Evidence of Overlapping Roles Between Clinics and Hospitals in Primary Care

Abstract

Background

Clinics and hospitals in South Korea are often perceived as competitors to each other. This study examines the overlapping roles in providing primary care provision between clinics and hospitals by analyzing the healthcare facility type where patients first receive diagnoses of hypertension (HTN) or diabetes mellitus (DM). We also explore the characteristics of patients that influence their choice of healthcare facility and compare healthcare utilization patterns in the first year post-diagnosis by facility type.

Methods

Using 2021 claims data from the Health Insurance Review and Assessment Service, we included data from 599,955 patients with newly diagnosed HTN and 195,668 patients with newly diagnosed DM. We analyzed the distribution of new diagnoses by facility type, characteristics of patients who chose hospitals, average number of facility visits and medical expenses in the first year following the initial diagnosis, and continuity of care index.

Results

Among the newly diagnosed patients, 82.5% of HTN and 66.6% of DM cases were diagnosed in clinics, whereas 17.5% and 33.4%, respectively, occurred in hospitals. Younger patients, those with comorbidities, and those residing outside Seoul were more likely to receive care at hospitals. Compared to those diagnosed in hospitals, patients diagnosed in clinics visited healthcare facilities more frequently and incurred greater total medical expenses but demonstrated higher continuity of care.

Conclusion

These findings indicate that both clinics and hospital-level healthcare facilities, excluding general hospitals, play significant roles in providing primary care. Strengthening the clinic-based primary care system and ensuring the quality of primary care provided by hospitals are crucial. Policies that promote transparency in the quality assessments of healthcare facilities can help patients make informed decisions when choosing between clinics and hospitals.

Graphical Abstract

INTRODUCTION

South Korea’s healthcare system is renowned for its high quality, particularly in acute care, which has shown excellent outcomes. Notably, the 30-day mortality rates for hemorrhagic and ischemic stroke in Korea are the lowest among Organisation for Economic Co-operation and Development (OECD) countries.1 However, its primary care system is considered one of the weakest among developed countries.2 Despite recommendations by the OECD in 2012 to strengthen the primary care infrastructure, substantial progress remains elusive.3,4

A major challenge in South Korea’s primary care system is the absence of clear functional distinctions between different types of healthcare facilities, along with a lack of a gatekeeping mechanism, which is essential for effective primary care.5 Although clinics are typically seen as the primary providers of primary care, hospitals also maintain large outpatient units that serve similar functions.6,7 Moreover, patients can freely choose their healthcare facility based on personal preference, irrespective of the facility level.8 This flexibility fosters a competitive environment between clinics and hospitals,5,8 potentially compromising care quality and leading to inefficient resource use. Addressing these issues is vital for improving the healthcare system.9

However, few studies have actually measured the level of competition between clinics and hospitals. We hypothesized that the level of competition between clinics and hospitals could be measured by identifying which healthcare facilities were used by those who were first diagnosed with hypertension (HTN) and diabetes mellitus (DM)—the most representative primary care diseases. Briefly, we used the distribution ratio of “first-contact” point by healthcare facilities, a fundamental element of primary care, as proxy indicator representing the level of competition between clinics and hospitals. This study aims to: 1) Identify the types of healthcare facilities where patients with newly diagnosed HTN and DM first receive their diagnoses; 2) Examine patient characteristics that influence choice of healthcare facility for initial diagnosis; and 3) Compare healthcare utilization patterns following initial diagnosis based on the chosen type of healthcare facility.

METHODS

Data source and study population

This study utilized Korea’s National Health Insurance and Medical Aid claims data, which included patients aged ≥ 19 years who were newly diagnosed with HTN or DM in 2021. The study population was identified using the International Classification of Diseases-10 codes and prescription records. To identify eligible claims, we initially extracted outpatient visits in 2021 with a primary diagnosis of HTN (I10–I15) or DM (E10–E14), and further refined patients based on their prescription records. For HTN, inclusion required claims with prescriptions for antihypertensive medications (Anatomical Therapeutic Chemical [ATC] codes: C02–C03, C07–C09). For DM, inclusion criteria were met if the prescription contained oral hypoglycemic agents or insulin (ATC codes: A10A, A10B). By aggregating claims at the patient level, we identified 6,606,593 individuals with HTN and 3,178,670 with DM. To identify newly diagnosed patients, we implemented a 3-year washout period by reviewing the diagnostic records. Patients who had been diagnosed with HTN or DM within the 3 years prior to their first outpatient visit in 2021 were excluded. Furthermore, given that chronic diseases predominantly affect adults, patients under the age of 19 were excluded from this study. The final cohort included 599,955 patients with newly diagnosed HTN and 195,668 with newly diagnosed DM (Fig. 1).

