The DSM-5 debate over the bereavement exclusion: Psychiatric diagnosis and the future of empirically supported treatment

•Fruitful treatment research requires valid, generalizable diagnostic categories.•Construct validity eludes us; focus on conceptual validity (normal vs. disordered).•Bereavement-excluded major depression lacks elevated recurrence and other validators.•The exclusion extended to other stressors has concurrent and predictive validity.•DSM-5's elimination of the bereavement exclusion undermined conceptual validity.

Valid diagnostic criteria support generalizations about treatment effectiveness, allowing progress in developing empirically supported treatments. The DSM-5 revision provides an opportunity to consider whether diagnostic changes are increasing validity. In this paper, I first offer broad suggestions for conceptually advancing diagnostic validity while awaiting greater etiological understanding. These include, for example, improving “conceptual validity” (disorder/nondisorder differentiation); extending diagnosis beyond disorders to include mismatches between normal variation and social demands (“psychological justice”); placing disorder etiology in evolutionary context as harmful failure of biologically designed functioning (“harmful dysfunction”); and taking an integrative theoretical approach to human meaning systems. The paper then examines the DSM-5's controversial decision to eliminate the major depression bereavement exclusion (BE), detailing the evidence and attendant debate. Elimination was defended by citing several hypotheses (e.g., excluded cases are similar to other MDD; exclusions risk missing suicidal cases; medication works with excluded cases), all of which were either empirically falsified or based on faulty arguments. Most dramatically, excluded cases were empirically demonstrated to have no more depression on follow-up than those who never had MDD. I conclude that BE elimination undermined rather than increased conceptual validity and usefulness for treatment research. Finally, I draw some general lessons from the DSM-5 BE debacle.