Medical travel patterns for hepatocellular carcinoma treatment in South Korea: National Health Insurance data from 2013 to 2021

Abstract

BACKGROUND

Hepatocellular carcinoma (HCC) remains a significant public health concern in South Korea even though the incidence rates are declining. While medical travel for cancer treatment is common, its patterns and influencing factors for patients with HCC are unknown.

AIM

To assess medical travel patterns and determinants and their policy implications among patients with newly diagnosed HCC in South Korea.

METHODS

This retrospective cohort study used the National Health Insurance Service database to identify patients with newly diagnosed HCC from 2013 to 2021. Medical travel was defined as receiving initial treatment outside one’s residential region. Patient characteristics and regional trends were analyzed, and factors influencing medical travel were identified using logistic regression analysis.

RESULTS

Among 64808 patients 52.4% received treatment in the capital. This proportion increased to 67.4% when including the surrounding metropolitan area. Medical travel was significantly more common among younger and wealthier patients. Patients with greater comorbidity burden or liver cirrhosis were less likely to travel. While geographic distance influenced travel patterns, high-volume academic centers in the capital attracted patients nationwide regardless of proximity.

CONCLUSION

This nationwide study highlighted the centralization of HCC care in the capital. This observation indicates that regional cancer hubs should be strengthened and promoted for equitable healthcare access.

Key Words: Hepatocellular carcinoma; Medical travel; Regional disparities; National Health Insurance Service; Healthcare centralization

Core Tip: This nationwide retrospective cohort study examined domestic medical travel patterns among patients newly diagnosed with hepatocellular carcinoma in South Korea. Despite universal health insurance and open provider access, over half of the patients with hepatocellular carcinoma traveled to the capital for initial treatment. Younger age, higher income, and residence in non-capital regions were key drivers of medical travel. These findings highlighted the persistent regional disparities and indicate that policy efforts to strengthen regional cancer centers should be made to ensure geographically equitable access to specialized care.

INTRODUCTION

The incidence of primary liver cancer in South Korea has been declining since 1999, with an annual reduction of 1.2% from 1999 to 2010 and followed by a more pronounced decrease of 4.0%. Despite this downward trend primary liver cancer remained the seventh most common cancer and the second-leading cause of cancer-related mortality in 2021, surpassed only by lung cancer[1]. Given its high incidence and mortality rates, liver cancer represents a substantial public health burden in South Korea. Although regional disparities in incidence and treatment have been reported, few studies have thoroughly examined inter-regional access to care[2].

Hepatocellular carcinoma (HCC) accounts for approximately 90% of primary liver cancers[3,4] and is the focus of this study. Globally, HCC represents a major oncological burden, particularly in East Asia where it accounts for a high proportion of liver cancer cases[3]. Its complex etiology and diverse treatment options, ranging from resection and locoregional therapy to transplantation and systemic therapy, require multidisciplinary management and highly specialized care settings[4]. Since the implementation of mandatory social health insurance in 1989, the National Health Insurance Service (NHIS) has provided coverage to nearly the entire South Korean population[5]. This system allows patients to freely select healthcare providers regardless of geographic location. Consequently, many cancer patients seek treatment outside their residential areas, particularly in Seoul[6-8]. During travel for medical purposes, patients typically pay additional costs for transportation and lodging, endure additional physical and mental fatigue, and must allow extra time. The expenses for medical travel are challenging for patients as well as the national health care system[9,10].

Currently, no large-scale studies have examined the trends of medical travel among South Korean patients with HCC. Although a prior investigation analyzing patients with prostate cancer revealed a mobility pattern concentrated in the capital region[8], HCC care is generally more complex and requires chronic disease management and multiple treatment phases. Therefore, evaluating travel patterns in this population may offer meaningful insights into access disparities and system-level inefficiencies. The aim of this study was to evaluate the extent of medical travel among patients with HCC across different regions from 2013 to 2021. We identified demographic and regional factors influencing medical travel behavior and assessed implications for healthcare policy in South Korea. Our observations can potentially inform other countries with centralized cancer care systems of their citizens’ medical travel patterns.

