ENM Endocrinology and MetabolismEndocrinol Metab (Seoul) 2093-596X 2093-5978 Korean Endocrine Society 10.3803/EnM.2026.2953 enm-2026-2953 Review Article Miscellaneous Nationwide Big Data Studies of Endocrine Diseases Using the Korean National Health Information Database: Research Trends and Standardization of Operational Definitions http://orcid.org/0000-0002-7394-3830 ChoSun Wook 1 * http://orcid.org/0000-0003-1932-0234 KimJung Hee 1 * http://orcid.org/0000-0001-7925-2515 KimKyoung Jin 2 * KimBeom-Jun 3 http://orcid.org/0000-0003-3205-9114 KimMee Kyoung 4 http://orcid.org/0000-0002-6108-7758 RheeEun Jung 5 Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea Division of Endocrinology and Metabolism, Department of Internal Medicine, Korea University Anam Hospital, Seoul, Korea Division of Endocrinology and Metabolism, Department of Internal Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea Division of Endocrinology and Metabolism, Department of Internal Medicine, Eunpyeong St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Korea Department of Endocrinology and Metabolism, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Korea Corresponding authors: Mee Kyoung Kim. Division of Endocrinology and Metabolism, Department of Internal Medicine, Eunpyeong St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, 1021 Tongil-ro, Eunpyeong-gu, Seoul 03312, Korea Tel: +82-2-2030-4435, E-mail: makung@catholic.ac.kr Eun Jung Rhee. Department of Endocrinology and Metabolism, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, 29 Saemunan-ro, Jongno-gu, Seoul 03181, Korea Tel: +82-2-2001-2485, Fax: +82-2-2001-1588, E-mail: hongsiri@hanmail.net

These authors contributed equally to this work.

 2 2026 26 2 2026 41 1 86 104 2 02 2026 4 02 2026 5 02 2026 Copyright © 2026 Korean Endocrine Society 2026 This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

The Korean National Health Information Database (NHID) is a large-scale dataset created through linkage of National Health Insurance claims with nationwide health screening records. Because it is released in a cohort-based format, the NHID enables longitudinal follow-up and allows investigation of rare endocrine conditions with low population prevalence. Mortality data, including both dates and causes of death, are additionally obtained through linkage with Statistics Korea. Over recent years, use of the NHID has expanded rapidly, establishing it as a major resource for epidemiological research in endocrinology. Nevertheless, because the database was originally developed for administrative and screening purposes rather than for research, investigators face several methodological limitations, particularly the need to construct and validate robust operational definitions of diseases. In this review, we describe the key features of the Korean NHID, summarize operational definitions of endocrine disorders that have been applied in prior research, and provide an overview of recent endocrine studies conducted using this database.

 Databases Endocrine system diseases Thyroid diseases Obesity Osteoporosis Diabetes mellitus INTRODUCTION

The Korean National Health Information Database (NHID) is a nationwide dataset constructed by the National Health Insurance Service (NHIS), the single public insurer that covers nearly the entire Korean population. Currently, approximately 97% of the Korean population is enrolled as subscribers in the NHIS [1,2]. The NHID consists of five component databases: an eligibility database, a healthcare utilization claims database, a national health screening database, a long-term care insurance database, and a healthcare provider database, and it can be further linked to national mortality statistics maintained by Statistics Korea. The health screening component includes detailed questionnaires on lifestyle factors, physical measurements, and laboratory test results obtained through periodic screening programs that have been progressively expanded to Korean adults beginning in early adulthood. In addition, the NHID incorporates data from the government-run Rare Incurable Disease program, in which physician-verified diagnoses are required to qualify for substantial copayment reductions, thereby providing a relatively reliable registry for these conditions. Collectively, these features make the NHID a distinctive longitudinal resource for population-based research on both common and rare endocrine and metabolic diseases in Korea.

Because the Korean NHID is generated for administrative reimbursement and billing purposes rather than for research, this design inherently introduces important constraints. As a result, diagnoses recorded in claims data may not always accurately reflect patients’ true clinical conditions. To mitigate potential misclassification, researchers must develop carefully designed operational definitions and rigorously evaluate their validity. In this review, we present commonly used operational definitions for endocrine-related diseases, including diabetes mellitus and thyroid disorders, and summarize recent research conducted using the Korean NHID.

 DIABETES MELLITUS Operational definitions of diabetes mellitus

Type 2 diabetes mellitus (T2DM) is commonly identified using NHID data based on either a fasting plasma glucose level of ≥126 mg/dL or at least one annual prescription claim for antidiabetic medications accompanied by International Classification of Diseases, 10th revision (ICD-10) codes E11–E14 (Table 1) [2,3]. The validity of this NHID-based definition was evaluated in a previous study by comparing claims-based diagnoses with data from the Korea National Health and Nutrition Examination Survey (KNHANES), which served as an external reference standard [3]. That study demonstrated that the accuracy of identifying T2DM in NHID claims data was substantially improved when prescription information was used in combination with diagnostic codes, rather than relying on diagnostic codes alone [3].

Type 1 diabetes mellitus (T1DM) was operationally defined by identifying patients with persistent insulin dependence among individuals assigned the ICD-10 code E10. To distinguish persistent insulin dependence characteristic of T1DM from transient insulin use in other clinical contexts, an additional insulin prescription occurring between 1 and 2 years after the initial prescription was required [2,4]. In a more recent study, T1DM was operationally defined using claims data as the presence of ICD-10 code E10 together with a documented history of insulin treatment, defined as more than two insulin prescription claims, with at least one claim occurring between 1 and 2 years after the first prescription [4].

Gestational diabetes mellitus (GDM) was operationally defined as cases in which pregnant women had at least two outpatient visits recorded with diagnostic codes for GDM. In Korea, diagnostic codes related to GDM are frequently applied at the time of ordering an oral glucose tolerance test, reflecting reimbursement and administrative practices rather than confirmed diagnoses. Accordingly, GDM was operationally defined as the presence of at least two outpatient visits with recorded GDM diagnostic codes to improve diagnostic specificity [5]. Pre-pregnancy body mass index (BMI), advanced maternal age, elevated fasting glucose levels before pregnancy, smoking, and fatty liver, or elevated γ-glutamyltransferase levels, were significantly associated with an increased risk of GDM, particularly insulin-requiring GDM [6].

Diabetic retinopathy was identified using ICD-10 code H36.0, and proliferative diabetic retinopathy was operationally defined by combining ICD-10 code H36.0 with procedure codes for pan-retinal photocoagulation, yielding a relatively specific definition of advanced diabetic retinopathy [2]. Diabetic foot amputation was operationally defined when procedure codes for footlevel amputation were recorded in individuals with concurrent ICD-10 diagnosis codes indicating diabetes with circulatory or multiple complications (E10.5, E11.5, E12.5, E13.5, E14.5, and E10.7–E14.7) [2,7]. A conservative definition relied on procedure codes N0572–0575, whereas a broader definition incorporating additional foot-level amputation codes (N0562, N0564–0566, and N0571–0575) was evaluated in sensitivity analyses (Table 1).

 Diabetes fact sheet in Korea

Diabetes-related fact sheets for the Korean population are primarily analyzed and reported using data derived from the Korean NHID [8-10]. The diabetes fact sheet in Korea 2021 provides an updated overview of diabetes prevalence, disease management, risk-factor control, and associated comorbidities in Korea. In 2020, diabetes affected 16.7% of adults aged ≥30 years; however, only 9.7% of individuals achieved simultaneous control of glycemia, blood pressure, and lipid levels [8]. Using nationwide data from the NHID and the KNHANES, the same report comprehensively characterized the epidemiology and management status of diabetic kidney disease (DKD) in Korea [9]. DKD was present in approximately one-quarter (25.4%) of adults with diabetes. Despite a gradual increase in sodium-glucose cotransporter 2 (SGLT2) inhibitor use among patients with impaired kidney function, reaching 5.94% in 2019, overall uptake remained low compared with the widespread use of renin–angiotensin system blockers (70%), highlighting substantial gaps in DKD management [9].

Using nationwide data from the NHID and KNHANES, another study examined population-level trends in cardiovascular disease (CVD) among Korean adults with T2DM between 2010 and 2019 [10]. Although the incidence of ischemic heart disease, ischemic stroke, and peripheral artery disease declined during this period, the incidence of heart failure increased over time, with a pronounced rise observed between 2015 and 2016. This increase may reflect population aging, improved survival following major cardiac events, and changes in diagnostic coding practices coinciding with the introduction of SGLT2 inhibitors. Despite these epidemiological trends, the use of evidence-based cardioprotective therapies, including high-intensity statins, SGLT2 inhibitors, and glucagon-like peptide-1 receptor agonists, remained limited, underscoring persistent deficiencies in comprehensive cardiovascular risk management.

Accordingly, SGLT2 inhibitor use remained low in 2019, with prescription rates below 12% across major CVD subgroups.

