Improving representativeness in trial recruitment: A data-driven approach

Persistent under-representation of certain racial, ethnic, or socio-demographic subgroups in clinical studies continues to put patients at risk for unfavorable health outcomes. In an environment where accrual increasingly relies on digital campaigns, we propose applying data-driven approaches to the management of campaign budgets. Intentional budget management approaches can integrate representativeness goals into research participant recruitment, thereby facilitating enrollment of diverse populations into clinical studies. Our concept consists of applying a reinforcement learning approach to allocate resources towards digital marketing channels dynamically, using a restless multi-armed bandit model. With this approach it is possible to generate sequential resource allocation strategies based on information about the size and parameter distribution of the target cohorts. Those strategies pursue the maximization of accrual numbers equitably across cohorts of interest. This new methodological approach could be a data-driven strategy to manage resource allocation in clinical study accrual efforts and improve equity and diversity in clinical research.