MATERIALS AND METHODS

Data source

Koreans are universally enrolled in the NHIS, a government-operated single-payer program. The Korean National Health Information Database combines information obtained from the NHIS collected for claims and reimbursement of health care services and general health examination results[11,12]. Accordingly, health care information on HCC was retrieved from the National Health Information Database.

Patient selection

This retrospective cohort included patients newly diagnosed with HCC between 2013 and 2021. They were identified by the International Classification of Diseases, 10^th Revision (ICD-10) code C22.0. Initial treatment was defined as the first recorded procedure code for HCC-related therapy, including hepatic resection (Q7221, Q7222, Q7223), liver transplantation (Q8040), transarterial embolization (M6644, OZ753), radiofrequency ablation (QZ841, Q7280), and systemic therapy agents reimbursed under anticancer drug codes.

Individuals with a prior diagnosis of any malignancy within the preceding 5 years were excluded to minimize confounding. To ensure accurate assessment of medical travel, patients without records of initial treatment were excluded. Patients who had only a single outpatient consultation without any procedural or therapeutic claims were assumed to have declined treatment and were therefore excluded. Cases with incomplete eligibility or missing core sociodemographic data were excluded. Figure 1 outlines the cohort selection process.

Figure 1 Flowchart of the study population selection based on the National Health Insurance Service data for hepatocellular carcinoma. HCC: Hepatocellular carcinoma; ICD: International Classification of Diseases; NHIS: National Health Insurance Service.

Medical travel

HCC management typically involves multiple treatment modalities and requires regular follow-up to evaluate treatment response and plan subsequent care. Because of this complexity, patients often prefer continuity of care at a single institution. The treatment region was determined by the administrative location of the facility where the initial treatment was provided. Medical travel was defined as receiving this initial treatment in a region other than the patient’s region of residence.

Regional categorization

South Korea is administratively divided into 17 regions. Chungcheong Province was consolidated to include both the southern and northern provinces as well as Sejong City to improve regional homogeneity. Similarly, the South and North Provinces of Jeolla and of Gyeongsang were merged into single regions. Therefore, the study population was categorized into 13 consolidated regions. The residential region and the region of the initial treatment facility were compared to identify interregional medical travel. Additionally, regional patterns of patient inflow and outflow were quantified.

Variables

Factors influencing medical travel consisted of age, sex, Charlson comorbidity index (CCI), year of diagnosis, income level, residential region, liver cirrhosis, and extrahepatic metastasis. The NHIS premium is principally based on household income, and the medical aid group (Medicare), which accounts for approximately 3% of the population, is exempt from this premium[12]. Income level was derived from NHIS premium data, which ranges from 0 to 20. Medicare beneficiaries correspond to level 0, while the remaining values were grouped into four income quintiles as follows: 1-5 (low income); 6-10 (low-middle income); 11-15 (middle-high income); and 16-20 (high income). CCI, a standardized measure of comorbidity severity designed to predict patient outcomes following critical illness, was calculated based on ICD-10 codes in the year prior to diagnosis (including liver cancer), resulting in no patients with a score of 0. Scores were categorized as 2, 3-4, and ≥ 5. Patients with missing values in key covariates such as income, region, or treatment status were excluded from the final cohort.

Statistical analysis

The trend of medical travel was deduced from the number of patients diagnosed and treated in a region. Net migration was calculated from the difference between the number of treated and diagnosed patients in each region and was analyzed using a t-test. The Cochran-Armitage trend test was employed to compare differences in trends in medical travel by year. Characteristics of patients treated either in or outside residential regions were analyzed using a t-test for continuous variables and the χ² test for categorical variables. A multivariable logistic regression model was used to calculate adjusted odds ratios (aORs) and 95% confidence intervals for receiving initial treatment at an out-of-region hospital. Variables with P < 0.05 in univariate analysis were included in the multivariate model. All analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc., Cary, NC, United States). Statistical significance was set at P < 0.05.

RESULTS

Patient characteristics by treatment regions

From 2013 to 2021, a total of 64808 patients were newly diagnosed and treated for HCC in South Korea. Table 1 presents the demographic and clinical characteristics stratified by treatment in the patients’ residential region or another region. The mean age of patients with HCC was 61.3 years, with 50988 (78.6%) males and 13820 (21.3%) females. The largest patient cohort comprised the 50-59 years age group, followed by the 60-69 years and 70-79 years age groups. Patients receiving treatment outside their residential region had a significantly lower mean age (60.3 years vs 62.4 years, P < 0.0001).