 Recent research in diabetes mellitus based on the Korean NHID

Several influential diabetes-related studies have been conducted using the Korean NHID. One nationwide investigation utilized NHID data to assess the combined and independent effects of diabetes mellitus and fatty liver disease on cardiovascular outcomes and mortality [11]. In this cohort of 7.8 million individuals, 5-year absolute risks of CVD and all-cause mortality increased in a stepwise manner with greater severity of non-alcoholic fatty liver disease (NAFLD) and were consistently higher among individuals with T2DM. The 5-year absolute risk of CVD ranged from 1.03 in individuals without NAFLD or T2DM to 4.66 in those with both T2DM and grade 2 NAFLD, with a corresponding increase in all-cause mortality from 1.25 to 5.91 [8]. Because even grade 1 NAFLD was associated with a clinically meaningful excess absolute risk among individuals with T2DM, systematic assessment of fatty liver disease should be considered in this population.

Another study investigated how longitudinal changes in income are associated with CVD risk in adults with T2DM [12,13]. Because the NHID does not directly capture household income, monthly health insurance premiums were used as a validated proxy for income status; these premiums are calculated based on wages for employee-insured individuals and on income plus property value for the self-employed and were categorized into quartiles [12,13]. Income fluctuations, income decline, and persistent low-income status over time were each independently associated with a significantly increased risk of incident CVD, even after adjustment for traditional cardiovascular risk factors. These findings indicate that socioeconomic instability represents an important and underrecognized determinant of cardiovascular risk among individuals with T2DM.

Young-onset diabetes, defined as diabetes diagnosed before 30 years of age, is associated with a markedly increased risk of cardiovascular complications, kidney failure, and all-cause mortality compared with the general population [4]. After multivariable adjustment, individuals with young-onset T1DM and T2DM exhibited substantially elevated risks of myocardial infarction, with hazard ratios of 6.76 and 5.07, respectively, as well as increased risks of kidney failure (hazard ratios, 20.92 and 2.78) and mortality (hazard ratios, 3.69 and 3.06) [4]. These findings underscore that early-onset diabetes confers a substantial long-term cardiometabolic burden, emphasizing the need for early identification, intensive risk stratification, and proactive preventive strategies.

 OBESITY Operational definitions of general and abdominal obesity

Obesity and abdominal obesity were defined using BMI and waist circumference (WC), respectively, in accordance with the 2024 Clinical Practice Guidelines for Obesity issued by the Korean Society for the Study of Obesity (KSSO) [14]. In the NHID, BMI data have been available since 2006, whereas WC measurements have been systematically collected since 2009 [15].

General obesity was defined as a BMI ≥25 kg/m², consistent with the 2024 KSSO guidelines [14]. According to these guidelines, obesity is further classified into class I (BMI 25.0–29.9 kg/m²), class II (BMI 30.0–34.9 kg/m²), and class III (BMI ≥ 35.0 kg/m²). Abdominal obesity was defined as a WC ≥90 cm in men and ≥85 cm in women.

 Obesity fact sheet in Korea: a decade of data and new trends

The KSSO has published the Obesity Fact Sheet periodically since 2015 to provide national statistics on obesity based on NHID and KNHANES data [16]. The most recent 2025 Obesity Fact Sheet, marking the 10th anniversary of its publication, analyzed data spanning 2014 to 2023 and revealed notable epidemiological shifts compared with earlier reports [16].

In contrast to the steadily rising trend observed over the previous decade, the overall prevalence of obesity, defined as BMI ≥25 kg/m², among adults has stabilized at approximately 38.4% over the most recent 3-year period from 2021 to 2023 [16]. However, this apparent stabilization conceals a pronounced sex-based disparity. Obesity prevalence among men continued to increase, reaching 49.8% in 2023, whereas prevalence among women remained relatively stable at 27.5% since 2021 [16]. Similarly, the prevalence of abdominal obesity stabilized at 24.3% overall but continued to rise among men, reaching 31.3%, while declining among women to 17.7% [16].

A major public health concern identified in recent big-data analyses is the disproportionate increase in severe obesity among young adults. Although overall obesity prevalence was highest among individuals in their 30s (42.0%) and 40s (42.2%), the steepest increase over the past decade occurred among individuals in their 20s, rising from 22.7% in 2014 to 32.0% in 2023 [16]. Notably, the prevalence of class III obesity, defined as BMI ≥35.0 kg/m², was highest among young men aged 20 to 24 years at 3.22%, indicating an urgent need for early interventions targeting this young-onset high-risk population.

Recent data also provide insight into childhood obesity trends in relation to the coronavirus disease 2019 (COVID-19) pandemic. Obesity prevalence among children and adolescents peaked in 2021 during the pandemic but subsequently declined to 13.8% in 2023, returning to levels comparable to those observed before the pandemic, similar to 2019 [16]. Furthermore, for the first time, the 2025 Obesity Fact Sheet leveraged nationwide big-data analyses to quantify intergenerational transmission of obesity, demonstrating that children of parents with class II obesity or higher had more than a fivefold increased risk of developing obesity themselves (Fig. 1).

 Recent research in obesity based on the Korean NHID Paradigm shift from NAFLD to metabolic dysfunction-associated steatotic liver disease

The transition from NAFLD to metabolic dysfunction-associated steatotic liver disease (MASLD) represents a fundamental shift toward affirmative diagnostic criteria centered on metabolic dysfunction [17]. Recent landmark studies using the Korean NHID have validated the clinical relevance and applicability of this revised definition specifically within the Korean population.

A cross-sectional analysis of 476,124 Korean adults, published in 2025, provided the first large-scale evidence directly comparing the clinical characteristics of NAFLD and MASLD in Korea [18]. The study demonstrated that although the prevalence of MASLD (29.8%) was similar to that of NAFLD (30.1%), the affirmative diagnostic framework of MASLD more effectively identified individuals with a substantial metabolic burden. Importantly, individuals meeting MASLD criteria exhibited a significantly higher risk of hepatic fibrosis, a key determinant of long-term prognosis, compared with those with steatotic liver disease (SLD) alone, confirming that the revised definition preferentially captures a high-risk phenotype [18]. The study also characterized the newly defined MetALD group, defined as MASLD with moderate alcohol consumption. This subgroup, which is often overlooked in traditional binary classifications, demonstrated a distinct clinical profile with elevated fibrosis scores, suggesting that metabolic dysfunction and alcohol exposure exert a synergistic adverse effect on liver health.

Moon et al. [19] examined a nationwide screening cohort of 351,068 adults aged 47 to 86 years who were followed for a median of 9.0 years. Compared with individuals without SLD, MASLD was associated with an increased subdistribution hazard ratio for CVD (subdistribution hazard ratio, 1.19). This study validated the MASLD, MetALD, and alcohol-related liver disease categorization against hard cardiovascular outcomes and demonstrated additive cardiovascular risk attributable to alcohol exposure when combined with metabolic dysfunction. In another nationwide cohort study involving nearly 9 million Korean adults, Lee et al. [20] evaluated MASLD prevalence and its association with CVD outcomes. Over a median follow-up of 12.3 years, individuals with MASLD and related disease entities had a significantly higher risk of incident CVD, including myocardial infarction, ischemic stroke, heart failure, and cardiovascular death, compared with individuals without SLD, even after multivariable adjustment (hazard ratio, 1.38; 95% confidence interval, 1.37 to 1.39). Among disease subtypes, MASLD conferred the highest cardiovascular risk, followed by MASLD with combined etiologies and MetALD. These findings support the clinical utility of the MASLD framework in identifying individuals with metabolically complicated SLD who are at elevated cardiovascular risk.

A key limitation of defining MASLD using NHID data is the absence of direct imaging or histological confirmation of hepatic steatosis, which is a core component of the original MASLD diagnostic framework and typically requires imaging modalities such as ultrasonography, computed tomography, magnetic resonance imaging–proton density fat fraction, or histological assessment. Consequently, most NHID-based studies rely on noninvasive surrogate indices, most commonly the fatty liver index, which is derived from BMI, WC, triglyceride levels, and γ-glutamyltransferase, for epidemiological identification of fatty liver in the absence of imaging [21]. Because these indices incorporate metabolic parameters that substantially overlap with the cardiometabolic criteria used to define MASLD, they may introduce circularity and misclassification bias relative to imaging-confirmed steatosis. In addition, some NHID analyses have applied the hepatic steatosis index, a validated alternative surrogate based on the aspartate aminotransferase-to-alanine aminotransferase ratio, BMI, diabetes status, and sex, to identify fatty liver in cohort research; however, these indices likewise lack confirmation against imaging or histology [22,23]. These limitations in steatosis ascertainment should be carefully considered when interpreting MASLD prevalence estimates and associations with clinical outcomes in NHID-based studies (Fig. 1).

