Cardiovasc Prev Pharmacother.  2020 Jul;2(3):99-102. 10.36011/cpp.2020.2.e13.

Causal Claims in Health Sciences and Medicine: a Difference-in-Differences Method

• Kim KN

• KMID: 2536955
• DOI: http://doi.org/10.36011/cpp.2020.2.e13

Abstract

The difference-in-differences (DID) method is a useful tool to make causal claims using observational data. The key idea is to compare the difference between exposure and control groups before and after an event. The potential outcome of the exposure group during the post-exposure period is estimated by adding the observed outcome change of the control group between the pre- and post-exposure period to the observed outcome of the exposure group during the pre-exposure period. Because the effect of exposure is evaluated by comparing the observed outcome and potential outcome of the same exposure group, unmeasured potential confounders can be cancelled out by the design. To apply this method appropriately, the difference between the exposure and control groups needs to be relatively stable if no exposure occurred. Despite the strengths of the DID method, the assumptions, such as parallel trends and proper comparison groups, need to be carefully considered before application. If used properly, this method can be a useful tool for epidemiologists and clinicians to make causal claims with observational data.

Keyword

Causality; Epidemiologic studies; Logic; Models, statistical; Observational study

Figure

• 

  Figure 1. Key concepts of difference-in-differences method.

Reference

1. Rubin DB. For objective causal inference, design trumps analysis. Ann Appl Stat. 2008; 2:808–40.

2. Butsic V, Lewis DJ, Radeloff VC, Baumann M, Kuemmerle T. Quasi-experimental methods enable stronger inferences from observational data in ecology. Basic Appl Ecol. 2017; 19:1–10.

3. Angrist JD, Pischke JS. Mostly Harmless Econometrics: An Empiricist's Companion. Princeton, NJ: Princeton University Press;2008.

4. Clair TS, Cook TD. Difference-in-differences methods in public finance. Natl Tax J. 2015; 68:319–38.