Serfling regression model for analyzing tuberculosis epidemiological characteristics in Jinhua City,Zhejiang Province

doi:10.19982/j.issn.1000-6621.20250523

Abstract

Abstract:

Objective: To establish an adjusted Serfling regression model for analyzing TB epidemic periods and excess cases, thereby providing evidence for control measures. Methods: Data of all clinically and laboratory-confirmed TB cases with disease onset between January 1, 2015, and October 31, 2025 and address in Jinhua City, was extracted from the Tuberculosis Information Management System, a subsystem of the China Disease Prevention and Control Information System. Descriptive analysis of TB epidemic characteristics was performed using X-13-ARIMA-SEATS, the Hodrick-Prescott (HP) filter, and the Joinpoint regression model. An adjusted Serfling regression model incorporating ARIMA error correction was established to identify the epidemic periods and estimate the number of excess cases, and predict the incidences from January to October 2025. Results: From 2015 to 2024, a total of 35659 TB cases were reported, with an average annual incidence rate of 51.73 per 100000 population (35659/68.936 million). The TB incidence rates in Jinhua City decreased from 63.44 per 100000 (4075/6.423 million) in 2015 to 43.72 per 100000 (3132/7.163 million) in 2024, showing a decline (AAPC=-5.05, P<0.001) trend. The adjusted Serfling regression model had a good fitting effect (AIC=521.19,BIC=533.66), a total of 46 epidemic months were identified, with March being a fixed epidemic month. The total number of excess cases was 2580 (95%CI: 2337-2822). The highest number of excess cases was recorded in 2017 (509 cases, 95%CI: 408-610), followed by 2018 (369 cases, 95%CI: 289-449). The prediction of TB incidence from January to October 2025 indicated that February and September 2025 would be epidemic months, with 49 (95%CI: 13-85) and 44 (95%CI: 8-80) excess cases, respectively. The peak of the predicted values was consistent with that of the actual incidence. Conclusion: From 2015 to 2024, the TB epidemic in Jinhua City showed a downward trend. Attention should be paid to TB prevention and control in spring. The adjusted Serfling regression model has a good fitting effect and can be used for early warning of TB and estimation of disease burden.

Key words: Tuberculosis, Epidemiologic study characteristics, Regression analysis, Forecasting

CLC Number:

R521

Chen Yina, Li Weidan, Zhu Kaiqiang, Tang Huiling. Serfling regression model for analyzing tuberculosis epidemiological characteristics in Jinhua City,Zhejiang Province[J]. Chinese Journal of Antituberculosis, 2026, 48(7): 941-947. doi: 10.19982/j.issn.1000-6621.20250523