 Obesity, cardiovascular disease, and heart failure

Large-scale analyses using the Korean NHID have provided robust evidence linking obesity to a broad spectrum of cardiovascular outcomes. In a nationwide cohort exceeding 400,000 individuals, obesity was independently associated with an increased risk of incident atrial fibrillation, with consistent findings observed across Korean NHID and UK Biobank populations, underscoring the generalizability of obesity-related arrhythmic risk [24].

Beyond conventional BMI-based definitions, recent NHID studies have emphasized the importance of body composition and fat distribution in stratifying heart failure risk. In a nationwide cohort of more than two million adults, longitudinal changes in predicted body fat and lean body mass indices were strongly associated with incident heart failure, suggesting that qualitative aspects of obesity may be as clinically relevant as absolute body weight [25]. Furthermore, among nearly one million individuals with newly diagnosed heart failure, the coexistence of abdominal obesity was associated with substantially worse cardiovascular outcomes, including all-cause mortality, myocardial infarction, atrial fibrillation, and heart failure hospitalization, even within identical BMI categories. These findings challenge the notion of a benign ‘obesity paradox’ in heart failure [26].

More recently, an integrated cardiometabolic framework has been applied to NHID data to contextualize obesity within a broader disease continuum. In a cohort of approximately 1.5 million adults, progressive stages of cardiovascular-kidney-metabolic syndrome, which incorporate obesity as a key upstream component, were associated with stepwise increases in mortality, myocardial infarction, stroke, and heart failure hospitalization [27]. Collectively, these landmark NHID-based studies establish obesity, particularly when characterized by adverse fat distribution or unfavorable body composition, as a major determinant of CVD and heart failure risk in the Korean population (Fig. 1).

 Obesity and cancer

Recent large-scale analyses using the Korean NHID have consistently demonstrated that obesity is associated with an increased risk of multiple site-specific cancers. Obesity contributes to carcinogenesis through several interrelated mechanisms, including insulin resistance and enhanced insulin-like growth factor-1 signaling, chronic low-grade inflammation with elevated pro-inflammatory cytokine production, impairment of antitumor immune surveillance, and hormonal dysregulation involving sex hormones [28].

Nationwide NHID studies have consistently reported a positive association between obesity and colorectal cancer risk, particularly when excess adiposity is sustained over time. A large population-based cohort demonstrated that persistent or repeated exposure to obesity across sequential health examinations was associated with a significantly increased risk of colorectal cancer, highlighting the importance of cumulative adiposity rather than single-time-point BMI measurements [29]. Complementing these findings, another NHID-based analysis showed that longitudinal changes in BMI influenced colorectal cancer risk, with sustained or worsening obesity conferring higher risk, underscoring the dynamic nature of obesity trajectories in colorectal carcinogenesis [30]. More recently, a nationwide NHID cohort study of older adults demonstrated that overweight and obesity were associated with increased incidence of both gastric and colorectal cancers, with particularly strong associations observed for colorectal cancer, further supporting excess adiposity as a major risk factor for gastrointestinal malignancies later in life [31].

Pancreatic cancer is closely linked to obesity-related metabolic disorders, particularly T2DM and pancreatitis. In a nationwide NHID cohort of young adults aged 20 to 39 years, overweight and obesity were associated with a significantly increased risk of young-onset pancreatic cancer, whereas underweight status was not, highlighting excess adiposity as an early-life risk factor for pancreatic carcinogenesis [32]. Complementing these findings, a nationwide nested case-control study using NHID data demonstrated that T2DM and pancreatitis markedly increased pancreatic cancer risk, particularly among individuals with post-pancreatitis diabetes or diabetes of short duration, underscoring the combined roles of metabolic dysfunction and pancreatic inflammation in pancreatic cancer development [33].

Emerging NHID-based evidence further suggests that obesity and metabolic dysfunction may play important roles in cancers traditionally considered less obesity-driven. Metabolic syndrome was associated with a significantly increased risk of thyroid cancer, with risk rising in proportion to the number of metabolic syndrome components, supporting a cumulative effect of metabolic dysregulation on thyroid carcinogenesis [34]. More recently, an NHID-based analysis incorporating the metabolic score for insulin resistance demonstrated a clear dose–response relationship between worsening metabolic dysfunction and thyroid cancer incidence, highlighting insulin resistance and adverse metabolic profiles, beyond BMI alone, as key contributors to thyroid cancer risk [35]. The biological mechanisms underlying the association between obesity-related metabolic dysfunction and thyroid cancer remain to be fully elucidated.

 BONE AND MINERAL METABOLISM

Research on bone and mineral metabolism using Korean nationwide data has expanded substantially over the past decade. The availability of large-scale administrative and cohort datasets, including the Korean NHID and the National Screening Program for Transitional Ages (NSPTA), has enabled population-level analyses that are not feasible using single-center or purely cross-sectional designs [36,37]. Osteoporosis-related fractures, a hallmark of skeletal aging, require long-term observation to adequately characterize their epidemiology, clinical consequences, and associated mortality [38,39]. Nationwide datasets facilitate extended follow-up and provide a platform for evaluating the cumulative effects of aging, comorbidities, medication exposure, and healthcare utilization [1]. As illustrated in Fig. 2, publications using Korean nationwide data for osteoporosis and fracture research have increased markedly since 2013, with particularly rapid growth observed in recent years. More recently, nationwide data have also been applied to parathyroid disorders, including primary hyperparathyroidism (PHPT), hypoparathyroidism, and parathyroid carcinoma, reflecting the expanding scope of mineral metabolism research [40-44].

Operational definitions of osteoporosis and osteoporosis-related fractures

Operational definitions of osteopenia, osteoporosis, and fracture outcomes require cautious application because bone mineral density (BMD) measurements are not available within the Korean NHID. Thus, osteoporosis is operationally defined using diagnostic codes in combination with medication prescription records. A commonly applied definition classifies individuals as having osteoporosis when any of the following conditions are met: (1) prescriptions for anti-osteoporosis medications (AOMs) (2) ICD-10 codes M80–M82 accompanied by prescriptions for AOMs; or (3) documentation of an osteoporosis-related fracture. In addition, individuals who visited a clinic at least once for osteoporosis-related diagnostic codes and received AOMs within 6 months before or 2 months after the baseline assessment are also classified as having osteoporosis [45-47].

Beyond claims-based definitions, additional phenotypic information is available through the NSPTA. The NSPTA incorporates BMD testing at ages 50 and 66 using dual-energy X-ray absorptiometry, quantitative computed tomography, peripheral dual-energy X-ray absorptiometry, or quantitative ultrasonography [37]. However, the NSPTA does not provide continuous T-scores; instead, each modality reports categorical outcomes, classified as normal, osteopenia, or osteoporosis, which function as standardized nominal variables [48]. When integrated with NHIS claims data, these categorical BMD results enhance operational definitions by enabling direct classification of bone health status and facilitating harmonized phenotyping across datasets.

Osteoporosis-related fractures are identified using ICD-10 diagnostic codes in combination with site-specific procedure claims to enhance diagnostic specificity. Major osteoporotic fractures are defined across six anatomical sites: vertebrae, hip, humerus, distal radius, pelvis, and ankle. A subsequent fracture is defined as a new fracture event occurring at least 6 months after the index fracture [49]. These coding strategies have been widely applied in Korean NHID-based cohorts and are summarized in Table 2. Nevertheless, operational definitions may vary across studies because investigators adapt diagnostic codes, exposure windows, and classification criteria according to specific research questions, study designs, and clinical contexts.

The NHID includes prescription information only for medications reimbursed under the national insurance system. Osteoporosis-related agents captured in the database include hormone replacement therapies, selective estrogen receptor modulators, bisphosphonates, denosumab, teriparatide, romosozumab, and calcium or vitamin D supplements.

 Operational definitions for parathyroid disorders

Operational definitions for parathyroid disorders in Korean nationwide datasets vary across studies, reflecting differences in research objectives, disease subtypes, and analytic approaches. For PHPT, surgically treated cohorts typically define PHPT using repeated ICD-10 codes (E21.0, E21.2, E21.3, or D35.1), parathyroidectomy procedure codes (P4541–P4543), and at least one hospitalization record [40]. Expanded definitions that include both surgical and nonsurgical PHPT further incorporate repeated diagnostic codes while excluding individuals with chronic kidney disease, dialysis, or kidney transplantation to minimize misclassification of secondary hyperparathyroidism [41]. Nonsurgical hypoparathyroidism is identified using repeated ICD-10 codes together with at least two prescriptions for active vitamin D analogs, while excluding individuals with prior thyroid or parathyroid surgery, neck irradiation, head and neck cancer, or stage 5 chronic kidney disease [43]. Postoperative hypoparathyroidism constitutes an additional category; permanent postoperative hypoparathyroidism is defined by at least three 90-day prescriptions for active vitamin D within 1 year after total thyroidectomy [44]. Parathyroid carcinoma is consistently defined using ICD-10 code C75.0 in combination with parathyroidectomy procedure codes and a hospitalization claim [42].

