Maternal mortality and accessibility to health services by means of transit-network estimated traveled distances

This study analyzed the relationship between maternal mortality and variables related to the use of health services (especially residence-hospital traveled distances estimated through transit networks). Deaths were identified for Rio de Janeiro and adjacent cities, from 2000 to 2002, and were matched by age and socio-economic level to birth admissions without maternal deaths (1 case to 3 controls). The variables used were: type of hospital (general × specialized maternity services), number of hospital beds, nature of hospital ownership (public × private-associated), main admission diagnostic, residence-hospital distance, age, income, and education. Distances were estimated by a geographic information system, and were based on most probable itineraries through the urban transit networks. The probability of death was estimated by conditional logistic regression models. 226 maternal deaths were studied, and another 10 were excluded due to incompleteness of information. The ROC area for the final model was 0.89 [95% CI (0.87-0.92)]. This model retained statistical significance for the variables admission diagnostic, type of hospital and residence-hospital distance. The death odds ratio for women who traveled 5-10 km (reference category: <5 km) was 3.84 [95% CI (1.96-7.55)]. The traveled distance measured through transit networks was an important risk factor for death in the studied population.