Development of indicators to evaluate the effectiveness of antimicrobial stewardship programs in Korea: a Delphi consensus study Authors: Sung Min Lim, Raeseok Lee, Jinnam Kim, Ji Young Park, Choseok Yoon, Myung Jin Lee, Song Mi Moon, Yong Chan Kim, Hong Bin Kim, Se Yoon Park, Bongyoung Kim Published: 2026 ABSTRACT BackgroundA pilot program launched in November 2024 to establish and operationalize Antimicrobial Stewardship Programs (ASPs) in Korean hospitals is expected to accelerate nationwide adoption. However, consensus on indicators for evaluating ASP performance remains lacking. This study aimed to develop indicators for evaluating the short- and long-term effectiveness of ASPs.MethodsAn expert panel comprised 18 specialists from academic societies and government agencies involved in infectious diseases and antimicrobial control in South Korea. Candidate indicators for evaluating ASP effectiveness were identified by five infectious disease specialists through a systematic literature review. The Delphi survey consisted of an initial open-ended round, followed by two multiple-choice rounds (rounds 2 and 3) with an intervening online panel meeting. Responses to the multiple-choice questions were rated on a 7-point Likert scale, and indicators with a content validity ratio of ≥ 0.47 were selected.ResultsThe systematic literature review yielded 59 candidate indicators of ASP effectiveness, which were classified into seven primary categories. After the third-round Delphi survey, the final set of indicators included the following. In the Antibiotic Use Indicators category: (1) days of therapy per 1,000 patient-days, (2) defined daily dose per 1,000 patient-days, (3) reduction rate of broad-spectrum antibiotic use, (4) intravenous-to-oral antibiotic conversion rate, (5) reduction rate of unnecessary combination therapy, and 6) change in antibiotic prescription ratio according to the World Health Organization Access, Watch, and Reserve classification. In the Process-Related Indicators category: (1) appropriate rate of preoperative surgical antibiotic prophylaxis, (2) implementation of therapeutic drug monitoring with dosage recommendations, and (3) establishment of institutional antibiotic use guidelines. In the Department-Specific Indicators related to operating rooms and surgery: (1) appropriate rate of preoperative surgical antibiotic prophylaxis, (2) appropriateness of postoperative antibiotic discontinuation timing, and (3) appropriateness of prophylactic antibiotic selection. Finally, among the Other Indicators, the implementation rate of core ASP elements was selected as an indicator for evaluating ASP effectiveness.ConclusionsThis study established consensus-based indicators and proposed short- and long-term targets for evaluating the effectiveness of ASPs in Korea. These indicators enable standardized assessment and help inform evidence-based policies for the sustainable operation of ASPs.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13756-026-01778-9. Background Antimicrobial resistance poses a growing global health crisis, driving substantial morbidity, mortality, and healthcare costs [1, 2]. Antimicrobial stewardship programs (ASPs) have been widely endorsed by major health organizations as an essential strategy to optimize antibiotic use, improve patient outcomes, and mitigate the emergence of resistance [3, 4]. To ensure that ASP efforts are effective and sustainable, they must be supported by robust performance metrics that demonstrate impact, guide continuous improvement, and enable benchmarking across institutions. Several countries have adopted indicators to evaluate ASP effectiveness, including days of therapy (DOT), defined daily dose (DDD), incidence of Clostridioides difficile infection (CDI), and adherence to prescription guidelines [5–7]. However, these measures are insufficient for comprehensive evaluation, as they capture only a subset of ASP activities. Recent international efforts have highlighted the importance of structured consensus building to identify relevant and practical indicators, with methods such as modified Delphi surveys used to reach agreement among multidisciplinary experts [8, 9]. In Korea, antibiotic consumption remains among the highest compared with other high-income countries, while antimicrobial resistance continues to rise [10]. The Korean National Action Plan on