Table 1 Patient characteristics according to treatment regions.

Characteristics	Different region		Residential region		P value
	n/N	%	n/N	%
Age in years
0-39	1258/1938	64.9	680/1938	35.1	< 0.0001a
40-49	5033/8745	57.6	3712/8745	42.4
50-59	11491/21062	54.6	9571/21062	45.4
60-69	9759/18575	52.5	8816/18575	47.5
70-79	5563/11775	47.2	6212/11775	52.8
≥ 80	1053/2713	38.8	1660/2713	61.2
Sex
Male	26852/50988	52.7	24136/50988	47.3	0.6841
Female	7305/13820	52.9	6515/13820	47.1
Residential region
Seoul	401/9879	4.1	9478/9879	95.9	< 0.0001b
Busan	1923/4777	40.3	2854/4777	59.7
Daegu	758/2780	27.3	2022/2780	72.7
Incheon	1392/3185	43.7	1793/3185	56.3
Gwangju	915/1194	76.6	279/1194	23.4
Daejeon	529/1212	43.6	683/1212	56.4
Ulsan	509/999	51.0	490/999	49.0
Chungcheong	4159/5122	81.2	963/5122	18.8
Gyeonggi	8909/15978	55.8	7069/15978	44.2
Gangwon	1552/2299	67.5	747/2299	32.5
Jeolla	4097/6194	66.1	2097/6194	33.9
Gyeongsang	8279/10042	82.4	1763/10042	17.6
Jeju	726/1147	63.3	421/1147	36.7
Year of diagnosis
2013	4008/7552	53.1	3544/7552	46.9	0.0045b
2014	3986/7382	54.0	3396/7382	46.0
2015	4047/7501	54.0	3454/7501	46.0
2016	3933/7436	52.9	3503/7436	47.1
2017	3894/7404	52.6	3510/7404	47.4
2018	3817/7204	53.0	3387/7204	47.0
2019	3775/7294	51.8	3519/7294	48.2
2020	3530/6872	51.4	3342/6872	48.6
2021	3167/6163	51.4	2996/6163	48.6
CCI
2	4063/7285	55.8	3222/7285	44.2	< 0.0001b
3-4	14207/26245	54.1	12038/26245	45.9
≥ 5	15887/31278	50.8	15391/31278	49.2
Income level, quintiles
Medicare	1450/3847	37.7	2397/3847	62.3	< 0.0001b
1 (low income)	5716/11297	50.6	5581/11297	49.4
2	4747/9474	50.1	4727/9474	49.9
3	5885/11406	51.6	5521/11406	48.4
4	7155/13047	54.8	5892/13047	45.2
5 (high income)	9204/15737	58.5	6533/15737	41.5
Liver cirrhosis
Yes	5611/11054	50.8	5443/11054	49.2	< 0.0001b
No	28546/53754	53.1	25208/53754	46.9
Extrahepatic metastasis
Yes	3912/7458	52.5	3546/7458	47.5	0.6442
No	30245/57350	52.7	27105/57350	47.3

^aP < 0.05 from t-test for continuous variables.

^bP < 0.05 from χ² test for categorical variables.

CCI: Charlson comorbidity index.

Seoul exhibited the highest in-region treatment rate at 95.9%, followed by Daegu and Busan. Patients receiving medical aid were the only socioeconomic group more likely to be treated in their residential region. Higher CCI scores were more prevalent among patients treated in their residential region. Treatment region was significantly associated with age, residential area, CCI, income level, and liver cirrhosis status (all P < 0.0001) and with the year of diagnosis (P = 0.0045).

