Multivariate vector autoregressive modelling of malaria with climate and vegetation factors in a remote hilly region of Northeast India

Abstract

Malaria remains a significant global health concern which continues to pose a life-threatening risk globally. The disease, transmitted by Anopheles mosquitoes acting as vectors, requires favorable environments for effective transmission. These environments are influenced by factors such as meteorological conditions and vegetation cover; a number of which have been examined in this study and incorporated into modeling the observed malaria incidence. This method provides a solution for common data inconsistencies encountered in healthcare and epidemiological research, while also offering predictions on incidence rates, thereby enabling more informed decision-making processes. A multivariate statistical modelling approach using the Vector Autoregressive (VAR) model has been employed, enabling dynamic analysis of all relevant parameters simultaneously. The environmental information obtained from satellite and reanalysis datasets, along with the recorded malaria cases in Dhalai district, Tripura, India, were evaluated for causality, refined, and subsequently utilized in the modelling process. The model’s reliability was assessed by comparing its short-term forecast with actual data using a number of accuracy metrics, revealing a mean absolute percentage error of 1.16% and a correlation coefficient of 0.721 between the testing and forecasted malaria incidence data. These observations highlight the model’s effectiveness in accurately capturing the variations in malaria incidence and its predictive capability. Notably, this model has yet to be widely utilized, which presents a unique opportunity for further exploration in other regions. Such studies could significantly contribute to the development of more targeted and effective control measures.