Epidemiological Models to Inform the Design And Evaluation of Official Plans for Risk-Based Surveillance of Foreign Hemorrhagic Fevers of Swine

Abstract

This project focused in two notifiable diseases to the World Organization for Animal Health (OIE), which are African swine fever (ASF) and classical swine fever (CSF), and how to calculate the risk of introduction, as well as measures for early detection in case of incursions in free areas. This document is divided into chapters. Chapter 1 includes the introduction of the thesis, where the characteristic of the diseases of concern are presented, also improvements and novelties in diagnostic and surveillance for ASF and CSF. Chapter 2 approached the risk assessment for introduction of ASF in Kazakhstan, with the objective of identifying high risk areas of possible incursions, in a country level perspective. We used Kazakhstan as a model because it is an ASF free country, surrounded by an infected neighborhood. To develop our model, we used conjoint analysis, which is a marketing tool for assessing and scoring clients preferences, linked to ordinal logistic regression to create a proxy-risk score for ranking the areas at higher risk of introduction of ASF in Kazakhstan, generating risk maps. Chapter 3 continued with the same focus of risk prediction; however, it was for the introduction of CSF in Mato Grosso, a State in Brazil, which has borders with CSF non-free areas in Brazil, and extensive dry border with Bolivia, which also has unknown status for CSF with episodes of this disease some years ago. With that, the approach for risk prediction model was developed to be applied in a state level of surveillance, where we combined stochastic quantitative risk assessment for commercial pig farms, and the methodology developed for Kazakhstan to assess the risk of CSF introduction through wild boars, which is related to backyard pig farms due to poor biosecurity. The results were displayed also in risk maps, but with more capillarity since we worked with states and their municipalities. In chapter 4, we worked with a surveillance approach to aid the early detection of ASF or CSF at farm-level. To accomplish this project, we developed a scoring system for enhanced passive surveillance for ASF and CSF, where through this surveillance activity, farms would be able to trigger alerts to get pigs tested in an early detection, since the scoring was built in a weekly basis, which allows an objective and adaptable surveillance. This protocol was piloted in pig farms in the Dominican Republic and in the United States. It is utmost to work with different perspectives of surveillance actions, where it is possible to have a macro vision of a disease-free area and being able to address particularities to the epidemiological model, increasing details, and being more specific in a dynamic and connected process. This Ph.D. project worked in different layers of surveillance and different data availability, being national, state, and farm-level, however with the same objective of calculating the risk of introduction ASF or CSF in free areas, and with that, we could deliver important material to subside Official Veterinary Services in actions related to early detection surveillance and disease control programs.