Spatio-temporal clustering characteristics of hepatitis B in China: an 18-year study Authors: Mengge Zhang, Chuancang Li, Juan Xu, Qian Cai, Huanyu Song, Liang Sun Published: 2026 ABSTRACT AbstractTo describe the spatiotemporal characteristics of hepatitis B in China and to identify its possible hotspots and potential high-risk areas. A retrospective, province-level ecological analysis of hepatitis B surveillance data. Global Moran’s I and local Moran’s I were calculated using GeoDa 1.16 to explore the spatial clustering characteristics from 2004 to 2021, and the results of the spatiotemporal scan analysis were obtained using SaTScan 10.1 to identify cluster areas from 2004 to 2018. From 2004 to 2021, the number of hepatitis B incident cases in China increased from 916,426 in 2004 to 976,233 in 2021 (AAPC = 0.38%, P = 0.613), which is not statistically significant. From 2018 to 2021, High-High cluster areas of hepatitis B incidence rates in China were concentrated in the southeast; from 2004 to 2019, High-Low cluster areas of hepatitis B mortality rates were concentrated in Beijing and Shanghai. During the 18-year study period, the crude mortality rate showed a statistically significant decline with an average annual decline of 4.03%. Hepatitis B high-risk areas were concentrated in northwestern and southeastern coastal regions of China, while death counts were clustered in the northwestern and eastern regions.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-57574-3.Keywords: Hepatitis B, Spatial epidemiology, Space–time scan statisticsSubject terms: Diseases, Gastroenterology, Health care, Medical research, Risk factors Introduction Hepatitis B remains a major global public health concern, with approximately 2.57 billion individuals worldwide having ever-exposed to the hepatitis B virus1. The World Health Organization (WHO) reported that over 354 million people worldwide live with chronic hepatitis, over 8000 new infections of hepatitis B and C occur every day, and more than one million deaths from advanced liver disease and liver cancer occur every year2. The global prevalence of chronic hepatitis is high, especially in certain developing countries, including China, Egypt and nine other countries where almost 50% of the population has the disease3. China has made significant progress in universal access to hepatitis B vaccination, with timely coverage of birth dose and full coverage of the vaccine series remaining > 95% since 20124. However, the disease burden of hepatitis B virus (HBV) in China is the highest among infectious diseases, with about 10 million deaths from chronic HBV expected by 20305. Currently, China still has the highest burden of HBV infection in the world, with one-third of the global 240 million chronic HBV carriers residing in China, many of whom are currently infected. The economic burden of HBV-related diseases has become a direct cause of poverty6. It should be noted that studies on hepatitis B mortality are scarce. Few have analyzed the mortality rate of hepatitis B from a comparative spatial and temporal perspective. Therefore, this study uses two disease frequency measures, incidence and mortality rates, to analyse the spatio-temporal characteristics of hepatitis B in China. Material and methods Data source The data of hepatitis B cases from 31 provinces (municipalities and autonomous regions) in the mainland of China from 2004 to 2018 were obtained from the Data Center of China Public Health Science of National Population and Health Science Data Sharing Platform in China (URL address: https://www.phsciencedata.cn/Share/). Data for 2019 to 2021 were obtained from the China Health Statistical Yearbook7. The population data used in this study are the year-end resident population figures released by the National Bureau of Statistics of the People’s Republic of China. Methods We applied ArcGIS 10.8 to create a hierarchical map of hepatitis B from 2004 to 2021 to visualise the spatial and temporal evolution of hepatitis B incidence and mortality rates in China. This study employed a second-order trend surface model to analyze the overall trends in hepatitis B incidence and mortality rates in China from 2004 to 2021. A coordinate system was created, with the x-axis representing the West–East direction and the y-axis representing the South-North direction. The fitted