Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Int J Equity Health . 2026 Mar 3;25:92. doi: 10.1186/s12939-026-02806-1 Search in PMC Search in PubMed View in NLM Catalog Add to search Mandated capacity reductions created new gender disparities in liver cancer surveillance: a population-based cohort study Li-Lin Liang Li-Lin Liang 1 Institute of Public Health, College of Medicine, National Yang Ming Chiao Tung University, No. 155, Sec. 2, Linong St., Beitou Dist, Taipei, 11221 Taiwan 2 Health Innovation Center, National Yang Ming Chiao Tung University, Taipei, Taiwan 3 Research Center for Epidemic Prevention and One Health, National Yang Ming Chiao Tung University, Taipei, Taiwan Find articles by Li-Lin Liang 1, 2, 3, ✉, # , I-Hua Chen I-Hua Chen 1 Institute of Public Health, College of Medicine, National Yang Ming Chiao Tung University, No. 155, Sec. 2, Linong St., Beitou Dist, Taipei, 11221 Taiwan 4 School of Medicine, College of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan Find articles by I-Hua Chen 1, 4, # , Chun-Ying Wu Chun-Ying Wu 2 Health Innovation Center, National Yang Ming Chiao Tung University, Taipei, Taiwan 3 Research Center for Epidemic Prevention and One Health, National Yang Ming Chiao Tung University, Taipei, Taiwan 5 Institute of Biomedical Informatics, National Yang Ming Chiao Tung University, Taipei, Taiwan 6 Microbiota Research Center, National Yang Ming Chiao Tung University, Taipei, Taiwan 7 Division of Gastroenterology, Department of Internal Medicine, Taipei Veterans General Hospital, Taipei, Taiwan Find articles by Chun-Ying Wu 2, 3, 5, 6, 7 Author information Article notes Copyright and License information 1 Institute of Public Health, College of Medicine, National Yang Ming Chiao Tung University, No. 155, Sec. 2, Linong St., Beitou Dist, Taipei, 11221 Taiwan 2 Health Innovation Center, National Yang Ming Chiao Tung University, Taipei, Taiwan 3 Research Center for Epidemic Prevention and One Health, National Yang Ming Chiao Tung University, Taipei, Taiwan 4 School of Medicine, College of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan 5 Institute of Biomedical Informatics, National Yang Ming Chiao Tung University, Taipei, Taiwan 6 Microbiota Research Center, National Yang Ming Chiao Tung University, Taipei, Taiwan 7 Division of Gastroenterology, Department of Internal Medicine, Taipei Veterans General Hospital, Taipei, Taiwan ✉ Corresponding author. # Contributed equally. Received 2025 Nov 22; Accepted 2026 Feb 24; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13067665 PMID: 41776508 Abstract Background Healthcare capacity constraints often force providers to prioritize acute conditions over preventive services such as cancer screening. However, few studies have quantified the effects of capacity-constraining policies, which is critical for emergency preparedness. We assessed how precautionary government-mandated hospital capacity reductions during the COVID-19 pandemic affected liver cancer surveillance, health equity, and clinical outcomes. Methods We conducted a population-based cohort study of 567,632 patients with hepatitis B/C in Taiwan’s National Health Insurance Database, using multiple-group interrupted time-series analysis (2021 vs. control period 2015–2019). The primary outcome was weekly abdominal ultrasound screening rates per 1,000 patients. A supplementary difference-in-differences analysis examined tumor size at diagnosis. Results At mandate onset, the surveillance rates dropped 48.8%. Cumulative surveillance deficits reached 22.7% (95% CI, − 34.7% to − 10.7%; P < 0.001) during the mandate, disproportionately affecting women, younger adults, and non-cirrhotic patients. While disparities among younger and non-cirrhotic patients were temporary, gender disparities persisted: women experienced 1.5-fold greater declines than men during the mandate (–25.3% vs. − 17.3%; P < 0.001) and 1.3-fold greater declines post-mandate (–47.0% vs. − 37.4%; P < 0.001). Mean tumor size at diagnosis increased 5.29 mm (a 9.1% growth; P = 0.001), with greater increases among women (12.8%) than men (8.5%). Conclusions Government-mandated capacity reductions, even with minimal COVID-19 caseload, led to substantial and prolonged surveillance deficits. These policies created new, persistent gender disparities and were associated with delayed cancer detection within a universal healthcare system. Crisis preparedness planning must incorporate health equity impact assessments to prevent unintended harm. Gender-sensitive strategies are urgently needed to re-engage women in preventive care. Supplementary Information The online version contains supplementary material available at 10.1186/s12939-026-02806-1. Keywords: Public health crisis, Gender inequality, Cancer screening, Health disparities, Health policy, COVID-19 Introduction Health system shocks, such as those caused by government regulations, great resignation, or public health crises, are occurring with increasing frequency worldwide. These events are often characterized by sudden reductions in healthcare capacity, including ward closures, workforce shortage, or supply chain disruptions. Faced with these constraints, hospitals and clinics must often prioritize patients with acute conditions, leading to delayed or deferred care for those with chronic illnesses. Crucially, these shocks can have heterogeneous effects across populations, widening pre-existing disparities in healthcare access and outcomes. To build equitable and resilient healthcare systems, it is critical to understand the impact of these shocks on high-risk patients, such as those requiring surveillance for liver cancer. Liver cancer, predominantly