Trends in inter-regional medical travel

Gyeonggi Province had the largest number of newly diagnosed patients, followed by the Gyeongsang Province and Seoul. This trend was consistent throughout the study period (Figure 2). When analyzing patients outside of Seoul and the Gyeonggi Province (which surrounds Seoul on three sides, to the north, east, and south), more than half of these patients traveled to Seoul for initial treatment. This proportion increased to 67.4% when including the Gyeonggi Province. This observation highlighted the significant concentration of treatment in the capital region. Among patients who received treatment outside their residential region, 71.6% selected Seoul as their treatment destination. When expanded to include the Gyeonggi Province, 79.4% of patients received treatment in Seoul, indicating a strong preference for this region. The Cochran-Armitage trend test indicated a statistically significant upward trend in medical travel over time (Z = 2.006, P = 0.0448).

Figure 2 Trends of patients with hepatocellular carcinoma. A: Diagnosed region; B: Treatment region.

Patients from the Gyeonggi, Gyeongsang, Chungcheong, and Jeolla Provinces were the most likely to leave their residential regions for medical treatment. Seoul showed the highest treatment inflow, receiving 24451 patients from other regions, accounting for 72.1% of all patients treated in the capital. Other regions with notable patient inflows included Gyeonggi, Daegu, and Busan. When we excluded Seoul from the analysis of patient inflow, Gwangju, Daegu, Daejeon, and Busan exhibited the highest proportions of out-of-region patients. A graphical representation of medical travel is shown in Figure 3. Figure 4 presents the treatment destinations of patients by their region of residence. Although the majority of medical travel was concentrated in Seoul, Gyeongsang Province had the highest proportion of patients who sought treatment outside of Seoul. Gwangju and Chungcheong Provinces also had high proportions of patients who traveled to regions other than Seoul for treatment.

Figure 3 Medical travel of patients with hepatocellular carcinoma. A: Out of the region (efflux) for treatment; B: Into the region (influx) for treatment; C: Proportion of patients who traveled out of diagnosed areas; D: Proportion of patients who traveled to different regions for management.

Figure 4  Distribution of healthcare service utilization across regions based on patient residency.

Factors associated with medical travel

This study examined factors influencing the likelihood of receiving initial treatment for HCC at a facility outside a patient’s residential region (Table 2). Multivariate logistic regression analysis revealed that the aOR of medical travel decreased with increasing age. It was significantly more likely that the 40-49 years and 50-59 years age groups traveled outside their residential region compared with the older cohorts. Compared with patients residing in Seoul (reference group), patients in all other regions had significantly higher aORs for receiving treatment outside their residential area. The highest aORs were observed in Gyeongsang Province, Chungcheong Province, and Gwangju. Higher income levels were associated with an increased likelihood of medical travel, whereas patients with liver cirrhosis or higher CCI were significantly less likely to travel for treatment.

Table 2 Factors affecting initial treatment at hospitals outside of residential regions.

Factors	aOR	95%CI
		Low	High
Age in years
00-39	Reference
40-49	0.733	0.661	0.812
50-59	0.649	0.589	0.715
60-69	0.598	0.542	0.659
70-79	0.484	0.438	0.535
≥ 80	0.343	0.304	0.387
Sex
Male	Reference
Female	1.008	0.971	1.047
Residential regions
Seoul	Reference
Busan	15.926	14.189	17.875
Daegu	8.861	7.779	10.093
Incheon	18.344	16.236	20.725
Gwangju	77.516	65.581	91.622
Daejeon	18.307	15.737	21.295
Ulsan	24.560	20.935	28.812
Chungcheong	102.079	90.350	115.330
Gyeonggi	29.790	26.829	33.078
Gangwon	49.107	43.005	56.074
Jeolla	46.179	41.247	51.700
Gyeongsang	110.994	99.196	124.194
Jeju	40.759	34.865	47.650
Year of diagnosis
2013	Reference
2014	1.038	0.973	1.107
2015	1.036	0.972	1.105
2016	0.993	0.931	1.059
2017	0.981	0.920	1.046
2018	0.996	0.934	1.063
2019	0.949	0.889	1.012
2020	0.934	0.875	0.997
2021	0.935	0.874	1.000
CCI
1-2	Reference
3-4	0.936	0.888	0.986
≥ 5	0.819	0.778	0.862
Income level, quintiles
Medicare	Reference
1 (low income)	1.693	1.571	1.825
2	1.66	1.537	1.793
3	1.762	1.635	1.899
4	2.007	1.864	2.161
5 (high income)	2.329	2.166	2.504
Liver cirrhosis
Yes	Reference
No	1.099	1.055	1.145
Extrahepatic metastasis
Yes	Reference
No	1.011	0.964	1.062

aOR: Adjusted odds ratio; CCI: Charlson comorbidity index; CI: Confidence interval.

