A Nationwide Study on Pretransfusion Testing for 10 Years in Korea Using Health Insurance Review and Assessment Claims Data Authors: Jae Hyeon Park, Yousun Chung, Dae-Hyun Ko, Hyungsuk Kim Published: 2026 ABSTRACT BackgroundAssessing pretransfusion testing utilization at medical institutions on a nationwide scale can be challenging. This study aimed to evaluate the status of pretransfusion testing in Korea by analyzing insurance claims data.MethodsWe identified all claims including pretransfusion testing or blood transfusion made between January 2011 and December 2020 from the Korean Health Insurance Review and Assessment Service database. ABO/RhD typing, antibody screening and identification, and crossmatching were analyzed according to the hospital type.ResultsA total of 42,968,177 claims from 20,088,744 patients were analyzed. The median ratios of ABO serum typing to cell typing were 1.000 for tertiary hospitals, 0.998 for secondary hospitals, 0.000 for primary hospitals, and 0.000 for clinics (P < 0.001). Between 2016 and 2020, there was a 10.3–32.3 percentage-point increase in the proportion of automated ABO typing in tertiary hospitals, secondary hospitals, and clinics, while there was only a 1.9 percentage-point increase in primary hospitals. The median ratios of antibody identification to screening were 0.459% for tertiary hospitals, 0.359% for secondary hospitals, 0.198% for primary hospitals, and 0.139% for clinics (P < 0.001). For the first red blood cell (RBC) transfusion for each patient at each hospital, 71.5% of the patients from primary hospitals and 68.3% of the patients from clinics received RBC transfusion after only crossmatching without antibody screening.ConclusionThe application of pretransfusion testing considerably differs according to the type of hospital in Korea. Primary hospitals and clinics need education and improvement in practice to pursue safe blood transfusion. INTRODUCTION Blood transfusion is an important means of replenishing a patient’s deficient blood components such as red blood cells (RBCs), platelets, and plasma. Proper pretransfusion testing, including ABO and RhD blood group testing, antibody screening/identification, and crossmatching, is essential to ensure safe transfusion.1 In addition to predictable ABO antibodies, other blood groups, including Rh, may develop clinically significant alloantibodies whose presence is unpredictable. Transfusing blood products that match only ABO and RhD without proper testing for these antibodies can lead to adverse reactions, such as potentially fatal hemolytic transfusion reactions. Appropriate pretransfusion testing is fundamental for patients requiring blood transfusion, but there is a lack of reporting on pretransfusion test utilization at the national level. Transfusion Comprehensive J-A 2023 Survey of the College of American Pathologists (CAP) included 3,401 organizations for ABO/RhD typing, 3,357 for antibody screening, and 3,260 for compatibility testing, confirming that more than 95% of participating organizations perform basic blood transfusion testing themselves.2 However, not all participants are US healthcare institutions, and CAP participation is voluntary, which may bias the results. In Korea, 2,487 institutions received blood in 2020,3 but only 375 (15.1%) used over 500 units annually, and 1,119 (45.0%) used fewer than 50 units. Low-blood-use healthcare institutions may struggle to perform all pretransfusion tests in-house. According to the 2020 Annual Report of the Korean Association of External Quality Assessment Service on Transfusion Medicine, only 897 (36.1%) institutions, less than half of all transfusion healthcare institutions, performed ABO/RhD blood group testing, 329 (13.2%) performed antibody screening, and 135 (5.4%) performed antibody identification.4 Therefore, it is estimated that more than half of the medical institutions performing transfusions in Korea do not participate in external proficiency testing and likely do not perform pretransfusion testing themselves. A survey of small- and medium-sized hospitals in Korea also demonstrated that a substantial proportion of institutions did not perform ABO serum typing or unexpected antibody screening and lacked essential blood bank equipment, underscoring the need for systematic training and institutional support for safe transfusion practices.5 Institutions unable to perform