Abstract

Whooping cough, a respiratory infection caused by the bacterium Bordetalla Pertussis, is re-emerging globally after decades of control through infant vaccination. In the UK we are experiencing the worst epidemic in 20 years, with over 5000 cases reported this year and 9 deaths. Immunity to re-infection with Pertussis has long been known to be imperfect. However, as reporting is heavily biased towards primary infections in children, there is considerable debate as to the extent to which adult carriage of Pertussis contributes to silent transmission.

Using an extensive data set of pre- and post-vaccination incidence in England and Wales we use transfer entropy to identify spatial hierarchies of transmission of pertussis in the pre-vaccination era. We use this information, and measures of stochastic persistence, to parameterise a mechanistic meta-population model of transmission using Approximate Baysian Compuation (ABC). We validate our fitted model through prediction of the impact of vaccination and cross-validation with a equivalent model for measles where immunity to reinfection is known to be life-long.

We find no evidence that adult carriage of pertussis contributed to the epidemic dynamics of pertussis before, or immediately after the introduction of vaccination. In addition, we also find that the newly proposed radiation law of human migration can perform as well, if not better, than traditional gravity models in explaining the spatial dynamics of childhood diseases.