Forecasting Tuberculosis Incidence in Iran Using Box-Jenkins Models

Abstract

Background: Predicting the incidence of tuberculosis (TB) plays an important role in planning health control strategies for the future, developing intervention programs and allocating resources. Objectives: The present longitudinal study estimated the incidence of tuberculosis in 2014 using Box-Jenkins methods. Materials and Methods: Monthly data of tuberculosis cases recorded in the surveillance system of Iran tuberculosis control program from 2005 till 2011 was used. Data was reviewed regarding normality, variance equality and stationary conditions. The parameters p, d and q and P, D and Q were determined, and different models were examined. Based on the lowest levels of AIC and BIC, the most suitable model was selected among the models whose overall adequacy was confirmed. Results: During 84 months, 63568 TB patients were recorded. The average was 756.8 (SD = 11.9) TB cases a month. SARIMA (0,1,1) (0,1,1)12 with the lowest level of AIC (12.78) was selected as the most adequate model for prediction. It was predicted that the total nationwide TB cases for 2014 will be about 16.75 per 100,000 people. Conclusions: Regarding the cyclic pattern of TB recorded cases, Box-Jenkins and SARIMA models are suitable for predicting its prevalence in future. Moreover, prediction results show an increasing trend of TB cases in Iran.