these tests may refer them to commercial laboratories, raising concerns that essential tests might be omitted. To the best of our knowledge, no comprehensive analysis of pretransfusion testing utilization using medical big data has been conducted in Korea, and existing reports focus only on the status of transfusion for specific disease groups.67 It was difficult to find similar nationwide studies from other countries reporting the utilization of pretransfusion testing. This study aims to identify pretransfusion testing utilization and transfusion status according to the type of medical institution in Korea using national claims data. METHODS Health Insurance Review & Assessment Service (HIRA) big data Korea’s National Health Insurance (NHI) service covers over 98% of the population, totaling about 50 million in 2014.8 HIRA big data is generated through reimbursement of healthcare providers under the NHI, which is mandatory for all medical institutions.9 The data include comprehensive information including age, sex, diagnosis, hospitalization, outpatient care, procedures, surgeries, tests, treatments, prescriptions, and costs. The database has been accessible to researchers since 2009, who can request customized data after review by HIRA.8 Study population Claims for all patients who received pretransfusion testing and transfusion prescriptions from January 1, 2011, to December 31, 2020, were extracted. The analysis included all claims from not only medical institutions, but also dental, Korean traditional medicine, and public healthcare centers. Relevant information was extracted from a secondary accessible research database (M20210924526) provided by HIRA derived from the original database. The views expressed are those of the authors and not necessarily those of the HIRA and the Ministry of Health and Welfare. Pretransfusion testing and transfusion In Korea, type and screen are the basic pretransfusion tests, with crossmatching recommended before releasing RBCs and/or platelets.10 Pretransfusion tests analyzed in the claims data included ABO typing, RhD typing, unexpected antibody screening, unexpected antibody identification, and crossmatching. Previously, ABO antigen (cell typing) and antibody (serum typing) tests had separate codes but have been integrated into a single code for ABO blood group typing since August 2015, requiring both tests for reimbursement. Additionally, separate codes for automated ABO typing and Rh typing have been newly categorized to distinguish them from conventional manual testing. The ratio of ABO serum typing to cell typing (Ab/Ag) was calculated based on whether a test was performed, not the number of tests per claim for each patient. The crossmatching rate for RBC transfusions was analyzed to assess pretransfusion testing appropriateness. We also identified antibody screening within 30 days preceding the first RBC transfusion for each patient at each healthcare institution. Statistical analysis Categorical variables are presented as frequencies, and continuous variables as medians and interquartile ranges (IQRs). Continuous variables were compared using the Mann-Whitney U and Kruskal-Wallis tests. In the pairwise Wilcoxon rank sum test, the P value was adjusted using the Bonferroni method. All statistical analyses were performed using R project (version 3.5.1; R Foundation for Statistical Computing, Vienna, Austria), with statistical significance set at P < 0.05. Ethics statement This study was approved by the Institutional Review Board of Seoul National University Hospital (E-2111-010-1267). Written informed consent was waived because the data were de-identified in the database used. RESULTS Claims statistics Over the 10-year study period, 42,968,177 claims for pretransfusion testing and transfusions were identified for 20,088,744 patients, accounting for 39.0% of the population (Supplementary Table 1). There were 44,775,272 ABO typing; 47,241,599 RhD typing; 21,046,483 antibody screening; 107,540 antibody identification; and 30,178,630 crossmatching. Additionally, 19,243,900 units of RBCs; 15,129,411 units of platelet concentrates; and 1,956,265 units of apheresis platelets were transfused. The overall rate of pretransfusion testing increased annually except in 2020. The 10-year average increases were 5.1% for ABO typing, 2.8% for RhD typing, 7.6% for antibody screening, and 10.5% for antibody