Designing Health Recommender Systems with a Health Equity Lens (Preprint)

Abstract

UNSTRUCTURED Health Recommender Systems (HRS), which use Artificial Intelligence (AI), have made great strides for human-centered care and prevention by providing personalized health advice on personal digital devices. HRS have demonstrated a unique role in the digital health field because they can offer relevant recommendations, not only based on what users themselves prefer and may be receptive to, but also using data about wider spheres of influence over human behavior, from peers, families, communities, and societies. Using the socioecological model, we identify how HRS could play a unique role in decreasing health inequities by targeting the interconnectedness of individuals and their environments. We then discuss the challenges and future research priorities. Despite the potential for targeting more complex systemic challenges in obtaining good health, current HRS are still focused on individual health behaviors, do not integrate lived experiences of users, and have had limited reach and effectiveness for individuals from low socioeconomic status (SES) and racial/ethnic minoritized backgrounds. In this perspective, we argue that a new design paradigm is necessary, in which future HRS focus on incorporating structural barriers to good health in addition to user preferences, and are designed from decolonial perspectives. If these steps are taken, HRS could play a crucial role in decreasing health inequities.