Herbal decoctions for dysmenorrhea under universal health coverage pilot project: Evidence from a nationwide claims database in the Republic of Korea.

Kyeore Bae: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writing – original draft, Visualization. Minjung Park: Methodology, Writing – review & editing. Junhyeok Yi: Conceptualization, Resources, Writing – review & editing, Funding acquisition.

This research was approved by the Public Institutional Review Board Designated by the Ministry of Health and Welfare (MoHW) in Seoul, Republic of Korea (P01–202,409–01–030). Participants in the pilot project provided consent for the use of their de-identified medical information and claims data for statistical analysis and policy research. 27

This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare, Republic of Korea (grant number: RS-2020-KH087602 ). The funders had no role in study design, investigation of data, statistical analyses, interpretation of data, and preparation of the manuscript.

This study was conducted following the REporting of Studies Conducted using Observational Routinely Collected Health Data (RECORD) statement. 26 This study used pre-merged NHI claims data from the Health Insurance Review and Assessment Service (HIRA) and the Ministry of Health and Welfare, combining regular claims and the pilot project checklist items at the patient level. Regular claims included data on patient demographics, diagnostic codes, and medical services utilized by the patients. To ensure appropriate pilot project claims, the Standardized Diagnostic Checklist for HDs—which includes information on symptoms, severity, onset date, and AEs—is submitted along with claims to assess patient conditions. TKM clinicians are required to document the symptoms for each HD prescription and monitor AEs during follow-up visits, even for single-cycle prescriptions. 27 As regular NHI claims databases do not capture non-claimable clinical outcomes, the evaluation of treatment effectiveness and safety requires the use of operational definitions based on diagnostic, medications, and procedure codes. However, operationally defining short-term pain reduction as an effectiveness outcome using analgesic prescription codes is nearly impossible due to the widespread use of over-the-counter painkillers for dysmenorrhea. 9 In this context, the checklist data collected for administrative purposes included items on symptom severity and AEs, which could be utilized for effectiveness and safety analyses. All data have been de-identified. Additionally, the monthly number of patient receiving insurance-covered medical services for dysmenorrhea (2014– 2023) were obtained from the HIRA big data portal. 28 Population statistics were sourced from the Ministry of the Interior and Safety. 29 A retrospective observational study design was employed to evaluate the effectiveness and safety of HD under the UHC pilot project. Additionally, time series analyses were conducted to identify the temporal trends and effects of the pilot project on access to TKM services. Patients were defined as those who agreed to participate in the pilot project conducted in primary care outpatient settings and had received at least one HD prescription for dysmenorrhea. Diagnoses were identified using the International Classification of Diseases, Tenth Revision, including primary dysmenorrhea [N944], secondary dysmenorrhea [N945], and unspecified dysmenorrhea [N946], with pilot project coding (S027), during the first phase of the pilot project, from November 1, 2020 to April 28, 2024. Since the target condition was dysmenorrhea, only female individuals were considered. Patients who did not receive HDs or participate in the pilot project were excluded. The study population ensured generalizability, as the pilot project participants were fully included in the NHI system’s database, where enrollment covers 98 % of the Korean population. 25 Because the pilot project guidelines specified different data collection criteria for symptom severity and AEs, the characteristics of the patients and effectiveness of HDs were analyzed in patients with at least two prescriptions, whereas the safety analysis included all patients prescribed HDs ( Fig. 1 ). Fig. 1 Recording criteria for symptom severity and definition of exposure period in the first phase of the pilot project. *Abbreviations: d, Days; Rx, Prescription course. X indicates the day of herbal decoction prescription. The black line represents the prescription days per cycle. The red line indicates the last course of prescription, while the red X marks the last day of herbal decoction prescription, which corresponds to the last recorded symptom severity. The exposure period ( blue dashed line ) represents the total number of days of herbal decoction prescriptions before the last recorded symptom severity. Fig 1 Recording criteria for symptom severity and definition of exposure period in the first phase