Clustering of malaria infections within an endemic population: risk of malaria associated with the type of housing construction.

The occurrence of malaria infections due to Plasmodium vivax and P. falciparum was monitored in a population of 3,023 people living in six contiguous villages in Kataragama, an area of endemic malaria in southern Sri Lanka, over a period of 17 months. The annual incidence of malaria in this population during the study period was 25.8%. Malaria attacks were clustered, occurring more frequently than expected in certain individuals and housing groups and less frequently than expected in others. In one of these villages, the distribution of cases was examined in relation to locality and to the type of house construction. There was a strong association between the malaria incidence and house construction, independent of location. The risk of getting malaria was greater for inhabitants of the poorest type of house construction (incomplete, mud, or cadjan (palm) walls, and cadjan thatched roofs) compared to houses with complete brick and plaster walls and tiled roofs. Houses that were better constructed had a significantly lower malaria incidence rate (10.5%) than those that were poorly constructed (21.2%; P less than 0.01, by Student's t-test). There was also a significantly higher number of indoor resting mosquitoes collected from the poorly constructed houses than from those better constructed; the average (geometric mean) of mosquito densities found in houses of better versus poor construction were 0.97 and 1.89 per collection in the dry season, and 1.95 and 3.42 per collection in the wet season, respectively (P less than 0.05 in both seasons). This indicated that the higher malaria risk associated with poorly constructed houses was at least partly due to higher human-mosquito contact among their inhabitants.