Figures/Tables 6

References 22

[1]	丁哲渊, 杨研, 傅天颖, 等. 2024年浙江省法定传染病疫情分析[J]. 预防医学, 2025, 37(5): 433-437,442. doi:10.19485/j.cnki.issn2096-5087.2025.05.001.
[2]	金华市发改委. 金华市人口发展中长期规划(2021—2030年)[EB/OL](2022-11-09)[2025-11-13]. http://fgw.jinhua.gov.cn/art/2022/11/9/art_1229542811_4042926.html.
[3]	Wang X, Wu S, MacIntyre C R, et al. Using an adjusted Serfling regression model to improve the early warning at the arrival of peak timing of influenza in Beijing[J]. PloS One, 2015, 10(3): e0119923. doi:10.1371/journal.pone.0119923.
[4]	Panahi M H, Parsaeian M, Mansournia M A, et al. Detection of influenza epidemics using hidden Markov and Serfling approaches[J]. Transbound Emerg Dis, 2021, 68(4): 2446-2454. doi:10.1111/tbed.13912. pmid: 33152160
[5]	董瑞. 湖南省手足口病流行特征、超额发病及间接经济负担研究[D]. 长沙: 中南大学, 2022. doi:10.27661/d.cnki.gzhnu.2022.001443.
[6]	赵丹, 吴凤彤. 不同流行标准下传染病预警界值[J]. 社区医学杂志, 2019, 17(20): 1241-1244. doi:10.19790/j.cnki.JCM.2019.20.01.
[7]	吴凤彤, 吴凤娟. 辽阳市四种常见传染病预警界值优选分析[J]. 中国预防医学杂志, 2019, 20(9): 871-875. doi:10.16506/j.1009-6639.2019.09.023.
[8]	中华人民共和国国家卫生和计划生育委员会. WS 288—2017 肺结核诊断[J]. 结核与肺部疾病杂志, 2024, 5(4): 376-378. doi:10.19983/j.issn.2096-8493.2024022.
[9]	梁瑞英, 张爱洁, 吴冬雪, 等. 基于调整Serfling回归模型的2014—2023年北京市朝阳区肺结核流行特征及超额病例分析[J]. 疾病监测, 2025, 40(11): 1442-1448. doi:10.3784/jbjc.202411130652.
[10]	World Health Organization. Global tuberculosis report 2025[M]. Geneva: World Health Organization, 2025.
[11]	许航, 马晓静, 贾梦, 等. 2023年我国公立医院传染病医防融合实施现状调查[J]. 中国公共卫生, 2025, 41(9): 1086-1090. doi:10.11847/zgggws1147561.
[12]	朱闵敏, 郭旭君, 黄垚, 等. 基于调整Serfling回归法的2011—2016年深圳市南山区肺结核流行期分析[J]. 医学与社会, 2017, 30(8): 38-40,43. doi:10.13723/j.yxysh.2017.08.011.
[13]	Kharwadkar S, Attanayake V, Duncan J, et al. The impact of climate change on the risk factors for tuberculosis: A systematic review[J]. Environ Res, 2022, 212(Pt C): 113436. doi:10.1016/j.envres.2022.113436.
[14]	李玉红, 梅金周, 李雪, 等. 2018—2022年全国跨省流动肺结核患者流行特征分析[J]. 中国防痨杂志, 2025, 47(6): 760-768. doi:10.19982/j.issn.1000-6621.20250062.
[15]	高洪, 汪静. 调整Serfling回归模型在北京市东城区其他感染性腹泻病的应用[J]. 首都公共卫生, 2025, 19(5): 279-283. doi:10.16760/j.cnki.sdggws.2025.05.004.
[16]	周慧娴, 邓春燕, 李闰冰, 等. 2010—2023年重庆市九龙坡区手足口病流行趋势及超额发病数估计[J]. 职业与健康, 2025, 41(16): 2245-2249. doi:10.13329/j.cnki.zyyjk.2025.0425.
[17]	林凯, 刘雅文, 古丽斯, 等. 应用调整Serfling回归模型估计2019—2023年深圳市盐田区流行性感冒超额发病情况[J]. 预防医学情报杂志, 2025, 41(5): 626-631. doi:10.19971/j.cnki.1006-4028.240132.
[18]	柳正卫, 彭颖, 张明五, 等. 浙江省推广基于GeneXpert MTB/RIF检测的结核病诊断流程的效果分析[J]. 中国防痨杂志, 2021, 43(8): 790-795. doi:10.3969/j.issn.1000-6621.2021.08.007.
[19]	Feng L, Zhang T, Wang Q, et al. Impact of COVID-19 outbreaks and interventions on influenza in China and the United States[J]. Nat Commun, 2021, 12(1): 3249. doi:10.1038/s41467-021-23440-1. pmid: 34059675
[20]	龚亚波, 刘礼亲, 徐祖辉, 等. 新型冠状病毒感染疫情对湖南省结核病定点医院医疗服务的影响研究[J]. 结核与肺部疾病杂志, 2023, 4(4): 277-282. doi:10.19983/j.issn.2096-8493.20230068.
[21]	王倪, 黄飞, 赵雁林. 新型冠状病毒肺炎疫情防控经验对结核病防治工作的启示[J]. 中华传染病杂志, 2022, 40(1): 1-5. doi:10.3760/cma.j.cn311365-20210902-00317.
[22]	陈梦安, 张玲. 金华市医疗机构法定传染病报告质量调查[J]. 预防医学, 2020, 32(5): 505-507. doi:10.19485/j.cnki.issn2096-5087.2020.05.018.

统计值	第1次回归	第2次回归	第3次回归
纳入月份数	120	59	59
R²值	0.634	0.820	0.820
F值	69.573	89.050	89.050
P值	<0.001	<0.001	<0.001
截距项	354.658	324.614	324.614
t值	-0.950	-0.865	-0.865
cos(2πt/12)	-33.057	-35.771	-35.771
sin(4πt/12)	-11.346	-14.576	-14.576
剔除月份数	61	0	0

统计值	估计值	标准误	Z值	P值
AR	0.283	0.131	2.163	0.031
截距项	323.635	7.021	46.095	<0.001
t值	-0.852	0.101	-8.447	<0.001
cos(2πt/12)	-35.395	3.784	-9.354	<0.001
sin(4πt/12)	-13.329	3.071	-4.340	<0.001

年份	流行月数	流行期(月)	超额发病数(例,95%CI)
2015	6	1,3,5,7,8,12	274(186~362)
2016	6	1,3,8,10,11,12	302(214~390)
2017	8	1,3,5,7,8,9,10,12	509(408~610)
2018	5	1,3,5,10,12	369(289~449)
2019	5	1,2,3,5,6	295(215~375)
2020	1	9	82(46~118)
2021	3	3,6,8	119(39~199)
2022	3	3,7,9	148(68~228)
2023	3	3,8,12	153(73~233)
2024	6	3,8,9,10,11,12	329(249~409)
合计	46		2580(2337~2822)