of the pilot project. *Abbreviations: d, Days; Rx, Prescription course. X indicates the day of herbal decoction prescription. The black line represents the prescription days per cycle. The red line indicates the last course of prescription, while the red X marks the last day of herbal decoction prescription, which corresponds to the last recorded symptom severity. The exposure period ( blue dashed line ) represents the total number of days of herbal decoction prescriptions before the last recorded symptom severity. The statistics of all patients with dysmenorrhea (with diagnostic codes N944, N945, and N946) who availed conventional or TKM services between 2014 and 2023 were utilized for the time series analysis. The baseline characteristics of the patients included age, diagnostic code, treatment initiation date, recent onset data, Traditional, Complementary and Integrative Medicine (TCIM) pattern differentiation, and prescription details. Owing to the absence of categorized symptom and severity items in the checklist before January 2022, effectiveness-related data were obtained only from patients who participated after the revision of the Guideline for the Herbal Decoction NHI Benefit Coverage Extension Pilot Program. 27 Symptom severity was rated on a scale of 1 to 5. Moderate severity was defined as a score of 3, and severe severity as scores of 4–5 on the symptom scale, corresponding to 5–6 and 7–10, respectively, on the visual analog scale (VAS, 0–10), 30 through linear transformation. The original checklist categorized AEs into gastrointestinal, neurological, dermatological, and other systems. However, multiple events were sometimes reported within a single entry and categorized inappropriately. To address this issue, the reported AEs were standardized using Common Terminology Criteria for Adverse Events terms and reclassified according to the Medical Dictionary for Regulatory Activities System Organ Class for broader organ system categorization. 31 Exposure was defined as the use of HDs and was quantified by the number of prescription days. As the actual days of intake were unavailable in the claims data, HD exposure was used as a surrogate indicator of HD intake. Since HDs were first recorded in claims data in November 2020, all patients were considered new users, making it infeasible to define a washout period. Additionally, because the pilot project guidelines prohibited the concurrent prescription of HDs with other herbal preparations, it was assumed that the observed effectiveness could be primarily attributable to HDs. 27 The obtained claims data did not include information on other conventional medications or procedures such as acupuncture. For continuous data, means and standard deviations (SDs) were calculated if the data followed a normal distribution; otherwise, medians and interquartile ranges (IQRs) were used. Categorical data were presented as frequencies and percentages. Changes in symptom severity were assessed using paired t -tests or Wilcoxon signed-rank tests. As an additional analysis, baseline characteristics between secondary and non-secondary dysmenorrhea were compared using chi-squared or Fisher’s exact test for categorical variables and Student’s t -test or Mann-Whitney U test for continuous variables. To estimate symptom improvement over the exposure period, a linear mixed-effects model (LMM) was employed, accounting for participant variability as a random effect. Based on previous research, age and the presence of secondary dysmenorrhea, identified as potential effect modifiers, were included as fixed effects. Because expert opinions indicated that the diagnostic codes N944 and N946 are not clearly distinguished in real-world clinical practice, only the presence of secondary dysmenorrhea (diagnostic code: N945) was considered. The rate of AEs was calculated based on the number of affected patients. To identify the impact of the UHC pilot project, time series data was analyzed using Seasonal-Trend decomposition using Locally estimated scatterplot smoothing and structural change point detection through F-statistics. 32 , 33 Monthly patient counts for insurance-covered medical services were standardized per 100,000 population. Statistical analyses were conducted using R version 4.4.1, with ‘lme4’ and ‘strucchange’ packages. 33 , 34

