An Observational Study With a Regression Discontinuity Design

Abstract
Background:
Observational studies using traditional research designs suggest that influenza vaccination reduces hospitalizations and mortality among elderly persons. Accordingly, health authorities in some countries prioritize vaccination of this population. Nevertheless, questions remain about this policy’s effectiveness given the potential for bias and confounding in observational data.

Objective:
To determine the effectiveness of the influenza vaccine in reducing hospitalizations and mortality among elderly persons by using an observational research design that reduces the possibility of bias and confounding.

Design:
A regression discontinuity design was applied to the sharp change in vaccination rate at age 65 years that resulted from an age-based vaccination policy in the United Kingdom. In this design, comparisons were limited to individuals who were near the age-65 threshold and were thus plausibly similar along most dimensions except vaccination rate.