ABSTRACT
OBJECTIVE: This report investigates the use of linear and travelled route (transit network-based) distances in estimating the accessibility of hospitals to patients, for some selected hospital admission diagnostics. METHODS: For patients admitted to 14 public hospitals in Rio de Janeiro City, during 1996, under the ICD-9 headings "Complications of Pregnancy, Childbirth and the Puerperium" and "Disorders of the Circulatory System", average distances between the patient's district of residence and hospital of admission were calculated (both as Euclidean and as network-based distances). Data were obtained from the country's public health data processing agency. Geographic co-ordinates were obtained for districts of residence from the postal codes of the patients' residences. Distances were estimated with the TransCAD Geographical Information System, based on a map of the city transit network. There were 8654 patients admitted under the "Complications of Pregnancy" heading and 3439 under "Disorders of Circulatory System". RESULTS: Variations of up to a factor of 5.3, and up to 34 km, could be identified between linear and network estimates. CONCLUSION: While recognizing that network estimates have advantages, the literature on accessibility frequently argues that aerial estimates are a good approximation for those. The present results show that this is not necessarily the case.