Peter Diggle (Lancaster/Liverpool)
River Blindness, Eyeworm and a Calibration Problem
Onchocerciasis (River blindness) is a major public health problem in the wet tropical regions of the world, including most of sub-Saharan Africa. A multi-national programme to control the disease by mass administration of a protective drug has been very successful, with more than 60 million treatments to date over 19 countries. However, the programme has been hampered by the recognition that people heavily infected with the Loa loa (eye worm) parasite. Before the drug is administered in a community, it is relatively easy to estimate the prevalence of eye worm infection, much harder under field conditions to estimate how many people are heavily infected, defined as carrying more than 8,000 parasites per ml of blood. This leads to the following prediction problem: given a sample of n people, of whom Y are infected with eye worm, how many people in the community from which the sample is drawn are carrying more than 8,000 parasites per ml of blood? To address this problem we first develop a model for the variation in parasite count between individuals within a village, and then explore how the parameters of this distribution vary between villages. The model will be used to calculate the predictive probability that an individual in a com­munity will be carrying more than 8,000 parasites per ml of blood, given an estimate of community-level prevalence and any other relevant information in the form of covariates and/or prevalence estimates from neighboring villages.