Representation of ethnic groups in dementia trials: systematic review and meta-analysis

Abstract Background Despite the projected increase in dementia cases affecting various ethnic groups worldwide, representation of these ethnic groups in randomized controlled trials (RCTs) of treatments to improve cognitive function in persons with dementia or mild cognitive impairment (MCI) is unclear. We aimed to quantify the inclusion of various ethnicities in dementia treatment RCTs. Methods RCTs published between January 1, 2000 and August 1, 2017 (inclusive) were included. Participants were community-dwelling adults with a diagnosis of either dementia or MCI randomized to receive either pharmacological or non-pharmacological interventions to improve cognitive function. Analyses were performed to determine study-level characteristics associated with recruitment of various ethnic groups. Random effects meta-analyses were conducted to determine the pooled prevalence for each ethnicity. Results A total of 96 RCTs consisting of 37,278 participants (57.2% female) were included in the final analysis. Only 39 (39.4%) trials reported the ethnicity of included participants. The pooled proportion of non-Caucasian trial participants was 11.4% (95% CI, 7.5 to 15.9%). Meta-regression results showed that there has been a slow increase in representation of non-Caucasian ethnic groups over time (0.6% per year, P value  = 0.041). Conclusions There is an underreporting of the ethnicity of trial participants and underrepresentation of non-Caucasian ethnic groups in RCTs designed to improve cognitive function in persons with dementia or MCI.