Wang Y, Liang Z, Qing S, Liu X, Xu C. Application of an ARFIMA Model to Estimate Hepatitis C Epidemics in Henan, China.
Am J Trop Med Hyg 2024;
110:404-411. [PMID:
38190747 PMCID:
PMC10859795 DOI:
10.4269/ajtmh.23-0561]
[Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/18/2023] [Accepted: 10/11/2023] [Indexed: 01/10/2024] Open
Abstract
Hepatitis C (HC) presents a substantial burden, and a goal has been established for ending HC epidemics by 2030. This study aimed to monitor HC epidemics by designing a paradigmatic autoregressive fractionally integrated moving average (ARFIMA) for projections until 2030, and evaluating its efficacy compared with the autoregressive integrated moving average (ARIMA). Monthly HC incidence data in Henan from January 2004 to June 2023 were obtained. Two periods (January 2004 to June 2022 and January 2004 to December 2015) were treated as training sets to build both models, whereas the remaining periods served as test sets to perform performance evaluation. There were 465,196 HC cases, with an escalation in incidence (average annual percentage change = 8.64, 95% CI: 3.71-13.80) and a peak in March and a trough in February. For both the 12 and 90 holdout data forecasts, ARFIMA generated lower errors than ARIMA across various metrics: mean absolute deviation (252.93 versus 262.28 and 235.37 versus 1,689.65), mean absolute percentage error (0.17 versus 0.18 and 0.14 versus 0.87), root mean square error (350.31 versus 362.31 and 311.96 versus 1,905.71), mean error rate (0.14 versus 0.15 and 0.11 versus 0.82), and root mean square percentage error (0.26 versus 0.26 and 0.24 versus 1.01). Autoregressive fractionally integrated moving average predicted 181,650 (95% CI: 46,518-316,783) HC cases, averaging 22,706 (95% CI: 5,815-39,598) cases annually during 2023-2030. Henan faces challenges in eliminating HC epidemics, emphasizing the need for strengthened strategies. Autoregressive fractionally integrated moving average can offer evidence-based insights for public health measures.
Collapse