Ivan Camilo Sanchez-Vega¹ , Alvaro David Orjuela-Cañon², Carlos Enrique Awad Garcia³, Erika Vergara⁴, Maria Angelica Palencia³ and Andres L. Jutinico¹, ¹Universidad Antonio NarinoColombia, ²Universidad del Rosario, Colombia, ³Subred Integrada de Servicios de Salud Centro Oriente, Colombia, ⁴Universidad Antonio Narino, Colombia

Antioquia is a Colombian department where 6.7 million people live. Currently, it is the region of the country with the newest cases of tuberculosis reported in 2021, about 18.8%. In addition, the incidence rate of tuberculosis was 36.8 per 100,000 inhabitants. Public government health policy regarding tuberculosis should aim to prevent the uninfected community, in addition to detecting and treating people with tuberculosis. In this sense, the study of algorithms to predict the epidemic trend should be promoted. This work addresses the prediction of tuberculosis cases in Antioquia, considering data from the health surveillance system between 2007 and 2021. For the prediction, the Kalman filter and the autoregressive model are considered. The results show a better performance using the Kalman filter for the prediction of tuberculosis cases at six weeks.

Tuberculosis, Forecasting, Autoregressive Model, Kalman Filter, Performance.