antimicrobial resistance has prioritised ASP at the policy level [11]; however, implementation at the hospital level has been heterogeneous. Although previous studies have reported favorable impacts on antibiotic consumption, resistance trends [12–14] and have provided valuable insights, they do not establish a harmonized framework for evaluating ASP performance at the national level. In November 2024, the Korean government launched a reimbursement-based pilot program involving 78 hospitals, representing a major step toward the formal institutionalization of ASPs [15]. Early implementation revealed substantial heterogeneity in infrastructure, staffing, and outcome measurements across institutions, reinforcing the need for a unified evaluation framework. As the program transitions toward broader implementation, a national consensus on evaluation indicators has become increasingly important to ensure fair performance assessment and sustainable reimbursement mechanisms. In the absence of standardized metrics, reliable national benchmarking and long-term policy evaluation remain challenging. In light of these considerations, we conducted a structured Delphi consensus process to develop a set of performance indicators tailored to the Korean healthcare system and potentially adaptable to other resource-limited settings. Drawing on published evidence and expert consensus from clinicians, pharmacists, microbiologists, and policymakers—including representatives from national public health and reimbursement agencies (the Korea Disease Control and Prevention Agency and the Health Insurance Review and Assessment Service)— we propose metrics that are practical within existing data systems and aligned with international standards. These indicators aim to provide a coherent framework for monitoring hospital-level ASPs, evaluating national initiatives, and informing ASP implementation in comparable healthcare contexts. Methods Overview and assembly of panel We conducted a systematic literature review to identify potential key quality indicators (QIs), three sequential rounds of online surveys, and one online expert panel meeting. Figure 1 presents an overview of the modified Delphi method used in this study. The consensus process synthesized the independent judgments of a multidisciplinary expert panel. A total of 18 panelists participated, representing the Korean Society of Infectious Diseases (n = 4), the Korean Society for Antimicrobial Therapy (n = 4), the Korean Society of Pediatric Infectious Diseases (n = 2), the Korea National Antimicrobial Use Analysis System (n = 2), the Korean Society of Health-System Pharmacists (n = 2), the Korea Disease Control and Prevention Agency (n = 2), and the Health Insurance Review and Assessment Service (n = 2). Fig. 1Flowchart of the study procedures Flowchart of the study procedures Search strategy A systematic literature search was performed to identify and analyze potential QIs used to evaluate the effectiveness of ASPs in countries with established stewardship systems, including the United States, Japan, the United Kingdom, and Australia. These countries were selected because they have formally institutionalized ASPs and have accumulated a relatively large body of research on ASP evaluation indicators, making their frameworks particularly relevant as references for developing a Korean ASP evaluation system. The review focused on methods for monitoring ASP implementation and assessing its short- or long-term effectiveness. Studies written in English that addressed the development or evaluation of ASP-related indicators were included. The following were excluded: (1) studies not involving humans; (2) studies primarily focused on human immunodeficiency virus infection; and (3) editorials, comments, case reports, errata, and conference abstracts (Additional file 1). The literature search was conducted on November 22, 2024, and the identified records were reviewed in full text by four infectious disease specialists and one pediatric infectious disease specialist (C.Y., J.K., M.J.L., R.L., and J.Y.P.) between October 28 and December 29, 2024. Any discrepancies regarding eligibility were resolved through discussion and consensus among the reviewers. Following full-text review, all potential QIs proposed or applied in the literature were extracted to develop a preliminary list of candidate indicators (Additional