identification, but −1.0% for crossmatching. ABO cell typing and serum typing From 2011 to 2015, 17,662,160 ABO cell typing and 13,771,995 ABO serum typing were performed. Data on the number of hospitals, tests, and the ABO Ab/Ag ratio per hospital are summarized in Table 1 and Fig. 1. The median ABO Ab/Ag ratio significantly differed by institution type: 1.0 (IQR, 0.998–1.000) in tertiary hospitals, 0.998 (IQR, 0.744–1.000) in secondary hospitals, 0.0 (IQR, 0.0–0.298) in primary hospitals, and 0.0 (0.0–0.0) in clinics (P < 0.001). The ABO Ab/Ag ratio in tertiary hospitals was significantly higher than in secondary hospitals, primary hospitals, and clinics (P = 0.002, < 0.001, and < 0.001, respectively). Secondary hospitals also had significantly higher ABO Ab/Ag ratios than primary hospitals and clinics (P < 0.001 and < 0.001), and primary hospitals had higher ABO Ab/Ag ratios than clinics (P < 0.001). Table 1ABO cell typing and serum typing based on hospital type from 2011 to 2015Hospital typesNo. of hospitalsNo. of ABO cell typingNo. of ABO serum typingMedian (IQR) ABO Ab/Ag ratio per hospitalTertiary hospital525,150,6935,971,1611.000 (0.998–1.000)Secondary hospital3335,671,1355,658,7000.998 (0.744–1.000)Primary hospital1,6354,621,1841,633,1250.000 (0.000–0.298)Nursing hospital1,21436,99912,8920.184 (0.000–0.915)Clinic7,4192,155,096483,7440.000 (0.000–0.000)Dental hospital/clinic2011,76912,1111.000 (0.775–1.000)Public healthcare center20914,9582250.000 (0.000–0.000)Korean traditional medicine43326370.000 (0.000–0.000)Total10,92517,662,16013,771,995-IQR = interquartile range, Ab/Ag = serum typing to cell typing. ABO cell typing and serum typing based on hospital type from 2011 to 2015 IQR = interquartile range, Ab/Ag = serum typing to cell typing. Histogram of ABO antibody to cell typing ratio based on hospital type. The histogram depicts the calculation of the ABO serum typing to cell typing ratio for each hospital, presented in ratio bins weighted by the number of ABO cell typing performed by each hospital. Automated and manual ABO typing From 2016 to 2020, 15,430,845 manual and 9,943,484 automated ABO tests were performed (Supplementary Table 2). The annual proportion of automated ABO testing is shown in Fig. 2. In 2016, the proportion of automated ABO testing was 50.6% in tertiary hospitals, 9.8% in secondary hospitals, 5.9% in primary hospitals, and 40.1% in clinics. By 2020, these rates increased to 78.2%, 42.1%, 7.8%, and 50.4%, respectively, with noticeable increases in tertiary and secondary hospitals. Annual proportion of automated ABO testing according to the hospital type. Unexpected antibody screening and identification We calculated the unexpected antibody identification to screening ratio for healthcare institutions that performed 500 or more antibody screenings over the 10-year period. This included 99.4% (20,921,658/21,046,483) of antibody screening and 96.1% (103,362/107,540) of antibody identification tests (Table 2). The median antibody identification to screening ratios were 0.459% (IQR, 0.339–0.689%) for tertiary hospitals, 0.359% (IQR, 0.085–0.650%) for secondary hospitals, 0.198% (IQR, 0.000–0.595%) for primary hospitals, and 0.139% (IQR, 0.000–0.708%) for clinics, with significant differences (P < 0.001). The antibody identification to screening ratio in tertiary hospitals was significantly different from secondary hospitals, primary hospitals, and clinics (P = 0.021, < 0.001, and = 0.001, respectively). Secondary hospitals differed significantly from primary hospitals (P = 0.015), but not from clinics (P = 0.064). No significant difference was found between primary hospitals and clinics (P = 1.000). Table 2Unexpected antibody screening and identification tests based on hospital typea Hospital typesNo. of hospitalsAntibody screeningAntibody identificationMedian (IQR) ratio (%) of antibody identification to antibody screening per hospitalTertiary hospital5310,593,29857,1650.459 (0.339–0.689)Secondary hospital3339,073,85740,9780.359 (0.085–0.650)Primary hospital2721,079,7154,3090.198 (0.000–0.595)Nursing hospital67,051140.212 (0.022–0.389)Clinic68142,3508530.139 (0.000–0.708)Dental hospital/clinic525,387430.090 (0.089–0.194)Total20,921,658103,362IQR = interquartile range.aOnly medical institutions with 500 or more unexpected antibody screenings over a 10-year period were included. Unexpected antibody screening and