values of the trend surface for incidence and mortality rates (z-axis) reflected the spatial variation trends across these two directions. In this study, global and local spatial autocorrelation analyses were conducted in GeoDa. The spatial weight matrix for Moran′s I was constructed using the K-Nearest neighbors (KNN), and the number of Monte Carlo permutations was set to the default value of 999. This study used space–time scan statistics to identify the primary and secondary clusters of hepatitis B incidence and mortality rates from 2004 to 2018. A retrospective spatiotemporal analysis using the discrete Poisson model was employed. The number of Monte Carlo permutation was set to the default value of 999, and the Gini coefficient was applied to determine the scanning window radius8–10. The 10% scan radius corresponding to the maximum Gini coefficient value was selected as the maximum spatial cluster size, while the maximum temporal cluster size was set to 50% of the study period, allowing the window to move across both temporal and spatial dimensions (Supplementary Fig. 1). All spatial analyses were performed at a provincial scale. Statistical analysis All numbers were tabulated and summarized in Excel (Office 2021), which were used to calculate the number of incident and death cases, incidence and mortality rates. The global Moran′s I and local Moran′s I were measured using GeoDa (Version 1.16). The space time scan statistics were calculated with the SaTScan (Version 10.1). All maps were developed with ArcGIS (Version 10.8). Results Time trends and basic information of hepatitis B From 2004 to 2021, the number of hepatitis B incident cases in China increased from 916,426 in 2004 to 976,233 in 2021 (AAPC = 0.38%, P = 0.613), which is not statistically significant, and the number of deaths decreased from 783 in 2004 to 413 in 2021 (AAPC = − 3.23%, P = 0.076, also non-significant). During the same period, the crude incidence rate of hepatitis B in China decreased from 70.50 per 100,000 population in 2004 to 69.25 per 100,000 population in 2021 (AAPC = − 0.09%, P = 0.901, non-significant), and the crude mortality rate decreased from 0.06 per 100,000 population in 2004 to 0.03 per 100,000 population in 2021 (AAPC = − 4.03%, P < 0.001). Among these, only the crude mortality rate presented a statistically significant downward trend (Table 1). Table 1Incidence and mortality trend of hepatitis B from 2004 to 2021.YearIncident casesCrude incidence per 100,000Mortal casesCrude mortality per 100,0002004916,42670.507830.062005982,29775.139080.0720061,109,13084.829950.0820071,169,94689.008540.0720081,169,56988.528310.0620091,179,60788.827920.0620101,060,58279.466890.0520111,093,33581.546370.0520121,087,08680.685820.042013962,97471.125500.042014935,70269.053600.032015934,21568.573520.032016942,26868.744050.0320171,001,95272.614250.032018999,98571.994130.0320191,002,29271.774470.032020902,47664.294640.032021976,23369.254130.03AAPC (%)0.38− 0.09− 3.23− 4.03P0.6130.9010.076< 0.001AAPC, average annual percent change. Incidence and mortality trend of hepatitis B from 2004 to 2021. AAPC, average annual percent change. Spatial distribution of hepatitis B The results of the 2004–2021 hepatitis B incidence rate grading map showed obvious spatial differences across China, and the high-risk areas for hepatitis B were mainly concentrated in the northwest and southeast coastal regions of China. The mortality rate grading map indicated that high-mortality rate areas were primarily distributed in the northwest and eastern regions of China (Fig. 1). 2004, 2010, 2016, and 2021 spatial distribution of hepatitis B. Trend surface analysis of hepatitis B In 2004, the hepatitis B incidence rate in China exhibited an inverted U-shaped distribution from east to west. In recent years, the incidence rate has shown a trend of decreasing from west to east and increasing from north to south. Overall, except for 2004, the hepatitis B mortality rate was characterized as high in the east and low in the west, and in the North–South direction, the hepatitis B mortality rate was higher in the north than in the south from 2004 to 2021 (Fig. 2). 