TY - JOUR T1 - Case validation in research using large databases JF - British Journal of General Practice JO - Br J Gen Pract SP - 160 LP - 161 DO - 10.3399/bjgp10X483472 VL - 60 IS - 572 AU - Luis Alberto García Rodríguez AU - Ana Ruigómez Y1 - 2010/03/01 UR - http://bjgp.org/content/60/572/160.abstract N2 - Computerised healthcare databases (CHCDs) have been increasingly used in epidemiologic research and have become the single most used source of information in pharmacoepidemiology. A key feature in the selection of a computerised database for research is completeness and validity of the data. As Khan et al highlight in the present issue of the BJGP,1 researchers should investigate their information source and how well it covers the diagnosis under study.The validation process of a database is complex, and the resources required to implement a study protocol using CHCDs will vary widely depending on the need and amount of validation required. It all starts with the selection of a database with a track record of acceptable internal and external validity supported by data provided by the database owners and peer-reviewed studies by external researchers. The General Practice Research Database and The Health Improvement Network are primary care UK databases that meet these criteria.2–5The second step in the validation process is to establish a good operational definition of the outcome of interest by constructing specific diagnostic algorithms using a list of codes from the corresponding clinical dictionary. This initial computerised search might also include objective eligibility (exclusion) criteria. This list may be … ER -