These operational definitions underscore that parathyroid phenotyping in Korean nationwide databases must be tailored to the clinical context, disease subtype, and specific analytic purpose.

 Osteoporosis and osteoporotic fracture fact sheets in Korea

Nationwide analyses spanning 2002 to 2022 demonstrate substantial changes in osteoporosis management in South Korea [49]. Use of AOMs increased markedly, accompanied by a clear shift from oral agents to injectable therapies. Oral prescriptions peaked in 2016 and declined thereafter, whereas injectable prescriptions surpassed oral therapy in 2020. Denosumab use increased sharply following its approval as a reimbursed first-line therapy in 2019, although anabolic agents have remained underutilized. Medication adherence also improved over time, with the 1-year medication possession ratio increasing from 35.4% in 2003 to 73.2% in 2021. Post-fracture treatment initiation rose from 31.1% to 39.9% but remained below 50%, and treatment rates varied substantially by sex and fracture type, with higher rates observed in vertebral and femoral fractures and lower rates in wrist, humerus, and ankle fractures.

Pelvic fractures increased markedly from 2006 to 2022, particularly among adults aged 80 to 89 years, while 1-year mortality declined during this period [50]. Ankle fractures also demonstrated rising incidence and refracture rates [51]. Hip fractures continued to increase with population aging, although overall incidence has plateaued in recent years; 1-year mortality declined until 2018 but increased during the COVID-19 period [52]. The growing burden of osteoporosis-related fractures in Korea, together with persistent gaps in post-fracture care, highlights the need for more precise fracture surveillance and continued epidemiological research. Because the clinical presentation and coding characteristics of fractures vary by anatomical site and treating specialty, operational definitions require periodic re-evaluation and methodological updating. Regular consensus updates involving endocrinologists, orthopedic surgeons, and other relevant experts are essential to maintain clinically accurate and reliable definitions for nationwide research.

 PITUITARY AND ADRENAL DISORDERS Operational definitions of pituitary and adrenal disorders

Nationwide big-data research using the Korean NHID has substantially advanced both epidemiological and prognostic investigations of pituitary and adrenal diseases. A critical prerequisite for these studies has been the establishment and validation of operational definitions based on ICD-10 codes, procedure codes, prescription records, and, for pituitary disorders, registration in the Rare Incurable Disease program (Table 3) [1]. Together, these methodological foundations have enabled reliable nationwide investigations of rare endocrine disorders that were previously difficult to study at the population level.

 Recent research in pituitary and adrenal disorders based on the Korean NHID Pituitary disorders

Among pituitary disorders, acromegaly has been the most extensively investigated condition using the NHID. Early nationwide studies estimated the annual incidence of acromegaly at approximately 0.36 cases per 100,000 persons and identified increased risks of malignancy and mortality, particularly among female patients [53]. Subsequent analyses confirmed sex-specific disparities in long-term outcomes, demonstrating a higher standardized mortality ratio in women, largely attributable to CVD and malignancy [54].

A growing body of NHID-based research has demonstrated that acromegaly is associated with a broad spectrum of systemic complications. Cardiovascular risks, including atrial fibrillation and congestive heart failure, were significantly increased, whereas the risks of myocardial infarction and stroke were comparable to those observed in matched control populations. Notably, the excess risk of atrial fibrillation was most pronounced within the first 4 years after diagnosis [55]. Beyond CVD, acromegaly has also been linked to neurodegenerative disorders, with increased risks of Parkinson’s disease and dementia [56]. More recent nationwide cohort studies further expanded the recognized disease burden of acromegaly, reporting elevated risks of depression [57], end-stage kidney disease [58], hip fractures [59], and malignancies, thereby reinforcing the need for long-term oncological surveillance [60]. The long-term prognosis and multisystem impact of acromegaly, based on NHID data, have been comprehensively summarized in recent review articles [61].

NHID-based research has also extended to other pituitary disorders. Nationwide epidemiological studies estimated the annual incidence of prolactinoma and Cushing’s disease at 1.6–2.4 and 0.23 cases per 100,000 person-years, respectively. Treatment was associated with reduced mortality in Cushing’s disease, whereas no significant mortality benefit was observed in prolactinoma [62,63]. In panhypopituitarism, patients exhibited increased risks of cardiovascular events, fractures, and mortality, with distinct sex-specific patterns [64,65]. During the COVID-19 pandemic, NHID analyses additionally provided important insights into infection-related outcomes among patients with pituitary diseases, highlighting altered susceptibility and prognosis in this vulnerable population [66].

 Adrenal disorders

In adrenal research, NHID-based studies initially focused on defining and validating operational criteria for major adrenal disorders, including pheochromocytoma and paraganglioma (PPGL), adrenal Cushing’s syndrome (CS), primary aldosteronism (PA), and congenital adrenal hyperplasia (CAH). Using these standardized definitions, the first nationwide epidemiological estimates for these conditions in Korea were generated. The prevalence and age-standardized incidence rate of PPGL were estimated at 2.13 per 100,000 persons and 0.18 per 100,000 person-years, respectively. Similarly, adrenal CS demonstrated a prevalence of 2.3 per 100,000 persons and an incidence of 0.13 per 100,000 person-years, whereas CAH showed a point prevalence of 5.3 per 100,000 persons in 2017, with a declining annual incidence over time [67-69].

Beyond descriptive epidemiology, subsequent studies expanded the analytic scope to long-term complications and prognosis. Nationwide cohort analyses revealed that metastatic disease was present in approximately one-fifth of patients with PPGL. Importantly, extended follow-up demonstrated that a substantial proportion of tumors initially classified as non-metastatic later progressed to metastatic disease over a median follow-up period exceeding 6 years, underscoring the need for prolonged surveillance even in apparently low-risk cases [67]. Long-term outcome data further indicated that patients with adrenal CS remain at increased risk of mortality for up to a decade following adrenalectomy, suggesting that surgical remission does not fully offset the persistent adverse cardiovascular effects of chronic hypercortisolism [68].

In patients with PA, Kim et al. [70] demonstrated that the risk of new-onset atrial fibrillation remained significantly elevated for up to 3 years after adrenalectomy or medical treatment, compared with individuals with essential hypertension. These findings highlight persistent cardiovascular vulnerability despite biochemical control. Furthermore, patients with PA were shown to have a significantly increased risk of all-cause dementia, including both Alzheimer’s disease and vascular dementia, relative to individuals with essential hypertension [71]. In CAH, large-scale population-based studies demonstrated a substantially increased burden of cardiometabolic and neuropsychiatric comorbidities, including CVD, cerebrovascular events, diabetes mellitus, and psychiatric disorders, across all age groups [69].

Collectively, these studies illustrate how NHID-based research has evolved from establishing foundational disease epidemiology to uncovering previously unrecognized long-term cardiovascular, neurocognitive, and systemic complications of adrenal disorders. Overall, nationwide big-data research using the Korean NHID has transformed pituitary and adrenal disease research, progressing from descriptive epidemiology to comprehensive evaluation of long-term outcomes and multisystem involvement. These findings underscore the value of population-level data in identifying residual risks, informing long-term surveillance strategies, and guiding personalized management for rare endocrine diseases.

 THYROID DISEASES

For the thyroid disease section, we systematically identified nationwide Korean big-data studies published between 2023 and 2025 that used either the Korean NHID or the Health Insurance Review and Assessment Service (HIRA) databases. Eligible studies were population-based analyses addressing thyroid cancer, functional thyroid diseases (hypothyroidism or Graves’ disease), or inflammatory thyroid disease, with outcomes derived from administrative claims, health screening databases, or linked national mortality records. To summarize the scope and volume of thyroid-related big-data research, we constructed an evidence map stratified by clinical topic and data source. Each study was assigned to a single category based on its primary research question to ensure mutually exclusive classification. Bubble size represents the number of studies within each category, with the exact count displayed in the figure, and colors indicate the predominant study design. Between 2023 and 2025, a total of 45 nationwide big-data studies on thyroid diseases were identified using NHID or HIRA databases (Fig. 3). Research volume was highest for thyroid cancer (n=17) and functional thyroid disease (n=17), followed by inflammatory thyroid disease (n=11). Although claims-only NHID remained the most frequently used data source, an increasing proportion of studies incorporated health screening databases and linked national mortality records, enabling analyses that extended beyond descriptive epidemiology toward long-term outcomes and disease trajectories.

Operational definitions of thyroid diseases and methodological convergence

Across recent nationwide studies, operational definitions of thyroid diseases have demonstrated substantial convergence, reflecting increasing methodological standardization in Korean thyroid big-data research (Table 4). Thyroid cancer was consistently defined using the ICD-10 code C73 in combination with thyroidectomy procedure codes or cancer-specific copayment reduction registration, with most studies applying a 1- to 2-year washout period to identify incident cases [1,72,73].