During the first phase of the pilot program, 39,574 patients received insurance-covered HDs for dysmenorrhea. Among them, 23,416 (59.2 %) received one prescription. Based on the available data, a descriptive characteristic analysis was conducted for 16,158 patients (40.8 %), and an effectiveness analysis was performed for 7016 patients (17.7 %) whose symptom severity data were collected via the checklist ( Fig. 2 ). Fig. 2 Flow diagram of subjects selection. Fig 2 Flow diagram of subjects selection. The median age of patients was 30 years (IQR: 21–40 years), with patients aged 20–29 representing the largest proportion (26.6 %). The median number of prescriptions was 3 (IQR: 2–5) ( Table 1 ). Table 1 Baseline characteristics of patients ( N = 16,158). Table 1 Characteristics Median (IQR); Number (Proportion) Age (years) 30.0 (21.0 - 40.0)  0–9 5 (0.0 %)  10–19 3504 (21.7 %)  20–29 4306 (26.6 %)  30–39 4204 (26.0 %)  40–49 3634 (22.5 %)  50–59 501 (3.1 %)  60+ 4 (0.0 %) Primary diagnostic code  N944 (Primary dysmenorrhea) 4100 (25.4 %)  N945 (Secondary dysmenorrhea) 488 (3.0 %)  N946 (Dysmenorrhea, unspecified) 11,570 (71.6 %) Time-to-treatment initiation (treatment initiation date-recent onset date) 11.0 (5.0 - 22.0) Pattern differentiation  Qi stagnation and blood stasis pattern 8393 (51.9 %)  Qi and blood deficiency pattern 4334 (26.8 %)  Cold dampness retention and obstruction pattern 1851 (11.5 %)  Deficiency of the liver and kidney pattern 772 (4.8 %)  Others 808 (5.0 %) Number of prescriptions 3.0 (2.0 - 5.0)  2–4 prescriptions 12,115 (75.0 %)  5–7 prescriptions 2213 (13.7 %)  8–10 prescriptions 905 (5.6 %)  ≥10 prescriptions 925 (5.7 %) Total prescription period (days) 30.0 (20.0 - 40.0) Herbal prescription   Hyeon-bu-i-gyeong-tang 2039 (12.6 %)   Modified Jo-gyeong-tang 1360 (8.4 %)   Modified Gung-gui-tang 1038 (6.4 %)   Modified O-jeok-san 1026 (6.3 %)   Dang-gui-jak-yak-san 987 (6.1 %)   Modified Gye-ji-bok-ryeong-hwan 844 (5.2 %)   Modified Dang-gui-jak-yak-tang 749 (4.6 %)   Bo-jung-ik-gi-tang 740 (4.6 %)   Dae-yeong-jeon 723 (4.5 %)   On-gyeong-tang 570 (3.5 %) Main symptoms *  Pain 5780 (82.4 %)  Systemic symptoms 949 (13.5 %)  Impairment in daily activities 228 (3.2 %)  Others 59 (0.8 %) Symptom severity (1–5) * 4.0 (3.0 - 4.0)  1 82 (1.2 %)  2 634 (9.0 %)  3 2216 (31.6 %)  4 2578 (36.7 %)  5 1506 (21.5 %) ⁎ The baseline main symptoms and symptom severity were recorded for 7016 patients after the revision of Guideline for Herbal Decoction NHI Benefit Coverage Extension Pilot Program was implemented in January 2022. †Continuous variables were non-normally distributed. IQR, interquartile range. Baseline characteristics of patients ( N = 16,158). The baseline main symptoms and symptom severity were recorded for 7016 patients after the revision of Guideline for Herbal Decoction NHI Benefit Coverage Extension Pilot Program was implemented in January 2022. †Continuous variables were non-normally distributed. IQR, interquartile range. Regarding the main symptoms, 82.4 % of the patients reported experiencing pain, 13.5 % reported systemic symptoms, and 3.2 % reported impairments in daily activities. Most patients (89.8 %) exhibited moderate to severe symptoms, suggesting that individuals with more severe conditions were more likely to participate in the pilot project. Baseline symptom severity was worse in secondary dysmenorrhea compared to non-secondary dysmenorrhea ( p < 0.001) (Supplementary Material). In 7016 patients with recorded symptom severity before and after treatment, the median severity score significantly decreased from 4 (IQR 3–4) at baseline to 3 (IQR 3–4) after treatment (Wilcoxon signed-rank test, p < 0.001). The results of the LMM analysis initially revealed a significant main effect of exposure time to HDs (10-day intervals) to HD, with dysmenorrhea severity decreasing by 0.050 points per 10 days (Model 1). After adjusting for age and secondary dysmenorrhea, the effect of the exposure time to HDs remained significant (β = –0.072, p < 0.001) (Model 3). This suggests that for every additional 10 days of HD exposure, the predicted dysmenorrhea severity decreased by 0.072 points corresponding to a reduction of approximately 0.18 points on the VAS (0–10). Secondary dysmenorrhea was significantly associated with more severe symptoms (β = 0.263, p < 0.001). Interaction effects revealed that the symptom improvement effect of the exposure period was attenuated as the age increased (β = 0.0007, p < 0.001). Conversely, patients with secondary dysmenorrhea demonstrated a greater reduction in symptoms over time (β = −0.048, p < 0.001) ( Table 2 , Fig. 3 ). Table 2 Result of a linear mixed-effects model estimating dysmenorrhea severity over the exposure period of herbal decoction. Table 2 Fixed effects Model 1: Exposure period only Model 2: + Age & Secondary dysmenorrhea Model 3: + Interaction terms Intercept 3.430 (0.010) *** 3.641 (0.028) *** 3.678 (0.028) *** Exposure period (10 days intervals) −0.050 (0.001) *** −0.050 (0.001) *** −0.072 (0.003) *** Age – −0.007 (0.001) *** −0.008 (0.001) *** Presence of secondary dysmenorrhea – 0.161 (0.060) ** 0.263 (0.062) *** Exposure period (10 days intervals) x Age – – 0.0007 (0.0001) *** Exposure period (10 days intervals) x Presence of secondary dysmenorrhea – – −0.048 (0.0068) *** Standard error values are presented in parentheses. *** p < 0.001, ** p < 0.01. Fig. 3 Prediction of symptom severity changes based on the linear mixed-effects model (LMM). The squares represent the mean symptom severities for each exposure period, while the black