[Epidemiology, research, and health services evaluation].

Health services epidemiological research generally involves testing the efficacy of new programs or strategies and the evaluation of health services quality. Repeating the evidence-based medicine model that has flourished in the clinical field, evaluations of public health programs' effectiveness has preferentially proposed and employed randomized studies as a way of guaranteeing a scientific standard of credibility in the results. A new trend among researchers of epidemiology applied to health services has challenged this premise and proposes that randomized studies be adapted to encompass the broad causal chain linking the implementation of programs to a given impact indicator. Others designs are proposed at different levels of causal inference, but sufficiently rigorous for their results to be reliable.