Similarly, functional thyroid diseases were operationally defined using diagnostic codes combined with sustained medication prescriptions, an approach widely adopted in nationwide studies evaluating long-term outcomes and treatment patterns in hypothyroidism and Graves’ disease [74-77]. Subacute thyroiditis (SAT) was identified using repeated diagnostic codes (E06.1) with supporting clinical context, a definition applied in nationwide population-based analyses, particularly in studies examining post–COVID-19 thyroiditis [78]. Collectively, this methodological convergence enhances reproducibility and comparability across NHID-based studies, thereby strengthening the reliability of evidence generated from nationwide thyroid research in Korea.

 Recent research in thyroid diseases based on the Korean NHID Survivorship and long-term complications after thyroid cancer

Recent thyroid cancer studies have increasingly focused on survivorship and long-term complications rather than incidence alone. Nationwide cohort analyses identified postoperative hypoparathyroidism as a key determinant of long-term outcomes, reporting elevated risks of skeletal complications, including fractures, as well as nonskeletal systemic complications [44,79]. This perspective reframes thyroid cancer from a condition characterized by excellent short-term prognosis to a chronic disease state associated with sustained health consequences. In addition, population-based studies evaluated mortality and healthcare burden after thyroid cancer using explicit linkage between NHID and Statistics Korea death records, enabling robust assessment of both all-cause and cancer-specific mortality at the national level [72].

 Pharmaco-epidemiological analyses in thyroid diseases

Beyond disease definitions, recent nationwide studies expanded into pharmaco-epidemiological research using large-scale claims data. In Graves’ disease, nationwide cohorts systematically classified patients according to initial treatment modality, including antithyroid drugs, radioactive iodine, and surgery, enabling comparative analyses of real-world treatment pathways and long-term outcomes [73-75]. Outcome-oriented studies further evaluated long-term risks associated with radioactive iodine exposure, including malignancy and metabolic outcomes, using clearly defined exposure windows and appropriate comparator groups [73]. In contrast, pharmaco-epidemiological analyses in hypothyroidism during this period primarily relied on prescription-based exposure definitions to characterize treatment patterns and duration, with increasing attention to potential misclassification related to thyroid cancer-associated thyroid-stimulating hormone suppression therapy in nationwide datasets [77]. Collectively, these studies provide clinicians with population-level estimates of treatment-associated risks that complement evidence derived from controlled trial settings.

 Metabolic axis and trajectory-based analyses using health screening data

The increasing integration of NHID-linked health screening databases has enabled large-scale analyses examining the metabolic context of thyroid disease. Recent nationwide cohort studies investigated associations between MASLD and incident thyroid cancer using baseline metabolic profiles with long-term follow-up [80]. This design permits population-level risk estimation through prospective time-to-event modeling within large thyroid cancer cohorts.

Looking ahead, incorporation of repeated metabolic measurements may further enhance these analyses by enabling trajectory-based modeling of dynamic metabolic changes over time. Such approaches, as demonstrated in large longitudinal resources including the UK Biobank [81], could be adapted to thyroid cancer susceptibility research in the Korean population.

 COVID-19–associated subacute thyroiditis and post-acute outcomes

Inflammatory thyroid disease research during this period was largely driven by nationwide investigations of SAT in the context of COVID-19. Using the Korean NHID, these studies demonstrated an increased incidence of SAT following severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection [78]. Although initial analyses primarily focused on incident disease, more recent studies emphasized careful definition of infection index dates and latency windows, thereby establishing a methodological framework for future investigations of postacute and long-term outcomes related to COVID-19–associated thyroiditis. While early work appropriately prioritized incident disease estimation, the increasing methodological emphasis on precise index-date and latency definitions provides a foundation for extending analyses toward post-acute and long-term outcomes [78]. This evolution illustrates how administrative data can rapidly generate hypothesis-generating evidence during public health emergencies while progressively supporting outcome-oriented investigation.

 Graves’ disease as a model for RCT-adjacent real-world evidence

Recent nationwide studies of Graves’ disease exemplify the methodological maturation of thyroid big-data research beyond descriptive epidemiology. Large-scale NHID-based cohorts explicitly defined treatment pathways, including antithyroid drugs, radioactive iodine, and surgery, and evaluated long-term outcomes and treatment failure using multivariable risk stratification models [73-76]. These analyses addressed clinical questions traditionally examined in randomized controlled trials (RCTs), such as comparative effectiveness and predictors of treatment failure, thereby generating trial-adjacent real-world evidence at a national scale. Although residual confounding inherent to observational study designs remains, these investigations demonstrate that contemporary Korean thyroid big-data research can meaningfully complement randomized trials, particularly in settings where RCTs are infeasible. Nationwide cohorts further addressed comparative effectiveness questions traditionally reserved for RCTs, including prediction of treatment failure and evaluation of long-term outcomes across therapeutic modalities [73-75]. By incorporating multivariable risk stratification and longitudinal follow-up at a national scale, these studies approximate several key elements of pragmatic trials. Although residual confounding cannot be fully eliminated, such analyses meaningfully complement randomized evidence, particularly in clinical scenarios where RCTs are infeasible, underpowered for rare outcomes, or limited in follow-up duration.

 RCTs and nationwide big data: complementary roles

Importantly, the advances highlighted in this review do not suggest that nationwide big-data studies should replace RCTs. Rather, they demonstrate that contemporary real-world evidence can complement and extend trial findings by addressing longterm outcomes, rare adverse events, and heterogeneous treatment effects within routine clinical practice [73-75]. When grounded in transparent operational definitions, careful exposure modeling, and appropriate data linkage, nationwide thyroid big-data research offers a robust framework for bridging the gap between efficacy demonstrated in trials and effectiveness observed in real-world settings.

 CONCLUSIONS

Research utilizing the Korean NHID in the field of endocrine disorders has progressed beyond descriptive epidemiology, generating valuable real-world evidence that can complement, and in selected contexts partially substitute for, RCTs when such trials are infeasible. In addition, the nationwide scale of the NHID provides an unparalleled opportunity to generate robust evidence for rare endocrine diseases, which are often underrepresented in conventional clinical research. With continued refinement of operational disease definitions and judicious use of administrative data, NHID-based research is expected to further advance understanding of endocrine disorders and contribute meaningfully to clinical decision-making and public health policy.

CONFLICTS OF INTEREST

Sun Wook Cho, Jung Hee Kim, Beom-Jun Kim, Mee Kyoung Kim and Eun Jung Rhee are deputy editors of the journal. But they were not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.