hepatocellular carcinoma (HCC), is a leading cause of cancer-related death globally [ 1 ]. Its high case-fatality rate often reflects late-stage diagnosis and limited treatment options for advanced disease [ 2 , 3 ]. Routine surveillance with semiannual abdominal ultrasound improves outcomes by enabling earlier detection when curative therapies are most effective [ 1 , 4 , 5 ]. International guidelines therefore recommend surveillance for high-risk individuals, including those with chronic hepatitis B or C, cirrhosis, or advanced fibrosis [ 1 ]. Despite these proven benefits, HCC surveillance remains critically underutilized. In the United States, for instance, fewer than 20% of patients with cirrhosis receive regular surveillance [ 6 ], and a systematic review reported a rate of only 18.4% [ 7 ]. Meta-analyses indicate that globally, just over half of high-risk patients adhere to surveillance recommendations [ 8 ], and only about one-third of HCC cases are detected through routine screening [ 9 ]. This study leverages a natural experiment in Taiwan, where a precautionary government mandate required medical institutions to reduce capacity during the COVID-19 pandemic. Before May 2021, Taiwan had maintained zero local transmission through stringent border controls. However, a small local outbreak in May 2021 prompted the Central Epidemic Command Center to issue a nationwide level-3 alert, mandating suspension of non-urgent services and allocation of at least 10% of ward capacity for COVID-19 patients. This mandate led to widespread cancellations of outpatient visits in anticipation of a severe outbreak that was ultimately averted. By year-end 2021, Taiwan’s cumulative infection rate remained remarkably low at 0.07%, compared to 12–19% in the US and Europe. We examined how this precautionary mandate affected liver cancer surveillance among patients with chronic viral hepatitis and whether effects varied by gender, age, socioeconomic status, or cirrhosis status. Previous research shows that the COVID-19 pandemic exacerbated underutilization of HCC surveillance [ 10 ], with sharp declines reported in North America [ 11 – 13 ], Europe [ 14 , 15 ], and Asia [ 16 ]. However, the specific impact of government-mandated capacity reductions remains poorly understood because most settings experienced concurrent severe outbreaks and control policies, making it impossible to disentangle policy effects from outbreak effect. Taiwan’s unique setting—where stringent capacity mandates were implemented despite minimal disease burden—allows us to isolate the impact of policy itself while minimizing confounding from outbreak severity. Critically, disparities in HCC surveillance were well-documented pre-pandemic, particularly in the United States, where older adults, racial and ethnic minorities, and individuals with lower socioeconomic status (SES) were less likely to undergo screening [ 6 , 7 , 17 – 19 ]. Women were more likely than men to receive surveillance in the US [ 19 , 20 ] and in rural China [ 21 , 22 ], and were also more often diagnosed through screening [ 23 ]. However, post-pandemic evidence on disparities is scarce and conflicting. Only a few studies have examined this issue, reporting contradictory findings on age-related differences [ 11 , 16 ] and limited data on gender disparities [ 24 ]. This study addresses several critical gaps. First, while prior work documented pandemic-era declines in HCC surveillance, few studies have quantified the causal effect of explicit capacity regulations. We specifically investigate how these policies reshape health inequities. Second, most research originates from the US and Europe; our study provides crucial evidence from Asia, which bears over 70% of the global HCC burden [ 25 ]. Finally, we extend our analysis beyond surveillance rates to examine downstream clinical consequences by assessing tumor size at diagnosis. In summary, by leveraging Taiwan’s natural experiment, this population-based study provides critical evidence on how healthcare capacity constraints affect cancer surveillance and alter health disparities. Because health systems worldwide may face policy- or crisis-related capacity shocks, our findings highlight a potentially transferable mechanism by which exogenous capacity constraints can disrupt preventive surveillance and reshape disparities. However, the magnitude and distribution of impacts are likely to be context-dependent, particularly in more fragmented systems with greater baseline access barriers. These insights underscore the need to anticipate and monitor distributional consequences when capacity-reduction policies are implemented, which we examine empirically in the analyses that follow. Methods Data source and study population We conducted a population-based study using data from Taiwan’s National Health Insurance Research Database (NHIRD), provided by the Health and Welfare Data Science Center, Ministry of Health and Welfare. Established under the single-payer National Health Insurance system, the NHIRD covers the entire population of 23 million and includes reimbursement claims from all hospitals and approximately 90% of clinics [ 26 ]. The database provides comprehensive information on diagnoses, procedures, prescriptions, cancer registries, and mortality [ 27 ]. The primary study cohort comprised patients diagnosed with HBV or HCV infection, identified by ≥ 3 outpatient visits or ≥ 1 hospital admission between Jan 1, 2010, and Dec 31, 2014. Patients were excluded if they had a prior cancer diagnosis, underwent liver-related procedures, or died before Dec 31, 2014. Diagnostic and procedural codes are listed in eTable S1 (Supplement 1 ). The final