ReviewThe exclusion of people with psychiatric disorders from medical research
Introduction
People with psychiatric disorders experience cancer, heart disease and other medical disorders at rates as high or higher than the rest of the population. However, they are underrepresented in trials of treatments for these disorders because investigators – with the best of intentions – design research protocols that do not permit people with psychiatric disorders to enroll (Humphreys et al., 2005, Stapleton, 2010). This pattern of exclusion is seen within psychiatric medical research (e.g., excluding patients with alcohol use disorders from enrolling in clinical trials of depression treatment) and outside of it (e.g., excluding patients with psychotic disorders from clinical trials of cancer treatment).
The most common ethical arguments for excluding people with psychiatric disorders from medical research – which may lack empirical support (Roberts et al., 2006) – are that they may be more vulnerable to exploitation due to decisional capacity deficits or greater susceptibility to coercive pressures, and, may be objectively at higher risk of harm in cases where the researched treatment (e.g., aggressive chemotherapy) could impose stress that might exacerbate their psychiatric disorder. Exclusion criteria can also be argued for on statistical grounds: If those with psychiatric disorders have markedly different outcomes than those without such disorders, the power of a trial to detect the effect of a medical treatment is lessened (Lipsey, 1990).
Yet from a different perspective, the exclusion of people with psychiatric disorders from medical treatment research may be considered unethical. Because psychiatric conditions are highly prevalent and often co-occur with other medical illnesses, declining to study people with psychiatric disorders could be considered a serious form of scientific neglect and therefore a social injustice. Results derived from clinical research guide state-of-the-art medical practice in the health care system, particularly if a trial is widely-cited and high profile (Rothwell, 2005). If people with psychiatric disorders respond differently to a treatment (e.g., a newly developed pain medication interferes with metabolization of their psychiatric medication) but are not included in the research evaluating it, the disparate impact will not be known until they receive that treatment on a broad scale in the health care system. In that sense, excluding people with psychiatric disorders from medical research may not so much reduce their risk of harm as shift it from a small number of people in a research study to a far larger number of people in the health care system. Because tens of millions of Americans have psychiatric disorders – including some highly vulnerable subpopulations (Martins et al., 2012) – the cumulative impact of this shifting of risk into everyday health care provision could exacerbate existing health disparities and erode public confidence in the value of the medical research enterprise.
As for whether exclusion is justified because it increases statistical power, this depends on whether the moderators of treatment outcome are known. With new treatments, whose main effect is not even established, knowing moderators in advance is logically impossible. Even for widely employed treatments, evidence suggests that researchers can guess wrong and find that an exclusion criterion has lowered rather than raised a clinical trial's statistical power (Humphreys et al., 2008). From an ethical perspective, one should further consider the question of why it is acceptable is to exclude people with psychiatric disorders from participation as the primary solution, rather than refining scientific designs or augmenting statistical power in other ways, such as improving reliability of measurement or increasing sample size (Kraemer and Thiemann, 1987).
Irrespective of where researchers come down in this debate, it would be highly useful to all parties to know the basic facts: How often are people with psychiatric disorders actually excluded from medical research, and which sorts of individuals are excluded? Accordingly, the present study engages these questions by examining the enrollment practices of the most widely-cited clinical trials of recent years, both within and without psychiatric medicine.
