patients. Interest consequently turned to decision support tools made up of multiple predictors, but the development of such tools has been hampered by small study sample sizes. Design recommendations are made here for future studies to address this problem.

Summary: Recommendations include using large prospective observational studies followed by pragmatic trials rather than smaller, expensive controlled treatment trials for preliminary development of decision support tools; basing these tools on comprehensive batteries of inexpensive self-report and clinical predictors (e.g., self-administered performance-based neurocognitive tests) versus expensive biomarkers; and reserving biomarker assessments for targeted studies of patients not well classified by inexpensive predictor batteries.