2004, 2010, 2016, and 2021 trend surface analysis of hepatitis B. Spatial autocorrelation analysis of hepatitis B The Moran′s I of the annual incidence rate of hepatitis B was positive, and the spatial autocorrelation of incidence was statistically significant in all years except 2013. The global Moran′s I was closer to 1 in 2004 and 2005, indicating more significant clustering features (Table 2). Table 2Global autocorrelation of hepatitis B from 2004 to 2021.Incidence rate of hepatitis BMortality rate of hepatitis BYearMoran′s IZ− valueP− valueSDE(I)Moran′s IZ− valueP− valueSDE(I)20040.3383.49290.0030.1058− 0.0333− 0.151− 1.37470.0570.0859− 0.033320050.3113.18250.0070.1077− 0.0333− 0.217− 1.80590.0160.1030− 0.033320060.2803.00670.0070.1050− 0.0333− 0.179− 2.12990.0030.0673− 0.033320070.2682.72540.0170.1100− 0.0333− 0.173− 1.97350.0090.0697− 0.033320080.2592.71630.0160.1068− 0.0333− 0.202− 1.89650.0080.0877− 0.033320090.2002.39700.0210.0965− 0.0333− 0.158− 2.23410.0040.0548− 0.033320100.2172.34580.0210.1059− 0.0333− 0.158− 2.27840.0020.0538− 0.033320110.2672.72920.0170.1093− 0.0333− 0.125− 2.49130.0080.0363− 0.033320120.2022.12960.0320.1081− 0.0333− 0.141− 1.45630.0240.0719− 0.033320130.1031.27850.1080.1024− 0.0333− 0.087− 1.38960.0780.0379− 0.033320140.1811.91590.0360.1079− 0.0333− 0.137− 2.35320.0040.0440− 0.033320150.1932.05240.0330.1072− 0.0333− 0.080− 1.17400.1330.0397− 0.033320160.1491.82650.0420.1012− 0.0333− 0.101− 1.08130.1130.0615− 0.033320170.1741.89180.0350.1063− 0.0333− 0.086− 0.96860.1350.0555− 0.033320180.1932.05620.0340.1075− 0.0333− 0.0180.24320.3480.0657− 0.033320190.1872.02890.0360.1065− 0.0333− 0.0220.22530.3810.0572− 0.033320200.2092.24190.0300.1062− 0.0333− 0.101− 0.92280.1780.0720− 0.033320210.2092.24190.0300.1062− 0.0333− 0.101− 0.92280.1780.0720− 0.0333 Global autocorrelation of hepatitis B from 2004 to 2021. The statistical results of the hepatitis B mortality rates in 2004, 2013, and 2015–2021 were not significant. The statistical results in 2005–2012 and 2014 were significant and the global Moran′s I were less than zero, indicating that the spatial distribution across China had a negative spatial correlation. The global Moran′s I was closer to − 1 in 2005 and 2008, indicating more significant dispersive features (Table 2). Local spatial autocorrelation analysis of hepatitis B The local spatial autocorrelation analysis showed different clusters according to LISA analysis. From 2004 to 2012, the High-High cluster areas of hepatitis B incidence rate in China were mainly concentrated in the northwest, including Xinjiang, Qinghai, Gansu and Ningxia. Low–High cluster areas were located in Tibet. High-Low cluster areas were distributed in Shandong and Anhui (Table 3). Table 3The result of local spatial autocorrelation analysis of hepatitis B from 2004 to 2021.YearHigh– high clusterLow– low clusterLow–high clusterHigh– low clusterIncidence rate of hepatitis B2004Xinjiang,Qinghai,Gansu,NingxiaHebei,ShandongTibet– 2005Xinjiang,Qinghai,Gansu,NingxiaShandong, ShanghaiTibet– 2006Xinjiang,Qinghai,Gansu,NingxiaJiangsu,ShanghaiTibet–2007Xinjiang, Gansu,NingxiaShandong,Jiangsu,ShanghaiTibet– 2008Xinjiang, GansuJilin,Liaoning.Shandong,Jiangsu,ShanghaiTibet– 2009XinjiangJilin,Liaoning.Shandong,Jiangsu,ShanghaiTibet–2010Xinjiang,NingxiaJilin,Liaoning.Shandong,Jiangsu,ShanghaiTibet– 2011Xinjiang,NingxiaJilin,Liaoning.Shandong,Jiangsu,ShanghaiTibet– 2012XinjiangJilin,Liaoning.Shandong,Jiangsu,ShanghaiTibet– 2013– Jilin,Liaoning.Shandong,Jiangsu,Anhui,ShanghaiTibet– 2014– Jilin,Liaoning.Shandong,Jiangsu,Anhui,Shanghai– – 2015– Jilin,Liaoning.Shandong,Jiangsu,Anhui,Shanghai– – 2016– Jilin,Liaoning.ShandongTibetAnhui2017– Jilin,Liaoning.Shandong– – 2018GuangdongHeilongjiang,Jilin,Liaoning– Shandong2019GuangdongHeilongjiang,Jilin,Liaoning,Sichuan– Shandong2020Guangdong HainanHeilongjiang,Jilin,Liaoning– Shandong, Anhui2021Guangdong HainanHeilongjiang,Jilin,Liaoning– ShandongMortality rate of hepatitis B2004– – – Beijing2005– – – Hubei,Shanghai2006– – – – 2007– – – – 2008– – – Beijing2009– – – Beijing2010– – – Beijing,Shanghai2011– – – Beijing2012– – – Shanghai2013– – HeilongjiangQinghai,Shanghai2014– – – – 2015– JiangxiJilinShanghai2016– – Heilongjiang– 2017– XinjiangJilin– 2018– – Jilin, Tianjin– 2019HeilongjiangXinjiang,QinghaiJilinShanghai2020– – – – 2021– – – – The result of local spatial autocorrelation analysis of hepatitis B from 2004 to 2021. There were no High-High, Low-Low or Low–High cluster areas for hepatitis B mortality rate in China from 2004 to 2012, while the