analytic cohort included 567,632 patients who were followed from Jan 1, 2015, to Dec 31, 2021 (cohort flowchart in eFigure S1 ). Exposure The exposure of interest was a government-mandated reduction in healthcare capacity, effective from May 19 to July 26, 2021. May 19 fell on the Wednesday of week 20, while July 26 fell on the Monday of week 30. Consequently, for the purpose of weekly aggregation, the exposure window is defined as weeks 20–29, with week 30 marking the beginning of the post-lift period. This mandate required the suspension or delay of nonurgent health services, including routine cancer surveillance [ 28 ]. Outcome measure The primary outcome was the weekly rate of abdominal ultrasound screening per 1,000 patients with HBV/HCV, including all screening records from outpatient, inpatient, and emergency department encounters. The cohort was updated weekly, excluding patients who were newly diagnosed with cancer, underwent liver-related procedures, or died, resulting in an overall 8.6% reduction in the baseline population during follow-up. Statistical analysis We used a multiple-group interrupted time-series analysis (ITSA) to estimate the impact of the 2021 government mandate on weekly cancer surveillance rates. This approach compared the observed rates in the exposure year (2021) to a counterfactual trend estimated from a pre-pandemic control period (2015–2019). The model included two intervention points to assess changes in the level and trend of surveillance: the start of the mandate period (week 20) and the start of the post-mandate period (week 30). The year 2020 was excluded because Taiwan had not yet experienced local COVID-19 transmission or associated public health restrictions [ 29 ]. We estimated the immediate effect as the level change in the surveillance rate at the onset of the mandate (week 20, 2021). We also calculated cumulative effects for both the mandate period (weeks 20–29) and the post-mandate period (weeks 30–52), which account for both immediate level shifts and subsequent changes in trend. All effects are reported as percentage differences relative to the predicted counterfactual rates. Analyses were performed in SAS and Stata 18.0 (StataCorp), with detailed model specifications available in Supplement 2 . Subgroup analyses We conducted subgroup analyses by age, gender, socioeconomic status (SES), and cirrhosis status. Age groups were defined as ≤ 40, 41–50, 51–60, 61–70, and > 70 years (as of 2021). SES was classified by employment sector and income level: formal-sector workers were divided into low, middle, and high income groups (based on monthly wages used for NHI premium calculations); informal-sector individuals were categorized as occupational union members, members of farmers’/fishermen’s associations, or “vulnerable groups” (low-income households and veterans). While premium-based income records in the NHIRD provide a precise measurement for the formal sector, equivalent data for the informal sector are often less accurate. Consequently, we categorized individuals in the informal sector by industry and social affiliations. Cirrhosis was defined as ≥ 3 outpatient visits or ≥ 1 hospital admission for cirrhosis during the baseline period (2010–2014). Between-group differences were assessed using the seemingly unrelated estimation (suest) method in Stata, which accounts for correlated errors across subgroup models. Lag-1 autocorrelation was detected in three subgroup models and corrected through model adjustment. Newey-West standard errors were applied to ensure robustness against autocorrelation and heteroskedasticity. Exploratory analysis of downstream clinical outcomes To assess potential clinical consequences of reduced surveillance, we performed a supplementary analysis of tumor size at diagnosis among patients newly diagnosed with HCC. This analysis used a distinct cohort of incident HCC cases ( N = 30,537) identified from the Taiwan Cancer Registry within the NHIRD. We applied a time-period difference-in-differences (DiD) design using daily data, comparing average tumor size in 2021 (exposure year) with that in 2015–2019 (baseline control years). The analysis was restricted to period from January 1 to July 26 of each year. The pre-mandate period was defined as January 1 to May 18, and the mandate period was defined as May 19 to July 26. Outcomes were evaluated overall and stratified by gender. Detailed DiD methods are provided in Supplement 3 . Ethics statement This study was approved by the Institutional Review Board of National Yang Ming Chiao Tung University (NYCU112108AE, NYCU112119AE and NYCU114016AE). The study adhered to the STROBE reporting guideline for cohort studies. Patients and the public were not involved in study design, conduct, reporting, or dissemination. Results Baseline characteristics and screening rates Table 1 shows demographic characteristics and mean weekly ultrasound screening rates during the control period (2015–2019). Among 567,632 patients with HBV or HCV, 56.2% were male, 51.5% were aged 51–70 years, 50.4% were employed in the formal sector, and 11.4% had cirrhosis. Table 1. Demographic characteristics and baseline weekly ultrasound screening rates per 1,000 HBV/HCV patients during the control period (2015–2019) Characteristic Patients, n (%) Weekly screening rate, mean (SD) Group difference, p -value Total 567,632 20.29 (2.72) Gender 0.42 Male 319,195 (56.2) 20.49 (2.70) Female 242,265 (42.7) 20.30 (2.82) Unknown 6,172 (1.1) n/a Age group (years) < 0.001 ≤ 40 42,258 (7.4) 15.05 (2.02) 41–50 120,203 (21.2) 17.83 (2.43) 51–60 149,738 (26.4) 20.62 (2.74) 61–70 142,424 (25.1) 22.66 (3.09) > 70 108,527 (19.1) 22.24 (3.24) Unknown 