vertical lines indicate the standard deviation. The red solid line shows the predicted symptom severity based on Model 3, which incorporates the exposure period, age, secondary dysmenorrhea, and interaction terms in the linear mixed-effects model. The x-axis represents the exposure period in 10-day intervals, with values outside the 0 to 100-day range omitted for clarity. Fig 3 Result of a linear mixed-effects model estimating dysmenorrhea severity over the exposure period of herbal decoction. Standard error values are presented in parentheses. *** p < 0.001, ** p < 0.01. Prediction of symptom severity changes based on the linear mixed-effects model (LMM). The squares represent the mean symptom severities for each exposure period, while the black vertical lines indicate the standard deviation. The red solid line shows the predicted symptom severity based on Model 3, which incorporates the exposure period, age, secondary dysmenorrhea, and interaction terms in the linear mixed-effects model. The x-axis represents the exposure period in 10-day intervals, with values outside the 0 to 100-day range omitted for clarity. Of the 39,574 patients, 231 (0.58 %) experienced AEs (345 cases). Gastrointestinal disorders including diarrhea, dyspepsia, and nausea were the most frequent AEs. Headache and rash acneiform were reported in 27 and 24 cases, respectively ( Table 3 ). Table 3 Types and frequency of reported adverse events. Table 3 System organ categorization Adverse event Number of cases Cardiac disorders 2 Palpitations 2 Gastrointestinal disorders 256 Diarrhea 81 Dyspepsia 77 Nausea 24 Abdominal pain 21 Flatulence 21 Constipation 17 Bloating 6 Gastrointestinal disorders - Other, specify 5 Vomiting 2 Dry mouth 1 Gastritis 1 General disorders and administration site conditions 13 Fatigue 3 Fever 3 Edema limbs 2 Generalized edema 2 General disorders and administration site conditions - Other, specify 1 Non-cardiac chest pain 1 Pain 1 Hepatobiliary disorders 1 Hepatobiliary disorders - Other, specify 1 Metabolism and nutrition disorders 4 Anorexia 4 Musculoskeletal and connective tissue disorders 2 Back pain 2 Nervous system disorders 27 Headache 17 Dizziness 8 Hypersomnia 2 Psychiatric disorders 8 Insomnia 8 Reproductive system and breast disorders 8 Dysmenorrhea 3 Irregular menstruation 3 Uterine hemorrhage/Irregular menstruation 1 Vaginal discharge 1 Skin and subcutaneous tissue disorders 24 Rash acneiform 12 Dry skin 2 Eczema 2 Hyperhidrosis 2 Pruritus 2 Urticaria 2 Skin and subcutaneous tissue disorders - Other, specify 1 Skin hyperpigmentation 1 Types and frequency of reported adverse events. In one patient, 'liver-specific allergy' was documented, but detailed information regarding the actual occurrence, laboratory results, or drug causality could not be confirmed. Owing to the limitations of the checklist system, the date of AE reporting and grade of AEs were not recorded, making it impossible to comprehensively evaluate the causality of the reported AEs. After excluding the decomposed seasonal and remainder components, a significantly steep increase in the number of patients with dysmenorrhea receiving TKM services was observed at the end of 2020. Conventional medical services showed a gradual upward trend, except for a temporary decline during the first half of 2020 ( Fig. 4 ). Regarding TKM services, structural change points were detected in October 2020, just before the implementation of the UHC pilot program, for primary dysmenorrhea ( F = 16.71) and unspecified dysmenorrhea ( F = 37.44). Additionally, a breakpoint was identified in September 2017 for secondary dysmenorrhea ( F = 34.40) (Supplementary Material). These findings suggest a potential shift in the utilization of TKM services following the introduction of the pilot project for dysmenorrhea. Fig. 4 10-year medical service utilization changes based on time series decomposition: (a) Traditional Korean medicine services, (b) Conventional medicine services. The black solid line represents N944 (primary dysmenorrhea ), the blue solid line represents N945 (secondary dysmenorrhea), and the red solid line represents N946 (unspecified dysmenorrhea). In the time-series plot of patients per 100,000 population (monthly), the black, blue, and red dashed vertical lines indicate the structural break points for the patient count trends of N944, N945, and N946, respectively. Fig 4 10-year medical service utilization changes based on time series decomposition: (a) Traditional Korean medicine services, (b) Conventional medicine services. The black solid line represents N944 (primary dysmenorrhea ), the blue solid line represents N945 (secondary dysmenorrhea), and the red solid line represents N946 (unspecified dysmenorrhea). In the time-series plot of patients per 100,000 population (monthly), the black, blue, and red dashed vertical lines indicate the structural break points for the patient count trends of N944, N945, and N946, respectively.

