This research aims to establish how behavior analytic and clinical trial data from vaccination programs can be integrated to enhance vaccination strategies in the retail sector in the United States. Due to vaccine skepticism and disparities in vaccine compliance among different populations, which continue to pose significant challenges for healthcare, there is a pressing need for the development of new strategies. These strategies must not only address the issue of one-size-fits-all messaging but also be supported by concrete evidence. 

We categorize consumer behavior datasets, including purchase history, store visits, and historical usage of health products from CVS and Target retail chain stores, into behavioral phenotypes. These phenotypes are then matched with clinical trial efficacy statistics for COVID-19 and influenza vaccines to create geographically and behaviorally targeted vaccination campaigns.

Evidence collected from an author-led multi-site A/B testing framework across 30 betting shops demonstrates that behavior-based messages are significantly more effective than generic ones, resulting in an over 11% increase in vaccine uptake. Furthermore, sharing localized clinical efficacy statistics, such as those from the CDC and FDA, increased vaccination intentions, particularly in conservative ZIP codes. 

The models generated in this study provided strong predictions of individuals in the targeted population, achieving a high AUC of 0.89. Analyzing vaccination data alongside consumer behavior enables the identification of gaps in vaccination coverage, thereby optimizing resources to improve community health through the retail setting.