A U.S. National Institutes of Health-sponsored study was initiated in 1998, ending in 2002, to examine the hypothesis that bovines, buffaloes in particular, were responsible for the persistence of schistosomiasis in human populations living in and around Poyang Lake, Jiangxi Province, P. R. China. The two villages are the focus of this paper, the experimental Jishan Village and the control Hexi Village. The prevalence and intensity of infection of humans and buffaloes were determined. Then, all inhabitants of all the villages were treated with praziquantel, while only the buffaloes of the experimental (intervention) village were treated. Following treatment, rates of reinfection of people in the two villages and buffaloes in the experimental village, as well as the ongoing prevalence of infections in the buffaloes of the control village were monitored annually.
Quantitative collections of snails were made pre- and post-flood for a total of nine seasons to determine the density of snails, density of infected snails, and prevalence of infected snails. Collections were made from ecologically different village zones and from hot spots for persistent snail infections within zones. The hypothesis was that treatments in the experimental village would drive down snail infections to 80% or more below pre-trial levels. This paper makes the case that the intervention did not achieve reducing snail infections village-wide by even 50% or more. The probable reasons for this failure are discussed.
The significant findings of this longitudinal study are several: (1) Foremost, the intervention did not achieve the goal of significantly reducing snail infections in Jishan. (2) To understand the dynamics of Schistosoma japonicum transmission throughout large areas, such as the administrative villages of Hexi and Jishan, it is important to partition the area into clearly definable ecological zones. (3) We have found that the greatest risk of infection is in very small areas within zones that we call hot spots. (4) The most important indicator for risk of infection is the density of infected snails, not the prevalence. (5) There are very significantly more infected snails per area in hot spots than in non-hot spots. (6) Density of infected snails is not correlated with density of snails overall. (7) The sampling strategy of zones, squares, and random selection of 20 4 m2 cells within squares has enabled robust statistical analyses of snail populations that have a strong negative binomial distribution. Sampling twice a year (before the floods and after the floods) for at least 8 or 9 seasons has provided a time series enabling a robust assessment of trends in the experimental and control villages. Paying close attention to environmental factors as they impact results is critical.