 REFERENCES ChoSW KimJH ChoiHS AhnHY KimMK RheeEJ Big data research in the field of endocrine diseases using the Korean National Health Information database Endocrinol Metab (Seoul) 2023 38 10 24 KimKS KimB HanK Big data research for diabetes-related diseases using the Korean National Health Information Database Diabetes Metab J 2025 49 13 21 BaekJH ParkYM HanKD MoonMK ChoiJH KoSH Comparison of operational definition of type 2 diabetes mellitus based on data from Korean National Health Insurance Service and Korea National Health and Nutrition Examination Survey Diabetes Metab J 2023 47 201 10 KimSE HanK ChoWK SuhBK Cardiovascular complications, kidney failure, and mortality in young-onset type 1 and type 2 diabetes: data from the Korean National Health Insurance Service Diabetes Care 2025 48 422 9 KimKS HongS HanK ParkCY The clinical characteristics of gestational diabetes mellitus in Korea: a national health information database study Endocrinol Metab (Seoul) 2021 36 628 36 LeeSH YuJ HanK LeeSW YouSY KimHS Predicting the risk of insulin-requiring gestational diabetes before pregnancy: a model generated from a nationwide population-based cohort study in Korea Endocrinol Metab (Seoul) 2023 38 129 38 YuJ KimJH KimB HanK LeeSH KimMK The severity of diabetes and the risk of diabetic foot amputation: a national cohort study Endocrinol Metab (Seoul) 2025 40 574 82 BaeJH HanKD KoSH YangYS ChoiJH ChoiKM Diabetes fact sheet in Korea 2021 Diabetes Metab J 2022 46 417 26 KimNH SeoMH JungJH HanKD KimMK KimNH 2023 Diabetic kidney disease fact sheet in Korea Diabetes Metab J 2024 48 463 72 KimJH LeeJ HanK KimJT KwonHS Cardiovascular disease & diabetes statistics in Korea: nationwide data 2010 to 2019 Diabetes Metab J 2024 48 1084 92 KimKS HongS HanK ParkCY Association of non-alcoholic fatty liver disease with cardiovascular disease and all cause death in patients with type 2 diabetes mellitus: nationwide population based study BMJ 2024 384 e076388 ParkYM BaekJH LeeHS ElfassyT BrownCC SchootmanM Income variability and incident cardiovascular disease in diabetes: a population-based cohort study Eur Heart J 2024 45 1920 33 LeeHS ParkJC ChungI LiuJ LeeSS HanK Sustained low income, income changes, and risk of all-cause mortality in individuals with type 2 diabetes: a nationwide population-based cohort study Diabetes Care 2023 46 92 100 BaeJH HaamJH JeonE KangSY SongS ParkCY 2024 Clinical practice guidelines for the diagnosis and pharmacologic treatment of overweight and obesity by the Korean Society for the Study of Obesity J Obes Metab Syndr 2025 34 322 43 KimYH HanK SonJW LeeSS OhSW KwonHS Data analytic process of a nationwide population-based study on obesity using the National Health Information database presented by the National Health Insurance Service 2006-2015 J Obes Metab Syndr 2017 26 23 7 Korean Society for the Study of Obesity. Obesity fact sheet in Korea 2025 [Internet]. Seoul: Korean Society for the Study of Obesity; 2025 [cited 2026 Feb 6]. Available from: https://www.kosso.or.kr Munk LauridsenM RavnskjaerK GluudLL SanyalAJ Disease classification, diagnostic challenges, and evolving clinical trial design in MASLD J Clin Invest 2025 135 e189953 LeeDY KoJH JangHN MoonSJ KwonHM ParkSE Characteristics of metabolic dysfunction-associated steatotic liver disease and its risk for hepatic fibrosis in 476,124 Korean adults: a cross-sectional study Endocrinol Metab (Seoul) 2025 40 648 52 MoonJH JeongS JangH KooBK KimW Metabolic dysfunction-associated steatotic liver disease increases the risk of incident cardiovascular disease: a nationwide cohort study EClinicalMedicine 2023 65 102292 LeeHH LeeHA KimEJ KimHY KimHC AhnSH Metabolic dysfunction-associated steatotic liver disease and risk of cardiovascular disease Gut 2024 73 533 40 BiciuscaT StanSI BalteanuMA CioboataR GheneaAE DanoiuS The role of the fatty liver index (FLI) in the management of non-alcoholic fatty liver disease: a systematic review Diagnostics (Basel) 2023 13 3316 KimGA OhCH KimJW JeongSJ OhIH LeeJS Association between non-alcoholic fatty liver disease and the risk of dementia: a nationwide cohort study Liver Int 2022 42 1027 36 LeeJ KwakJ KimMH LeeSH ChoJH LimDJ Association between metabolic dysfunction-associated steatotic liver disease and the risk of thyroid cancer: a nationwide cohort study Korean J Intern Med 2025 40 813 22 ChoiSH YangPS KimD SungJH JangE YuHT Association of obesity with incident atrial fibrillation in Korea and the United Kingdom Sci Rep 2023 13 5197 WooHG KimDH LeeH KangMK SongTJ Association between changes in predicted body composition and occurrence of heart failure: a nationwide population study Front Endocrinol (Lausanne) 2023 14 1210371 KimJ ChunS ParkJK LeeH JangJY KimH Poor cardiovascular outcomes of underweight abdominal obesity in the entire population of newly diagnosed heart failure ESC Heart Fail 2025 12 4242 54 KimHJ KimBS KimH LeeJ JoHH KimDW Long-term clinical outcome and risk stratification across stages of cardiovascular-kidney-metabolic syndrome in a nationwide cohort Korean J Intern Med 2025 40 975 89 LimJS KimS ChoIY ShinDW Obesity and cancer: mechanisms, epidemiological evidence, and potential risk reduction J Obes Metab Syndr 2026 35 14 37 SeoJY JinEH ChungGE KimYS BaeJH YimJY The risk of colorectal cancer according to obesity status at four-year intervals: a nationwide population-based cohort study Sci Rep 2023 13 8928 NamSY JoJ ChunH JeonSW BMI changes and presence of cofactors influence CRC risk in middle-aged and older Adults Dig Liver Dis 2025 57 1838 44 ChoiJ ShinCM HanK JungJH KimSY JoH Impact of overweight and obesity on gastric and colorectal cancer incidence in the older adults: a nationwide cohort study BMC Geriatr 2026 26 184 KimHJ ParkJH HongJY HanK ShenJJ ParkJO Increased risk of young-onset pancreatic cancer among adults aged 20-39 years with overweight or obesity, but not underweight: a nationwide cohort study Eur J Cancer 2026 232 116111 KimYE YuMH NamCM KangES Effects of pancreatitis and type 2 diabetes mellitus on the development of pancreatic cancer: a nationwide nested case-control study Diabetes Metab J 2025 49 252 63 ParkJH ChoiM KimJH KimJ HanK KimB Metabolic syndrome and the risk of thyroid cancer: a nationwide population-based cohort study Thyroid 2020 30 1496 504 KimHR SonM HuhSJ MoonSY MoonH KangYW Relationship between METS-IR and thyroid cancer incidence in Korea: a nationwide population-based study Front Oncol 2024 14 1383864 SeongSC KimYY ParkSK KhangYH KimHC ParkJH Cohort profile: the National Health Insurance Service-National Health Screening Cohort (NHIS-HEALS) in Korea BMJ Open 2017 7 e016640 KimHS ShinDW LeeWC KimYT ChoB National screening program for transitional ages in Korea: a new screening for strengthening primary prevention and follow-up care J Korean Med Sci 2012 27 Suppl S70 5 ClynesMA HarveyNC CurtisEM FuggleNR DennisonEM CooperC The epidemiology of osteoporosis Br Med Bull 2020 133 105 17 MorinSN LeslieWD SchousboeJT Osteoporosis: a review JAMA 2025 334 894 907 KongSH KimJH ParkMY KimSW ShinCS Residual risks of comorbidities after parathyroidectomy in a nationwide cohort of patients with primary hyperparathyroidism Endocrine 2023 79 190 9 KimKJ BaekS YuMH ShinS ChoS RheeY Secular trends in the incidence and treatment patterns of primary hyperparathyroidism in Korea: a nationwide cohort study JBMR Plus 2024 8 ziae065 KongSH KimJH ParkMY KimSW ShinCS Epidemiology and prognosis of parathyroid carcinoma: real-world data using nationwide cohort J Cancer Res Clin Oncol 2021 147 3091 7 KimSH RheeY KimYM WonYJ NohJ MoonH Prevalence and complications of nonsurgical hypoparathyroidism in Korea: a nationwide cohort study PLoS One 2020 15 e0232842 AhnSH LeeYJ HongS ParkJW JeonYJ YooBN Risk of fractures in thyroid cancer patients with postoperative hypoparathyroidism: a nationwide cohort study in Korea J Bone Miner Res 2023 38 1268 77 AhnSH ParkSM ParkSY YooJI JungHS NhoJH Osteoporosis and osteoporotic fracture fact sheet in Korea J Bone Metab 2020 27 281 90 HaYC KimHY JangS LeeYK KimTY Economic burden of osteoporosis in South Korea: claim data of the National Health Insurance Service from 2008 to 2011 Calcif Tissue Int 2017 101 623 30 ShimYB ParkJA NamJH HongSH KimJW JeongJ Incidence and risk factors of subsequent osteoporotic fracture: a nationwide cohort study in South Korea Arch Osteoporos 2020 15 180 BaekYH ChoSW JeongHE KimJH HwangY LangeJL 10-Year fracture risk in postmenopausal women with osteopenia and osteoporosis in South Korea Endocrinol Metab (Seoul) 2021 36 1178 88 KimKJ KimKM LeeYK KimJ JangH KimJ Twenty-year trends in osteoporosis treatment and post-fracture care in South Korea: a nationwide study J Bone Metab 2025 32 57 66 YoonBH KimTY ChungH KimJ KimJ JangH Epidemiological (secular) trends of pelvic and acetabular fractures (2006-2022) in South Korea using a nationwide cohort data over ages 50: is it osteoporosis related? Arch Osteoporos 2025 20 67 ParkJW KimHY KimKM KimJ JangH KimJ Epidemiology of osteoporotic ankle fractures in South Korea: a nationwide retrospective cohort study (2006-2022) Osteoporos Int 2025 36 801 9 ChungH KimTY KimJ HaYC LeeYK KimJ Long-term trends of osteoporotic hip fractures in South Korea: analysis of National Health Insurance Data from 2006 to 2022 Osteoporos Int 2025 36 1999 2007 ParkKH LeeEJ SeoGH KuCR Risk for acromegaly-related comorbidities by sex in Korean acromegaly J Clin Endocrinol Metab 2020 105 dgz317 KimJ HongN ChoiJ MoonJH KimEH HongJW Sex differences in mortality in patients with acromegaly: a nationwide cohort study in Korea Eur J Endocrinol 2023 189 225 34 HongS KimKS HanK ParkCY Acromegaly and cardiovascular outcomes: a cohort study Eur Heart J 2022 43 1491 9 HongS HanK KimKS ParkCY Risk of neurodegenerative diseases in patients with acromegaly: a cohort study Neurology 2022 99 e1875 85 MoonS HongS HanK ParkCY Risk of depression in patients with acromegaly in Korea (2006-2016): a nationwide population-based study Eur J Endocrinol 2023 189 363 71 HongS KimKS HanK ParkCY A cohort study found a high risk of end-stage kidney disease associated with acromegaly Kidney Int 2023 104 820 7 KimJ HongN ChoiJ MoonJH KimEH LeeEJ Increased risk of hip fracture in patients with acromegaly: a nationwide cohort study in Korea Endocrinol Metab (Seoul) 2023 38 690 700 KimYS YunJS KimH JeunSS KimB LeeSW Acromegaly and the risk of cancer: a nationwide population-based cohort study in Korea Eur J Endocrinol 2025 192 220 7 HongS HanK ParkCY Long-term prognosis and systemic impact of acromegaly: analyses utilizing Korean National Health Insurance Data Endocrinol Metab (Seoul) 2025 40 1 9 OhJS KimHJ HannHJ KangTU KimDS KangMJ Incidence, mortality, and cardiovascular diseases in pituitary adenoma in Korea: a nationwide population-based study Pituitary 2021 24 38 47 ParkJS YunSJ LeeJK ParkSY ChinSO Descriptive epidemiology and survival analysis of prolactinomas and Cushing’s disease in Korea Endocrinol Metab (Seoul) 2021 36 688 96 ParkSS JeongH AhnCH ParkMJ KimYH KimK Sex-specific cardiovascular risks and mortality in patients with panhypopituitarism: a nationwide cohort study Endocrinol Metab (Seoul) 2025 40 469 83 ParkSS JeongH AhnCH ParkMJ KongSH KimSW Fracture risks in patients aged 50 years and older with panhypopituitarism: a nationwide cohort study Osteoporos Int 2025 36 1951 63 HaJ KimKM LimDJ SongK SeoGH Pituitary diseases and COVID-19 outcomes in South Korea: a nationwide cohort study J Clin Med 2023 12 4799 KimJH MoonH NohJ LeeJ KimSG Epidemiology and prognosis of pheochromocytoma/paraganglioma in Korea: a nationwide study based on the National Health Insurance Service Endocrinol Metab (Seoul) 2020 35 157 64 AhnCH KimJH ParkMY KimSW Epidemiology and comorbidity of adrenal Cushing syndrome: a nationwide cohort study J Clin Endocrinol Metab 2021 106 e1362 72 KimJH ChoiS LeeYA LeeJ KimSG Epidemiology and long-term adverse outcomes in Korean patients with congenital adrenal hyperplasia: a nationwide study Endocrinol Metab (Seoul) 2022 37 138 47 KimKJ HongN YuMH LeeH LeeS LimJS Time-dependent risk of atrial fibrillation in patients with primary aldosteronism after medical or surgical treatment initiation Hypertension 2021 77 1964 73 HongN KimKJ YuMH JeongSH LeeS LimJS Risk of dementia in primary aldosteronism compared with essential hypertension: a nationwide cohort study Alzheimers Res Ther 2023 15 136 KimWG Trends in thyroid cancer mortality rates in Korea: insights from National Health Database Endocrinol Metab (Seoul) 2024 39 853 5 KimKJ ChoiJ KimKJ SongE YuJH KimNH Cancer risk in Graves disease with radioactive 131I treatment: a nationwide cohort study J Nucl Med 2024 65 693 9 KimKJ ChoiJ ShinSM KimJA KimKJ KimSG Treatment patterns and preferences for Graves’ disease in Korea: insights from a nationwide cohort study Endocrinol Metab (Seoul) 2024 39 659 63 KimJA KimKJ ChoiJ KimKJ SongE YuJH Unveiling risk factors for treatment failure in patients with Graves’ disease: a nationwide cohort study in Korea Endocrinol Metab (Seoul) 2025 40 125 34 AhnHY ChoSW LeeMY ParkYJ KooBS ChangHS Prevalence, treatment status, and comorbidities of hyperthyroidism in Korea from 2003 to 2018: a nationwide population study Endocrinol Metab (Seoul) 2023 38 436 44 ChungCW AhnHY ChoSW YiKH Rising incidence and comorbidities of endogenous hypothyroidism in Republic of Korea from 2004 to 2018: a nationwide population study Endocrinol Metab (Seoul) 2024 39 891 8 AhnHY ChoiHS HaS ChoSW Incidence of subacute thyroiditis during the COVID-19 pandemic in South Korea using the National Health Insurance Service Database Thyroid 2022 32 1299 306 KuEJ LeeJ YooWS BaeJ LeeEK AhnHY Long-term nonskeletal complications in patients with thyroid cancer and hypoparathyroidism post total thyroidectomy J Clin Endocrinol Metab 2025 110 3538 45 MoonSY SonM ChoJH KimHI HanJM BaeJC Association between metabolic dysfunction-associated steatotic liver disease and thyroid cancer Thyroid 2025 35 79 86 LuF YangH SheB LuQ BaoY SetoWK Investigating multimorbidity trajectories in people living with MASLD diagnosis: a trajectory analysis using the UK Biobank Diabetes Obes Metab 2025 27 6968 78 Figures and Tables