DISCUSSION

HCC remains the second-leading cause of cancer-related deaths in South Korea[1]. It is the primary cause of death among people in their 40s and 50s, highlighting the significant public health implications of this disease[13]. Although the clinical outcomes of HCC treatment have been studied extensively, there is a paucity of research on the determinants and implications of patients’ choices for treatment location. These types of investigations have been conducted for other cancer types[7,14], but this study represented the first focused analysis of domestic medical travel among patients with HCC in South Korea.

The most striking finding was the pronounced concentration of patients traveling to Seoul for medical treatment. A substantial 52.4% of patients with HCC sought treatment in Seoul, and this figure increased to 67.4% when considering the entire Seoul metropolitan area (which included the surrounding Gyeonggi Province). This geographic pattern remained relatively consistent from 2013 to 2021. Of note, however, the Cochran-Armitage trend test revealed a significant upward trend in medical travel over the years, indicating that the centralization of care in Seoul has become even more pronounced over time. A similar trend has been observed for other cancer types. One study analyzing the locations for diagnosis and treatment of patients with prostate cancer in South Korea revealed that 31% of patients were diagnosed in Seoul and 55% of patients received treatment there[14]. This disparity could be attributed to the number of hospitals and the population size of Seoul. Although only 15% of patients with HCC were initially diagnosed in Seoul, there are factors beyond hospital accessibility and population density that are influencing this geographic concentration of treatment.

A multitude of factors influence a patient’s choice of healthcare facility. Logistic regression analysis revealed that age, residential area, year of diagnosis, comorbidity burden, socioeconomic status, and liver cirrhosis significantly influenced the decision to seek treatment at a hospital in a different region. Other studies[9,15,16] have also found associations between younger age, higher socioeconomic status, fewer comorbidities, and higher education levels with a greater likelihood of traveling farther for medical care. The clinical complexity and relatively early age of onset of HCC[13] make access to specialized, continuous care even more critical[4].

Previous studies have consistently highlighted distance as a significant factor influencing medical travel[9,10,15]. For radiotherapy, which requires daily hospital visits, patients are less likely to choose distant facilities[10]. However, domestic medical travel in South Korea exhibited distinct patterns compared with other countries. A study on patients with prostate cancer in South Korea reported that distance was not a significant factor impacting medical travel to Seoul. In fact, the highest proportion of patients traveling to Seoul for surgery originated from Jeju Island, which is the most distant province from the capital[8]. Our study found that the highest aOR for medical travel was observed among patients residing in Gyeongsang Province, a region relatively far from Seoul.

These findings suggest that the Korean transportation infrastructure, which includes high-speed rail and domestic air travel, may mitigate geographic barriers and support greater patient mobility[17]. In addition, the tendency to seek treatment in Seoul reflects the perceptions of more medical expertise, access to advanced technology, and better treatment outcomes that are associated with high-volume academic hospitals[18]. The socialized health insurance policies in South Korea allow full coverage for patients regardless of their residential region and treatment region. This has likely reduced financial deterrents and facilitated medical centralization.

The concentration of patients seeking medical treatment in Seoul is also closely associated with structural and institutional dynamics of the healthcare system. It reflects systemic imbalances in healthcare infrastructure and workforce distribution. The majority of nationally designated cancer centers and liver disease specialists are located in Seoul[18]. Recent research highlighted that regional disparities in the availability of cancer clinical trials persist in South Korea with a vast majority being concentrated in the Seoul metropolitan area. This disparity limits access to innovative treatments for patients in regions outside the capital and reinforces the structural centralization of cancer care delivery[19]. Because HCC therapies rapidly evolve (e.g., the recent adoption of atezolizumab plus bevacizumab as a first-line systemic treatment), access to clinical trials is becoming increasingly critical for ensuring timely and equitable care.

