Predicting influenza vaccinations


The influenza (flu) virus contributes significantly to morbidity and mortality each year. Although the vaccine has been shown to effectively reduce disease and economic burdens, vaccination rates remain low at under 50 percent nationally. Prior research has focused on patient perceptions of vaccination, but few studies have explored other factors related to flu vaccination during primary care visits.


We leveraged publicly available and objective electronic health record data to identify patients in our health system’s primary care practices who were eligible for vaccination at the time of a visit during flu season. We then designed prediction models to identify patient, physician, and environmental factors associated with flu vaccination.


We found that various patient, visit-level, temporal, provider, and weather variables significantly predicted vaccination status. These findings could help to inform the design of new, more targeted interventions.