4,482 (0.8) n/a Socioeconomic status < 0.001 Formal sector Q1 (highest income) 95,189 (16.8) 20.96 (2.89) Q2 91,924 (16.2) 20.03 (2.82) Q3 (lowest income) 99,255 (17.5) 19.35 (2.87) Informal sector Occupational union 1 101,712 (17.9) 20.78 (3.02) Farmers and fishermen 85,768 (15.1) 22.50 (3.17) Others 2 89,001 (15.7) 18.86 (2.52) Unknown 4,783 (0.8) n/a Cirrhosis < 0.001 Yes 64,929 (11.4) 30.28 (4.47) No 502,703 (88.6) 19.09 (2.55) Open in a new tab SD: standard deviation. p < 0.05 indicates statistically significant between-group differences 1 Includes foreign crew members 2 Includes low-income households and veterans The national weekly screening rate was 20.29 per 1,000 patients. This value corresponds to a screening intensity of 52.8% (20.29 × 52 / 2,000), which means the healthcare system delivered about half of the total screening volume recommended by clinical guidelines (two screenings per patient per year). Screening rates did not differ by gender but were lower among younger individuals. The lowest rates were seen in the “others” category of the informal sector (18.86) and in the lowest-income formal group (19.35). Patients with cirrhosis had significantly higher screening rates than those without cirrhosis (30.28 vs. 19.09; P < 0.001). Visualization of ITSA-estimated changes in weekly ultrasound screening Figure 1 illustrates the weekly rate of ultrasound screening and the changes estimated by the ITSA model. The 2021 mandate was associated with a sharp, immediate decline in screening, followed by a sustained downward trend after the mandate was lifted. The validity of this ITSA approach is supported by the parallel pre-intervention trends observed between the exposure year and the control period (Supplement 4 ). Fig. 1. Open in a new tab Observed and Predicted Weekly Ultrasound Screening Rates Among Patients with Hepatitis B/C. The plot illustrates weekly screening rates per 1,000 patients before, during, and after a mandated capacity reduction in 2021. Intervention points are marked at the start of mandate (week 20) and the start of post-mandate period (week 30). Blue dots represent observed weekly rates in the exposure year (2021); hollow dots are weekly averages from the control period (2015–2019). The solid line shows the ITSA-predicted trend for 2021, while the dashed line represents the counterfactual trend based on the control period Immediate effects of the 2021 mandate Figure 2 shows the immediate effect of the 2021 mandate on weekly ultrasound screening rates, with corresponding numerical estimates in eTable S3 (Supplement 5 ). Fig. 2. Open in a new tab Immediate Effects of the Government Mandate on Weekly Ultrasound Screening Rates. The plot displays the immediate percentage change (level shift) in HCC surveillance at the onset of the mandate (week 20, 2021) for the overall population and by subgroups. Estimates represent the ITSA-estimated deviation from the predicted counterfactual rate based on the 2015–2019 control period. Error bars indicate 95% confidence intervals. Between-group differences were assessed using seemingly unrelated estimation At the onset of the mandate (week 20), the screening rate immediately dropped by 48.8% (95% CI, − 62.7% to − 35.0%; P < 0.001), or 10.44 fewer screenings per 1,000 patients, compared with the same week in control years. This reduction was significantly greater for women than for men (–55.1% vs. − 44.0%; P < 0.001). Steeper declines were also observed among younger adults aged ≤ 40 years (–61.6%), individuals in the highest-income formal sector group (–61.1%), and patients without cirrhosis (–51.1% vs. − 37.2% for those with cirrhosis; all P < 0.001). Cumulative effects of the 2021 mandate Figure 3 illustrates the cumulative effects of the mandate, with detailed estimates in eTables S5– S6 (Supplement 6 ). Fig. 3. Open in a new tab Cumulative Effects of the Government Mandate on Weekly Ultrasound Screening Rates. The figure displays the cumulative percentage change in HCC surveillance for the overall population and by subgroups.( A ) Effects during the mandate period (weeks 20–29, 2021). ( B ) Effects during the post-mandate period (weeks 30–52, 2021). Estimates represent the cumulative deviation from the predicted counterfactual rate based on the 2015–2019 control period. Horizontal bars denote the estimated effects, and error bars indicate 95% confidence intervals. Between-group differences were assessed using seemingly unrelated estimation During the mandate itself (weeks 20–29), the cumulative screening deficit was 22.7% (95% CI, − 34.7% to − 10.7%; P < 0.001), equivalent to 48.5 fewer screenings per 1,000 patients. This deficit widened in the post-mandate period (weeks 30–52), reaching 41.6% (95% CI, − 68.7% to − 14.5%; P = 0.003), or 193.6 fewer screenings per 1,000 patients. While surveillance rates reduced for almost all subgroups, women experienced larger cumulative reductions than men during both the mandate (–25.3% vs. − 17.3%; P < 0.001) and post-mandate periods (–47.0% vs. − 37.4%; P < 0.001). During the mandate, greater deficits were also seen in younger adults (–32.2%), the highest-income group (–25.2%), and patients without cirrhosis (–22.1% vs. − 14.2% for those with cirrhosis; all P < 0.01). Notably, these disparities were no longer statistically significant after the mandate was lifted. Exploratory analysis of tumor size at diagnosis In an exploratory analysis, we used a difference-in-differences approach to examine changes in tumor size among patients newly diagnosed with HCC. The validity of the DiD model was supported by parallel pre-mandate trends between the exposure (2021) and control years (2015–2019) (Supplement 3 ). During the mandate period (May 19–July 26), the mean tumor size at diagnosis increased by 5.29 mm (95% CI, 2.17 to 8.42; P = 0.001) relative to the control period (Table 2 ). This represents a 9.1% increase compared to the pre-mandate baseline. In gender-stratified analyses, tumor size increased by 5.20 mm (an 8.5% growth; 95% CI, 1.31 to 9.09; P = 0.009) in men and by 6.53 mm (a 12.8% growth; 95% CI, 1.47 to 11.59; P = 0.011) in women. Table 2. Difference-in-differences analysis of tumor size at HCC diagnosis: pre-mandate (January 1–May 18) vs. during mandate (May 19–July 26) All patients January 1–May 18 May 19–July 26 n = 25,995 n = 4,542 DiD 1 [95% CI 2 ] % Change 3 2015–2019 57.95 57.06 2021 61.29 65.70 Difference 3.35 8.64 5.29 [2.17–8.42] 9.13% (p-value) (0.0002) (< 0.0001) (0.001) Males January 1–May 18 May 19–July 26 n = 18,114 n = 3,160 DiD [95% CI] % Change 2015–2019 60.94 60.02 2021 63.91 68.19 Difference 2.97 8.17 5.20 [1.31–9.09] 8.53% (p-value) (0.007) (< 0.0001) (0.009) Females January 1–May 18 May 19–July 26 n = 7,787 n = 1,365 DiD [95% CI] % Change 2015–2019 51.08 49.98 2021 54.97 60.40 Difference 3.89 10.42 6.53 [1.47–11.59] 12.8% (p-value) (0.008) (< 0.0001) (0.011) Open in a new tab Tumor size is reported as diameter in millimetres. n refers to the number of HCC patients (the sample size) ¹DiD = difference-in-differences estimate comparing change from pre-mandate to mandate in 2021 vs. 2015–2019 2 95% CI = 95% Confidence Interval 3 Percentage change = DiD estimate divided by mean tumor size during pre-mandate period in 2015–2019 Sensitivity analyses The study’s main findings were robust in a sensitivity analysis where we used observed data from the control years instead of model-predicted counterfactual values (Supplement 7 ). Discussion This study found that a precautionary government-mandated 10% reduction in hospital capacity was associated with a substantial 22.7% decline in HCC surveillance. Critically, this disruption had prolonged collateral effects, with the surveillance deficit worsening to 41.6% even after the mandate was lifted. The mandate also exacerbated existing disparities among younger and non-cirrhotic patients, though these effects were temporary. However, a significant new and persistent gender disparity emerged, with women remaining the most affected group both during and after the mandate. This gender disparity was a novel finding as a consequence of emergency policy. Before the pandemic, men and women in Taiwan had comparable screening rates under the universal National Health Insurance (NHI) program. However, during the mandate, women experienced a 1.5-fold greater decline in surveillance than men (–25.3% vs. − 17.3%) and a 1.3-fold greater decline post-mandate (–47.0% vs. − 37.4%), highlighting their unique vulnerability. This new disparity in Taiwan likely reflects a combination of behavioral and social factors, as women may have exhibited greater risk aversion [ 30 ], shouldered heavier caregiving burdens [ 31 , 32 ], and faced greater mental health challenges during crises [ 33 , 34 ], all of which constrain attendance for routine care. This finding contrasts sharply with studies from the United States, where women typically have higher surveillance rates both before and during the pandemic [ 19 , 20 , 24 ]. This suggests that gendered vulnerabilities are context-dependent and can be magnified by health system shocks, challenging Western-centric assumptions about healthcare utilization. This also highlights the need for an intersectional approach, suggesting that future research should assess how gendered vulnerabilities are shaped by ethnicity in multicultural settings. Multiple factors likely explain the substantial nationwide decline in screening. Beyond the mandated minimums, many hospitals further reduced services due to workforce shortages or proximity to urban outbreak hotspots in northern Taiwan. This geographic concentration may explain why patients in the highest-income formal employment sector—who are more likely to live in these areas and seek care at high-tier medical centers [ 35 ]—experienced a more pronounced decline during the mandate. While supply-side constraints were the primary driver, reduced patient demand likely contributed to the screening deficit and its persistence. After the mandate, fear of infection probably discouraged patients from resuming non-urgent care, as COVID-19 vaccination coverage remained low (around 30%) [ 36 ]. Simultaneously, health systems faced a backlog of deferred visits and likely prioritized urgent cases, further displacing routine surveillance. The potential clinical consequences of these disruptions were suggested by an observed increase in tumor size at diagnosis, which rose by 8.5% in men and 12.8% in women. This pattern is consistent with delayed detection and a reduction in early-stage diagnoses, where smaller tumors may have gone undetected, thereby shifting the average tumor size among diagnosed cases upward. While similar findings were reported in England [ 14 ], a Japanese study found no significant change [ 37 ], highlighting the need for long-term monitoring to fully capture the downstream effects of interrupted care. Younger adults and non-cirrhotic patients were disproportionately affected, widening pre-existing gaps in care for these groups. This pattern mirrors findings for other cancers [ 38 – 40 ], where younger individuals—who often have a lower perceived risk of cancer [ 39 ] and heightened concerns about infection and job security [ 41 , 42 ]—experienced greater screening declines. The larger impact on non-cirrhotic patients may also reflect clinical triage, where providers prioritized