The volume-outcome relationship suggests that high-volume centers have better outcomes, supporting centralization[20,21]. However, benefits may plateau or reverse beyond a certain threshold[22]. Nallamothu et al[23] found that high-volume hospitals generally had better outcomes for percutaneous coronary interventions. However, this advantage diminished after a certain volume threshold. Other studies argue that hospital resources, surgeon expertise, and care coordination are equally or more important than volume alone[24,25]. High-volume hospitals may face overcrowding, long wait times, and provider burnout, compromising the quality of care[26]. Moreover, centralization may weaken the role and viability of local hospitals, particularly in rural or underserved regions, exacerbating disparities and marginalizing healthcare access[18].

Our analysis of travel for medical treatment revealed an interesting pattern. While there was a significant outflow of patients from major metropolitan cities such as Busan and Daegu to Seoul, there was also a substantial inflow into these cities from surrounding regions. Gwangju exhibited the highest proportion of patients from other regions (64.0%), followed by Daegu (50.3%), Daejeon (49.1%), and Busan (33.4%). The primary inflow for Gwangju was from the Jeolla Province, while Daegu and Busan received most of their patients from the Gyeongsang Province and Daejeon from the Chungcheong Province. Regional hubs may serve as secondary centers of excellence and alleviate patient pressure on Seoul-based hospitals[7]. Policy initiatives including establishing regional liver cancer centers, implementing telemedicine-based tumor boards, and providing travel subsidies for underserved patients could serve as effective decentralization strategies[16].

This study has implications beyond South Korea. Other countries with centralized healthcare systems face similar challenges in balancing efficiency and equity[27-29]. Our Korean experience highlighted the need for proactive policy interventions that enhance regional capacity and promote spatial equity in healthcare delivery. Countries with universal health coverage may experience comparable trends of patients concentrating their medical care in capital or metropolitan areas[30]. The trade-offs between the pursuit of quality through centralization and the need to ensure geographically equitable access must be navigated by countries experiencing this trend[31]. Even with universal coverage and advanced transportation networks, regional disparities in access and care quality may persist, underscoring the need for targeted policy support[32]. The phenomenon of patients traveling for medical treatment remains prevalent and is influenced by a complex interplay of demographic, structural, and institutional factors. Policymakers must weigh the benefits of high-volume care against the potential to exacerbate regional disparities[22]. Further research should be carried out to develop a more sophisticated approach to guarantee high-quality care in all healthcare environments.

Several limitations of this study should be acknowledged. First, the health insurance data used in this study did not include detailed information such as patient stage or Child-Pugh score. This limited our ability to assess the impact of patient condition and severity on medical travel. Although liver cirrhosis was included as a covariate in the logistic regression analysis, it may not fully capture the heterogeneity in patient severity. With more granular data on patient conditions, further analysis on medical travel would have been possible. Second, the CCI was calculated using ICD-10 codes that included liver cancer, resulting in a minimum score of 2 for all patients. This approach will affect comparability with other studies that assess comorbidity independently of the index cancer. Third, this study used claims data limited to reimbursed care. This method excludes privately paid treatments or procedures performed outside the NHIS framework. Fourth, the dataset did not distinguish between a curative intent and a palliative intent of treatment. We hypothesize that this parameter would influence the choice of treatment location. Fifth, the lack of address-level data restricted our ability to assess precise spatial patterns, including urban–rural disparities. Sixth, model diagnostics such as the area under the curve or the Hosmer-Lemeshow test could not be performed due to analysis constraints. While some regional variables showed high aORs, we believe this reflects the strong centralization of care in Seoul rather than model overfitting.

CONCLUSION

This nationwide retrospective cohort study revealed a striking centralization of HCC treatment in the capital, Seoul where over half of all patients received care regardless of where they were diagnosed. Medical travel was more common among younger, wealthier patients, reflecting persistent regional disparities in healthcare access. To address this disparity, policymakers should implement decentralization strategies including strengthening regional cancer centers, redistributing healthcare resources, and improving care coordination across regions. Our observations offer valuable insights for South Korea as well as other countries with centralized healthcare systems where similar equity-efficiency tradeoffs may arise. Ensuring regionally balanced access to high-quality cancer care remains essential for sustainable and equitable health system performance.