Fig. 1. Data analysis flowchart.

HTN = hypertension, DM = diabetes mellitus.

Measures and analysis

Number of newly diagnosed patients by healthcare facility type

Newly diagnosed patients were those who received their initial diagnosis of chronic conditions (HTN or DM) and were prescribed medications during an outpatient visit within the study period. In 2021, we analyzed the proportion of newly diagnosed patients by facility type and calculated the number of new patients per physician (internist, family medicine specialist) based on facility capacity and provider availability. Healthcare facilities were broadly categorized into clinics and hospitals. Hospitals were further classified into small, general, and tertiary hospitals. To prevent confusion between the term “hospital” as an umbrella category and small-scale hospitals with 30–100 beds, we specifically adopted the term “small hospital” to refer to the latter.

Factors influencing healthcare facility choice for initial diagnosis

Logistic regression analysis was conducted to identify patient characteristics associated with the choice of healthcare facility. Facility choice was categorized into clinics (0) vs. hospitals (small, general, and tertiary hospitals) (1). The analysis included variables such as sex, age, area of residence, income level, and comorbidity status, which were hypothesized to influence the choice of healthcare facility. Due to limitations in claims data, the area of residence was determined based on the location of the healthcare facility visited. Income level was divided into two groups: National Health Insurance (NHI) enrollees and Medical Aid recipients (low-income group). Areas of residence were classified into Seoul, metropolitan, cities, and counties. Comorbidities were assessed using the Charlson Comorbidity Index (CCI), which assigns scores based on the relative risk of 17 disease groups.10 Patients were categorized into three groups: no comorbidities (score of 0), a score of 1, and score of ≥ 2.

Healthcare utilization by healthcare facility type

We tracked outpatient visits for 1 year following initial diagnosis of HTN or DM to evaluate healthcare costs and continuity of care (COC). The healthcare costs included consultation fees, diagnostic tests, and medications. The analysis compared average visits per patient, total expenses, and cost per visit across facility types.

COC is a key indicator of primary care quality that reflects the ongoing relationship between patients and providers.11,12,13 In this study, we assessed COC using the COC index,14 which measures both the concentration of visits (a higher proportion of visits to a specific provider indicates higher COC) and dispersion of visits (visiting a greater number of providers indicates lower COC). The index ranges from 0 (all outpatient visits were made to different providers) to 1 (all visits were made to the same provider). Previous research has identified COC scores of ≥ 0.75 as “high continuity” and scores below 0.75 as “low continuity.”15,16 We analyzed average COC scores for patients with HTN and DM by healthcare facility type. Additionally, χ² tests were used to compare the proportions of patients with high and low continuity across different facility types. To ensure reliability, the COC index was calculated only for patients who had at least four outpatient visits within the first year following their initial diagnosis.17,18

All analyses were conducted using SAS Enterprise Guide version 7.1 (SAS Institute Inc., Cary, NC, USA), with P values below 0.05 being deemed statistically significant.

Ethics statement

This study was reviewed and approved by the Institutional Review Board (IRB) of the Health Insurance Review and Assessment Service, under IRB No. 2023-073-001. The IRB waived the requirement for written informed consent since all datasets used were anonymous and had been de-identified.

RESULTS

Participant characteristics

As of 2021, there were 599,955 newly diagnosed patients with HTN and 195,668 with DM as shown in Table 1. Among these patients, males comprised 55.7% of those with HTN and 61.0% of those with DM, exceeding the number of females in both categories. The largest number of new diagnoses for both conditions occurred in the 50–59 year age group. In terms of healthcare facility locations, about 50% of the diagnoses for both conditions took place in cities, followed by metropolitan, Seoul, and counties. The majority of participants were covered by NHI, with only 2.2–3.8% being covered Medical Aid. Regarding comorbidities, as classified by the CCI, 94% of patients newly diagnosed with HTN had no comorbidities, whereas 73.0% of those with DM had one or more comorbidities.

Table 1. Baseline characteristics.