identification tests based on hospital typea IQR = interquartile range. aOnly medical institutions with 500 or more unexpected antibody screenings over a 10-year period were included. Unexpected antibody screening and crossmatching for the first RBC transfusion We identified whether unexpected antibody screening and crossmatching were performed on the first RBC transfusion per patient at each healthcare institution (Table 3, Fig. 3). The number of RBC units included in the first transfusion was 4,618,104 in tertiary hospitals; 4,887,243 in secondary hospitals; 2,229,491 in primary hospitals; 121,856 in nursing hospitals; and 140,953 in clinics, accounting for 24.0%, 25.4%, 11.6%, 0.6%, and 0.7% of the total RBC concentrate shipments from blood centers to hospitals, respectively. The proportions for RBC transfusions with both antibody screening and crossmatching were 94.2% in tertiary hospitals, 85.4% in secondary hospitals, 26.2% in primary hospitals, 27.3% in nursing hospitals, and 11.4% in clinics. The proportions for RBC transfusions with crossmatching alone were 71.5% in primary hospitals, 65.7% in nursing hospitals, and 68.3% in clinics. Although claiming errors could not be precluded, 16,675 (2.1%) patients in primary hospitals; 3,951 (6.2%) in nursing hospitals; and 10,400 (18.3%) in clinics were transfused without both antibody screening and crossmatching. Table 3Pretransfusion test at the first red blood cell transfusion for individual patients at each hospitalHospital typesTFTF, SCRTF, XMTF, SCR, XMTotal patientsRBC units in the first claimsaRBC units in total claimsaTertiary hospital9,990 (0.8)9,101 (0.7)56,333 (4.4)1,217,389 (94.2)1,292,8134,618,104 (24.0)8,455,954 (43.9)Secondary hospital9,557 (0.6)11,243 (0.7)202,968 (13.2)1,308,210 (85.4)1,531,9784,887,243 (25.4)7,704,759 (40.0)Primary hospital16,675 (2.1)1,729 (0.2)562,347 (71.5)205,765 (26.2)786,5162,229,491 (11.6)2,710,713 (14.1)Nursing hospital3,951 (6.2)448 (0.7)41,538 (65.7)17,282 (27.3)63,219121,856 (0.6)187,714 (1.0)Clinic10,400 (18.3)1,133 (2.0)38,920 (68.3)6,517 (11.4)56,970140,953 (0.7)179,330 (0.9)Values are presented as number (%).TF = transfusion, SCR = unexpected antibody screening, XM = crossmatching, RBC = red blood cell.aProportion among total RBC releases over 10 years of 19,243,900 units. Pretransfusion test at the first red blood cell transfusion for individual patients at each hospital Values are presented as number (%). TF = transfusion, SCR = unexpected antibody screening, XM = crossmatching, RBC = red blood cell. aProportion among total RBC releases over 10 years of 19,243,900 units. Histogram of antibody screening to crossmatching rate in the first RBC transfusion based on hospital type. Calculating the antibody screening to crossmatching rate in the first RBC transfusion for each hospital, the histogram is presented based on bins of the rate, weighted by the number of transfused patients at each hospital. RBC = red blood cell. DISCUSSION By analyzing medical big data from an NHI system, this study found significant differences in pretransfusion testing application across hospital types in Korea. Primary hospitals and clinics were identified as needing education and improvement in clinical practice. The findings of this study emphasize the importance of basic pretransfusion testing and can guide future policies for safe blood transfusion. In Korea, the government periodically assesses medical practices, and in 2020, blood transfusion adequacy was evaluated for the first time.11 While this audit focused on general blood transfusion processes, our study specifically emphasized pretransfusion testing utilization. The 2020 government assessment (adult patients; tertiary/secondary/primary hospitals; 6-month window) implemented detailed indicators solely for unilateral total knee replacement arthroplasty, whereas our analysis spans all ages, all facility types, and a 10-year horizon, focusing specifically on pretransfusion testing utilization. The results of our study can help identify and propose evaluation indices for future government medical practice assessments and provide a basis for other researchers interested in evaluating clinical laboratory test utilization based on claims data. The current study results show that the number of patients, RBC and platelet transfusions, ABO/RhD typing, and crossmatching have increased consistently, except for