file 2). Each QI was categorized into one of seven domains: antibiotic use, clinical outcome, microbiological, economic, process-related, department-specific, or other indicators. Process-related indicators were defined as measures reflecting hospital-wide ASP monitoring activities, capturing the overall adequacy of antibiotic management practices across the institution. Department-specific indicators, in contrast, were defined as measures targeting granular, component-level evaluation of antibiotic use within high-priority clinical settings, such as the operating room and surgical ward, and were designed to provide more operationally targeted feedback at the point of care. Although certain department-specific indicators share a common focus with process-related indicators, they were retained as distinct categories to enable complementary evaluation at different levels of institutional practice. Duplicate items were removed and conceptually overlapping indicators were consolidated into unified descriptions. Development of survey items and the process of Delphi method Each survey was distributed to the expert panel via email, and the preliminary candidate QIs derived from the literature review were shared with the panelists in advance. The first-round survey (February 12–24, 2025) used an open-ended questionnaire, allowing panel members to suggest revisions or additions to indicators and provide free-text comments. The second-round survey (March 11–19, 2025) consisted of a structured, closed-ended questionnaire containing the same items as those in the first round. Panelists evaluated each indicator according to three criteria: (1) expected outcomes reflecting appropriate antibiotic use, (2) the degree of improvement anticipated when ASP activities are led by infectious disease specialists compared with other clinicians, and (3) feasibility of assessment within domestic healthcare institutions (Additional file 3). After the second-round survey, an online expert meeting was held on March 21, 2025, to review the results and discuss items requiring further clarification or consensus. The third-round survey (April 1–7, 2025) incorporated revisions based on the preceding discussion and aimed to finalize indicator selection. Evaluation criteria for this round included: (1) expected outcomes reflecting appropriate antibiotic use; (2) feasibility of measurement in domestic institutions; and (3) the need for generating evidence to support ASP implementation. Items in the second- and third-round surveys were rated on a 7-point Likert scale ranging from 1 (“strongly disagree”) to 7 (“strongly agree”), with scores of 6 or 7 defined as essential responses. In the third-round survey, an additional question item regarding “an appropriate timeframe for achieving measurable outcomes” was included, requiring respondents to select one of four options: short-term (3–4 years), mid-term (5–9 years), long-term (≥ 10 years), or not applicable. Validation of items Responses from the second and third Delphi survey rounds were analyzed using the content validity ratio (CVR). The CVR was calculated using the formula CVR = (Ne − N/2) / (N/2), where Ne represents the number of experts who rated an item as “essential,” and N denotes the total number of experts [16]. Items with a CVR ≥ 0.47 were accepted, whereas those with a CVR ≥ 0.40 were further reviewed through panel discussion. The threshold of CVR ≥ 0.47 was based on the corrected critical value for a panel of 18 experts at a significance level of p < 0.05, as derived by Ayre and Scally [17], who identified a computational error in Lawshe’s original method and provided revised critical values. The secondary threshold of CVR ≥ 0.40 corresponds to a significance level of p < 0.10 and was applied to flag borderline items for structured panel discussion, rather than automatic rejection [16, 17]. Alternatively, items with a lower limit of the 95% confidence interval of ≥ 6.5 were accepted, and those with a lower limit of ≥ 5.5 were considered for discussion. The two thresholds were established a priori through deliberation among the research team before data collection, as no established reference exists for these specific cutoffs. A lower bound of ≥ 6.5 indicates that the conservative estimate of the confidence interval falls within the “essential” range of the 7-point Likert scale (score ≥ 