Identifying high-risk areas for dog-mediated rabies using Bayesian spatial regression.

Area of Science:

• Veterinary Public Health
• Spatial Epidemiology
• Zoonotic Disease Control

Background:

• Rabies remains a significant zoonotic threat globally, necessitating effective surveillance and control strategies.
• Understanding the spatial and temporal dynamics of rabies spread is crucial for targeted interventions.

Purpose of the Study:

• To quantify the location-based risk of dog-mediated rabies in Thailand using a Bayesian spatial regression model.
• To identify key epidemiological factors associated with rabies risk in both human and animal populations.

Main Methods:

• A conditional autoregressive (CAR) Bayesian zero-inflated Poisson (ZIP) regression model was applied to human and animal rabies case data (2012-2017).
• Predictor variables included dog bites, dog/cat populations, Buddhist temples, garbage dumps, vaccination, prophylaxis, poverty, and shared borders.
• Model performance was evaluated using cross-validation.

Main Results:

• Key predictors for rabies risk included dog bites, owned/un-owned dog populations, shared borders, Buddhist temples, and poverty.
• The spatial regression model demonstrated adequate performance (AUC = 0.81).
• The model could have predicted 2016/17 cases with 71% sensitivity and 80% specificity if applied in 2015.

Conclusions:

• This spatial regression approach, integrating multiple data sources, effectively estimates rabies risk in Thailand.
• Findings can inform rabies surveillance, identify high-risk zones, and guide the deployment of control measures like vaccination and post-exposure prophylaxis.