Emulating Target Trials to Study Perioperative and Critical Care Interventions with Observational Data: Promise and Limitations

Estimating effects of interventions is a central task in perioperative and critical care outcomes research. While randomized trials remain the accepted standard for causal inference, trial data are not always available to inform clinical decisions, and some questions cannot be answered feasibly or efficiently with trials. In these settings, studies using observational healthcare data may be used to inform practice. Causal inference from observational data has been reconsidered in recent years, challenging the prevailing notion among clinical researchers that causal conclusions cannot be drawn from observational studies. The “target trial framework” is one contribution within a growing methodologic field that helps investigators avoid common pitfalls in observational study design and analysis. Importantly, researchers must understand which biases this framework can—and cannot—help avoid. The authors present an overview of target trial emulation and describe the promise and limitations of this framework for improving observational perioperative and critical care outcomes research.