Comparison of Nationwide Volumes for Five Major Gastrointestinal Surgical Procedures in Japan: Implications From Publicly Available NCD and NDB Open Data Authors: Atomu Kiriyama, Masakazu Fujii, Toshimichi Asano, Satoshi Hirano Published: 2026 ABSTRACT ABSTRACTAimIn Japan, the National Clinical Database (NCD), a clinician‐reported registry, and the National Database of Health Insurance Claims and Specific Health Checkups of Japan Open Data (NDB Open Data), an administrative claims database, are both widely used as nationwide surgical databases, but their consistency in capturing national surgical volumes has not been systematically evaluated.MethodsWe compared publicly available data from the NCD annual surgical statistics and the NDB Open Data from 2014 to 2023. Annual case numbers were extracted for five major gastrointestinal procedures: cholecystectomy, inguinal hernia repair, appendectomy, colorectal cancer surgery, and gastric cancer surgery. Consistency was evaluated using three metrics: the median NCD/NDB ratio with interquartile range, Pearson's correlation coefficient for absolute annual volumes, and the proportion of years with concordant year‐to‐year change.ResultsThe NCD generally reported higher case volumes than the NDB Open Data, with ratios ranging from close to unity (1.00) for colorectal cancer surgery to 1.10 for appendectomy. Temporal associations between the two databases (Pearson's r) ranged from 0.70 to 0.99 across the five procedures. The proportion of years with concordant year‐to‐year change ranged from 44% for appendectomy to 100% for colorectal and gastric cancer surgery.ConclusionDespite relying on clinician‐reported data, the NCD demonstrated comparable temporal patterns to the NDB Open Data across the five major gastrointestinal procedures. The stability of procedure‐specific NCD/NDB ratios and the concordance of temporal trends support the use of NCD annual statistics for monitoring broad temporal trends in these selected gastrointestinal procedures in Japan.Keywords: database, general surgery, health insurance, Japan, registries Introduction Nationwide databases with near‐complete national coverage of surgical procedures are essential for accurately characterizing past surgical demand, informing future healthcare needs, and enabling nationwide quality assessment and performance benchmarking [1, 2, 3]. In rapidly aging societies such as Japan, where regional disparities in the surgical workforce are becoming increasingly important healthcare policy challenges, accurate nationwide surgical volume data are also critical for understanding national surgical trends and informing future healthcare demand projection. Two nationwide databases currently serve as major data sources for surgical research in Japan: the National Clinical Database (NCD) and the National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB). Previous projections of future gastrointestinal surgical demand in Japan have relied on nationwide surgical volume data, highlighting the importance of understanding the structural characteristics and comparability of these data sources [4]. The NCD is a Japanese nationwide clinician‐reported web‐based data entry system for surgical procedures that was established in 2010 by the Japan Surgical Society, the Japanese Society of Gastroenterological Surgery, and eight other professional surgical societies, and has been reported to achieve high national coverage, with capture rates exceeding 95% reported for selected procedures [5, 6]. The NDB is a comprehensive administrative database based on medical insurance claims collected and published by the Ministry of Health, Labour and Welfare, covering health insurance claims from approximately 99% of all hospitals in Japan [7, 8]. Its publicly available aggregated dataset (NDB Open Data) is released annually and has been widely used for epidemiological and health services research. While both databases serve as powerful data foundations for understanding national surgical volumes, they differ substantially in their data collection methods and target scope, and the degree of consistency between them has not been systematically evaluated. In the present study, we focused on major gastrointestinal surgical procedures because they account for a substantial proportion of the national surgical workload in Japan and are closely linked to population aging and future healthcare demand [4, 5]. This study aimed to compare annual surgical volumes for major gastrointestinal procedures between the NCD and the NDB Open Data in Japan, and to evaluate whether this clinician‐reported registry captures national surgical trends comparable to those derived from a near‐complete administrative claims database. Methods Study Design This descriptive analysis compares publicly available data from the NCD annual surgical statistics and the NDB Open Data, focusing on the annual number of major gastrointestinal surgical procedures in Japan. The comparison covered the period from 2014 to 2023, for which data from both sources were publicly available. The NCD reports surgical case numbers on a calendar‐year