Background: Investigation in health care services by data bases (DB) is undergoing an important increase in numerous countries. Several of the most relevant epidemiologic studies on the postsurgical morbidity and mortality of benign hypertrophy of the prostate (BHP) have used DB. The aim of this study was to prepare and optimize the DB of a health cooperative (Asistencia Sanitaria Colegial).
Methods: Since 1981 the DB contains sociodemographic variables, data concerning visits to general practitioners and specialists, complementary examinations and hospitalizations. The coding system was specifically developed by the cooperative. The study base was made up of all the policies of the members between January 1981 and December 1988 (101,400 males). To identify the policies 14 codes potentially related with BHP were initially chosen and all the policies containing any of these codes were selected, with the manual analysis of 204 policies being thereby performed.
Results: The total number of policies initially selected for the study was 3,157. Manual revision discarded the use of 9 of the 14 codes used in the first selection. In agreement with the 5 codes able to define diseases of prostatic gland and to the file in which they appeared the policies were classified as probable diseases of prostatic gland (n = 1,007, 31.9%), possible (n = 805, 25.5%) and improbable (n = 1,345, 42.5%). Likewise, the strategy to follow for the selection of a cohort of BHP cases was defined.
Conclusions: The manual analysis of policies evidenced a remarkable coherence and exhaustiveness of the information registered. The principal difficulties found with the data base used were the lack of a sole diagnosis and the coding system used in addition to the impossibility of automatically controlling the assistance received by the patients outside the cooperative system. The availability of a data base such as that herein described with a well defined large population and with accessible sociodemographic and health care information justifies collaborative efforts among health care administrators, clinicians, computer scientists and epidemiologists.