Spatial and temporal heterogeneities of district-level typhoid morbidities in Ghana: A requisite insight for informed public health response.

Typhoid fever is estimated to cause between 9.9-24.2 million cases and 75,000-208,000 deaths per year globally. Low-income and middle-income countries report the majority of cases, especially those in sub-Saharan Africa. The epidemiology of typhoid fever is poorly understood, particularly in Ghana where there has been no study of the within-country variation. Our objective was to explore and analyze the spatial and temporal patterns of typhoid fever morbidities in Ghana. We used the global and local Moran's indices to uncover the existence of global and local spatial patterns, respectively. Generalized linear autoregressive moving average (glarma) models were developed to explore the overall and regional level temporal patterns of morbidities. The overall index of spatial association was 0.19 (p < 0.001). The global Moran's monthly indices of clustering ranged from ≈ 0 - 0.28, with few non-significant (p > 0.05) estimates. The yearly estimates were all significant (p < 0.001) and ranged from 0.1-0.19, suggesting spatial clustering of typhoid. The local Moran's maps indicated isolated high contributions of clustering within the Upper West and Western regions. The overall and regional level glarma models indicated significant first and second-order serial correlation as well as quarterly trends. These findings can provide relevant epidemiological insight into the spatial and temporal patterns of typhoid epidemiology and useful to complement the development of control strategies by public health managers.