Regression discontinuity designs in epidemiology: causal inference without randomized trials.

Regression discontinuity designs in epidemiology: causal inference without randomized trials.