patients with cirrhosis due to their higher risk of HCC. Our findings have critical implications for health equity and crisis preparedness. Capacity-constraining policies, even when precautionary and well-intentioned, can produce lasting collateral damage to essential health services and create new inequities within universal healthcare systems. The emergence and persistence of gender disparities demonstrate that emergency response planning must incorporate health equity impact assessments as standard practice. Policymakers should implement gender-sensitive strategies—including flexible screening hours, mobile outreach, and integration of surveillance with women’s health services—alongside targeted re-engagement campaigns for younger and non-cirrhotic patients who already face lower baseline uptake. Beyond Taiwan, our results speak to a transferable mechanism: when policy or crisis response reduces healthcare capacity, providers may triage toward acute care, preventive services may be deferred, and backlogs may prolong disruption even after restrictions are lifted. However, Taiwan’s centralized, single-payer system with near-universal coverage shapes both baseline surveillance patterns and the implementation of capacity policies; therefore, the magnitude—and even the direction—of subgroup disparities (including gender differences) may differ in fragmented systems with larger financial and organizational barriers. Future studies should test how system fragmentation, referral pathways, and baseline inequities modify the distributional impacts of capacity shocks. This study has several strengths, including its large, population-based cohort within a universal healthcare system and a unique natural experiment isolating the effects of capacity-constraining policies from outbreak severity. Methodologically, multiple-group ITSA with historical controls strengthens causal inference by accounting for unobserved confounders, while the complementary difference-in-differences analysis of tumor size illuminates downstream clinical consequences of policy interventions. This integrated approach provides a methodological template for evaluating health system resilience during emergencies. Several limitations should be noted. First, we could not directly measure the psychosocial or behavioral mechanisms underlying the observed disparities, including individual-level risk perceptions and care-seeking behaviors. Future research using individual-level data is needed to disentangle these factors and inform tailored interventions. Second, SES classification in the NHIRD is primarily based on premium-derived income records; thus, economic status may be measured less precisely than with direct financial reporting. We cannot exclude misclassification, particularly in subgroups for whom administrative records may not fully capture actual SES. Conclusion Government-mandated hospital capacity reductions led to substantial and sustained declines in HCC surveillance, disproportionately affecting women, younger adults, and non-cirrhotic patients. Although the disproportionate impact on younger and non-cirrhotic patients was temporary, the new gender disparity persisted post-mandate, identifying women as uniquely vulnerable. Concurrent increases in tumor size at diagnosis may reflect the clinical consequences of delayed surveillance. These findings demonstrate that capacity-constraining policies, even when precautionary, can create lasting inequities within a universal healthcare setting. Future capacity planning should integrate health equity impact assessments and gender-sensitive strategies to safeguard continuity of care and strengthen health system resilience during public health emergencies. However, while the mechanisms of disruption may be broadly relevant, the magnitude and distribution of impacts will depend on local system structure and baseline barriers to surveillance. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 (856.7KB, pdf) Acknowledgements We thank the National Science and Technology Council and the Ministry of Health and Welfare in Taiwan for their financial support. We also appreciate the valuable comments from participants of the Taiwan Public Health Joint Annual Conference. Abbreviations HCC Hepatocellular carcinoma HBV Hepatitis B virus HCV Hepatitis C virus SES Socioeconomic status ITSA Interrupted time-series analysis DiD Difference-in-differences NHIRD National Health Insurance Research Database Author contributions LLL had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. LLL and IHC contributed to the concept and design of the study. LLL was responsible for data acquisition, analysis, and interpretation. The manuscript was drafted by LLL and IHC, and critically revised for important intellectual content by LLL and CYW. Statistical analysis was performed by IHC. Funding was obtained by LLL. LLL provided administrative, technical, and material support and also supervised the study. The corresponding author (LLL) attests that all listed authors meet authorship criteria, that no others meeting the criteria have been omitted. The corresponding author also affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained. Funding This work was funded by National Science and Technology Council (grant numbers: 111-2314-B-A49A-502-MY3; 113-2321-B-A49-011; 114-2628-B-A49-007-MY3) and Ministry of Health and Welfare (grant numbers:113-TDU-B-221-134007; 114-TDU-B-221-144007). The grants were awarded to the lead author (LLL). Data availability The data employed in this study are subject to access restrictions and were used under a specific license for this research. Summary data and a data dictionary, which defines each