Risk assessment of the entry of canine-rabies into Papua New Guinea via sea and land routes.

Canine-rabies is endemic in parts of Indonesia and continues to spread eastwards through the Indonesian archipelago. Papua New Guinea (PNG) has a land border with Papua Province, Indonesia, as well as logging and fishing industry connections throughout Asia. PNG has a Human Development Index of 0.505; therefore, an incursion of canine-rabies could have devastating impacts on human (7.5 million) and animal populations. Given the known difficulties of rabies elimination in resource-scarce environments, an incursion of rabies into PNG would also likely compromise the campaign for global elimination of rabies. A previous qualitative study to determine routes for detailed risk assessment identified logging, fishing and three land-routes (unregulated crossers ["shopper-crossers"], traditional border crossers and illegal hunters) as potential high risk routes for entry of rabies-infected dogs into PNG. The objective of the current study was to quantify and compare the probability of entry of a rabies-infected dog via these routes into PNG and to identify the highest risk provinces and border districts to target rabies prevention and control activities. Online questionnaires were used to elicit expert-opinion about quantitative model parameter values. A quantitative, stochastic model was then used to assess risk, and parameters with the greatest influence on the estimated mean number of rabies-infected dogs introduced/year were identified via global sensitivity analysis (Sobol method). Eight questionnaires - including 7 online - were implemented and >220 empirical distributions were parameterised using >2900 expert-opinions. The highest risk provinces for combined sea routes were West Sepik, Madang and Western Province, driven by the number of vessels and the probability of bringing dogs. The highest risk border districts for combined land routes were Vanimo-Green River and South Fly, driven by the number of people crossing the border and the number of dogs (with hunters). Overall, the risk posed by land routes was much higher than the risk of rabies introduction by sea routes. This study provides a foundation to develop targeted border control measures, surveillance and response strategies for canine-rabies for the highest risk routes and regions in PNG. Sensitivity analysis using the Sobol method played a key role in this study and directed further data collection to refine risk estimates. The ease of expert-elicitation using online methods demonstrates the feasibility of using such methods for animal and human disease surveillance in PNG.