Workshop - Predicting Zoonotic Outbreaks: Building on the Plague Threshold Model

Real-world studies of host-pathogen dynamics have often failed to keep pace with the development of theoretical concepts. One area where this shortfall is especially regrettable is the application of those concepts in the medical and veterinary sciences. A specific example is the development of predictive models of times and/or locations of high risk of a disease being passed from a wildlife reservoir to human (or domestic animal) populations / so-called zoonotic diseases. This would be the subject of this workshop. A rare example of a model predicting zoonotic outbreaks used long-term data on bubonic plague in Kazakhstan. There, and throughout much of Central Asia, where plague remains a considerable public health concern, the main reservoir host is the great gerbil, Rhombomys opimus. That study demonstrated the existence of a 'critical abundance threshold' in the gerbils, which had to be exceeded for plague to become established in its reservoir (and be a threat), Furthermore, because it was based on past gerbil abundances, the model was not only descriptive but could also be used to predict future outbreaks, i.e. an early warning system. Subsequently, the output from that model has been used to develop a predictive 'expert system' which is being piloted by public health workers in Kazakhstan. The purpose of the proposed workshop is to use the plague model as a springboard to ask (i) What are the prospects of, and what data will be necessary for, utilising critical abundance thresholds, and other key concepts, in developing early warning models for other disease systems in wildlife that are a threat to public health? Indeed, how widely applicable are simple 'threshold' models likely to be? Under what circumstances would different (types of) models be necessary? (ii) How best can such models be turned into tools that are useful to public health practitioners? What do public health workers want/need? Does this differ between the developed and the developing world, and if so how? (iii) How best can predictions such as these be incorporated into a larger risk analysis and/or cost-benefit analysis? 'Ecological' models / at best / generate a probability (with confidence attached) that a disease will be present at a particular place at a particular time. But what is the best way to incorporate the costs of inaction if the disease appears and action has not been taken? Likewise the costs of action that turns out to be unnecessary? And to what extent is it necessary to go beyond knowing what is optimal on purely economic grounds in order to take 'social' factors into account (e.g. loss of public confidence following an untreated outbreak).