Categories		Hypertension	Diabetes mellitus
Total		599,955 (100.0)	195,668 (100.0)
Sex
	Male	334,161 (55.7)	119,401 (61.0)
	Female	265,794 (44.3)	76,267 (39.0)
Age, yr
	19–39	66,233 (11.0)	25,963 (13.3)
	40–49	130,050 (21.7)	40,498 (20.7)
	50–59	178,907 (29.8)	57,481 (29.4)
	60–69	143,013 (23.8)	46,017 (23.5)
	≥ 70	81,752 (13.6)	25,709 (13.1)
Location of healthcare facility visiteda
	Seoul	109,292 (18.2)	36,775 (18.8)
	Metropolitan	137,541 (22.9)	48,207 (24.6)
	City	305,571 (50.9)	97,996 (50.1)
	County	47,551 (7.9)	12,690 (6.5)
Type of health insurance
	NHI	582,721 (97.1)	188,146 (96.2)
	Medical Aid	17,234 (2.9)	7,522 (3.8)
CCI
	0	565,296 (94.2)	52,897 (27.0)
	1	10,089 (1.7)	55,018 (28.1)
	≥ 2	24,570 (4.1)	87,753 (44.9)

All values are presented as number (%).

NHI = National Health Insurance, CCI = Charlson Comorbidity Index.

^aLocation of healthcare facility visited by the patient.

Proportion of newly diagnosed HTN and DM by healthcare facility type

In 2021, 82.5% (494,911 patients) of newly diagnosed HTN cases were seen at clinics, whereas 17.5% (105,044 patients) received their diagnosis at hospitals. For DM, 66.6% (130,311 patients) were diagnosed at clinics, and 33.4% (65,357 patients) at hospitals (Table 2). Among patients with HTN, the highest proportion of cases was diagnosed at clinics, followed by small, general, and tertiary hospitals. In contrast, for DM, compared to small hospital, general hospitals accounted for a larger number of new diagnoses, and the proportion of patients diagnosed at tertiary hospitals was notably higher.

Table 2. Newly diagnosed hypertension and diabetes mellitus by healthcare facility type.

Categories		Total	Hospitals-subtotald	Tertiary hospital	General hospital	Small hospital	Clinic
Healthcare resourcesa
	No. of healthcare facilitiesb	35,673 (100.0)	1,761 (4.9)	45 (0.1)	319 (0.89)	1,397 (3.9)	33,912 (95.1)
	No. of internists/family medicine specialist available	84,128 (100.0)	41,344 (49.1)	13,714 (16.3)	17,407 (20.7)	10,223 (12.2)	42,784 (50.9)
Hypertension
	No. of newly diagnosed patients	599,955 (100.0)	105,044 (17.5)	6,168 (1.0)	45,638 (7.6)	53,238 (8.9)	494,911 (82.5)
	No. of newly diagnosed patients per physicianc	7.13	2.54	1.62	9.8	30.27	40.77
Diabetes mellitus
	No. of newly diagnosed patients	195,668 (100.0)	65,357 (33.4)	11,684 (6.0)	33,398 (17.1)	20,275 (10.4)	130,311 (66.6)
	No. of newly diagnosed patients per physicianc	2.33	1.58	3.06	7.17	11.53	10.73

Values which are including number of healthcare facility and medicine specialist available, newly diagnosed patients are presented as number (%).

^aData adapted from Annual Statistical Report 2021.19

^bExcluding long-term care hospitals, dental hospitals/clinics, Korean medicine hospitals/clinics.

^cNumber of newly diagnosed patients per internist/family medicine specialist.

^dIncluding tertiary hospital, general hospital, small hospital.

When analyzed according to the number of physicians (specialists in internal medicine and family medicine), clinics recorded an average of 40.8 newly diagnosed HTN patients per physician, whereas in small hospitals it was 30.3, which was considerable higher than the 9.8 at general hospitals and 1.6 at tertiary hospitals. Similarly, for DM, both clinics and small hospitals had higher numbers of newly diagnosed patients per physician. However, the number was slightly greater in small hospitals, with 11.5 patients per physician, whereas in clinics it was 10.7.

