Intraurban variations in adult mortality in a large Latin American city

Urbanization is high and growing in low- and middle-income countries, but intraurban variations in adult health have been infrequently examined. We used spatial analysis methods to investigate spatial variation in total, cardiovascular disease, respiratory disease, and neoplasm adult mortality in Buenos Aires, Argentina, a large city within a middle-income country in Latin America. Conditional autoregressive models were used to examine the contribution of socioeconomic inequalities to the spatial patterning observed. Spatial autocorrelation was present in both men and women for total deaths, cardiovascular deaths, and other causes of death (Moran's Is ranging from 0.15 to 0.37). There was some spatial autocorrelation for respiratory deaths, which was stronger in men than in women. Neoplasm deaths were not spatially patterned. Socioeconomic disadvantage explained some of this spatial patterning and was strongly associated with death from all causes except respiratory deaths in women and neoplasms in men and women [relative rates (RR) for 90th vs 10th percentile of percent of adults with incomplete high school and 95% confidence intervals: 1.23 and 1.09-1.39 vs 1.24 and 1.08-1.42 for total deaths in men and women, respectively; 1.36 and 1.15-1.60 vs 1.22 and 1.01-1.47 for cardiovascular deaths; 1.21 and 0.97-1.52 vs 1.07 and 0.85-1.34 for respiratory deaths; 0.94 and 0.85-1.04 vs 1.03 and 0.87-1.22 for neoplasms; and 1.49 and 1.20-1.85 vs 1.63 and 1.31-2.03 for other deaths]. There is substantial intraurban variation in risk of death within cities. This spatial variability was present for multiple causes of death and is partly explained by the spatial patterning of socioeconomic disadvantage. Our results highlight the pervasive role of space and social inequalities in shaping life and death within large cities.