basis (January–December), whereas the NDB Open Data are compiled by Japanese fiscal year (April–March); this structural difference in the definition of an annual reporting period was taken into account when interpreting year‐to‐year comparisons between the two databases. Data Collection and Processing Publicly available data from the NCD annual surgical statistics [9, 10, 11, 12, 13, 14, 15, 16, 17, 18] and the NDB Open Data [19, 20, 21, 22, 23, 24, 25, 26, 27, 28] were obtained for the years 2014–2023. For each database, annual surgical case numbers corresponding to the target procedures were extracted using the relevant procedure identifiers or medical fee codes. Differences in procedure names and definitions between the two databases were reviewed, and items were standardized to ensure comparability. For each procedure group, annual case counts were calculated as the sum of all corresponding NCD or NDB codes across surgical approaches and subcategories (e.g., open and laparoscopic, or simple and complicated), in order to capture the total surgical volume for that operation regardless of the specific technique used (Table 1). For colorectal and gastric cancer surgery, procedure codes explicitly designated for malignant disease were selected where available; for a subset of procedures for which the coding system does not distinguish between benign and malignant indications (e.g., rectal resection, proximal gastrectomy), the corresponding codes were nonetheless included, as these operations are performed predominantly for malignant disease in clinical practice. In the NCD, all procedures performed within a surgical episode are individually registered by the operating surgeon, whereas in the NDB, the capture of multiple or concurrent procedures is subject to billing rules governing which procedure combinations can be claimed together. Because both the NCD annual report and the NDB Open Data are published as pre‐aggregated annual case counts, the treatment of multiple or concurrent procedures reflects the registration and reporting conventions inherent to each source database, rather than a processing decision made in this study. After harmonization, annual case numbers from both databases were organized and integrated, and absolute values and year‐to‐year trends were compared. TABLE 1NCD procedure identifiers and NDB medical fee codes used for each procedure group.Procedure nameNCD codeNDB open data codeCholecystectomyLaparoscopic cholecystectomyOQ0181150254110Open cholecystectomyOQ0182150172410AppendectomyLaparoscopic appendectomy (simple)OQ0306150337610Laparoscopic appendectomy (complicated)OQ0307150272050Open appendectomyOQ0305150181610, 150337510Inguinal hernia repairLaparoscopic inguinal hernia repairOQ0073150263610Inguinal hernia repairOQ0074150159010Laparoscopic femoral hernia repairOQ0075150361310Femoral hernia repairOQ0076150159110Colorectal cancer surgeryLaparoscopic colectomyOQ0323150324910, 150442110Laparoscopic rectal resectionNQ0665, NQ0685, OQ0373150325210, 150407210, 150337810, 150407310, 150429910, 150430110, 150337910, 150407410Open colectomyNQ0548, NQ0550150181910Open rectal resectionNQ0469, NQ0744, OQ0366, OQ0367, OQ0368, OQ0369, OQ0370, OQ0371, OQ0374150187110, 150187210, 150245410, 150297510, 150420610, 150245310Gastric cancer surgeryLaparoscopic distal gastrectomyOQ0140150323510, 150406710Laparoscopic total gastrectomyOQ0141, NQ0741150323710, 150407110, 150429110Open distal gastrectomyOQ0134, OQ0135150168010Open total gastrectomyOQ0136, OQ0137, OQ0138150168110, 150429010Proximal gastrectomyNQ0605, NQ0740150337310, 150377910, 150406910 NCD procedure identifiers and NDB medical fee codes used for each procedure group. Target Surgeries The selected procedures represented high‐volume gastrointestinal surgeries in Japan and included a mixture of benign and malignant diseases, elective and emergency procedures, and upper and lower gastrointestinal surgery. Procedures requiring complex harmonization because of substantial coding instability or low annual volume were not included. The following procedure groups were selected for comparison: cholecystectomy, inguinal hernia repair, appendectomy, colorectal cancer surgery, and gastric cancer surgery. Table 1 summarizes the corresponding NCD procedure identifiers and NDB medical fee codes used to extract cases for each procedure group; the official definition of each individual code contained in these groups is listed in Table S1. Analysis Methods Given the aggregate nature of the publicly available data and the limited number of annual observations, the analysis was designed as a descriptive consistency assessment rather than a formal validation or inferential modeling study. Therefore, emphasis was placed on descriptive comparison of temporal patterns and relative consistency between databases. Annual surgical case numbers from the NCD and the NDB Open Data were compared for each procedure group. To evaluate the consistency of temporal patterns between the two databases, three complementary descriptive metrics were applied. First, the NCD/NDB ratio was calculated for each year and