[Epidemiological evaluation of 4 public health interventions]

Methods for the epidemiological evaluation of interventions in Public Health are divided into experimental studies with randomization of "exposed" groups and "controls", and experimental studies without randomization where comparisons are made between periods of time ("before" and "after") or groups differing in their localization ("here" and "elsewhere"). Four attempts to evaluate interventions in Public health are presented. They were designed to appraise the efficiency of pregnancy survey by midwives acting at the home of pregnant women for the prevention of prematurity, to evaluate a program of prevention of ill-treatment of very young children, to assess a program of prevention of sexually transmitted diseases and of the impact of their complications on sterility, and to determine the effects of a campaign to support women having decided to breast-feed their babies. Despite the difficulties encountered in the definition of the protocols and their implementation these four surveys show that it is feasible to organize a network of professionals of various origins and to take into account professional constraints, ethical issues and the necessities of scientific evaluation.