Effect of topography on the risk of malaria in the Usambara Mountains, Tanzania

Abstract

There has been a progressive rise in malaria in parts of the African highlands over the last 50 years. In this area of unstable malaria, devastating epidemics are experienced at irregular intervals. Altitude plays a very important role in determining malaria transmission and infection. However, other landscape features may also influence this relationship. This research investigates whether the risk of malaria is related to the shape of the surrounding land, at various altitudes. We hypothesized that households situated close to flat areas where water is expected to accumulate, and are thus potential mosquitoes breeding sites, are at greater risk from malaria than those further away. Cross-sectional clinical surveys were carried out in seven villages along an altitudinal transect rising from 300 m to 1650 m in the western Usambara Mountains, Tanzania. Each village was mapped and incorporated within a geographical information system (GIS). Univariate analysis showed that the risk of an enlarged spleen was positively correlated with decreasing altitude. Other influential topographic variables identified were: water accumulation, flatness and swampiness. Logistic regression analysis produced two models and their equations were used in the GIS to map the risk of malaria infection within each village area. Model 1 included only altitude and correctly predicted the malaria status of 73% of households, whereas Model 2 incorporated altitude and the amount of swampiness within 400 m radius of each household to predict with 76% accuracy whether households were positive or not. We have identified that between 750 m and 1200 m, characteristics of the landscape play an important role in governing malaria risk. At these elevations malaria is highly unstable, and favourable meteorological conditions can cause malaria epidemics. This novel approach of exploring how topography affects the risk of malaria could be used to identify epidemic-prone areas m other African highland regions and help to improve the targeting of control activities in high-risk areas.