National Health Insurance Data Analysis for the Second-Wave Development of Korean Medicine Clinical Practice Guidelines in South Korea

South Korea has a unique dual medical system, which includes traditional medicine similar to China and Taiwan [ 1 ]. Korean medicine (KM), which provides medical care with treatments such as acupuncture, herbal medicine, moxibustion, cupping therapy, and chuna manual therapy, is used to treat diseases and enhance public health in South Korea [ 2 ]. However, the “lack of scientific evidence” reduces the public trust in KM, even though the prevalence of KM use in Korea is more than 60% [ 3 ]. The relatively limited evidence versus the high usage rate is a challenge that complementary and integrative medicine (CIM) also faces worldwide [ 4 ]. In this context, a national project (Korean Medical Standard Clinical Practice Guideline Project, 2016-2022) to develop and implement the evidence-based KM clinical practice guidelines (KM-CPGs) was initiated in 2016, funded by the Ministry of Health and Welfare, Republic of Korea, based on the third National Comprehensive Plan for KM Development (2016-2020). At the time of project implementation, 30 targeted diseases for the KM-CPGs development were selected by comprehensively reviewing the strengths and limitations of KM, the public use of KM, and the possibility of public or private insurance coverage. The KM-CPGs for the 30 diseases were developed in 2021 [ 5 ], the first-wave development of KM-CPGs. Health authorities decided to develop novel CPGs in the newly funded Korean Medicine Innovative Technology Development Project. Therefore, new candidate diseases have been proposed for the second-wave development of KM-CPGs by the National Institute for Korean Medicine Development (NIKOM). Prioritizing the diseases for CPGs is a crucial issue for both medical professionals and policymakers; however, there is no evidence regarding the status of KM healthcare utilization for those diseases in South Korea. In this study, we analyzed the demand and economic importance of the candidate diseases for the second-wave development of KM-CPGs in real-world clinical settings using the Health Insurance Review and Assessment Service (HIRA)’s National Patient Sample (NPS) data, which can be used as a criterion for the second-wave development of future KM-CPGs. This analysis may elucidate KM usage in South Korea, which is one of the countries where CIM is most popular. It can be referenced for policymaking in countries that use the same treatments. This study was conducted with support from the Korean Medicine Innovative Technology Development Project under NIKOM, which supervises the second-wave development of KM-CPGs. This study aimed to investigate the development priorities of the candidate diseases in the second-wave development of KM-CPGs.

The top five diseases with the highest annual number of total visits were: sciatica (135,894.5), adhesive capsulitis shoulder (29,768), tinnitus (13,256.5), total cancer (12,241), and cervical herniated intervertebral disk (c-HIVD) (11,495). The top five diseases with the highest annual number of outpatient visits were similar: sciatica (135,335.5), adhesive capsulitis shoulder (29,610.5), tinnitus (13,232.5), c-HIVD (11,256), and total cancer (10,410.5). However, the top five diseases with the highest annual number of inpatient visits were as follows: cerebral palsy (3,058), total cancer (1,830.5), sciatica (559), fracture (524), and breast cancer (503) ( Fig. 1 ). The top five diseases with the highest number of outpatients and the highest total number of patients per year were as follows: sciatica (19,595 and 19,818), adhesive capsulitis shoulder (5,724.5 and 5,769), gastritis (2,243.5 and 2,255.5), c-HIVD (1,909.5 and 2,085.5), and tinnitus (1,719.5 and 1,725). However, the top five diseases with the highest annual inpatients were: cerebral palsy (581.5), total cancer (513.5), sciatica (223), fracture (206), and c-HIVD (176) ( Fig. 2 ). The top five diseases with the highest annual total healthcare expenditure per patient were: cerebral palsy (4,393.42); kidney, bladder, and urinary tract cancer (3,890.95); total cancer (3,702.13); breast cancer (3,532.49); and fracture (3,424.21). The top five diseases with the highest annual outpatient expenditure per patient were: cerebral palsy (970.73); head and neck cancer (874.96); female genital cancer (778.11); kidney, bladder, and urinary tract cancer (752.93); and liver and bile duct cancer (743.33). The top five diseases with the highest annual inpatient healthcare expenditure per patient were: cerebral palsy (3,422.69); kidney, bladder, and urinary tract cancer (3,138.03); fracture (3,018.88); total cancer (3,010.36); and breast cancer (2,806.56) ( Fig. 3 ). The top five diseases with the highest annual total healthcare expenditure per institution were: sciatica (3,027,944.64), adhesive capsulitis of the shoulder (670,129.15), total cancer (602,798.11), c-HIVD (494,857.70), and cerebral palsy (429,471.7). The top five diseases with the highest annual outpatient treatment expenditure per institution were: sciatica (2,851,550.41), adhesive capsulitis of the shoulder (639,588.84), c-HIVD (282,119.59), tinnitus (271,948.11), and total cancer (234,424.62). However, the top five diseases with the highest annual inpatient treatment expenditure per institution were: total cancer (368,373.49), cerebral palsy (367,273.64), c-HIVD (212,738.11), fracture (188,514.92), and sciatica (176,394.23) ( Fig. 4 ).