High-Low cluster areas were mainly concentrated in Beijing and Shanghai from 2004 to 2019. The High-High cluster areas were concentrated in Heilongjiang, and Low–High cluster areas were primarily distributed in Jilin from 2015 to 2019 (Table 3). Space–time scan analysis of hepatitis B The results showed that the incidence rate presented one primary clustering area and two secondary clustering areas. The primary clustering area was located in the west of China, including Tibet, Qinghai, Xinjiang and Gansu. The secondary clustering areas were identified in Hainan and Guangdong from 2011 to 2017, and in Sichuan and Chongqing from 2004 to 2007 (Table 4) (Fig. 3). Table 4Space–time clustering of reported hepatitis B from 2004 to 2018.Cluster typeLocation IDs includedCoordinates / radius(km)Time frameN*Observed*Expected*RRLLRPIncidence rate of hepatitis BPrimary ClusterTibet, Qinghai, Xinjiang, GansuTibet (31.540900 N, 88.614700 E) / 1322.282006/1/1 to 2012/12/314810,086301,426.022.78300,786.490.000Secondary Cluster 1Hainan, GuangdongHainan (17.366600 N, 110.605800 E) / 723.612011/1/1 to 2017/12/3121,096,917633,115.001.79146,359.660.000Secondary Cluster 2Sichuan, ChongqingSichuan (30.594800 N, 102.696000 E) / 498.822004/1/1 to 2007/12/312463,912339,034.491.3821,116.240.000Mortality rate of hepatitis BPrimary ClusterBeijingBeijing (40.184000 N, 116.410300 E) / 02007/1/1 to 2013/12/31172564.2012.141120.1979070.000Secondary Cluster 1Qinghai, Gansu, Tibet, SichuanQinghai (35.670000 N, 96.032100 E) / 837.812004/1/1 to 2008/12/314547274.822.05108.3602770.000Secondary Cluster 2Fujian, JiangxiFujian (26.055900 N, 117.974900 E) / 282.052005/1/1 to 2009/12/312349190.211.8754.3871810.000(The radius is reported only when the cluster areas include more than 1 units, and the criteria for Reporting Secondary Clusters is “no geographical overlap”. N*: number of provinces in a cluster; Observed*: Number of Observed cases in a cluster; Expected*: Number of Expected cases in a cluster; RR: Relative risk; LLR: Log-likelihood Ratio). Space–time clustering of reported hepatitis B from 2004 to 2018. (The radius is reported only when the cluster areas include more than 1 units, and the criteria for Reporting Secondary Clusters is “no geographical overlap”. N*: number of provinces in a cluster; Observed*: Number of Observed cases in a cluster; Expected*: Number of Expected cases in a cluster; RR: Relative risk; LLR: Log-likelihood Ratio). Space–time clustering of reported hepatitis B from 2004 to 2018. The results showed that the mortality rate contained one primary clustering area and two secondary clustering areas. The primary clustering area was located in Beijing. The secondary clustering areas were located in the west of China, including Qinghai, Gansu, Tibet and Sichuan, from 2004 to 2008, and in Fujian and Jiangxi from 2005 to 2009 (Table 4) (Fig. 3). Discussion The results of the study found that over the 18-year period, the overall crude incidence rate of hepatitis B in China showed a downward trend, though the difference was not statistically significant; the crude mortality rate also showed a downward trend with an average annual decline of 4.03%, which is consistent with the studies11,12. The decline in incidence rate may be partly associated with the immunization strategy pursued in China.In 1992, the Chinese government incorporated hepatitis B vaccination into its routine immunization program. In 2002, the hepatitis B vaccine was formally included in the National Immunization Program, with free vaccination provided to newborns and a requirement that the first dose be administered within 24 h of birth. By 2005, free vaccination coverage had been extended to all newborns.Studies have shown that since 1992, 90 million cases were prevented against the HBV infection and 24 million HBV carriers were decreased13.The reduction in mortality rate can be attributed to both the long-term effects of universal vaccination and, more directly, the significant improvement in access to antiviral treatment in recent years.Most oral antiviral medications for hepatitis B (such as entecavir and tenofovir) are now covered by medical insurance, greatly improving their accessibility and affordability.The number of hepatitis B incident cases is on the rise, but the crude incidence rate is declining, which is not statistically significant. As