Exclusion criteria and generalizability in bipolar disorder treatment trials.

Jessie J Wong,Christine Timko,Keith Humphreys,Nev Jones

doi:10.1016/j.conctc.2018.01.009

Abstract

ObjectiveThe current paper reviews the English-language research on exclusion criteria in bipolar disorder treatment trials and discusses how study samples compare to the general bipolar patient population.Methods& Results: Across 8 identified studies of exclusion criteria and their impact, between 55% and 96% of people with bipolar disorder would be excluded from treatment research. The number of exclusion criteria varies across bipolar disorder treatment research, with one study estimate of a median of 7 criteria used across studies. The criteria that excluded the greatest number of potential participants were comorbid substance use disorder, suicidal risk, and comorbid medical conditions. Both studies that compared treatment responses among participants who met and did not meet exclusion criteria found no statistically significant differences.ConclusionsMost potential participants are excluded from outcome research, which creates challenges for recruitment and limits generalizability of study findings. Common exclusionary practices lead to unrepresentative samples that limit generalizability and reduce the confidence of clinicians that findings can be translated to front-line practice with bipolar disorder patients.

Highlights

Bipolar disorder affects 4.4% of the population at some point in their lifespan [1] and often causes significant disruptions to work, social, and family life domains [2] as well as increased suicidal risk [3]
Clinical trials across a range of psychiatric disorders have traditionally attempted to recruit samples of individuals with symptoms that emanate exclusively from their primary diagnosis [5]. This approach to sample selection has raised concerns regarding the generalizability of research samples to ‘real-world’ community patient populations, most whether “evidencebased” treatments are effective for the severely troubled patients who tend to be excluded from clinical trials [6]
This paper reviews the literature on the exclusion criteria that have been employed in bipolar disorder treatment research, the proportion of patients excluded, and how exclusion criteria may affect the generalizability of results

Summary

Introduction

Bipolar disorder affects 4.4% of the population at some point in their lifespan [1] and often causes significant disruptions to work, social, and family life domains [2] as well as increased suicidal risk [3]. The nature of bipolar disorder and its comorbidities present unique challenges to treatment researchers, including how to select exclusion criteria that balance rigor and relevance [4]. Clinical trials across a range of psychiatric disorders have traditionally attempted to recruit samples of individuals with symptoms (and related impairment) that emanate exclusively from their primary diagnosis [5]. This approach to sample selection has raised concerns regarding the generalizability of research samples to ‘real-world’ community patient populations, most whether “evidencebased” treatments are effective for the severely troubled patients who tend to be excluded from clinical trials [6]. This paper reviews the literature on the exclusion criteria that have been employed in bipolar disorder treatment research, the proportion of patients excluded, and how exclusion criteria may affect the generalizability of results

Methods

Results

Conclusion