A Closer Look at Exclusion Criteria in Randomized Controlled Psychotherapy Trials for Adult Depression

A common concern regarding clinical trials evaluating treatments for depression is findings may not generalize to the population of interest due to the use of restrictive exclusion criteria. To evaluate the exclusion criteria used in such research, we examined a large body of reports of randomized controlled trials of psychotherapies for adult depression. We analyzed the rates of four major types of exclusions: suicide related exclusions, an exclusion for when depression is not deemed the primary or principal diagnosis, exclusions due to comorbid diagnoses, and exclusions related to substance use. Data were drawn from a compilation of randomized controlled and comparative trials of psychotherapy on adult depression assembled by Professor Pim Cuijpers. Our analyses include 332 of the original 358 studies, as we excluded prevention studies, meta analyses, reports not available in English, and those that we could not access. Two raters coded each of these studies. Reliability for the raters' judgements ranged from moderate to high (Cohen's kappa of .52 to 1.00). Of the studies examined, 42% used a suicide related exclusion, 72% excluded for the presence of specific comorbid disorders, 63% excluded participants who did not have depression as a primary / principal diagnosis, and 52% excluded participants on the basis of substance use. Within those broader categories, the following specific exclusions were found to be used increasingly over time: use of a measure for operationalizing a suicide-related exclusion, self-harm, any bipolar spectrum disorder, current alcohol abuse, and current drug abuse. In contrast, four exclusions were found to be used with reduced frequency over time: requiring hospitalization, antisocial personality disorder (ASPD), having no disorder-based exclusions, and past alcohol abuse. These changes show that exclusion criteria have changed over the course of the last fifty years. To our knowledge, these changes have been unidentified to date. While researchers may disagree about what exclusion criteria should be used in trials, specific differences have gotten little attention in the literature, perhaps preventing a consensus about under what circumstances various criteria should be used. We hope our work can help to prompt greater consideration of the use of specific exclusion criteria.No embargoAcademic Major: Psycholog