Effectiveness of Population-Based Hypertension Screening: A Multidimensional Regression Discontinuity Design

Population-based screening can prevent disease but also induce false positives to use low-value healthcare. Using data on individuals aged 40+ in rural South Africa and a multidimensional regression discontinuity design, we estimate effects of clinical referral based on blood pressure (BP) above diagnostic thresholds for hypertension. Referral increases hypertension treatment but has no effect on BP after four years, on average. However, for screens that are less likely to be false positives-based on the time of day and air temperature at which BP was measured-we estimate that referral reduces mean systolic BP by 5 mm Hg (3.6%) and raises the probability of achieving BP control by 22 percentage points (44%). These results demonstrate the potential for false positives to lower the average effect of screening.