The literature on vaccine scares and vaccine risk perception has focused on the individual, however there is a group context to individual vacinating behaviour that should not be ignored: individuals are influenced by the opinions of other members of the population, and the vaccinating behaviour of other others determines the rates of vaccine uptake and hence the risk that an individual becomes infected. The mathematical (game theory) tools required to capture the interaction between group behaviour and individual behaviour have been successfully applied in fields such as economics, ecology and evolution, but are almost entirely undeveloped in public health and vaccination policy in particular.

Methods

We developed a mathematical (dynamic game theory) model to capture the dynamics of vaccinating behaviour and the interaction between vaccine uptake, disease incidence, vaccine risk perception and individual vaccinating behaviour. The model is a synthesis of standard compartmental (dynamic) models (Anderson and May 1991) and evolutionary game theory models (Hofbauer and Sigmund 1998), and builds upon our previous work on static (non-temporal) models (Bauch and Earn, PNAS 101: 13391-13394, 2004). The model was parameterized using time series data on vaccine uptake and pertussis incidence before the start of the England & Wales whole-cell pertussis vaccine scare of the 1970s, and validated against the time series data after the start of the vaccine scare.

Results

After the pertussis vaccine scare in England & Wales, vaccine coverage levels recovered approximately three times as slowly than they dropped during the scare. The mathematical model also predicts a slower recovery of vaccine uptake after the scare, although the recovery is approximately five times slower than the initial decline. This pattern is predicted to be a common but not universal feature of vaccine scares for which the decline in vaccine uptake is sufficiently steep.

Conclusion

The group context of individual vaccinating behaviour is crucial and should not be ignored. Using established methods from the ecological and epidemiological sciences, we have shown that it is possible to build models for vaccinating behaviour that seem to predict qualitative aspects of real-world vaccine scares. Therefore, game theoretical models for predicting and understanding vaccinating behaviour should be further developed.