Application of univariate forecasting models of tuberculosis cases in Kelantan

Introduction: Tuberculosis (TB) remains an important public health problem worldwide with major cause of illness and death. Objective: This study was designed to determine the pattern and the best time series model of TB cases 2003 to 2010 in Kelantan and to forecast the number of TB cases for the next 2 years. Methodology: This was a survey using registered cases which have been notified to Kelantan Department of Health. All the registered TB cases from 2003 to 2010 have been reviewed and included in the series data as monthly data. Univariate modelling (Naïve Method, Average Forecast, Exponential Smoothing Techniques, and Box-Jenkins Method) has been applied to the data series in order to determine the best forecasting model for the data series. Results: Based on 72 monthly data series of TB cases, increasing in trend pattern was noted. Double Exponential Smoothing technique was found to be the best time series model comparing to Single Exponential Smoothing, Holt's Method, ARRES, and Holt Winter's Trend and Seasonality for both multiplicative and additive effect assumption. The future values estimated to be gradually increased from 115 cases in January 2011 to 121 cases in December 2012. Conclusion: The increasing trend of the TB cases should be taken seriously in term of controlling and preventing the cases in order to achieve the Millennium Development Goals (MDG) by 2015.