The authors declare that they have no conflicts of interest.

The present study, using NHI claims data from the entire first phase of the pilot project, demonstrated a tendency for patients with moderate-to-severe dysmenorrhea to receive HD prescriptions. HDs exhibited a manageable safety profile, with gastrointestinal symptoms being the most commonly reported AEs. This study demonstrated that HD significantly reduced dysmenorrhea symptom severity over time. The effect remained significant after adjusting for age and the presence of secondary dysmenorrhea, with HD exposure being associated with a 0.072-point decrease in symptom severity every 10 days, corresponding to a 0.18-point reduction in the VAS. Given that dysmenorrhea treatments, including NSAIDs, hormonal therapy, and TCIM interventions, are typically applied and evaluated over three consecutive menstrual cycles, 17 , 35 , 36 a cumulative 1.62-point reduction in the VAS can be expected over a 3-month period. In this study, customized HDs improved symptoms of both secondary and non-secondary dysmenorrhea symptoms, but patients with secondary dysmenorrhea demonstrated a greater reduction in symptoms over time. This variation may be explained by the distinct pathophysiological characteristics and the potential mechanisms of HDs in secondary and non-secondary dysmenorrhea. The main pathology of primary dysmenorrhea is associated with elevated levels of prostaglandin F2α and prostaglandin E2 in the endometrium, leading to uterine hypercontraction and resulting in ischemia and hypoxia of the uterine muscle. 37 , 38 , 39 A previous review proposed that the analgesic effects of HDs on menstrual pain may be mediated through the reduction of prostaglandin levels and the inhibition of excessive uterine contractions. This review also suggested that herbs with blood-activating and stasis-resolving properties may improve uterine blood flow. 40 In secondary dysmenorrhea associated with underlying anatomic abnormalities, the pathology is more complex. Increased production of reactive oxygen species, prostaglandins, cytokines, chemokines, and nerve growth factors induces an inflammatory environment in the peritoneal cavity and enhances nerve sensitivity. 41 , 42 Regarding endometriosis-associated dysmenorrhea, reductions in cytokine and prostaglandin levels have been observed in both clinical studies and in vivo studies. Herbal medications have been shown to inhibit angiogenesis, adhesion, and cell proliferation pathways. 43 A Cochrane review demonstrated that NSAIDs significantly relieved pain in primary dysmenorrhea (odds ratio: 1.29, 95 % CI: 1.11–1.51), but 11–14 % of women experienced AEs. 44 Another Cochrane review on oral contraceptive pills (OCPs) showed a moderate reduction in pain compared with placebo (standardized mean difference: −0.58, 95 % CI: −0.74 to −0.41), corresponding to an improvement of 0.8 points on the Total Dysmenorrhea Score (0–6). However, OCPs increased the risk of AEs compared with placebo or no treatment (relative risk: 1.31, 95 % CI: 1.20–1.43), with 71–86 % of users experiencing AEs and 0.5–6.8 % experiencing serious AEs. 45 A previous cross-sectional study found that among undergraduate students with dysmenorrhea, the desire to reduce analgesic use was significantly associated with the current use of complementary and alternative therapies, such as heating pads, massage, and natural herbs (adjusted odds ratio: 4.066, 95 % CI: 2.136–7.739). 46 Although the magnitude of symptom severity improvement with HDs in this study was not substantial, a 3-month course of HDs may serve as a viable alternative for patients who do not desire, or are concerned about AEs associated with NSAIDs and hormonal therapies. Also, this provides healthcare providers with an additional treatment option, particularly for patients with contraindications to conventional medications. The significant shift in the trend of patients utilizing TKM services implies that unmet needs existed among both patients and healthcare providers before the implementation of the UHC pilot project. One possible explanation for this is that the project potentially facilitated the affordability of HDs. In 2020, household expenditures on herbal medications, including HDs, showed an approximately 2.8-fold difference across income quartiles. Specifically, households in the first income quartile spent approximately USD 183 (KRW 265,178), whereas those in the fifth income quartile spent approximately USD 504 (KRW 732,177) annually. 47 A 2022 survey of TKM institutions revealed that the costs of a single 10-day dose of non-reimbursable HDs for disease treatment ranged from approximately USD 87 (KRW 126,000) to USD 191 (KRW 277,000). 