Whose Past is Prologue? Evidence of Learning in Flu Shot Decisions

 Getting a flu shot is broadly recommended, particularly for at-risk populations such as senior. Using medical claims for a 5% Sample of Medicare FFS beneficiaries, we document that actual utilization of flu shots is much lower than the recommendation, and lower than self-reported measures of receiving the flu shot. We then study the dynamic relationship between getting a flu shot, getting the flu, and subsequent decisions, both at an individual level and a geographic level. We find that both getting the flu shot and getting the flu are associated with increases in subsequently getting a flu shot, both at the individual and county/CBSA level. However, getting the flu while also getting the flu shot (i.e., the flu shot not working) has the reverse effect, lowering the likelihood of subsequent preventative care. Our research highlights the importance of flu shot effectiveness on subsequent take-up, and how individual decision-making relies on both personal and broader-market experience