A behavioral approach to personalizing public health

Abstract

Behavioral policies are increasingly popular in a number of health care contexts. However, evidence of their effectiveness, specifically in low-income and highly disadvantaged populations, is limited. Some positive effects have been found for adaptive interventions, which merge more personalized approaches with advances in data collection and modern analytical methods. These approaches have only recently become feasible, as their implementation requires a confluence of large-scale datasets, contemporary machine learning, and validated behavioral insights. Such methods have considerable potential to improve outcomes without requiring substantial increases in effort on the part of individuals. Using examples from health insurance choice, clinical attendance rates, and prescription of medicines, we present an argument for how adaptive approaches, especially those considering disadvantaged populations explicitly, offer an opportunity to generate equity in public health.