Target Trial Emulation to Improve Causal Inference from Observational Data: What, Why, and How?

Abstract

Target trial emulation has drastically improved the quality of observational studies investigating the effects of interventions. Its ability to prevent avoidable biases that have plagued many observational analyses has contributed to its recent popularity. This review explains what target trial emulation is, why it should be the standard approach for causal observational studies that investigate interventions, and how to do a target trial emulation analysis. We discuss the merits of target trial emulation compared with often used, but biased analyses, as well as potential caveats, and provide clinicians and researchers with the tools to better interpret results from observational studies investigating the effects of interventions.

Figure 1

Alignment of meeting eligibility criteria, treatment assignment and start of follow-up. (A) the randomized trial (target trial) and (B) the observational study emulating the randomized trial. A_S, treatment assignment selected by investigator (on the basis of whether data are consistent with being assigned to that treatment strategy); A_R, randomized treatment assignment; E, meeting all eligibility criteria; T₀, time zero (=start of follow-up).

Figure 2

Workflow for conducting observational studies in the target trial emulation framework.

Figure 3

Merits and caveats of the target trial emulation framework.