[Data quality of a breast cancer database in Naples, Italy. Do the necessary conditions exist for implementing a disease registry?].

Data quality is one of the fundamental aspects of health information systems and is influenced both by the type and accuracy of the data sources used. The aim of this study was to assess the quality of data in a breast cancer database held by a local health authority in Naples (Italy) and to evaluate the possibility of implementing a disease registry. The database collects information from different sources and was evaluated in terms of completeness, accuracy and comparability. Data linkage of all district data sources was performed; these included medical charts of the Integrated Territorial Oncology Service (SOTI), hospital discharge abstract forms, disability registry, list of subjects affected by specific disorders and therefore entitled to exemption from healthcare costs related to their disease, and the Nominal Registry of Causes of Death (ReNCaM). From 1 January 1995 to 31 December 2003, 475 incident cases of breast cancer were identified; 313 (65.9%) of which through the Integrated Territorial Oncology Service, and 71 (14,9%) through discharge abstract records, while only 3.6%, 4.0% and 11.6%, respectively, through the disability registry, lists of subjects with healthcare costs exemptions, and the ReNCaM. Medical charts of the Oncology Service were found to be the most complete data source for demographic information, hospital name, staging and treatments given, histologic diagnosis, and for most risk factors. By linking 178 patients registered both in the discharge abstract forms and in the Oncology medical charts, discordant demographic data, name of hospital, and date of diagnosis were found respectively in 7.3%, 16.3%, and 42.1% of cases. Overall the quality of data was found to be good and comparable to other registries. However the database cannot yet be considered as a disease registry in view of the small geographical area involved and the lack of continuous and systematic data flow; the latter can be provided only by facilities with sufficient economic resources and personnel specifically assigned to the task.