According to the analysis of data from the HIRA-NPS, 2017-2018, musculoskeletal disorders, including sciatica and adhesive capsulitis of the shoulder, were the leading topics in terms of the number of visits and patients and treatment expenditure per institution. However, cerebral palsy was a more important topic in terms of inpatient clinical settings than musculoskeletal conditions or cancer and had the highest treatment expenditure per patient. Furthermore, fractures are highly important in inpatient clinical settings. This study highlights the gap between the real-world clinical setting and the research field. This study guides the second-wave development of future KM-CPGs.

According to the analysis of the HIRA-NPS, musculoskeletal conditions, including sciatica and adhesive capsulitis of the shoulder, yielded the highest number of KM visits and patients during 2017-2018. Among them, sciatica was a high proportion of the total number of visits and patients (52.05% of the total number of visits and 48.34% of the total number of patients). The total number of visits and patients were mostly correlated with outpatient settings. Based on the amount of healthcare utilization, musculoskeletal disorders ranked high in terms of healthcare expenditure per institution. That is, the treatment expenditure per institution for sciatica accounted for 42.12% of the total amount. This is consistent with previous studies that indicated CIM treatments are commonly used for musculoskeletal conditions [ 7 ]. However, in addition to musculoskeletal disorders, some important topics were presented at KM institutions. Cerebral palsy ranked the highest for inpatient healthcare utilization (36.03% of the total number of inpatient visits and 24.55% of the total number of inpatient patients), followed by total cancer. Based on the healthcare expenditure per patient, cerebral palsy ranked highest, followed by some cancer types, including kidney, bladder, and urinary tract cancer; breast cancer; and stomach cancer. A recent cross-sectional Canadian study has found that 27% of patients with cerebral palsy used CIM [ 8 ], which was supported by our findings, especially in the inpatient setting in South Korea. Fractures ranked high in terms of the number of inpatient visits and inpatient patients (fourth place in both cases) and inpatient treatment expenditure per patient (third place) and institution (fourth place). In a 2007 study, Canadian researchers found that the rate of CIM use in patients with fractures was 35% [ 9 ]. However, the use of CIM in patients with fractures is not well studied, with more research focusing on osteoporosis [ 10 ], a major cause of fractures. Therefore, our findings highlight this gap between real-world clinical and research in fracture. In this sample, no patients who visited the KM medical institution of interest had influenza A virus infection or posttraumatic stress disorder (PTSD). Because the HIRA-NPS includes 3% of all patients in South Korea [ 6 ], the use of KM for influenza A virus infection or PTSD may have been omitted. Nevertheless, the use of KM for these conditions is thought to be small compared with that for other diseases. There are many studies on CIM, particularly herbal medicine, for acute infectious diseases, including influenza A virus infection [ 11 ], and their importance is emphasized in the COVID-19 era [ 12 ]. PTSD is also a popular research topic in the CIM field [ 13 ]. However, the use of KM for these purposes in KM institutions is rare, indicating another gap between the real-world clinical and research fields. In this study, we analyzed data from the HIRA-NPS from 2017 to 2018 to investigate the demand and economic importance of disease candidates for the second-wave development of KM-CPGs in real-world clinical settings in South Korea. We were able to prioritize the second-wave development of future KM-CPGs. However, this study had the following limitations. First, because this study is based on data from 2017 to 2018, its generalizability to other years is limited. Additionally, since this data was generated before the ICD code for COVID-19 was generated, the HIRA-NPS analysis for COVID-19 was impossible in this study. Second, we did not perform an analysis according to the type of KM treatment. Therefore, the importance and demand for KM treatments, such as acupuncture, moxibustion, and herbal medicine, performed by KM doctors were not revealed in this study. Finally, since this study did not analyze the types of KM institutions, including KM clinics, KM hospitals, long-term care hospitals, and public health centers, the demand and economic importance of candidate topics in each KM institution type were not analyzed.

