Participant exclusion criteria in treatment research on neurological disorders: Are unrepresentative study samples problematic?

ObjectiveExclusion criteria are an important determinant of the external validity of treatment research findings, yet the prevalence and impact of exclusion criteria have not been studied systematically. Our objective was to describe prevalent exclusion criteria in treatment research on neurological disorders and to analyze their impact on sample representativeness and generalizability of findings.DesignNarrative literature review of studies focusing on treatment for neurological disorders. Studies were identified from PubMed and bibliographies.ResultsEight studies were included in the narrative review: 3 studies focused on Alzheimer's disease/dementia, 2 each focused on traumatic brain injury (TBI) and epilepsy, and 1 focused on amyotrophic lateral sclerosis (ALS). The total number of patients screened across all studies was 20,018, of which 14,721 (73.5%) were excluded. An average of 6 exclusion criteria was applied. The criteria that contributed most to exclusion were the presence of comorbid psychiatric conditions, a history of alcohol or other substance misuse, and cognitive impairments. Women and the elderly were underrepresented among included samples. Race/ethnicity proportions were seldom reported.ConclusionExclusion criteria are used extensively in neurological treatment research and prevent about 3 in 4 patients from participating in research. This limits the generalizability of current findings. Further, because excluded individuals are disproportionately from vulnerable populations, extensive exclusion also raises ethical concerns. Exclusion criteria should be used only in cases where there is a strong rationale so that neurological treatment research can make a greater impact on clinical care.