Section snippets
Materials and methods
Using standardized search terms across databases in Web of Science (see Humphreys et al., 2013), we identified the 20 top-cited randomized controlled trials with results published from 2002 to 2010 for each of a varied group of 20 prevalent chronic disorders (i.e., a total sample of 400 trials). A starting date of 2002 was chosen because the CONSORT criteria were revised in 2001 to clarify that reporting of enrollment procedures was required for clinical trials (Moher et al., 2001). To avoid a
Results
Across the 400 trials, 38% reported at least one definite psychiatric exclusion criterion, 20% reported at least one possible psychiatric exclusion criterion, and 50% reported use of either or both. The prevalence of psychiatric exclusion criteria varied across disorders (F[19,380] = 9.55, p < .001; see Fig. 1). The disorders with the five highest rates of possible or definite psychiatric exclusion criteria were all psychiatric: alcohol use disorder (95%), depression (95%), schizophrenia (85%),
Discussion
Half of the most-widely cited recent clinical trials in medicine employed exclusion criteria that definitely and/or possibly prevented enrollment by individuals with psychiatric disorders. Because highly cited trials influence medical practice, this finding raises the question of whether clinical care is evidence-based and attuned to the true health concerns of people living with mental disorders. This is particularly the case for medical conditions which are unusually prevalent among
Contributors
All authors participated in critically revising all sections of the manuscript, and have approved of the final version. In addition, Dr. Humphreys and Dr. Roberts conceptualized and designed the study; Dr. Humphreys directed the study implementation, including designing literature search processes and data analyses, drafting the manuscript and designing the table and figure; Dr. Roberts contributed to the table design and drafting of the Discussion section of the text; and Ms. Blodgett
Role of the funding source
The Greenwall Foundation had no involvement in the writing or review of this manuscript.
Conflicts of interest
None.
Acknowledgment
This project was made possible by a grant from The Greenwall Foundation. We are grateful to Leena Bui for help with coding and Phil Lavori for assistance with conceptualizing the project.
References (21)
- et al.
Alternative decision makers' perspectives on assent and dissent for dementia research
Am. J. Geriatr. Psychiatry
(2013) External validity of randomised controlled trials: “to whom do the results of this trial apply?”
Lancet
(2005)- et al.
Physical illness in patients with severe mental disorders. I. Prevalence, impact of medications and disparities in health care
World Psychiatry
(2011) - et al.
Eligibility criteria for HIV clinical trials and generalizability of results: the gap between published reports and study protocols
AIDS
(2005) - et al.
The MacArthur treatment competence Study. III: abilities of patients to consent to psychiatric and medical treatments
Law Hum. Behav.
(1995) - et al.
Reporting the recruitment process in clinical trials: who are these patients and how did they get there?
Ann. Intern. Med.
(2002) - et al.
Are participants in pharmacological and psychotherapy treatment trials for social anxiety disorder representative of patients in real-life settings?
J. Clin. Psychopharmacol.
(2014) - et al.
Indicators for Chronic Disease Surveillance—United States, 2013
(2015) - et al.
Subject eligibility criteria can substantially influence the results of alcohol-treatment outcome research
J. Stud. Alcohol Drugs
(2008) - et al.
Extent and reporting of patient nonenrollment in influential randomized clinical trials, 2002 to 2010
JAMA Intern. Med.
(2013)
Cited by (39)
Exclusion of patients with psychiatric disorders from early phase oncology trials: A limited access to innovation?
2024, General Hospital PsychiatryOlder adults with clinically relevant depressive symptoms have equal mobility benefit from a chronic physical activity intervention
2023, Mental Health and Physical ActivityPalliative Care Interventions Effects on Psychological Distress: A Systematic Review & Meta-Analysis
2023, Journal of Pain and Symptom ManagementStudy protocol for a randomized trial of bridge: Person-centered collaborative care for serious mental illness and cancer
2022, Contemporary Clinical TrialsReconsidering research exclusion for serious mental illness: Ethical principles, current status, and recommendations
2021, Journal of Psychiatric ResearchCitation Excerpt :Results indicated that individuals with psychiatric disorders were excluded from medical trials about half of the time. Research on many common medical conditions which differentially affect people who manage psychiatric disorders were also likely to exclude those with psychiatric disorders (e.g., HIV/AIDS: 17%, low back pain: 55%, hypertension: 10%) (Humphreys et al., 2015). Psychiatric exclusion for studies of diabetes ranged from 5–28%, and for studies of ischemic heart disease it was 3–12%.