Objective: Gonorrhea has a focused geographic distribution characterized by high incidence rates in defined "core" areas and decreased incidence as the radial distance from the central core increases. Dense cor group transmission has long been hypothesized.
Methods: We have previously mapped sexually transmitted disease (STD) rates in Baltimore census tracts using STD morbidity data interfaced with a geographic information system. Core areas were defined using a standard definition based on gonorrhea distribution. We studied spatial distance patterns between sexual partners, using the residential addresses of 572 individuals, representing 286 dyad partnerships recruited as part of an epidemiology and behavioral study. To determine if partners lived closer together than would be expected, a modified bootstrap algorithm using Monte Carlo models was developed to compare the distances between partners' residences and all other possible residences.
Results: Two distinct (east and west) core areas were previously identified. Compared with randomly selected Baltimore addresses, partners tended to reside closer to one another than would be expected by chance (z = -1.8), with a median distance of 1.7 kilometers. Within the core areas, women resided a median of 547 meters from their partner, and men resided a median of 339 meters from their reported partners. When all partnerships were considered, the median distance was 1,699 meters. Of the 500 simulation models, the minimum median distance was 4,889 meters.
Conclusions: Partners of patients in core areas in Baltimore live remarkably close to one another, and the partner selection patterns in general indicate nonrandom distribution. Geographic information system-determined patterns of STD patients residing in hyperendemic census tracts support the core theory of disease transmission. In these areas, targeted geographically based interventions may be warranted.