Overview of obesity trends, mechanistic pathways, and health outcomes based on data from the Korean National Health Information Database (NHID). NHIS, National Health Insurance Service; MASLD, metabolic dysfunction-associated steatotic liver disease.

Annual number of publications on osteoporosis and fractures using the Korean National Health Information Database (PubMed search, 2008–2025). This graph shows the annual number of publications identified through a PubMed search using the terms (osteoporosis[Title/Abstract] OR “osteoporotic fracture”[Title/Abstract] OR fracture[Title/Abstract]) AND (Korea[Title/Abstract] OR Korean[Title/Abstract]) AND (“NHIS”[All Fields] OR “National Health Insurance Service”[All Fields] OR “Health Insurance Review”[All Fields]).

Evidence map of nationwide Korean big data studies on thyroid diseases (TD) (2023–2025). This evidence map summarizes nationwide Korean big-data studies on TD published between 2023 and 2025, stratified by clinical topic (thyroid cancer, functional TD, and inflammatory TD) and data source (National Health Insurance Service [NHIS]/National Health Information Database [NHID] claims-only, NHID with health screening database, NHID-linked to Statistics Korea mortality data, and Health Insurance Review and Assessment Service [HIRA] claims). Circle size is proportional to the number of studies, and the number inside each circle indicates the exact study count (n) for each category. Circle color denotes the predominant study design, including retrospective cohort, nested case-control or matched cohort, and time-varying exposure or landmark analyses.

Operational Definitions of Diabetes-Related Diseases

Disease Operational definition ICD-10 code Reference
Type 2 diabetes mellitus ≥1 Prescription claim per year for antidiabetic medications under ICD-10 E11–14 or fasting glucose concentration ≥126 mg/dL in the health check-up database E11–14 [2,3]
Type 1 diabetes mellitus (1) ≥1 Claim under ICD-10 E10, (2) ≥3 claims for the prescription of insulin, and (3) ≥1 additional insulin prescription claim occurring between 1 and 2 years after the first insulin prescription E10 [2,4]
Exclusion: patients who had claims under ICD-10 E11–14 within 730 days after the first prescription of insulin or who underwent total or partial pancreatectomy
Gestational diabetes mellitus ≥2 Outpatient claims with ICD-10 O24.4 or O24.9 during pregnancy O24.4, O24.9 [5]
Exclusion: individuals with antidiabetic medication prescriptions based on ICD-10 E10–14 before pregnancy or a fasting glucose level ≥126 mg/dL at the pre-pregnancy health check-up
Impaired fasting glucose Fasting glucose concentration ≥100 and <126 mg/dL in the health check-up database - [2]
Exclusion: Individuals who had prescription claims for antidiabetic medications before the health examination
Diabetic retinopathy, proliferative ICD-10 H360+procedural code S5160 or S5161 (pan-retinal photocoagulation) H360 [2]
Diabetic retinopathy, non-proliferative ICD-10 H360+without procedural code S5160–S5161 H360 [2]
Diabetic foot with amputation ICD-10 codes for complicated diabetes (E10.5–14.5, E10.7–14.7)+Procedural codes N0572-0575 (broader definition incorporating additional foot-level amputation code: N0562, N0564–6, and N0571–5) E10.5–14.5 [2,7]
E10.7–14.7

ICD-10, International Classification of Diseases, 10th Revision.

