Using Healthcare Databases To Learn What Works When No Randomized Trials Exist by Miguel Hernan

Making clinical decisions among several courses of action requires knowledge about their causal effects. Randomized trials are the preferred method to quantify those causal effects. When randomized trials are not available, causal effects are often estimated from observational data. Therefore, causal inference from observational data can be viewed as an attempt to emulate a hypothetical randomized trial—the target trial—that would quantify the causal effect of interest. Contrary to what is generally believed, many well-known failures of observational studies were the result of not adequately emulating a target trial rather than limitations of the observational data. This talk explains those methodological failures in non-technical language and describes several examples of how observational data can be used to inform clinical guidelines when randomized trials do not exist.