Uses of clinical databases.

Clinical databases consist of observational data collected on patients who meet specific criteria. The uses of these databases depend on whether the observations are drawn from a single institution, multiple clinical centers, or are population-based. Single institution databases frequently are used to profile patient accrual. In cases of rare diseases or unusual procedures, multicenter databases are required to amass sufficient information for study. Multicenter databases can be used to address issues related to intercenter variation and to develop statistical models to predict outcome based on prognostic factors. Population-based databases are required to assess incidence, prevalence, and mortality rates of disease. Although the inferences that can be drawn from observational data are limited by selection bias, clinical databases are valuable tools in planning clinical research. Clearly, however, resources are required to develop and maintain clinical databases. In an era when much research funding is directed at hypothesis-driven research, the importance of these clinical databases in developing clinical research hypotheses should not be overlooked.