Regression discontinuity designs for time‐to‐event outcomes: An approach using accelerated failure time models

Abstract Regression discontinuity designs (RDDs) have been developed for the estimation of treatment effects using observational data, where a is administered an externally defined decision rule, linked to continuous assignment variable. Typically, RDDs applied situations outcome interest and non-temporal. Conversely, time-to-event outcomes received less attention, despite such being common in many applications. We explore subject right censoring. An accelerated failure time (AFT) approach used establish effect estimate fuzzy RDD (where not always strictly according rule). This robust different levels fuzziness unobserved confounding, assessed simulation studies compares favourably established structural AFT models. A motivating example presented which models are fitted metformin on mortality cardiovascular disease rate real data from UK Primary Care.