a decline in 2020, likely due to a decrease in healthcare utilization during the coronavirus disease 2019 pandemic.12 Nevertheless, antibody screening and identification continued to increase during this period. The medical practice assessment on blood transfusion in 2020 and the heightened awareness of patient blood management are assumed to have influenced an increase in unexpected antibody testing despite the overall decrease in blood utilization.13 One of the major policy changes in pretransfusion testing in Korea was the integration of cell typing and serum typing in one reimbursement code in 2015. Prior to the integration of ABO cell typing and serum typing into a single code, claims data identified institutions not performing ABO serum typing, notably in primary hospitals and clinics, which constitute a considerable proportion of the healthcare system. ABO cell typing alone can raise patient safety issues due to unrecognized ABO discrepancies. Particularly, cis-AB, a rare mutation in the ABO gene that produces both A and B antigens, is more common in the southwest region of Korea than in other parts of the world, where it is found in 0.0354% of blood donors, necessitating caution in ABO blood group testing.14 Using automated blood bank analyzers can enhance compliance in conducting both ABO cell typing and serum typing in clinical laboratories. After persuading the Korean government about the need for a separately increased reimbursement for automated blood bank testing to improve patient safety, it was officially recognized and covered by the NHI in 2015, leading to increased use of automated blood bank analyzers. The UK National External Quality Assessment Service (NEQAS) also observed an increase in automated core hours from 74% in 2011 to 90% in 2016 after insurance inclusion.15 Implementation of automated ABO/RhD testing enhances throughput and reduces errors, particularly those arising from clerical mistakes.1617 A review of ABO-incompatible transfusions in Korea found a 9% error rate during testing, the third most frequent cause after administration (50%) and sampling (13%).18 Automated ABO testing can significantly reduce this issue and prevent fatal hemolytic transfusion reactions. Particularly, primary hospitals had high rates of crossmatching only (71.5%) among 786,516 patients. Although some degree of error may be inevitable in claims-based data, Table 3 demonstrates a clear trend that the proportion of transfusions apparently performed without crossmatching increases as institutional size decreases, being highest in clinics. This suggests that actual cases of transfusion without crossmatching may indeed have occurred, which underscores the significance of our findings. Furthermore, it has been previously reported that some institutions do not perform crossmatching at all.5 The first quartile value of the antibody identification to screening ratios for all hospitals was estimated to be less than 0.1%, indicating many healthcare providers do not conduct antibody identification testing despite positive screening results. The antibody screening positivity rate generally ranges from 1.0% to 1.5% and varies between 0.5% to 4% in Korea.19 To exclude very small-sized hospitals with extremely low testing volumes, this study analyzed only institutions performing 500 or more antibody screenings over the 10-year study period. Our results showed that small institutions with infrequent cases may struggle to conduct antibody identification independently due to the high probability of being unable to utilize purchased reagents within their expiration date, resulting in significant wastage. Although confirming antibody specificity at a referral laboratory is an alternative, hospitals may find it compelling to delay blood transfusion while waiting for the results. However, relying on crossmatching to find compatible units, even if including the IAT phase, has its limits, such as failing to detect heterozygous antigen-positive units that show dosage effect. Lack of knowledge may also lead to improper utilization of pretransfusion testing. Multinational evaluations of transfusion medicine knowledge among residents in internal medicine, anesthesiology, and pediatrics revealed a consistent lack of training and competency.202122 Acquiring knowledge about evidence-based practices beyond one’s specialty after completing medical school and