6), reflecting strong consensus with minimal inter-rater variability, and was therefore adopted as the acceptance criterion. A lower bound of ≥ 5.5 was set as the threshold for panel discussion, representing a midpoint between scores of 5 (“somewhat agree”) and 6 (“agree”); when some experts assigned lower scores, the resulting lower CI bound tended to fall near 5.5, indicating that the item warranted structured re-evaluation. Items falling between these two thresholds were subject to panel discussion rather than automatic acceptance or rejection. Results Literature review and initial identification of candidate indicators A total of 2,066 articles were reviewed following the systematic literature search (Fig. 1), and 161 studies were selected according to predefined inclusion and exclusion criteria. Preliminary candidate indicators were extracted from the included studies. In total, 59 specific indicators were identified across seven domains: antibiotic use (n = 7), clinical outcome (n = 8), microbiological (n = 7), economic (n = 7), process-related (n = 8), department-specific (n = 14), and other indicators (n = 8) (Additional file 2). Refinement of indicators A total of 17 of 18 (94.4%) panel members responded to the first survey. Based on the free-text responses, panelists proposed revisions, additions, and subdivisions of the preliminary candidate indicators. Subindicators within each major category were further refined and supplemented through this process, and consolidated feedback was shared during an online meeting. As a result of the first-round survey, the initial set of 59 candidate indicators was expanded to 113 indicators, reflecting the addition of newly proposed items and the subdivision of existing indicators into more granular subindicators across all seven domains. All panelists (18/18, 100%) completed the structured questionnaire (second Delphi round, Additional file 3), and an online expert meeting was convened to review and clarify the survey results, with 13 of the 18 panelists (72.2%) participating. Through these procedures, several indicators from the process-related, department-specific, clinical outcome, microbiological, and economic categories were excluded because of limited applicability, feasibility, interpretability, or measurement challenges, whereas the CDI rate and implementation of therapeutic drug monitoring were retained. In contrast, antibiotic consumption indicators and the ASP core element implementation rate were prioritised for further refinement, and clinician satisfaction was advanced for additional consideration. Consequently, 113 indicators remained, including antibiotic use (n = 23), clinical outcome (n = 19), microbiological (n = 9), economic (n = 11), process-related (n = 21), department-specific (n = 19), and other indicators (n = 11) (Additional file 4). Final consensus indicator The third-round Delphi survey included all 18 panelists. Based on the consensus, a final set of indicators for evaluating ASP effectiveness was established (Table 1). The indicators were organized into four categories: antibiotic use, process-related, department-specific, or other indicators. Consensus was not achieved for the clinical outcome, economic, and microbiological domains. In the antibiotic use category, six indicators were finalized: (1) antibiotic consumption measured as DOT per 1,000 patient-days; (2) DDD per 1,000 patient-days; (3) reduction rate of broad-spectrum antibiotic use (e.g., carbapenems); (4) rate of intravenous-to-oral antibiotic conversion (e.g., fluoroquinolones); (5) reduction rate of unnecessary combination therapy (e.g., overlapping anaerobic coverage); and (6) change in antibiotic prescription ratio according to the World Health Organization Access, Watch, and Reserve (WHO AWaRe) classification. All indicators were considered suitable for short-term evaluation except for DDD per 1,000 patient-days, which was classified as short- to mid-term. This distinction reflects the inherent methodological complexity of DDD as a metric: because DDD is based on the WHO-defined average adult maintenance dose, institutions with a higher proportion of pediatric or dialysis patients may systematically yield lower DDD values, making short-term cross-institutional comparisons more challenging. In contrast, DOT counts the number of days a patient receives any dose of a given antibiotic