field in the dataset, will be provided by the corresponding author ([email protected]) upon reasonable request. Declarations Ethical approval This study was approved by the Institutional Review Board of National Yang Ming Chiao Tung University (NYCU112108AE, NYCU112119AE and NYCU114016AE). The study adhered to the STROBE reporting guideline for cohort studies. Patients and the public were not involved in study design, conduct, reporting, or dissemination. Role of the Funder/Sponsor The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Li-Lin Liang and I-Hua Chen Co-first authors. References 1. Galle PR, Forner A, Llovet JM, Mazzaferro V, Piscaglia F, Raoul J-L, et al. EASL clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol. 2018;69(1):182–236. [ DOI ] [ PubMed ] [ Google Scholar ] 2. Choi DT, Kum HC, Park S, Ohsfeldt RL, Shen Y, Parikh ND, et al. Hepatocellular carcinoma screening is associated with increased survival of patients with cirrhosis. Clin Gastroenterol Hepatol. 2019;17(5):976–e874. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Tan D, Yopp A, Beg MS, Gopal P, Singal AG. Meta-analysis: underutilisation and disparities of treatment among patients with hepatocellular carcinoma in the United States. Aliment Pharmacol Ther. 2013;38(7):703–12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 4. Ferrante ND, Pillai A, Singal AG. Update on the diagnosis and treatment of hepatocellular carcinoma. Gastroenterol Hepatol (NY). 2020;16(10):506–16. [ PMC free article ] [ PubMed ] [ Google Scholar ] 5. Wu C-Y, Hsu Y-C, Ho HJ, Chen Y-J, Lee T-Y, Lin J-T. Association between ultrasonography screening and mortality in patients with hepatocellular carcinoma: a nationwide cohort study. Gut. 2016;65(4):693–701. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Davila JA, Morgan RO, Richardson PA, Du XL, McGlynn KA, El-Serag HB. Use of surveillance for hepatocellular carcinoma among patients with cirrhosis in the United States. Hepatology. 2010;52(1):132–41. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Singal AG, Yopp A, C SS, Packer M, Lee WM, Tiro JA. Utilization of hepatocellular carcinoma surveillance among American patients: a systematic review. J Gen Intern Med. 2012;27(7):861–67. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Zhao C, Jin M, Le RH, Le MH, Chen VL, Jin M, et al. Poor adherence to hepatocellular carcinoma surveillance: a systematic review and meta-analysis of a complex issue. Liver Int. 2018;38(3):503–14. [ DOI ] [ PubMed ] [ Google Scholar ] 9. Zhao C, Xing F, Yeo YH, Jin M, Le R, Le M, et al. Only one-third of hepatocellular carcinoma cases are diagnosed via screening or surveillance: a systematic review and meta-analysis. Eur J Gastroenterol Hepatol. 2020;32(3):406–19. [ DOI ] [ PubMed ] [ Google Scholar ] 10. Muñoz-Martínez S, Sapena V, Forner A, Nault JC, Sapisochin G, Rimassa L, et al. Assessing the impact of COVID-19 on liver cancer management (CERO-19). JHEP Rep. 2021;3(3):100260. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Mahmud N, Kaplan DE, Goldberg DS, Taddei TH, Serper M. Changes in hepatocellular carcinoma surveillance and risk factors for noncompletion in the veterans health administration cohort during the coronavirus disease 2019 pandemic. Gastroenterology. 2021;160(6):2162–e643. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Kim NJ, Rozenberg-Ben-Dror K, Jacob DA, Berry K, Ioannou GN. The COVID-19 pandemic highlights opportunities to improve hepatocellular carcinoma screening and diagnosis in a national health system. Am J Gastroenterol. 2022;117(4):678–84. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Makuza JD, Wong S, Morrow RL, Binka M, Darvishian M, Jeong D, et al. Impact of COVID-19 pandemic on hepatocellular carcinoma surveillance in British Columbia, Canada: an interrupted time series study. J Viral Hepat. 2024;31(10):592–600. [ DOI ] [ PubMed ] [ Google Scholar ] 14. Geh D, Watson R, Sen G, French JJ, Hammond J, Turner P, et al. COVID-19 and liver cancer: lost patients and larger tumours. BMJ Open Gastroenterol. 2022;9(1):e000794. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 15. Furumaya A, van Delden OM, de Wilde RF, de Vos-Geelen J, van der Geest LG. Impact of COVID-19 on incidence, treatment, and survival of patients with hepatocellular carcinoma in the Netherlands. Disaster Med Public Health Prep. 2024;18:e243. [ DOI ] [ PubMed ] [ Google Scholar ] 16. Toyoda H, Huang DQ, Le MH, Nguyen MH. Liver care and surveillance: the global impact of the COVID-19 pandemic. Hepatol Commun. 2020;4(12):1751–57. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 17. Parikh ND, Tayob N, Al-Jarrah T, Kramer J, Melcher J, Smith D, et al. Barriers to surveillance for hepatocellular carcinoma in a multicenter cohort. JAMA Netw Open. 2022;5(7):e2223504–04. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 18. Singal AG, Li X, Tiro J, Kandunoori P, Adams-Huet B, Nehra MS, et al. Racial, social, and clinical determinants of hepatocellular carcinoma surveillance. Am J Med. 2015;128(1):e901–7. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Perumalswami PV, Wyatt B, Bowman CA, Patel K, Mageras A, Lewis SC, et al. Hepatocellular carcinoma surveillance, incidence, and tumor doubling times in patients cured of hepatitis C. Cancer Med. 2022;11(9):1995–2005. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 20. Palmer LB, Kappelman MD, Sandler RS, Hayashi PH. Surveillance for hepatocellular carcinoma in a Medicaid cirrhotic population. J Clin Gastroenterol. 2013;47(8):713–8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. Cao M, Li H, Sun D, He S, Yang F, Yan X, et al. Assessment of the compliance, influencing factors, and yielding results of liver cancer screening in a high-risk population: a cross-sectional study. Cancer. 