Factors influencing healthcare facility choice among patients with newly diagnosed HTN and DM

Table 3 presents the results from logistic regression analyses aimed at identifying factors influencing healthcare facility choice among patients newly diagnosed with HTN and DM. For HTN, female patients were more likely than males to be diagnosed at hospitals, whereas for DM, the odds were lower for females compared to males. In HTN cases, younger patients tended to prefer hospitals, with the likelihood decreasing as age increased. For DM, compared to those in their 70s, patients under 30 were more likely to choose hospitals, though no significant differences were observed among other age groups. Geographically, the likelihood of choosing a hospital was highest in the metropolitan areas for patients with HTN, followed by city, county, and Seoul. For patients with DM, the odds were also higher in metropolitan areas compared to Seoul, but lower in counties than in Seoul. Regarding income level, NHI enrollees had a higher likelihood of selecting hospitals compared to Medical Aid recipients for both conditions, although these differences were not significant. The presence of comorbidities generally increased the likelihood of selecting hospitals for both conditions. Specifically, patients with HTN and single low-severity comorbidity (CCI score of 1) showed higher odds than those with multiple comorbidities (CCI score ≥ 2). Conversely, for DM, patients with multiple comorbidities (CCI score ≥ 2) were more likely to choose hospitals compared to those with a single comorbidity.

Table 3. Logistic regression analyses of the choices of health providers made by patients.

Variables		Hypertension		Diabetes mellitus
		OR (95% CI)	P value	OR (95% CI)	P value
Sex
	Male	Ref.	-	Ref.	-
	Female	1.05 (1.03–1.06)	< 0.001	0.89 (0.87–0.90)	< 0.001
Age, yr
	19–39	1.62 (1.58–1.66)	< 0.001	1.36 (1.31–1.41)	< 0.001
	40–49	1.25 (1.22–1.28)	< 0.001	1.06 (1.02–1.09)	0.002
	50–59	1.15 (1.12–1.17)	< 0.001	1.05 (1.01–1.08)	0.007
	60–69	1.06 (1.04–1.09)	< 0.001	0.97 (0.94–1.00)	0.076
	≥ 70	Ref.	-	Ref.	-
Location of healthcare facility visiteda
	Seoul	Ref.	-	Ref.	-
	Metropolitan	1.51 (1.47–1.54)	< 0.001	1.14 (1.11–1.18)	< 0.001
	City	1.43 (1.40–1.46)	< 0.001	0.97 (0.95–1.00)	0.021
	County	1.30 (1.26–1.34)	< 0.001	0.63 (0.60–0.66)	< 0.001
Type of health insurance
	NHI	0.99 (0.95–1.03)	0.758	1.07 (1.02–1.12)	0.010
	Medical Aid	Ref.	-	Ref.	-
CCI
	0	Ref.	-	Ref.	-
	1	1.20 (1.14–1.26)	< 0.001	1.13 (1.10–1.16)	< 0.001
	≥ 2	1.13 (1.10–1.17)	< 0.001	1.30 (1.27–1.33)	< 0.001

Dependent variable: healthcare facility choice (0 = clinic, 1 = hospital). All variables in the table were simultaneously adjusted in the logistic regression model.

OR = odd ratio, CI = confidence interval, NHI = National Health Insurance, CCI = Charlson Comorbidity Index.

^aLocation of healthcare facility visited by the patient.

Healthcare utilization in the first year following initial diagnosis by healthcare facility type

In 2021, patients newly diagnosed with HTN and DM visited healthcare facilities an average of 7.6 and 8.2 times, respectively, in the first year after their diagnosis (Supplementary Table 1, Fig. 2). For both conditions, those diagnosed at clinics visited more frequently compared to those diagnosed at hospitals (HTN: 7.9 visits at clinics; DM: 9.0 visits at clinics).

Fig. 2. Number of outpatient visits and healthcare expenditure in the first year following initial diagnosis by healthcare facility type. (A) Average number of visits by healthcare facility type (HTN). (B) Average number of visits by healthcare facility type (DM). (C) Total medical expense per patient by healthcare facility type (HTN). (D) Total medical expense per patient by healthcare facility type (DM). (E) Total medical expense per visit by healthcare facility type (HTN). (F) Total medical expense per visit by healthcare facility type (DM).

HTN = hypertension, DM = diabetes mellitus.

The average total healthcare expenditure per patient over the year was approximately 190,000 KRW for HTN and 280,000 KRW for DM (Supplementary Table 1, Fig. 2). For both conditions, total expenses per patient were the highest at tertiary hospitals, followed by general hospitals, clinics, and small hospitals. Conversely, costs per visit increased with the level of healthcare facilities, rising from clinic to tertiary hospital for both conditions.

COC for newly diagnosed patients by healthcare facility type

The average COC index for patients newly diagnosed with HTN and DM in the first year after diagnosis was 0.87 and 0.89, respectively (Table 4). For HTN, the COC index was highest for patients diagnosed at clinics, followed by general hospitals, small hospitals, and tertiary hospitals. Moreover, the proportion of patients achieving high continuity (COC index ≥ 0.75) was highest at clinics (78.6%). For DM, the highest COC index was recorded for patients diagnosed at tertiary hospitals, followed by clinics, general hospitals, and small hospitals. Nevertheless, clinics maintained the largest proportion of patients with high continuity, at 80.6%.

