Contact Surface Models for Infectious Diseases

Controlling of infectious diseases requires information about the rates at which individuals make contact. We propose a novel approach to modeling contact rates via a continuous contact surface. This provides a more realistic and flexible representation of contact rates than currently used methods. Our approach allows modeling of sources of heterogeneity due to age, individual effects, and gender. The models are fitted to serologic survey data by maximum likelihood. This involves solving an integral equation linking the contact surface to the infection hazards. The method is illustrated with two datasets, on mumps and rubella and on Epstein–Barr virus and herpes simplex virus type 1 infection. The advantages and shortcomings of the method, particularly the identifiability of the contact surface, are discussed.