procedure, and temporal stability was summarized using the median and interquartile range (IQR) across the study period. Second, the temporal association of annual surgical volumes between the two databases was assessed using Pearson's correlation coefficient (r). Pearson's correlation coefficients were used descriptively to summarize similarity in temporal patterns rather than to infer statistical agreement. Third, as an intuitive descriptive indicator of temporal consistency, the proportion of years in which the two databases showed concordant directions of year‐to‐year change (i.e., both increasing and both decreasing) was calculated for each procedure. In addition, to evaluate whether the systematic differences between the two databases remained stable over time, the annual NCD/NDB ratio, the annual percentage difference between the two databases [(NCD − NDB)/NDB × 100], and the year‐to‐year percentage change in case volumes recorded separately by each database were plotted for each procedure (Figure S1). Results Cholecystectomy The median annual number of cholecystectomies was 133 012 (IQR: 129 780–134 705) in the NCD and 123 027 (IQR: 121 276–125 981) in the NDB Open Data. Annual case numbers in both databases increased gradually from 2014, peaking in 2019, before declining in 2020 during the COVID‐19 pandemic and subsequently recovering (Figure 1A). The NCD consistently reported higher case volumes than the NDB Open Data throughout the study period. Descriptive consistency metrics are summarized in Table 2. Annual surgical volumes for five major gastrointestinal procedures in Japan, comparing the NCD and NDB Open Data. (A) Cholecystectomy (2014–2023), (B) inguinal hernia repair (2018–2023), (C) appendectomy (2014–2023), (D) colorectal cancer surgery (2016–2023), (E) gastric cancer surgery (2014–2023; NCD value for 2020 omitted). Filled circles with solid lines, NCD; open squares with dashed lines, NDB Open Data. For inguinal hernia repair (B), NDB Open Data values prior to 2018 were classified under a separate aggregated billing category and are not directly comparable to subsequent years; these years are therefore omitted from the panel. For colorectal cancer surgery (D), the NCD procedure codes for laparoscopic rectal resection (NQ0665, NQ0685) were introduced in 2016, precluding comparison for 2014–2015; these years are therefore omitted from the panel. For gastric cancer surgery (E), the NCD value for 2020 is omitted because one major procedure code (OQ0140, laparoscopic distal gastrectomy for malignancy) was absent from the published NCD annual report for that year; the corresponding NDB Open Data value for 2020 is retained. Y‐axes are truncated (i.e., do not begin at zero) to facilitate visual comparison of temporal trends between databases. TABLE 2Descriptive consistency metrics for annual surgical volumes between the NCD and NDB Open Data for five major gastrointestinal procedures.Median NCD/NDB ratio (IQR)Temporal association of annual volumes (Pearson's r)Concordance of year‐to‐year changeCholecystectomy1.08 (1.06–1.08)0.7078% (7 of 9)Inguinal hernia repair a 1.08 (1.07–1.10)0.9680% (4 of 5)Appendectomy1.10 (1.08–1.12)0.7844% (4 of 9)Colorectal cancer surgery b 1.00 (1.00–1.01)0.77100% (7 of 7)Gastric cancer surgery c 1.07 (1.05–1.08)0.99100% (8 of 8) a Data restricted to 2018–2023. b Data from 2014 to 2015 excluded. c Data from 2020 excluded. Descriptive consistency metrics for annual surgical volumes between the NCD and NDB Open Data for five major gastrointestinal procedures. Data restricted to 2018–2023. Data from 2014 to 2015 excluded. Data from 2020 excluded. Inguinal Hernia Repair For inguinal hernia repair, detailed procedure‐level data in the NDB Open Data were classified under a separate aggregated category prior to 2018, precluding procedure‐specific comparisons for those years; therefore, all analyses were restricted to the period from 2018 to 2023 (Figure 1B). The median annual number of inguinal hernia repairs over this period was 142 445 (IQR: 136 050–144 969) in the NCD and 133 244 (IQR: 125 285–134 483) in the NDB Open Data. Annual case numbers in both databases declined in 2020 during the COVID‐19 pandemic, followed by a gradual recovery and subsequent increase through 2023. Descriptive consistency metrics are summarized in Table 2. Appendectomy The median annual number of appendectomies was 61 254 (IQR: 60 711–61 536) in the NCD and 54 875 (IQR: 54 233–55 634) in the NDB Open Data. Annual case numbers in both databases remained relatively stable from 2014 to 2021, followed by a decline in 2022 and 2023 (Figure 1C). The year‐to‐year concordance rate was notably lower than that of other procedures. Descriptive consistency metrics are summarized in Table 2. Colorectal Cancer Surgery For colorectal cancer surgery, two procedure codes (NQ0665 and NQ0685) were introduced in the NCD in 2016, with no equivalent codes available for prior years; thus, data from 2014 and 2015 were excluded from all analyses for this procedure. The median annual number of colorectal cancer surgeries over the 2016–2023 period was 107 354 (IQR: 