To prioritise the list of 86 zoonoses, a multi-criteria analysis (MCA) was used (for details see Priority Setting Method) consisting of five different phases. Following these phases, an assembled score is assigned to each agent. By ordening the scores, the priority of pathogens is defined. Quantitative data on the different phases of the MCA are presented in the priority system of EZIPs. A special feature of EZIPs is the possibility to modify the data input and assess its effect on the assembled score, and therefore the priority, of individual pathogens.
Priority Setting Method
The quantitative method is based on multi-criteria analysis (MCA) method. MCA offers methods and techniques to structure complex decision-making processes. Generally, a MCA consists of the following five phases:
- List and structure appropriate criteria (aspects of risk) to assess pathogens
- Score pathogens on the selected criteria
- Determine the relative importance (weight) of each criterion
- Aggregate the scores and weights of the criteria into one overall value per pathogen
- Sensitivity analysis
After completing the different phases, information can be added or modified without the necessity to completely redo the analyses. This is especially valuable in the priority setting of emerging zoonoses, where information changes constantly.
Phase A. Criteria for priority setting
The decision rules help to appoint natural numbers to the scores of the criteria for each pathogen. For each, the criterion unit, definition of criteria, point estimates of each class and the decision rules are described.
The following seven criteria are defined:
- Probability of introduction of the pathogen in the Netherlands
- Fraction of animal reservoir infected
- Economic damage, animal
- Fraction of human population, infected by animal sources
- Fraction of human population, infected by human to human transmission
- Morbidity, human
- Mortality, human
Phase B. Criteria for priority setting
The pathogens are scored in view of the current situation in the Netherlands using available literature completed with expert opinion. This entails that factors like effectiveness of existing surveillance systems and preventive measures with respect to the animal reservoirs and the effectiveness of the existing public health service is taken into account in the evaluation.
Phase C. Weighting criteria
In the present study a quantitative approach, developed in collaboration with Professor R.M. Cooke (Delft University of Technology) is used, analysing expert opinions and quantifying the uncertainty of the results caused by insufficient or incomplete data. In collaboration with TU Delft, panel sessions were organised in which respondents were presented with a number of scenarios reflecting hypothetical zoonoses in which the values of the seven criteria in different combinations are present. The scenarios were computer generated random combinations of one the four (five) levels of each criterion. The respondents were asked to arrange the scenarios from least to most threatening. By analyses of the answers of the responders, a weight can be assigned to each criterion using a mathematical technique known as probabilistic inversion.
Phase D. Aggregation of information
By combining the scores of the selected zoonotic agents with the weights, an aggregated score is assigned to each agent. This score was normalised by comparison with a hypothetical agent with the highest possible threat (score 1) and a hypothetical agent with the lowest possible threat (score 0).This score defines the priority. An advantage of this procedure is that the risk of future zoonotic threats can be evaluated using the developed system and that new or changed information can be processed in real time.
Phase E. Sensitivity analysis
The website has dynamic features that allow the user to change scores for any pathogen on each criterion to evaluate the possible impact of uncertain or modified information. It is also possible to compute unweighted scores or to introduce a new pathogen and compare it with pathogens already in the database. The website does not store data from scenario analyses.
In summary, more important than the ranking of the listed zoonoses so far is the concept of this new approach to prioritize emerging zoonoses. The results already have shown the potential of the method. Firstly, the method has a large discriminative power; the normalized scores of the agents prioritized ranged from 0.02 to 0.65. Secondly, the method has already proven to make priority setting processes more transparent and objective. And last but not least, the dynamic, flexible and interactive database resulting from this method provides a flexible knowledge management system.