Synthesized Difference in Differences

TL;DR

This paper introduces Synthesized Difference in Differences (SDD), a method combining RCT and observational data to accurately estimate treatment effects despite confounding and non-parallel trends.

Contribution

The paper proposes SDD, a novel approach that adjusts Difference in Differences using RCT data to handle non-parallel confounding shifts, improving causal inference.

Findings

SDD outperforms existing methods on synthetic datasets.

SDD maintains accuracy even with limited RCT data.

SDD effectively integrates RCT and observational data for causal estimation.

Abstract

We consider estimating the conditional average treatment effect for everyone by eliminating confounding and selection bias. Unfortunately, randomized clinical trials (RCTs) eliminate confounding but impose strict exclusion criteria that prevent sampling of the entire clinical population. Observational datasets are more inclusive but suffer from confounding. We therefore analyze RCT and observational data simultaneously in order to extract the strengths of each. Our solution builds upon Difference in Differences (DD), an algorithm that eliminates confounding from observational data by comparing outcomes before and after treatment administration. DD requires a parallel slopes assumption that may not apply in practice when confounding shifts across time. We instead propose Synthesized Difference in Differences (SDD) that infers the correct (possibly non-parallel) slopes by linearly adjusting a conditional version of DD using additional RCT data. The algorithm achieves state of the art performance across multiple synthetic and real datasets even when the RCT excludes the majority of patients.