Po 8430 a Geospatial Approach to Predicting Diarrhea Prevalence in Nigeria

Abstract

BackgroundNigeria ranks second globally only behind India in under-five mortality prevalence. In Nigeria, 108.8 children die per 1000 live births before their 5th birthday. It is of note that diarrhoea (15.3% prevalence) is the second leading cause of under-five mortality in Nigeria after pneumonia. General poor hygiene and nutritional status are contributory factors to diarrhea.MethodsData was collected for severe acute malnutrition (SAM) using the weight for height z-value (WHZ) and/or oedema criteria. In addition, data on diarrhoea prevalence, oral rehydration salt therapy (ORST), improved source of drinking water and improved sanitation were collected. These were obtained for 36 states and federal capital territory (FCT) from the National Bureau of Statistics headquarters in FCT, Abuja for 2015. Correlation analysis was first carried out to determine relationships followed by geographically weighted regression analysis (GWR). GWR was used to predict under-five mortality pattern and accuracy mapped.ResultsObserved correlation coefficients to diarrhoea prevalence were 0.59,–0.49, −0.35 and −0.63 for SAM, ORST, improved drinking water access, and improved sanitation, respectively. R2 varied across states, though positive, from 0.29 in Akwa Ibom to 0.95 in Kebbi states. Standard deviation of residuals in the regression model ranged from −3.89 to 3.33 in Borno and Gombe states respectively, while Sokoto and Bauchi had 0.006 and 0.024 respectively, thus having the best accuracy in predictions across all states in the country. Both correlation and GWR were at p<0.05.ConclusionThe results obtained support literature, confirming the inverse relationship between ORST prevalence, improved drinking water access and improved sanitation to diarrhea prevalence. It also supports the already confirmed positive relationship between poor nutrition of children and susceptibility to diarrhoea. The study however expanded knowledge by incorporating geocomputation to predict diarrhoea prevalence.