The Brazilian Hospital Information System and the acute myocardial infarction hospital care

Objective To analyze the applicability of the Brazilian Unified Health System national hospital database to evaluate the quality of acute myocardial infarction hospital care. Methods It was evaluated 1,936 hospital admission forms having acute myocardial infarction (AMI) as primary diagnosis in the municipal district of Rio de Janeiro, Brazil, in 1997. Data was collected from the national hospital database. A stratified random sampling of 391 medical records was also evaluated. AMI diagnosis agreement followed the literature criteria. Variable accuracy analysis was performed using kappa index agreement. Results The quality of AMI diagnosis registered in hospital admission forms was satisfactory according to the gold standard of the literature. In general, the accuracy of the variables demographics (sex, age group), process (medical procedures and interventions), and outcome (hospital death) was satisfactory. The accuracy of demographics and outcome variables was higher than the one of process variables. Under registration of secondary diagnosis was high in the forms and it was the main limiting factor. Conclusions Given the study findings and the widespread availability of the national hospital database, it is pertinent its use as an instrument in the evaluation of the quality of AMI medical care.