48 The Ministry of Health and Welfare reported that, through the first phase of the pilot project, the cost per patient for insured HD was reduced by approximately USD 58 (KRW 84,860) compared with non-reimbursable HDs. 49 The findings of the descriptive analysis showed that various customized HDs had been prescribed to patients who participated in the pilot project. Given that only six prescriptions— Gamisoyo-san, Daewhangmokdanpi-tang, Doinseunggi-tang, Sosiho-tang, Ojeok-san, and Palmul-tang —among the 56 pre-formulated herbal preparations had been covered by the NHI for dysmenorrhea as one of their indications between 1990 and 2020, 50 the pilot program expanded the range of covered HDs for the treatment of dysmenorrhea. This retrospective observational study has several limitations that stem from the constraints of real-world data (RWD) used. First, this study employed a single-arm design. A control group without HD prescriptions was unavailable because symptom severity and AEs were derived from the pilot project checklist. Due to the lack of a control group, it was not possible to adjust for baseline differences in symptom severity between patients who received HDs and those who did not. Although patients with more severe symptoms may have been more likely to seek HDs, this potential self-selection bias could not be controlled in the current study design. Second, the NHI claims database, primarily designed for billing purposes, lacks clinical variables such as body mass index, dietary habits, menstrual characteristics, smoking status, psychological factors, and the use of over-the-counter painkillers. 51 , 52 As important effect modifiers that may impact on the magnitude of dysmenorrhea symptoms could not be included as covariates in our LMM analysis, the omission of these variables may affect the precision of estimating treatment effects of HDs in dysmenorrhea. Additionally, unmeasured health behaviors may have led to residual confounding. Future clinical studies with comprehensive data collection are warranted to further validate and elaborate on the impact of HDs. Furthermore, due to the limitations of the pilot project data utilized for analyses, concurrent treatments such as conventional medications or procedures, including acupuncture, could not be considered. Although the results of LMM analysis showed greater improvements in secondary dysmenorrhea, the findings should not yet be interpreted as conclusive. Potential pharmacological and surgical treatments for underlying medical conditions could also not be adjusted for. Lastly, according to the administrative procedures, symptom severity was recorded at the last prescription date rather than after all HD intakes had ceased. This approach may have led to disregard of the effects of the final HD cycle and prolonged effects. Finally, the current checklist system does not capture the dates of AE occurrences, relying solely on complaint entries. This lack of information renders causal inference assessments between HDs and AEs unfeasible. All reported AEs in the analysis were included to maintain conservative estimates; however, the complaints of patients who did not revisit clinics would not be captured. Despite these limitations, the significance of this study is underscored by its nationwide scope and use of RWD, thereby enhancing its generalizability. This study highlights how policy changes in TCIM could potentially improve access to women’s health services and promote health equity. Although a recent report from Japan indicated increased outpatient visits and hormonal therapies for secondary dysmenorrhea following the implementation of a financial incentive policy for medical facilities in April 2020, 53 this is the first study to examine the impact of such policies in the TCIM field. Moreover, the inclusion of previously non-covered interventions has fostered the establishment of unprecedented large-scale RWD on HDs, which will support further evidence generation. In the first phase of the pilot project, which was limited to primary care institutions, linking electronic medical records (EMRs) to claims data was not feasible due to the risk of re-identification. However, as the second phase expands to include hospital-level institutions, 49 future studies will be able to conduct pseudonymized data linkage across claims data, national health screening data, and EMRs. This will enable the inclusion of important clinical variables such as body mass index and smoking status, thereby improving the robustness of effectiveness analyses. The findings of this study suggested that prolonged HD intake is associated with reduced dysmenorrhea symptom severity, and that the majority of AEs were gastrointestinal disorders. The observed significant increase in the number of patients utilizing TKM services indicated that the enhancement of service coverage through the HDs pilot project contributed to improved accessibility for patients with dysmenorrhea. Importantly, this nationwide-level observational study provides generalizable insights into the impacts of HDs on dysmenorrhea. However, interpretive caution is needed, as data limitations of the pilot project claims data constrained control group assessment, sufficient adjustment for clinical variables, and analysis of AE causality. Future research leveraging expanded nationwide RWD from the second phase pilot project is necessary to validate these findings.

