Application of ARIMA, hybrid ARIMA and Artificial Neural Network Models in predicting and forecasting tuberculosis incidences among children in Homa Bay and Turkana Counties, Kenya

Abstract

AbstractBackgroundTuberculosis (TB) infections among children (below 15 years) is a growing concern, particularly in resource-limited settings. However, the TB burden among children is relatively unknown in Kenya where two-thirds of estimated TB cases are undiagnosed annually. Very few studies have used Autoregressive Integrated Moving Average (ARIMA), hybrid ARIMA, and Artificial Neural Networks (ANNs) models to model infectious diseases globally. We applied ARIMA, hybrid ARIMA, and Artificial Neural Network models to predict and forecast TB incidences among children in Homa bay and Turkana Counties in Kenya.MethodsThe ARIMA, ANN, and hybrid models were used to predict and forecast monthly TB cases reported in the Treatment Information from Basic Unit (TIBU) system for Homa bay and Turkana Counties between 2012 and 2021. The data were split into training data, for model development, and testing data, for model validation using an 80:20 split ratio respectively.ResultsThe hybrid ARIMA model (ARIMA-ANN) produced better predictive and forecast accuracy compared to the ARIMA (0,0,1,1,0,1,12) and NNAR (1,1,2) [12] models. Furthermore, using the Diebold-Mariano (DM) test, the predictive accuracy of NNAR (1,1,2) [12] versus ARIMA-ANN, and ARIMA-ANN versus ARIMA (0,0,1,1,0,1,12) models were significantly different, p<0.001, respectively. The 12-month forecasts showed a TB prevalence of 175 to 198 cases per 100,000 children in Homa bay and Turkana Counties in 2022.ConclusionThe hybrid (ARIMA-ANN) model produces better predictive and forecast accuracy compared to the single ARIMA and ANN models. The findings show evidence that the prevalence of TB among children below 15 years in Homa bay and Turkana Counties is significantly under-reported and is potentially higher than the national average.