Eligibility of emergency psychiatry patients for clinical trials studying depression

Abstract

Background: Differences often exist between people with depression who are eligible for clinical trials and those seen in clinical practice. The impact of exclusion criteria on eligibility has been previously reported for inpatients and outpatients, but has not been assessed for emergency psychiatry patients; a group that overlaps with inpatients and outpatients but also has important distinctives. Understanding the frequencies of commonly used exclusion criteria in this population could inform interpretation of existing data (generalizability) and highlight opportunities/needs for future trials.

Methods: We reviewed 67 clinical trials studying depression using Qualitative Content Analysis to identify common and recurring exclusion criteria. We examined the frequency of these exclusion criteria among a clinical sample of emergency psychiatry patients.

Results: Most clinical trials had exclusions for basic research requirements, age, symptom severity, psychosis, and substance use. Applying 9 commonly used exclusion criteria to the clinical population resulted in a 3.3 % eligibility rate (95 % CI 1.2 %-7.0 %). Exclusions for psychosis (85.1 % of trials), substance use (83.6 % of trials), and suicide risk (65.7 % of trials) would likely exclude 93 % of emergency psychiatry patients. The prevalence of psychosis, substance use, and suicide risk was much higher among emergency psychiatry patients than among previously studied populations.

Limitations: Some eligibility criteria could not be measured. The Qualitative Content Analysis consolidated similar exclusion criteria, losing potentially important nuances in wordings.

Conclusions: Exclusion criteria commonly used in contemporary clinical trials of depression limit generalizability to emergency psychiatry patients, due in large part to exclusions for psychosis, substance use, and suicide risk.

Keywords: Clinical trial; Depression; Emergency psychiatry; Exclusion criteria.