Forecasting Scabies Trends in Ghana using Seasonal Autoregressive Integrated Moving Average and Generalized Linear Model

Abstract

Background: Scabies, a neglected tropical disease, poses a significant public health challenge in resource-limited settings such as Kokofu in Ghana’s Ashanti Region. Accurate forecasting of scabies incidence is crucial for effective allocation and management of healthcare resources. Objective: This study aimed to develop and validate predictive models of scabies incidence in the Ashanti Region of Ghana using time series analysis and generalized linear models. The goal was to demonstrate the potential utility of these models in enhancing infectious disease control by generating accurate predictions to inform healthcare resource allocation and planning. Methods: Monthly scabies case data from January 2016 to May 2023 were extracted from the Kokofu Hospital Information System. We applied Seasonal Autoregressive Integrated Moving Average (SARIMA) models for time-series analysis and compared Poisson and Negative Binomial regression models. The model selection was based on the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). Forecasting accuracy was assessed using Mean Absolute Percentage Error (MAPE) and Mean Absolute Scaled Error (MASE). Results: The SARIMA(1,1,2)(0,1,1) model has emerged as the optimal forecasting tool (AIC: 866.79, BIC: 878.44). This model revealed a significant upward trend in scabies incidence and distinct seasonal patterns, with peak incidence during the dry season. The Negative Binomial regression outperformed the Poisson model (AIC: 1039.6 vs. 2496.4), identifying significant increases in scabies cases in June and July; and decreases in November and December. Conclusion: Our study demonstrated the efficacy of combining time-series analysis with generalized linear models in forecasting the incidence of scabies. The developed models provide a robust framework for predicting outbreaks and seasonal variations, offering valuable insights into healthcare planning and resource allocation. Our approach has the potential to enhance scabies management and can be integrated into existing health information systems to support evidence-based decision making in similar resource-limited settings.