residency is challenging, raising concerns about inappropriate practices in transfusion medicine, especially outside training hospitals. In this study, both the ratios of antibody identification to screening tests and ABO Ab/Ag were higher in tertiary hospitals. This may be attributed to the presence of specialists in laboratory medicine. In these institutions, appropriate pretransfusion testing is ensured through external quality assessment, and peer review of laboratory service by accreditation auditors.423 One of the interesting findings in our study was that the sum of the rates of performing both antibody and crossmatch tests before the first transfusion, and performing only crossmatch tests, does not reach 100% in all institutions. These findings are consistent with a previous nationwide survey of small- and medium-sized hospitals, which reported that many institutions omitted ABO serum typing and antibody screening and had insufficient equipment to conduct basic pretransfusion testing.5 Possible explanation for the observed discrepancy is that some medical institutions may not perform crossmatching on site. Instead, they either transfuse without performing a crossmatch in rare cases or refer specimens to referral laboratories for crossmatching.5 In addition, factors such as emergency release of uncrossmatched blood, coding errors, incomplete claim submissions, and delays between the actual clinical procedure and claim entry may also contribute to these discrepancies. Based on our findings, we propose several policy actions to improve pretransfusion safety in Korea: 1) revising reimbursement structures to provide greater incentives for antibody screening and automated testing, 2) expanding participation in external quality assessment, 3) establishing regional referral laboratories for antibody identification, 4) strengthening education and training programs for non-specialist institutions, and 5) extending transfusion adequacy evaluation to include small-sized clinics. We believe that these policy measures, tailored to the Korean healthcare context, will enhance transfusion safety nationwide. This study has some limitations. Medical research based on claims data may suffer from incomplete information, coding errors, selection bias, and a lack of contextual details, potentially leading to inaccuracies and limitations in the generalizability and depth of findings. Misclassification of test codes may lead to underestimation or overestimation of certain pretransfusion tests. Delays between the actual clinical event and the submission of claims may introduce discrepancies in the temporal alignment of transfusion events and associated pretransfusion tests. As crossmatch claims do not distinguish the types of various phases of crossmatch tests, we cannot precisely determine pretransfusion practice of a medical institution. The analysis of our study was restricted to data up to December 2020, as the customized HIRA dataset available to us (M20210924526) did not include claims beyond this period. Consequently, our findings do not reflect the impact of more recent national initiatives, such as the first National Blood Management Plan (2021–2025) and the subsequent transfusion adequacy evaluations. Nevertheless, this study provides valuable baseline information on pretransfusion testing practices in Korea prior to the implementation of these national policies, highlighting the need for future investigations to assess their impact on transfusion safety and practice patterns. Although claims data have inherent limitations, their advantage lies in being readily available at the national level and showing longitudinal trends.82425 Blood transfusion is a common and vital medical procedure widely performed in many healthcare centers. Although institutions generally do not require specific certifications to perform blood transfusion, the complex process of donor recruitment, blood collection, processing, testing, storage, administration, and patient monitoring necessitates careful consideration to ensure patient safety.26 Our study highlights that healthcare centers, particularly those of smaller scale, in Korea face certain challenges in achieving optimal practices in pretransfusion testing. To improve transfusion safety, education and policy support are needed to improve the appropriate utilization of pretransfusion testing on a nationwide scale.