Applying Time Series Analysis (Box-Jenkins) to Predict the Number of Patients with Cancer at Benghazi Medical Center

Abstract

Background and objective. Cancer has a negative impact on human health in the world, leading to illness and death. This study designed to apply time series analysis using Box-Jenkins methodology to describe the behavior and find an appropriate model for predicting the time series of patients with cancer in Benghazi City. Methods. The data were collected from the medical records of the Oncology department at Benghazi Medical Center (BMC) from January 2011 to December 2022, of 11266 patients distributed on 144 time series of observations. Results. The study found that the time series of data was non-stationary and it had a fluctuation pattern around the mean that changed into a stationary series by performing the first difference. We generated several ARIMA models and compared them based on different criteria, including the R-squared=0.423, root mean square error (RAMSE) =14.045 and Bayesian information criterion (BIC) = 5.447. The suitable model chosen to represent the data series was ARIMA (2,1,3) which, also used for predicting new cases in the next four years. As a result, the estimated model established a similarity between the predicted values and the real values of the time series. In addition, the results indicated that a progressive increase in the total number of people with cancer from January 2023 to December 2026, reaching up to 5344 patients. Conclusion. The ARIMA (2,1,3) model is a good tool for predicting the number of patients with cancer in Benghazi Medical Center. Finally, the importance of this study depending on results to raise awareness and knowledge of risks the cancer in the Libyan community.