Invited Commentary on Quasi-Experiments: Going Beyond the Observational and Experimental Dichotomy in Epidemiological Study Design

Abstract

ᅟClassically, epidemiological investigations are restricted to two dichotomous options: experimental studies on one side and observational studies on the other. Experimental studies, particularly randomized trials, are powerful instruments for producing knowledge in health sciences. However, in several situations, such studies are not feasible. On the other hand, observational studies that investigate the effect of “exposures,” even when using prospective approaches frequently lose the opportunity to observe, capture, and report the effects of policies and other interventions. In situations in which the use of experimental studies is not possible, quasi-experimental studies have great potential. They can produce solid causal evidence when randomized controlled trials cannot be used for different reasons. Given that health is affected by several human-made interventions (i.e., technologies, processes, policies, laws, and regulations), quasi-experiments create a significant opportunity for investigating the role of social determinants of health, with the advantage of the best articulating research results and policy-making processes, given their high external validity and the fact that they are conducted as the intervention is implemented.