Table 4. Differences in COC among patients with newly diagnosed hypertension and diabetes mellitus across healthcare facility types.

Healthcare facilitya		No. of newly diagnosed patientsb	Mean COC indexc	Level of COCd
				High	Low	P value
Hypertension						< 0.001
	Total	439,896	0.87	339,010 (77.07)	100,886 (22.93)
	Tertiary hospital	4,175	0.80	2,625 (62.87)	1,550 (37.13)
	General hospital	32,128	0.84	22,595 (70.33)	9,533 (29.67)
	Small hospital	34,705	0.83	23,725 (68.36)	10,980 (31.64)
	Clinic	368,888	0.88	290,065 (78.63)	78,823 (21.37)
Diabetes mellitus						< 0.001
	Total	151,487	0.89	120,114 (79.29)	31,373 (20.71)
	Tertiary hospital	9,036	0.90	7,274 (80.50)	1,762 (19.50)
	General hospital	25,511	0.88	19,904 (78.02)	5,607 (21.98)
	Small hospital	14,317	0.85	10,247 (71.57)	4,070 (28.43)
	Clinic	102,623	0.89	82,689 (80.58)	19,934 (19.42)

COC = continuity of care.

^aCategorized based on the level of the healthcare facility where the initial diagnosis was made.

^bOnly included patients who had four or more outpatient visits within the first year following their initial diagnosis.

^cCOC index; ranges from 0–1.

^dCOC index of ≥ 0.75 was categorized as “high” and index below 0.75 categorized as “low.”

DISCUSSION

This study is the first to explore the overlapping roles of different healthcare facilities in providing primary care, focusing on their function as the first contact point for patients diagnosed with HTN and DM. We analyzed where these diagnoses were made to assess the overlap between clinics and hospitals.

Among the newly diagnosed patients, 82.5% of HTN and 66.6% of DM cases were diagnosed in clinics, whereas 17.5% and 33.4%, respectively, occurred in hospitals. Younger patients, those with comorbidities, and those residing outside Seoul were more likely to receive care at hospitals. Compared to those diagnosed in hospitals, patients diagnosed in clinics visited healthcare facilities more frequently and incurred greater total medical expenses but demonstrated higher COC.

In other words, hospitals were the first contact point for 18% of newly diagnosed HTN patients and one-third of newly diagnosed patients with DM. This indicates that hospitals play the role of primary care up to 18–34%. These figures are notably uncommon in countries where primary care physicians typically function as gatekeepers.20,21 The competition between clinics and hospitals was particularly evident in the comparison of the number of new diagnoses per physician. For DM, small hospitals recorded a higher number of new diagnoses per physician than clinics, suggesting an important role in primary care delivery. This is somewhat disappointing considering government policies that encourage clinics to manage milder conditions within communities.22 The overlap arises from the lack of clearly defined roles among different healthcare facilities, enabling patients to freely select their preferred providers. In Korea, hospital-based care is often perceived as superior to that offered by clinics, potentially leading to the excessive utilization of higher-level facilities.5,7 To address this issue, policymakers promoting a robust primary care system may consider implementing targeted policies to redirect patient flow from hospitals back to clinics.6

However, the overlap in roles between clinics and hospitals should not solely be viewed as a failure of the healthcare delivery system but as an inherent feature of Korea’s unique healthcare structure. Often, small hospitals in Korea have grown from clinics that expanded after accumulating sufficient capital, positioning them in a transitional space between clinics and hospitals in terms of size and function.23 Our findings further underscore that clinics primarily compete with small hospitals, which represent 94% of all hospitals (41,344 facilities, including 10,223 small hospitals), rather than with general or tertiary hospitals.

Patient characteristics play a significant role in influencing the choice of hospitals. In this study, patients with comorbidities tend to prefer hospital care, possibly stemming from diagnoses made during treatment for other conditions. Thus, selecting a hospital that offers multiple specialties may be a logical choice for these patients. Similarly, the availability of diagnostic equipment, specialized personnel, and infrastructure in hospitals could be a critical factor influencing this preference. In this study, the COC index of patients with DM initially diagnosed at hospitals was higher than those diagnosed at clinics. This tendency differs from that observed with HTN. This may be attributed to the complex nature of diabetes care, which necessitate regular and specialized examinations for eye or kidney complication. These demands make hospitals a more advantageous choice for DM management.

