Epidemiological Modeling of Equine Colic Surgical Risk Across Northern Nigerian States: A Predictive Analytics Approach

Equine colic requiring surgical intervention is a major cause of horse mortality in northern Nigeria, and the development of predictive epidemiological models capable of identifying high-risk populations and settings has the potential to significantly improve preventive management strategies and resource allocation. This doctoral study employs predictive analytics and epidemiological modeling to investigate the surgical colic risk profile across equine populations in northern Nigerian states encompassing Kano, Katsina, Zamfara, Sokoto, Kaduna, Kebbi, and Adamawa between 2020 and 2024. A retrospective case record review identified 314 confirmed surgical colic cases across veterinary facilities in the target states. Predictor variables including housing type, feeding practices, breed, seasonality, water access, management category, and geographic zone were encoded and analyzed using multivariate logistic regression, classification and regression tree analysis, and a gradient boosting machine learning model. Spatial clustering of surgical colic cases was identified using geographic information system mapping, with high-risk zones concentrated around semi-arid Sahelian areas with limited water availability. Feeding systems characterized by high dry feed loads and low green forage availability were the strongest nutritional risk predictors. The gradient boosting model achieved the highest predictive accuracy with an area under the receiver operating characteristic curve of 0.87. A three-tier risk stratification classification emerged from model outputs, enabling targeted preventive intervention prioritization. This doctoral study contributes the first epidemiological predictive model for equine surgical colic in northern Nigeria, advancing data-driven resource planning for equine veterinary services across the region. Keywords: equine colic, predictive modeling, epidemiology, northern Nigeria, machine learning.