Medical data bases have become an increasingly popular tool in clinical research. The observational data can be used to generate hypotheses for the development of clinical prediction rules and confirmation or extension of findings from other studies. When the analyses are carefully performed, the results from observational data bases generally have confirmed and complemented the findings of major randomized trials. In recent years, important clinical insights have been derived from analyses of data bases of large scale epidemiologic investigations; such insights may provide significant information on the impact of health policy decisions on patient care. To maximally and carefully use observational data, certain principles in both study design and the data collection and analysis phases should be applied.