M adenci et al. 1 used the target trial emulation framework to estimate the causal effect of bariatric surgery on the 7-year risk of cardiovascular disease (CVD). They highlighted the pitfalls in the design and analysis of previous observational studies conducted on this topic, which found overall that bariatric surgery lowers the risk of CVD. 2,3 In particular, a central limitation of those studies was the consideration of bariatric surgery as a "point intervention," overlooking the potential role of preoperative components, which could directly affect the risk of CVD. By explicitly specifying and emulating two target trials, Madenci et al. clearly articulated how current recommendations related to bariatric surgery and CVD may be misguided due to being based on potentially flawed observational studies. The use of target trial emulation also enabled the authors to uncover potential limitations of healthcare databases to emulate trials of bariatric surgery.
A FEW WORDS ON THE TARGET TRIAL EMULATION FRAMEWORKRobins, Hernán and their team have formally introduced the target trial emulation framework 4,5 and highlighted the importance of harmonizing the protocol of the hypothetical (target) trial with that of the corresponding observational study. This process has added transparency to the design of observational studies, improved their quality and reproducibility, 6 and has uncovered important and avoidable flaws in previous analyses. There are many examples of observational studies that led to different conclusions than those of corresponding randomized controlled trial (RCTs) due to differences in study design, rather than violations in the standard identifiability assumptions for causal inference (conditional exchangeability, positivity, and consistency). For example, an important flaw in these observational studies was the comparison of prevalent users versus nonusers, instead of initiators versus noninitiators of a treatment (as considered in the RCTs). Two of the most notorious examples are the studies on the effect of hormone therapy on coronary heart disease 7 and statins on cancer. 8 The studies that used the target trial emulation framework for the aforementioned examples 9,10 produced very similar results with the corresponding RCTs and illustrated the importance of well-designed observational studies for health research.The target trial emulation framework is a useful tool in the arsenal of epidemiologists and medical researchers. However, the use of this framework does not guarantee that the results are accurate, as the validity of the findings strongly depends on the quality and the features of