2022;128(20):3653–62. [ DOI ] [ PubMed ] [ Google Scholar ] 22. Li J, Li H, Zeng H, Zheng R, Cao M, Sun D, et al. Trends in high-risk rates and screening rates for the population-based cancer screening program on esophageal, stomach and liver cancer in China, 2010–2016. J Natl Cancer Cent. 2021;1(3):101–07. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 23. Ladenheim MR, Kim NG, Nguyen P, Le A, Stefanick ML, Garcia G, et al. Sex differences in disease presentation, treatment and clinical outcomes of patients with hepatocellular carcinoma: a single-centre cohort study. BMJ Open Gastroenterol. 2016;3(1):e000107. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Reiche WS, Cooper S, Destache CJ, Sidhu S, Schutte B, Keirns D, et al. Sex and race disparities in hepatocellular carcinoma surveillance in patients with chronic hepatitis B during COVID-19: a single-center retrospective review. Gastroenterol Res. 2023;16(4):203–08. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Singal AG, Lampertico P, Nahon P. Epidemiology and surveillance for hepatocellular carcinoma: new trends. J Hepatol. 2020;72(2):250–61. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 26. NHIA. 2022–2023 National health insurance annual report. Taipei: National Health Insurance Administration (NHIA). 2022. 27. Hsing AW, Ioannidis JP. Nationwide population science: lessons from the Taiwan National Health Insurance Research Database. JAMA Intern Med. 2015;175(9):1527–9. [ DOI ] [ PubMed ] [ Google Scholar ] 28. Liang L-L, Cambia JM. Taiwan case study: balancing health and economy during COVID-19. Asian Pac J Health Econ Policy. 2024;7(1):2–16. [ Google Scholar ] 29. Lu J-FR, Liang L-L. The role of digital health under Taiwan’s national health insurance system: progress and challenges. Health Syst Reform. 2024;10(2):2375433. [ DOI ] [ PubMed ] [ Google Scholar ] 30. Ferrín M. Reassessing gender differences in COVID-19 risk perception and behavior. Soc Sci Q. 2022;103(1):31–41. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Li S, Xu Q, Xie J, Wang L, Li H, Ma L, et al. Associations of parenting daily hassles with parents’ mental health during the COVID-19 school closure. Soc Sci Med. 2022;311:115301. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Yamamura E, Tsustsui Y. The impact of closing schools on working from home during the COVID-19 pandemic: evidence using panel data from Japan. Rev Econ Househ. 2021;19(1):41–60. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Abbas S, Nasir JA. Estimating the social burden of COVID-19 among caregivers of COVID-19 patients in Punjab, Pakistan. Sci Rep. 2024;14(1):29614. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 34. Dal Santo T, Sun Y, Wu Y, He C, Wang Y, Jiang X, et al. Systematic review of mental health symptom changes by sex or gender in early-COVID-19 compared to pre-pandemic. Sci Rep. 2022;12(1):11417. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Yang CJ, Tsai YC, Tien JJ. Patients with minor diseases who access high-tier medical care facilities: New evidence from classification and regression trees. Int J Health Plann Manage. 2019;34(2):e1087–97. [ DOI ] [ PubMed ] [ Google Scholar ] 36. Mathieu E, Ritchie H, Rodés-Guirao L, Appel C, Gavrilov D, Giattino C, et al. Coronavirus (COVID-19) vaccinations [Internet]. Oxford: OurWorldinData.org. 2020 [cited 2025 Apr 15]. Available from: https://ourworldindata.org/covid-vaccinations 37. Okushin K, Tateishi R, Hirakawa S, Tachimori H, Uchino K, Nakagomi R, et al. The impact of COVID-19 on the diagnosis and treatment of HCC: analysis of a nationwide registry for advanced liver diseases (REAL). Sci Rep. 2024;14(1):2826. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 38. Doan C, Li S, Goodwin JS. Breast and lung cancer screening among Medicare enrollees during the COVID-19 pandemic. JAMA Netw Open. 2023;6(2):e2255589. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Miller MM, Meneveau MO, Rochman CM, Schroen AT, Lattimore CM, Gaspard PA, et al. Impact of the COVID-19 pandemic on breast cancer screening volumes and patient screening behaviors. Breast Cancer Res Treat. 2021;189(1):237–46. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 40. Barclay NL, Pineda Moncusí M, Jödicke AM, Prieto-Alhambra D, Raventós B, Newby D, et al. The impact of the UK COVID-19 lockdown on the screening, diagnostics and incidence of breast, colorectal, lung and prostate cancer in the UK: a population-based cohort study. Front Oncol. 2024;14:1370862. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Pasion R, Paiva TO, Fernandes C, Barbosa F. The age effect on protective behaviors during the COVID-19 outbreak: sociodemographic, perceptions and psychological accounts. Front Psychol. 2020;11:561785. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Bruine de Bruin W. Age differences in COVID-19 risk perceptions and mental health: evidence from a national U.S. survey conducted in March 2020. J Gerontol B Psychol Sci Soc Sci. 2021;76(2):e24–29. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary Material 1 (856.7KB, pdf) Data Availability Statement The data employed in this study are subject to access restrictions and were used under a specific license for this research. Summary data and a data dictionary, which defines each field in the dataset, will be provided by the corresponding author ([email protected]) upon reasonable request. Articles from International Journal for Equity in Health are provided here courtesy of BMC ACTIONS View on publisher site PDF (2.6 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top