Abstract
The relationship between paired hookworm prevalence and mean intensity of infection data from geographically defined communities was examined. The results show that, in spite of major socio-economic and environmental differences between communities, the relationship is consistent and non-linear. A generalized value of k (the exponent of the negative binomial distribution) for hookworms was estimated to be 0.34, which is consonant with previous estimates from cross-sectional data. Maximum likelihood analysis indicates that the severity of hookworm aggregation in humans has an inverse relationship to mean worm burden which is less marked than for Ascaris lumbricoides. A simple model, based on published estimates of hookworm burdens associated with hookworm anaemia, was used to predict prevalence of morbidity from prevalence of infection data for Tanzania, Kenya and Zambia. Predictions correspond to the observation that hookworm anaemia is highly focal, and largely coastal, in distribution. These analyses suggest that locality-targeting of chemotherapy is particularly appropriate for the control of hookworm morbidity.