Operational Definitions of Commonly Used Outcomes and Covariates in the Field of Bone Metabolism and Fracture Research

Operational definition ICD-10 code Procedure codes Reference
Osteoporosis and osteoporosis-related fractures
 Osteoporosis Defined using claims-based criteria because routine T-scores are not available in NHIS. Osteoporosis is classified when ≥1 of the following six criteria is met: ICD-10 M80–M82 [45-47]
 (1) prescription of medications used exclusively for anti-osteoporosis purposes (bisphosphonates, SERMs, denosumab, teriparatide, and romosozumab); Osteoporosis-related fractures: see below per anatomical site
 (2) ICD-10 codes M80–M82 combined with prescriptions for anti-osteoporosis medications or hormone therapy;
 (3) older adults (men ≥70, women ≥65) with ICD-10 osteoporosis codes;
 (4) history of medications known to induce secondary osteoporosis plus ICD-10 codes;
 (5) history of diseases known to induce secondary osteoporosis plus ICD-10 codes;
 (6) osteoporosis-related fracture requiring site-specific procedure claims.
 NSPTA provides device-derived categorical BMD variables (normal/osteopenia/osteoporosis), which can be used directly as nominal bone health phenotypes.
Osteoporosis-related fractures Defined across six skeletal sites: vertebrae, hip, humerus, distal radius, pelvis, ankle. Each fracture diagnosis must be accompanied by a site-specific procedure claim (N-codes) to ensure specificity
 Vertebral fracture Vertebral fracture diagnosis+vertebral fracture-related procedure ICD-10 M48.4, M48.5, M49.5, S22.0, S22.1, S32.0 Operational codes (N0471, N0472, N0473, N0474) and imaging code (G430, G440, G450, G460) [49]
Inpatients with a principal diagnosis corresponding to vertebral fracture
Vertebral fracture diagnosis within 1 day before or after the imaging date
  Hip fracture Hip fracture diagnosis+hip fracture-related procedure code ICD-10 S72.0, S72.1 N0601, N0611, N0991, N0641, N0652, N0654, N0711, N2070, N0715, N2710
Inpatient admission with a principal hip fracture diagnosis
Emergency department visit with hip fracture diagnosis+hip fracture-related procedure code
  Pelvis fracture Pelvic fracture diagnosis+pelvic fracture-related procedure ICD-10 S32.1–S32.5, S32.7, S32.8 Operational codes (N0592, N0593, N0594, N0981, N0641, N0475) and imaging code (G460, G470, G500, G510, G520)
Inpatient admission with principal pelvic fracture diagnosis
Pelvic fracture diagnosis within 1 day before or after a pelvic imaging procedure
  Humerus fracture Inpatient admission with a principal diagnosis of humerus without any fracture-related surgery code ICD-10 S42.2, S42.3 N0602, N0612, N0992, N0642, N0982, N0986, T6010, T6110, T6151, T6020, N0521
Humerus fracture diagnosis related procedure code
  Distal radius fracture Inpatient admission with a principal diagnosis of distal radius without any fracture-related surgery code ICD-10 S52.5, S52.6 N0607, N0603, N0993, N0994, N1601, N1611, N1603, N1613, N0996, N0998, N0983, N0643, T6020, T6030, T6151, T6152
Distal radius fracture diagnosis+related procedure code
  Ankle fracture Inpatient admission with a principal diagnosis of ankle without any fracture-related surgery code ICD-10 S82.3, S82.5, S82.6, S82.8 N1604, N1614, N1605, N1615, N1616, N1606, N0642, N0999, N1000, N1001, N0982, N0986, N0642, T6060, T6061, T6154, T6051, T6052, T6153
Ankle fracture diagnosis+related procedure code
Parathyroid disorders
 Primary hyperparathyroidism Varies across studies; ICD-10 E21.0, E21.2, E21.3, D35.1 P4541–P4543 [40]
Surgical PHPT cohorts: ≥2 ICD-10 codes (E21.0, E21.2, E21.3, D35.1)+parathyroidectomy code+hospitalization requirement; exclusion of CKD or dialysis
Surgical+medical PHPT cohort: ≥2 ICD-10 codes (E21.0), ≥2 PTH measurements, exclusion of secondary HPT, renal failure, dialysis, kidney transplant
 Idiopathic hypoparathyroidism ≥2 ICD-10 codes plus ≥2 prescriptions for active vitamin D analogs; excludes prior thyroid/parathyroid surgery, head and neck cancer, neck irradiation, CKD stage 5 ICD-10 D82.1, E20.0, E20.8, E20.9, E31.0, E31.8, E31.9 [43]
 Postsurgical hypoparathyroidism ≥3 Prescriptions of active vitamin D (each covering approximately 90 days) within 1 year after total thyroidectomy for thyroid cancer (C73); excludes patients with prior hypoparathyroidism or parathyroidectomy, head and neck cancer or irradiation, CKD stage 5 and active vitamin D or levothyroxine use before total thyroidectomy [44]
 Parathyroid carcinoma Defined as ICD-10 C75.0+parathyroidectomy procedure code+≥1 hospitalization ICD-10 C75.0 P4541–P4543 [42]

ICD-10, International Classification of Diseases 10th Revision; NHIS, National Health Insurance Service; SERM, selective estrogen receptor modulator; NSPTA, National Screening Program for Transitional Ages; BMD, bone mineral density; PHPT, primary hyperparathyroidism; CKD, chronic kidney disease; PTH, parathyroid hormone; HPT, hyperparathyroidism.

Operational Definitions of Pituitary and Adrenal Diseases

Operational definition
 Procedural/Measurement code Reference
Inclusion Exclusion
Acromegaly ≥2 Acromegaly (E22.0) Acromegaly-related treatment (medical therapy, operation, or radiotherapy) within 2 years of the first medical claim for acromegaly [53,54]
≥1 Acromegaly (E22.0) or V112 [56,57]
Cushing’s disease E24.0 and D35.2 [62,63]
V162 and V114
Prolactinoma E22.1+D35.2 [62,63]
PA: V162 Hyperprolactinemia: V112
Panhypopituitarism E23, E23.0–E23.7, E893, V165 E22, E240 Thyroid hormones and glucocorticoids for ≥180 days within 1 year, prescribed after or within 30 days before the diagnosis of panhypopituitarism, the initial prescription of both hormones within a 180-day interval [64,65]
D352, D443, C751, D353, D444, C752, V162
Pheochromocytoma/Paraganglioma ≥2 D350, D441, I1522, C741, or C749; D356, D446, D447, D487, or C755 (primary diagnosis) E260, EI1520, I1521, E240, E248, E249, C740 (primary or secondary diagnosis) P4571, P4572, P4581, P4582, Q2501, Q2502, R3512, Q1591, Q1592 [67]
Measurement code: ≥2 C3211, C3212, C3213, C3231, C3232, C3233, C3234, C3235, C3239 (≥ preoperative)
Adrenal Cushing’s syndrome Cushing’s syndrome (E240, E248, E249, E270) C741 (malignant neoplasm of adrenal gland medulla) ectopic CS (E243) or pituitary CS (E240) Unilateral or bilateral adrenalectomy (P4571, P4572) [68]
Exclusion: Pituitary gland surgery (S4633, S4743)
Primary aldosteronism ≥2 E26, I15.20, or I15.21 (primary or secondary diagnosis) E26.1 (secondary hyperaldosteronism) (Saline infusion test or the captopril challenge test) or adrenalectomy or prescription of spironolactone for >6 months [70]
Congenital adrenal hyperplasia ≥2 V115, E25/E25.0/E25.8/E25.9 (primary or secondary diagnosis) Glucocorticoid or fludrocortisone for more than 6 months [69]

Operational Definitions of Thyroid Diseases

Disease Operational definition (based on 2023 EnM review) Recommended refinements based on verified 2023–2025 studies ICD-10 code Reference
Overall thyroid cancer Defined by ICD-10 code C73 combined with thyroidectomy procedure codes or registration in the cancer-specific copayment reduction program (1) Incident case definition using a washout period (≥1–2 years) is recommended to exclude prevalent cases in nationwide cohorts. (2) Survivorship analyses should consider postoperative hypoparathyroidism as a key stratification factor, given its association with long-term systemic complications. (3) When mortality is evaluated, explicit linkage to Statistics Korea national death records should be stated. C73 [1,72,79,80]
Hypothyroidism Defined by ICD-10 codes E02 or E03 with levothyroxine prescriptions on ≥2 occasions or ≥180 days (1) To reduce misclassification, patients receiving levothyroxine for thyroid cancer-related TSH suppression should be excluded or analyzed separately. (2) Claims-based definitions require transparent reporting and sensitivity analyses due to the absence of biochemical data. E02, E03 [77]
Graves’ disease/ hyperthyroidism Defined by ICD-10 code E05 with antithyroid drug prescriptions for ≥60–180 days (1) Nationwide NHIS studies emphasize clear classification of initial treatment modality (antithyroid drugs, radioactive iodine, or surgery). (2) For outcome studies involving radioactive iodine, explicit exposure definitions and appropriate non-RAI comparators are essential to minimize bias. E05 [73-76]
Subacute thyroiditis Defined by repeated ICD-10 code E06.1 plus erythrocyte sedimentation rate testing (1) In post–COVID-19 analyses, the infection index-date and latency window should be clearly defined. (2) Requiring repeated diagnostic codes improves specificity in claims-based epidemiologic studies. E06.1 [78]

EnM, Endocrinology and Metabolism; ICD-10, International Classification of Diseases 10th Revision; TSH, thyroid-stimulating hormone; RAI, radioactive iodine ablation; COVID-19, coronavirus disease 2019.