Diversity in Alzheimer's disease drug trials: Reflections on reporting and social construction of race

WeappreciateDr.Hu’s thoughtful commentsonour recentlypublished systematic review that critically assessed the Alzheimer’s disease (AD) drug trial landscape across the past two decades. Dr. Hu raises some concerns related to targeting diversity in trial participants for recruitment and poses several hypothetical questions regarding (1) the basis of inclusion of a diverse sample if biological correlates do not vary, (2) the intention behind inclusionwhile cautioning aimless overenrollment of diverse populations, and (3) the value of assessing social determinants of health (SDOH) if there is no influence on biological outcomes. We agree the problem is complex and value the space for conversation. However,wehave some conjectures on several vital points in response. In the United States, despite race–and ethnicity to a lesser extent– being a social construction, both are integral parts of our identity (individual to population) since they are externally imposed/reinforced via formal structures (census, laws, policies) and informally through group/culture dynamics. Daily, both constructs innervate our social interactions, may impact access to health-care services, play a latent role in where we live, and how we perform and participate in society. As a result, it is essential to consider race and ethnicity in trials, beyond demographic characterization. A strand of scientific argument supports the notion that it is counterproductive to include more diverse participants in drug trials if the biological mechanisms have slight variance. However, if non-Hispanic, White, middle-class, and relatively healthy participants are continually enrolled in drug trials, then we only know how those drugs work within that population. We see this reasoning as defeatist and in violation of the Belmont Report’s core principles, which originate from a deep history of unethical practices toward ethnic and racially minoritized research participants. As a result, there is a scientific, ethical, and social mandate to move toward representation. To clarify, our study did not recommend baseless overenrollment of racially and ethnically diverse populations into clinical trials. Clinical trials will not be able to tell us how drugs work in all groups or how SDOH may affect treatment response. Randomization creates statistical independence between the treatment assignment and potential outcomes (that is, it establishes exchangeability between groups). Thus, these trials give us population-level inferences. Given this design limitation, we want the trial participants to look like the underlying population that wewish tomake inferences about, because it gives us better information about how thedrugwillwork in the target population. The overreliance on molecular correlates of disease has severely skewed the landscape of biomedical science as a field and extends into AD and related dementias. One can argue that the lack of progress in AD in disease-modifying drugs is due to an overt, siloed reliance on biological processes to the exclusion of environmental and social processes. In theUnited States, it is established that contributions ofmedical care explain 10% to 20% of an individual’s health outcomes, while SDOH constitutes the remaining 80% to 90%.1 Dr. Hu mentioned that appropriate measures of SDOH are underdeveloped in biomedicine, thus making consideration of such measures in trial recruitment problematic. One can argue that this is a result of a narrow scientific focus that prioritized biological mechanisms. Disciplines of sociology and anthropology, including theirmedical, biological, or environmental subspecialties, have validated measures that have existed for decades and can and should be adopted in biomedicine. Finally, from our global viewpoint, these issues seem to vary across regions and continents. For example, the events in Europe during World War II have led to strict legislation on the collection of race/ethnicity data in several European countries–for example, in France, publicly funded census/statistics bureaus donot collect data on race/ethnicity, generally only collecting nationality data.2 Topics specific to race do not seem to be as deeply ingrained in European organizations, policy, and identity as they are in the United States. Discussions in medical and health research thoughtfully revolve around aspects such as country of origin, religion, ethnicity, socioeconomic status, and/or language proficiency, and race per se is rarely discussed in themedia or other contexts in Europe.