The effect of human mobility on the geographic spread of dengue fever in Mexico.

Abstract

Investigations of three dengue epidemics in Mexico were conducted to study the geographic patterns of epidemic spread, and to identify the role of human mobility in contributing to that spread. Information was collected on local exposure and mobility patterns in each community. In addition, clinical history and serologic samples were obtained to determine recent infection with dengue fever. A total of 269 households were sampled overall, consisting of 1471 persons. Serologic samples were successfully processed for 655 persons. A geographic risk score was computed to determine the risk of infection resulting from time spent in one's home, in the geographic sector of residence, and time spent outside of one's home visiting other sectors of the community. Deterministic simulations of a geographically stratified transmission model were conducted to examine the sensitivity of the epidemic patterns to the rates of movement among geographic areas. The geographic risk score was significantly different between cases and non-cases in one epidemic. No significant differences were found in the risk score in the other two epidemics. Deterministic simulations showed that in low risk sectors the time to the peak of the epidemic could be accelerated by contact with the high risk sectors. In addition, the overall rates of movement and time spent in different sectors was seen to influence the epidemic pattern in any one sector. Within a single community, understanding both the risk by area and the structure of contact among areas will help in focusing control measures where they will stop the most chains of transmission. Risk areas with which there is a great deal of contact may serve to disseminate infection to other areas of the community. However, examination of the effect of human mobility deserves further exploration at a national level. The degree of mobility at a national level may be much lower than that observed within a single community, and its importance may be more relevant for increasing the probability of epidemic in an area than for accelerating the dissemination of an epidemic.