[Review: hybrid designs of cohort studies and case control studies].

Hybrid epidemiologic designs combining elements from case-control studies and cohort studies are reviewed. Firstly, the characteristics of selection of both index group and reference group are commented on. The selection of cases required a population-based disease registry. The type of selection of reference group permits to identify the main hybrid designs: the case-cohort approach, sampling from the base at the beginning of risk period, and the nested (within a cohort) case-control studies, usually with a matched density sampling. The assumptions needed for estimating without bias relative risk or rate ratio via odds ratio and without needing the condition of disease rarity are revised. Advantages, uses, and drawbacks are summarized. Lastly, general guidelines for the analysis of these designs are displayed.