infectious diseases reporting system continues to improve, the timeliness and completeness of case detection and reporting have gradually increased, resulting in more cases that may have previously been underreported being included in the statistics, which has objectively led to a rise in the number of incident cases. According to official data from the National Bureau of Statistics, China’s population grew from 1.299 billion in 2004 to 1.413 billion in 2021, a net increase of 114 million people. Although the number of hepatitis B cases has risen, its growth rate has been lower than the population growth rate during the same period, resulting in a downward trend in the incidence rate. The results of spatial autocorrelation showed that the Moran′s I for hepatitis B annual incidence rate in China was greater than zero, and the spatial autocorrelation of incidence rates was statistically significant in all years except 2013, which was consistent with previous studies14. The spatial autocorrelation statistics for hepatitis B mortality rates in 2004, 2013, and 2015–2021 were not statistically significant. This may be primarily because the spatial autocorrelation analysis in this study was performed at the provincial level, where the spatial scale was not precise enough, thereby obscuring the spatial clustering of hepatitis B mortality rates15. The global Moran′s I for hepatitis B mortality rates was negative and statistically significant in 2005–2012 and 2014, indicating spatial dispersion rather than clustering, while incidence rates showed consistent positive spatial autocorrelation during these years. This suggests that high-risk areas of hepatitis B incidence are not necessarily areas with high mortality rates, indicating a spatial “decoupling” between incidence and mortality rates. While the underlying mechanism cannot be fully clarified in this study, this phenomenon merits attention and warrants further investigation. Combined with the spatial distribution and trend surface analysis of hepatitis B, the high-risk areas of hepatitis B were mainly concentrated in the northwestern region of China and the southeastern coastal region, and the number of deaths is mainly concentrated in the northwestern and eastern regions of China, which is consistent with previous studies16. The results of local spatial autocorrelation showed that the high incidence rate areas of hepatitis B gradually shifted from the northwest to the east. Starting from 2017, the high incidence rate areas have shifted from Xinjiang, Qinghai, Gansu and Ningxia to Guangdong, Hainan, Shandong and Anhui, which is consistent with related studies17,18. Based on existing evidence, it can be hypothesized that the rapid economic development of eastern China and large-scale population migration may have intensified population mobility, which might potentially facilitate the spread of the hepatitis B epidemic in the southeastern coastal regions. High-High cluster areas were mainly distributed in the northwest showing a patchy distribution.It is speculated that a combination of factors—including insufficiently targeted prevention strategies, frequent inter-provincial population mobility due to geographical proximity, relatively limited healthcare resources, and a relatively underdeveloped economy—may have contributed to the high hepatitis B incidence rate in the northwest19,20. Notably, for Hainan and Guangdong, the high incidence rate without correspondingly high mortality rate in recent years may be partially explained by the long disease course of chronic hepatitis B, which typically progresses to cirrhosis and liver cancer over decades, thereby delaying mortality. The mortality rate of hepatitis B was primarily concentrated in Beijing and Shanghai.In these megacities, which have abundant medical resources, several non-mutually exclusive hypotheses may explain this High-Low cluster pattern. First, referral bias may have contributed to the observed clustering, as severely ill patients from surrounding areas were referred to these cities for advanced treatment, resulting in deaths being registered at the treatment site rather than the place of residence21,22. Second, differences in disease reporting quality across regions may have artificially inflated mortality in areas with more robust surveillance systems. Third, demographic