regardless of the amount, and is therefore less susceptible to such case-mix confounding, allowing for more immediate interpretation of changes in antibiotic consumption. In the process-related category, three indicators were selected: (1) appropriate rate of preoperative surgical antibiotic prophylaxis; (2) implementation of therapeutic drug monitoring and dose adjustment recommendations; and (3) establishment of institutional antibiotic use guidelines. All were classified as short-term evaluation measures. In the department-specific category, three indicators—primarily related to surgical and perioperative care—were finalized: (1) appropriate rate of preoperative surgical antibiotic prophylaxis; (2) appropriateness of antibiotic discontinuation timing after surgery; and (3) appropriateness of antibiotic selection for surgical prophylaxis. The first two were classified as short-term evaluation indicators, whereas the latter was considered mid-term. Finally, in the other indicators category, the implementation rate of core ASP elements—defined as a standardized set of structural and organizational components essential for establishing effective ASPs, including leadership commitment, accountability, drug expertise, tracking, reporting, and education [7]—was identified as a comprehensive structural indicator reflecting overall ASP performance and was considered appropriate for short-term assessment. Table 1Outcome indicators related to antimicrobial stewardship programs in Korea: results of the third Delphi roundIndicatorExpected ASP outcomeFeasibility in Korean hospitalsNeed for ASP evidence generationTimeframeConsensusCVR95% CILower95% CIUpperCVR95% CILower95% CIUpperCVR95% CILower95% CIUpper Antibiotic use indicators Days of therapy (DOT)/patient days0.8756.176.830.8756.176.830.8006.076.93ShortAcceptDefined daily dose (DDD)/patient days0.7146.006.780.5005.386.500.7145.896.78Short/MidAcceptReduction rate in broad-spectrum antibiotic use (e.g., carbapenem)0.5005.436.460.8005.826.630.7145.576.76ShortAcceptRestricted antibiotic use0.0004.576.090.5004.886.340.3644.856.37ShortRejectIV-to-oral antibiotic conversion rate (e.g., fluoroquinolones)0.2005.316.14− 2.5003.845.160.8755.456.55ShortAcceptInappropriate antibiotic prescription rate(by syndrome or time point)0.2005.236.22− 16.0002.633.930.3644.526.03ShortRejectAppropriateness of antibiotic duration(limited to representative syndromes)0.0005.126.22− 4.0003.124.550.2004.546.02MidRejectUnnecessary combination therapy(e.g., anti-anaerobic duplication)0.7145.526.59− 0.5714.395.720.7144.956.50ShortAcceptAntibiotic use by WHO AWaRe classification0.2004.956.050.2004.495.960.5004.506.05ShortAccept Clinical outcome indicators Length of stay− 7.0003.354.650.3645.076.26− 1.6004.115.33LongRejectICU length of stay− 16.0002.743.930.0004.876.02− 7.0003.484.74N/ARejectICU admission rate during hospitalization− 16.0002.703.970.0004.776.01− 7.0003.334.67N/ARejectC. difficile infection rate†0.5005.336.34− 0.5714.415.700.3645.326.46MidRejectInfection-related mortality− 16.0003.594.74− 7.0002.964.27− 1.6003.985.36ShortRejectLength of stay (limited to specific infection syndromes)− 4.0004.165.06− 0.5713.745.48− 2.5003.965.27LongRejectAdverse drug reaction rate related to antibiotics− 16.0003.434.57− 7.0003.104.56− 1.6003.685.10N/ARejectSurgical site infection rate− 2.5003.354.98− 2.5003.855.04− 1.6003.234.89N/ARejectAcute upper respiratory infection antibiotic prescription rate–––0.5004.966.27− 0.2504.325.79ShortReject Economic indicators Annual total antibiotic costs (excluding antifungals)– 4.0003.076.260.0004.256.98– 1.0003.476.42Short/MidRejectAntibiotic cost per patient– 2.5003.596.52– 0.2503.996.79– 0.5713.626.49MidRejectAntibiotic cost per hospital day– 4.0003.076.26– 0.2503.926.75– 0.5713.506.39MidRejectCost of operating ASP (limited to personnel expenses)– 7.0001.715.400.2003.576.65– 1.6002.025.32ShortRejectCost savings from IV-to-oral conversion0.0004.026.87– 2.5002.926.08– 1.0003.786.66ShortReject Process indicators Appropriate preoperative prophylactic antibiotic use rate0.8004.657.690.3644.167.060.7144.577.54ShortAcceptTDM implementation and dose intervention rate0.8004.677.660.3643.986.910.5004.477.30ShortAcceptPresence of institutional antibiotic use guidelines0.5004.537.360.5004.537.360.5004.327.24ShortAcceptIntervention rate for prolonged antibiotic prescriptions0.5004.467.32– 