We selected the candidate diseases proposed for the second-wave development of KM-CPGs based on the previous two planning projects conducted by NIKOM in 2019 and 2020 funded by the Ministry of Health and Welfare. The ‘Advance Planning Study for Korean Medicine Innovative Technology Development Project’ conducted in 2019 suggested 29 candidates, including frozen shoulder and asthma. The ‘Demand Survey of Priority Diseases for the Development of Korean Medicine Clinical Practice Guidelines’ conducted in 2020 suggested 10 candidates, including coronavirus disease 2019 (COVID-19). Disability (e.g., sequelae of cerebral palsy), was added as a candidate by NIKOM based on policy necessity, and a total of 34 candidates were listed, excluding overlapping candidates ( Table 1 ). The HIRA-NPS is a secondary data source stratified by patient units according to gender and age group for one year from the date of commencement of care after removing information about individuals and corporations from the National Health Insurance (NHI) claims raw data [ 6 ]. These are datasets that sampled 3% of all patients in Korea and contain data on approximately 1.4 million individuals per year [ 6 ]. This sample is representative of the entire patient population in South Korea and is generalizable to the entire South Korean population [ 6 ]. In this study, we analyzed data from the HIRA-NPS from 2017-2018. The dataset was provided by HIRA at the request of NIKOM in October 2020, with the purpose of establishing the development priorities of the candidate diseases into the second-wave development of KM-CPGs. This study was approved by the Institutional Review Board (IRB) of Dongeui University Korean Medicine Hospital (IRB No. DH-2022-04; approved on April 18, 2022). The need for patient consent was waived as this dataset includes untraceable and deidentified secondary claims. We matched all disease candidates with the related codes of the Tenth Revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10) before the data analysis. Among the suggested 34 disease candidates, we classified 20) cancer further as 20-1) total cancer; 20-2) liver and bile duct cancer; 20-3) thyroid cancer; 20-4) bone cancer; 20-5) brain tumor; 20-6) head and neck cancer; 20-7) lymph cancer; 20-8) esophageal cancer; 20-9) kidney, bladder, and urinary tract cancer; 20-10) female genital cancer; 20-11) stomach cancer; 20-12) breast cancer; 20-13) prostate cancer; 20-14) colorectal cancer; 20-15) pancreatic cancer; 20-16) skin cancer; 20-17) hematological cancer; and 20-18) respiratory system cancer. Some KM-CPGs for cancers from the first-wave development of KM-CPGs have been published for both specific and total cancers (numbers 21 and 29 on the first-wave development in Table 1 ). Therefore, in this analysis, both total cancer and specific cancer were analyzed to support both development strategies. Since COVID-19 was listed in the ICD-10 from March 2020, we excluded it from this analysis because it was not included in the 2017-2018 HIRA-NPS. Finally, we analyzed 50 topics in this study. The ICD codes of the topics were defined for the HIRA-NPS analysis ( Table 2 ). The criteria for the analysis of the 50 topics were as follows: 1) period: 2017-2018; 2) diagnosis: only allowed if the topics of interest are included in the primary diagnosis, but not the secondary diagnosis; 3) institutions: KM primary clinics, KM hospitals, long-term care hospitals, and public healthcare centers where KM medical services are provided; 4) patient type: outpatient, inpatient, and total; and 5) treatment type: any type of insured KM treatment, including acupuncture, moxibustion, and herbal medicine. The following four outcomes were analyzed: annual total number of visits, annual total number of patients, healthcare expenditure (1,000 KRW) per patient, and healthcare expenditure (1,000 KRW) per institution. At this time, the “healthcare expenditure” used in the cost analysis means out-of-pocket money and the NHI’s contribution due to the characteristics of health insurance data. The annual results for all outcomes were calculated using the mean values of two years’ data (2017-2018). We performed all statistical analyses using R for Windows 4.1.3 (Ross Ihaka and Robert Gentleman, New Zealand). Finally, we visualized the results with Excel heatmaps and data bars (Microsoft Corp., Redmond, WA, USA).