Interestingly, the protective effect was much stronger in women than in men (for men, the effect is statistically insignificant).
Future studies should investigate why!
What did we do?
We used a natural experiment in Wales where shingles vaccine (Zostavax) eligibility was based on an birthdate cutoff, separating eligible and ineligible cohorts. This gave us a great RDD setting to causally estimate its effect on dementia risk.
Why do economists study this?
This was a perfect setting for combining medical data and econometric methods! Policy evaluation of large public health intervention is rarely possible, but here was a unique chance to learn something causal about a vaccine policy at a large scale.