composition—including a large floating population or differential age structures—may influence local mortality rates. Fourth, despite the overall strong healthcare infrastructure, disparities in access to treatment for chronic hepatitis B within these megacities cannot be ruled out. The western region of China, including Qinghai, Gansu, Tibet, and Sichuan, was at high risk for hepatitis B mortality from 2004 to 2008, consistent with the results of an observational population-based study. Based on the available evidence, it can be speculated that the lower level of regional economic development and relatively limited medical resources may be potential influencing factors for this mortality distribution pattern23,24. Notably, as an ecological and descriptive study, the present work cannot quantitatively verify causal relationships. All influencing factors discussed above are proposed as reasonable hypotheses and contextual explanations based on existing literature, rather than confirmed causal conclusions. Therefore, further targeted empirical studies are required to verify these potential driving mechanisms. Limitations This study has several limitations. First, variations in the coverage and sensitivity of hepatitis B surveillance systems across different regions may introduce confounding factors; furthermore, regional economic disparities may affect surveillance quality, and underreporting may occur in some areas. Second, this study did not perform a space–time scan analysis for 2019–2021. Third, this study used aggregated provincial-level data for ecological analysis, which cannot reflect individual-level characteristics and thus makes it difficult to support causal inferences. Fourth, only crude mortality and crude incidence rates were used, without age-standardized rates, which may affect the comparability of cross-regional comparisons. Fifth, the use of dual data sources may be prone to data heterogeneity and systematic bias. Supplementary Information Supplementary Information 1. Acknowledgements All authors read and approved the final version of the manuscript. Author contributions MZ: methodology and writing, formal analysis-original draft. MZ, CL, JX, QC, and HS: data curation. LS: review and supervision, project administration. All authors contributed to the article and approved the submitted version. Funding This research was supported by the Jinan Municipal Health Commission (No. 2025306008) and the Health Commission of Henan Province (No. LHGJ20240677). Data availability The datasets generated and/or analysed during the current study are available in the following repositories: (1)Hepatitis B case data from 2004 to 2018 in the Data Center of China Public Health Science, URL:https://www.phsciencedata.cn/Share/. (2)Hepatitis B case data from 2019 to 2021 in the China Health Statistical Yearbook. (3)Population data in the National Bureau of Statistics of China,URL:https://www.stats.gov.cn/. (4)Map data in the National Catalogue Service for Geographic Information, URL:https://www.webmap.cn/main.do?method=index. • The datasets generated and/or analysed during the current study are not publicly available due to concern about potential plagiarism, as the manuscript has not yet been published,but are available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Ethics approval and consent to participate Not applicable. Consent for publication All authors have read and agreed to the published version of the manuscript. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Information 1. Data Availability Statement The datasets generated and/or analysed during the current study are available in the following repositories: (1)Hepatitis B case data from 2004 to 2018 in the Data Center of China Public Health Science, URL:https://www.phsciencedata.cn/Share/. (2)Hepatitis B case data from 2019 to 2021 in the China Health Statistical Yearbook. (3)Population data in the National Bureau of Statistics of China,URL:https://www.stats.gov.cn/. (4)Map data in the National Catalogue Service for Geographic Information, URL:https://www.webmap.cn/main.do?method=index. • The datasets generated and/or analysed during the current study are not publicly available due to concern about potential plagiarism, as the manuscript has not yet been published,but are available from the corresponding author on reasonable request.