Dysmenorrhea is one of the most common gynecological disorders, affecting 35 % to 92.9 % of women of reproductive age worldwide. 1 , 2 , 3 , 4 , 5 , 6 Dysmenorrhea is often inadequately treated, because menstrual pain is often perceived as normal and inevitable. 7 , 8 A Korean national report identified reasons for not seeking treatments for severe dysmenorrhea including reluctance due to feelings of shame (12.0 %), financial burden (4.7 %), and concern about social stigma (3.4 %). 9 Unmanaged dysmenorrhea interferes with daily activities, education, and work productivity. 8 , 10 , 11 In Japan, direct total health care costs for primary and secondary dysmenorrhea were estimated at USD 1917 and USD 2465, respectively 12 In contrast, the direct medical costs for primary dysmenorrhea in Korea were estimated to be only USD 39.80 in 2018, attributed to low healthcare utilization. 13 Although the indirect productivity losses caused by dysmenorrhea has not been clearly established, a previous study estimated productivity losses due to endometriosis at USD 10,177.54 per person per year in the United States. 14 Another Australian study estimated an annual per-person loss of 16,970 international dollars from chronic pelvic pain without endometriosis, with 75 % attributed to productivity losses. 15 Non-steroidal anti-inflammatory drugs (NSAIDs) and hormonal contraceptives, either independently or in combination, are first-line treatment options for primary dysmenorrhea for three months. For secondary dysmenorrhea, NSAIDs or hormonal therapies may be considered, but treatment of the identified abnormalities is also required. 16 , 17 , 18 Studies have demonstrated that HDs, either alone or in combination with conventional medications, improve symptoms in both primary and secondary dysmenorrhea. 19 , 20 , 21 , 22 , 23 A meta-analysis focusing on adverse events (AEs) concluded that HDs are safer compared to active controls including conventional medications. 24 In November 2020, the Korean government launched the Pilot Project for Expanding National Health Insurance (NHI) Coverage to Herbal Decoctions (HDs) to enhance public health rights and improve access to herbal medicines for dysmenorrhea. Although the NHI system already provides population coverage of approximately 98 %, service coverage and financial protection have remained limited in certain areas, particularly in traditional Korean medicine (TKM). This pilot project was expected to enhance one dimension of universal health coverage (UHC) —service availability—by lowering the financial burden of high out-of-pocket costs, as HD had not previously been reimbursed. Although NHI claims data are readily available for research, 25 nationwide assessments of HDs were not feasible before 2020 because of their non-reimbursed status in the Republic of Korea. However, the initiation of the HD pilot project enabled the collection of relevant NHI claims. 24 Therefore, this retrospective observational study aimed to assess the patient characteristics, short-term effectiveness and safety of customized HDs for dysmenorrhea using NHI claims data from the pilot project. Additionally, we analyzed the monthly healthcare service utilization data over a 10-year period to explore the impact of the UHC pilot project on healthcare accessibility for patients with dysmenorrhea.

Monthly patient counts for dysmenorrhea can be obtained from the HIRA big data portal [ https://opendata.hira.or.kr/ ]. Resident registration population statistics can be obtained from the Ministry of the Interior and Safety [ https://jumin.mois.go.kr/ ]. The pre-merged claims dataset, which combines regular NHI claims and pilot project checklist items, is not publicly available. Access is restricted to the MoHW and HIRA of the Republic of Korea for statistical analysis and policy research upon request.