Disease Transmission Models for Public Health Decision Making: Toward an Approach for Designing Intervention Strategies for Schistosomiasis Japonica. (Articles)
Spear, Robert C., Hubbard, Alan, Liang, Song, Seto, Edmund, Environmental Health Perspectives
Mathematical models of disease transmission processes can serve as platforms for integration of diverse data, including site-specific information, for the purpose of designing strategies for minimizing transmission. A model describing the transmission of schistosomiasis is adapted to incorporate field data typically developed in disease control efforts in the mountainous regions of Sichuan Province in China, with the object of exploring the feasibility of model-based control strategies. The model is studied using computer simulation methods. Mechanistically based models of this sort typically have a large number of parameters that pose challenges in reducing parametric uncertainty to levels that will produce predictions sufficiently precise to discriminate among competing control options. We describe here an approach to parameter estimation that uses a recently developed statistical procedure called Bayesian melding to sequentially reduce parametric uncertainty as field data are accumulated over several seasons. Preliminary results of applying the approach to a historical data set in southwestern Sichuan are promising. Moreover, technologic advances using the global positioning system, remote sensing, and geographic information systems promise cost-effective improvements in the nature and quality of field data. This, in turn, suggests that the utility of the modeling approach will increase over time. Key words: disease transmission, mathematical models, parameter estimation, schistosomiasis. Environ Health Perspect 110:907-915 (2002). [Online 12 August 2002] http://ehpnet1.niehs.nih.gov/docs/2002/110p907-915spear/abstract.html
In a companion article, Eisenberg et al. (2002) present an approach to the analysis of infectious disease transmission for waterborne pathogens using dynamic models studied via computer simulation techniques. Here we present an application of this approach to designing local control strategies for the parasitic disease schistosomiasis. The schistosomiasis transmission cycle involves mammals and freshwater snail species linked through contact with different forms of the parasite in surface waters. Our work focuses on agricultural villages in the southwestern part of Sichuan Province in China, where schistosomiasis is endemic. The challenge is to determine whether a dynamic modeling approach can be a useful tool in specifying effective intervention strategies. We propose to use the model to integrate general knowledge of the factors controlling transmission of the disease, quantitative data specific to the transmission of schistosomiasis in China, and site-specific data of the sort typically available in these settings.
This report is of work in progress in that our activities to date have been concerned with model formulation and its parameterization, particularly in light of the kind of field data commonly generated in rural China. We have not yet designed and implemented an intervention program. However, much of our work has been devoted to analysis of data from a study that culminated in a successful intervention program carried out by our colleagues at the Sichuan Institute of Parasitic Disease over 1987-1995. Regrettably, that intervention was not sustainable because of recurrent annual costs of drug treatment. This underscores that the search is for an intervention strategy that is not only effective but also sustainable in a local context.
It is important to point out at the outset that we are not designing intervention trials in a traditional epidemiologic context. Our objective is not to determine whether a particular intervention is effective when all other factors are controlled. For schistosomiasis, there is a considerable body of knowledge about the array of methods of controlling transmission that have been employed in different settings. The task is to determine which blend of the subset of feasible interventions should be used in a particular setting and to predict its probable effectiveness in diminishing disease transmission. …