Furthermore, younger patients, particularly those under 30, are more likely to select hospitals compared to those aged ≥ 70. HTN and diabetes are not conditions that typically occur as primary disease in individuals in their 20s and 30s but are often secondary to underlying health issues. Consequently, young adults may visit hospitals for additional examinations related to these conditions. Younger adults also tend to exhibit a delayed recognition of chronic conditions compared to older adults,24,25 often detecting the condition only when severe symptoms emerge or through routine check-ups. In this context, many individuals under 30 are likely diagnosed with HTN or DM through Korea’s mandatory General Health Screening Program. Additionally, many companies collaborate with hospitals to provide on-site health check-ups for their employees, which may lead to higher hospital utilization among this age group. Therefore, initial diagnoses made in hospitals may be justified based on regional characteristics or individual patient circumstances, and not necessarily indicative of inappropriate healthcare utilization.

We also monitored healthcare utilization during the first year following diagnosis. Clinics followed up with patients an average of 7.9–9.0 times per year (approximately every 6–7 weeks), which is higher whereas in hospitals patients were followed up an average of 5.8–6.7 times per year (approximately every 8–9 weeks). Consequently, total healthcare expenditure was highest at clinics, despite the lower cost per visit. Determining whether clinics or hospitals provide a more appropriate approach remains challenging, as clear guidelines for optimal follow-up intervals are lacking. However, Korea’s National Quality Assessment Program recommends that medical institutions follow up with patients with HTN and DM at least once every 3 months.26 The visit frequencies observed in both clinics and hospitals in our study align with these recommendations. However, frequent visits may not equate to effective disease management. A prior study found that only 13.3–28.4% of patients with DM receiving care at clinics underwent screening for complications,27 suggesting that the lower cost per visit at clinics might be due to fewer diagnostic tests being performed. Thus, further studies are required to elucidate the relationships among patient visit frequency, healthcare expenditure, and quality of care provided. Nevertheless, patients diagnosed at clinics show higher COC than those at other facilities, indicating better ongoing care from the same provider and improved health outcomes.13,28,29 Previous studies have shown that patients with HTN or DM who primarily use clinics or receive their first prescriptions at clinics exhibit better medication adherence.30,31 These findings underscore the importance of continuing to strengthen the role of clinics in primary care while taking steps to improve the quality of care they provide.

This study has several limitations. First, healthcare facilities were classified based on the initial diagnosis location. Therefore, the number of visits and average healthcare expenditure could vary if patients attended multiple facility types. However, according to the report of National Quality Assessment Program, over 80% of patients primarily use a single facility, suggesting that multi-facility utilization has minimal impact.26 Second, this study calculated the COC index for patients with at least four outpatient visits within 1 year of their initial diagnosis to ensure robustness and validity, consistent with prior research.17,18 However, this approach excluded patients with fewer visits, potentially limiting the representation of those with low treatment compliance. Therefore, the findings should be interpreted with caution when generalizing to populations with irregular or infrequent healthcare engagement. Third, while we tracked healthcare utilization for 1 year following the initial diagnosis, we did not examine the specific quality of services provided. Regular screenings and consistent medication adherence are crucial for managing HTN and DM.32,33 Future research should explore service quality and its effect on health outcomes in newly diagnosed patients. Despite these limitations, this study provides valuable insights into the role overlap between clinics and hospitals in primary care and offers a detailed analysis of healthcare utilization patterns based on facility choice.

This study confirmed that both clinics and small hospitals play a significant role in providing primary care in Korea. In this context, it is desirable to redirect patients who use hospitals to clinics, aligning with the government’s policy objectives; however, there are numerous challenges in implementing this approach immediately. Therefore, it is imperative to reassess the policy direction to enhance the overall quality of primary care. This can be achieved by reinforcing the clinic-centered primary care system and managing quality of services provided by hospitals.

This approach emphasizes the need for quality control in primary care provided by small hospitals. One possible strategy is to expand the scope of public reporting under National Quality Assessment Program, which currently discloses the results of clinics only, to include small hospitals. This could help patients make informed decisions between clinics and hospitals based on reliable information. Ultimately, this approach can contribute to shaping primary care policies that better address how patients select their first-contact point.

SUPPLEMENTARY MATERIAL

Supplementary Table 1

Number of outpatient visits and healthcare expenditure in the first year following initial diagnosis by healthcare facility type