106 775–108 178) in the NCD and 107 820 (IQR: 105 915–109 202) in the NDB Open Data (Figure 1D). The NCD/NDB ratio was closest to unity among all procedures examined. Descriptive consistency metrics are summarized in Table 2. Figure S1 additionally shows that the NCD/NDB ratio for colorectal cancer surgery underwent a transient reversal in 2023, coinciding with a marked increase in the NDB‐recorded volume of robotic‐assisted colectomy following the expansion of insurance coverage for this procedure in April 2022. Prior to this expansion, robotic‐assisted colectomy was not covered by public insurance and therefore lacked a corresponding NDB reimbursement code, so these cases were not captured in the NDB Open Data; because the NCD records laparoscopic and robotic‐assisted colectomy under a single procedure code regardless of insurance status, this divergence likely reflects the prior absence of NDB billing data for robotic cases rather than a change in NCD completeness. Gastric Cancer Surgery For gastric cancer surgery, the procedure code OQ0140 (laparoscopic distal gastrectomy for malignancy) was absent from the NCD annual report for 2020. As the reason for this omission was not stated in the published report and could not be independently verified from publicly available sources, data from 2020 were excluded from all analyses for this procedure to avoid potential misrepresentation of national surgical volumes. The median annual number of gastric cancer surgeries over the 2014–2023 period excluding 2020 was 48 357 (IQR: 40 028–53 805) in the NCD and 44 833 (IQR: 37 552–49 728) in the NDB Open Data. Annual case numbers in both databases showed a consistent declining trend throughout the study period (Figure 1E). Descriptive consistency metrics are summarized in Table 2. Discussion This study provides a nationwide comparison of annual surgical volumes for major gastrointestinal procedures derived from the NCD and NDB Open Data datasets, two complementary components of Japan's nationwide surgical surveillance infrastructure. It is widely recognized that no single database captures 100% of all surgical procedures at the national level. Nevertheless, the NCD, despite relying on clinician‐reported data, has been shown to achieve exceptionally high national coverage. This high level of completeness may partly reflect its integration with professional accountability and quality assurance frameworks in surgical practice. This comparison also carries policy‐relevant implications. The NCD's high national coverage has largely rested on internal validation within the surgical community [5, 6], but whether this coverage agrees with an independent, administratively derived data source used directly by policymakers had not been demonstrated. By benchmarking the NCD against the NDB Open Data, this study provides external, quantitative evidence that a clinician‐reported registry approximates trends captured by a nationwide administrative system—offering policymakers and health administrators an evidentiary basis for the credibility of NCD‐derived statistics beyond the surgical community's own assessment. The stable, procedure‐specific NCD/NDB ratios observed may also serve as practical reference values for reconciling case volumes between the two databases. In our analysis, the NCD generally reported higher surgical volumes compared with the NDB Open Data, although the ratio was close to unity for colorectal cancer surgery. Several structural characteristics of the NDB Open Data likely explain these differences. As an insurance‐based database, the NDB Open Data inherently excludes population subgroups not enrolled in standard health insurance schemes, including recipients of public assistance. In addition, because the NDB Open Data is derived from reimbursement claims, the capture of multiple or concurrent surgical procedures is constrained by billing rules and permitted claim combinations, potentially leading to underrepresentation of complex cases. In contrast, the clinician‐reported NCD records procedures at the operative level, enabling more comprehensive documentation of multiple procedures performed within a single surgical episode. Furthermore, an additional factor is that various professional societies actually conduct their evaluation of case numbers for board certification based on the cases registered in NCD. These findings suggest that the NCD captures national surgical activity at a level broadly comparable to that reflected in administrative claims data. Importantly, neither database should be regarded as an absolute reference standard, as both the NCD and NDB Open Data have inherent structural limitations related to data collection, reporting frameworks, and coverage. Rather than formal validation, the present study aimed to evaluate the degree of temporal consistency and structural comparability between these nationwide data sources. Despite these volume differences, temporal trends between the two databases were broadly concordant across major gastrointestinal procedures. This stability supports the utility of both data sources for monitoring relative