A stochastic model for analyzing prevalence surveys of hepatitis A antibody.

Epidemiologic surveys of the age-specific prevalence of antibody to hepatitis A virus (anti-HAV) provide information on the spread of infection such as the infection rate and age-dependent characteristics. However, the data on prevalence are confounded with the mortality and diminished immunity of surveyed individuals. Through modeling, the age-specific prevalence of an individual can be separated from these confounding factors. A Markov chain is used to model the process of acquisition of anti-HAV by an individual and to derive the age-specific prevalence. Data from Frösner et al. [Am. J. Epidemiol. 110:63-69 (1979)] are used for illustration and estimation of parameters. The model offers an explanation of the well-known phenomenon of a decline in prevalence in older age. In addition to hepatitis, the framework of the model can be adapted to analyzing seroepidemiologic surveys of other diseases.