Objective: The aim of our study is to present the methods used to observe clinical practices, to evaluate their advantages and limits, and to specify the context in which they are preformed.
Methods: We conducted a bibliographical search on MEDLINE using key words, completed by an offline search. We selected articles published in French or English between 1985 and 2003.
Results: Thirty articles, of which 3 were reviews, as well as an additional 5 reports and 9 books, were selected. We distinguished direct or objective approaches to clinical practices as well as indirect or declarative approaches. We also noted whether the approach was individual with data collected from individual practitioners, or collective, with data coming from databases. Most methods of collecting data used medical records, often in the context of internal or external clinical audits or peer reviews. The other sources of data were self-administered questionnaires, written case simulations and simulated patients. Data was also extracted from previously existing databases: claims databases, registries and ad hoc databases.
Conclusion: An objective individual approach insures the highest quality of information, but possible high costs can limit the use of this approach. It is necessary to continue researching the best means of collecting data on clinical practices, because of the obligation of each practitioner to evaluate individual clinical practices.