[The advantages and limitations of observational studies]

Randomized clinical trials are considered the gold standard for establishing treatment efficacy and generate evidence-based medicine. Nonetheless, because of the stringent exclusion criteria used in selecting study populations, concerns are raised about the limited generalizability of evidence they provide. Indeed, randomized clinical trials assess treatment efficacy for an "average" patient, quite often far from older adults characterized by chronic comorbidities of different severity, or by functional and/or cognitive impairment. Observational studies have been proposed as alternative means of testing intervention effectiveness in older populations with multifaceted problems. Unlike randomized clinical trials, they assess outcomes in regular clinical practice, thereby reflecting real adherence to treatment/intervention. The availability of huge, high quality databases offers the potential to bring research closer to practice and audit. Databases provide fertile grounds for observational studies, and can generate hypotheses and provide ready access to trialists as well, setting new possibilities for epidemiological research. They must include complete data on all consecutive patients, use standard definitions of conditions and outcomes, and include all clinical characteristics likely to affect outcomes. In addition, their potential for research and audit is greatly enhanced by linking to other databases, like the census ones, which allow evaluation of geographical and contextual information.