OBJECTIVE: To identify the quality of external-cause hospital morbidity admission data in the city of São José dos Campos, São Paulo, Brazil. METHOD: Admissions via the Unified Health System to the Municipal Hospital, a reference trauma care hospital in the city, resulting from external-cause injuries were analyzed during the first semester of 2003, by comparing the data entered in the Hospital Information System (SIH) to 990 patient admission charts. Agreement between variables regarding victim, admission, and the offence was analyzed by rate of gross agreement and by the Kappa coefficient. Injuries and external causes were coded according to the International Classification of Diseases, 10th revision, chapters XIX and XX respectively. RESULTS: The rate of crude agreement was very good for variables regarding victim and admission with an 89.0% to 99.2% variation. Injuries had very good agreement, except for neck traumas (k=0.73), multiple trauma (k=0.50), and chest fractures (k=0.49). External causes had very good agreement for land transportation accidents (k=0.90) and falls (k=0.83). Reliability was smaller for assaults (k=0.50), undetermined intent events (k=0.37), and medical care complications (k=0.03). There was very good agreement for transportation accidents involving pedestrians, cyclists, and motorcyclists. CONCLUSION: Most of the variables in the study were good in the aggregation level analyzed. Some variables regarding the victim and some types of external cause require data quality improvement. The profile of hospital morbidity found in this study has confirmed that land transportation accidents are an important external cause for admission in the city.
Accidents; External Causes; Hospitalization; Information Systems; Injuries; Morbidity; Reliability; Violence