changes in national surgical activity, even when absolute case counts differ. The observed concordance suggests that, although the NDB Open Data may undercount procedures in certain structural contexts, it nonetheless reflects genuine temporal changes in surgical practice at the population level. Among the five procedures examined, the year‐to‐year concordance rate varied substantially, ranging from 44% for appendectomy to 100% for colorectal and gastric cancer surgery. The notably low concordance rate for appendectomy warrants specific discussion. Unlike elective cancer surgeries, appendectomy is performed as an emergency procedure whose incidence is subject to monthly variation driven by seasonal and climatic factors [29, 30]. Given that the NCD aggregates data on a calendar‐year basis whereas the NDB Open Data are compiled by fiscal year (April to March), cases occurring in January through March are counted in the current calendar year by the NCD, but are attributed to the preceding fiscal year by the NDB Open Data. Because monthly case volumes are not constant, even modest fluctuations during this three‐month boundary window can cause the same cases to be attributed to different years across the two databases—an increase in appendectomy incidence during this period would appear as an increase in NCD year N but as an increase in NDB fiscal year N − 1, leaving NDB fiscal year N to appear relatively decreased by comparison, and thereby producing opposing year‐to‐year directional trends for the same underlying change in surgical activity. Furthermore, the absolute year‐to‐year change in appendectomy volume was small relative to other procedures, which may amplify the effect of such timing differences; when the magnitude of change is small, even minor discrepancies in case attribution between databases can result in opposite directional trends. In contrast, the high year‐to‐year concordance observed for colorectal and gastric cancer surgery may reflect the larger absolute changes in these procedure volumes over time, which are less susceptible to such timing‐related noise. This study has several limitations inherent to analyses based solely on publicly available data. First, we were unable to assess case‐level concordance between the two databases because individual‐level linkage was not possible; future studies directly linking patient‐level records would provide more definitive validation. Second, our analysis was limited to major gastrointestinal surgical procedures, and the generalizability of our findings to other surgical specialties or health care settings remains uncertain. Finally, the specific mechanisms underlying differences in coverage between the databases—such as the impact of insurance exemptions or billing rules—could not be fully elucidated and warrant further investigation. In conclusion, despite relying on clinician‐reported data, the NCD demonstrated broadly comparable temporal patterns to the NDB Open Data across the five major gastrointestinal procedures examined. The observed stability of procedure‐specific NCD/NDB ratios supports the use of NCD annual statistics for analyses of temporal trends in these selected gastrointestinal procedures in Japan. Author Contributions Atomu Kiriyama: data curation, formal analysis, investigation, visualization, writing – original draft, writing – review and editing. Masakazu Fujii: conceptualization, methodology, data curation, formal analysis, investigation, visualization, writing – original draft, writing – review and editing. Satoshi Hirano: conceptualization, supervision, writing – review and editing. Toshimichi Asano: conceptualization, methodology, supervision, writing – review and editing. Funding The authors have nothing to report. Ethics Statement The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Table S1: Individual NCD procedure identifiers and NDB medical fee codes with official code definitions. Figure S1: Temporal stability of the systematic differences between the NCD and NDB Open Data for five gastrointestinal surgical procedures. Data Availability Statement The data that support the findings of this study were derived entirely from publicly available resources. NCD annual surgical statistics are available from the Japan Surgical Society website. NDB Open Data are available from the Ministry of Health, Labour and Welfare website. The specific annual data files used in this study are cited in the references. References Associated Data Supplementary Materials Table S1: Individual NCD procedure identifiers and NDB medical fee codes with official code definitions. Figure S1: Temporal stability of the systematic differences between the NCD and NDB Open Data for five gastrointestinal surgical procedures. Data Availability Statement The data that support the findings of this study were derived entirely from publicly available resources. NCD annual surgical statistics are available from the Japan Surgical Society website. NDB Open Data are available from the Ministry of Health, Labour and Welfare website. The specific annual data files used in this study are cited in the references.