1.0003.356.310.2003.936.85ShortRejectIntervention rate for combination (duplicate) therapy0.7144.337.33– 0.5713.436.350.3643.916.86ShortRejectConcordance of antibiotic use with organism identification0.8004.727.73− 7.0002.686.210.3644.157.08ShortReject Department-specific indicators (OR/Surgery appropriate rate of preoperative surgical antibiotic prophylaxis0.8005.826.630.7145.476.300.5004.856.37ShortAccept(OR/Surgery) Appropriate timing of postoperative antibiotic discontinuation0.8005.716.620.6155.106.240.5004.696.31ShortAccept(OR/Surgery) Appropriateness of prophylactic antibiotic selection0.8005.826.630.7145.576.320.5004.746.26MidAccept(ED) Appropriate empirical antibiotic selection for major infections(e.g., UTI, pneumonia)– 0.5714.545.68– 1.0003.274.95– 0.2504.575.76MidReject Other indicators Number of ASP training sessions conducted0.2005.096.350.8005.866.690.2004.706.19ShortRejectUse of ASP education indicators[e.g., (training hours × attendance)/staff]0.2004.796.320.7145.686.650.2004.576.09ShortRejectHealthcare provider satisfaction– 1.6003.965.59– 2.5003.645.02– 1.0004.475.64ShortRejectCore elements implementation rate0.5005.286.610.3644.776.230.6155.146.53ShortAcceptAbbreviations: ASP, Antimicrobial Stewardship program; CVR, Content validity ratio; CI, Confidence interval; DOT, Days of therapy; DDD, Defined daily dose; TDM, Therapeutic drug monitoring; ICU, Intensive care unit; ED, Emergency department; OR, Operating room; UTI, Urinary tract infection; WHO, World Health Organization; AWaRe, Access, watch, and reserve.Timeframe: Short = 3–4 years; Mid = 5–9 years; Long = ≥10 years; N/A = Not applicable.† C. difficile infection rate: Expected ASP Outcome CVR met the threshold (0.500); however, Feasibility CVR did not meet the acceptance criterion (CVR = −0.571). The Reject judgment was retained.‡ (ED) Appropriate empirical antibiotic selection: Not adopted primarily due to concerns regarding operational feasibility (systematic tracking not yet widely available in Korean hospitals), case heterogeneity, and reliance on microbiological confirmation data unavailable at the time of ED prescription. Table 1Outcome indicators related to antimicrobial stewardship programs in Korea: results of the third Delphi roundIndicatorExpected ASP outcomeFeasibility in Korean hospitalsNeed for ASP evidence generationTimeframeConsensusCVR95% CILower95% CIUpperCVR95% CILower95% CIUpperCVR95% CILower95% CIUpper Antibiotic use indicators Days of therapy (DOT)/patient days0.8756.176.830.8756.176.830.8006.076.93ShortAcceptDefined daily dose (DDD)/patient days0.7146.006.780.5005.386.500.7145.896.78Short/MidAcceptReduction rate in broad-spectrum antibiotic use (e.g., carbapenem)0.5005.436.460.8005.826.630.7145.576.76ShortAcceptRestricted antibiotic use0.0004.576.090.5004.886.340.3644.856.37ShortRejectIV-to-oral antibiotic conversion rate (e.g., fluoroquinolones)0.2005.316.14− 2.5003.845.160.8755.456.55ShortAcceptInappropriate antibiotic prescription rate(by syndrome or time point)0.2005.236.22− 16.0002.633.930.3644.526.03ShortRejectAppropriateness of antibiotic duration(limited to representative syndromes)0.0005.126.22− 4.0003.124.550.2004.546.02MidRejectUnnecessary combination therapy(e.g., anti-anaerobic duplication)0.7145.526.59− 0.5714.395.720.7144.956.50ShortAcceptAntibiotic use by WHO AWaRe classification0.2004.956.050.2004.495.960.5004.506.05ShortAccept Clinical outcome indicators Length of stay− 7.0003.354.650.3645.076.26− 1.6004.115.33LongRejectICU length of stay− 16.0002.743.930.0004.876.02− 7.0003.484.74N/ARejectICU admission rate during hospitalization− 16.0002.703.970.0004.776.01− 7.0003.334.67N/ARejectC. difficile infection rate†0.5005.336.34− 0.5714.415.700.3645.326.46MidRejectInfection-related mortality− 16.0003.594.74− 7.0002.964.27− 1.6003.985.36ShortRejectLength of stay (limited to specific infection syndromes)− 4.0004.165.06− 0.5713.745.48− 2.5003.965.27LongRejectAdverse drug reaction rate related to antibiotics− 16.0003.434.57− 7.0003.104.56− 1.6003.685.10N/ARejectSurgical site infection rate− 2.5003.354.98− 2.5003.855.04− 1.6003.234.89N/ARejectAcute upper respiratory infection antibiotic prescription rate–––0.5004.966.27− 0.2504.325.79ShortReject Economic indicators Annual total antibiotic costs (excluding antifungals)– 4.0003.076.260.0004.256.98– 1.0003.476.42Short/MidRejectAntibiotic cost per patient– 2.5003.596.52– 0.2503.996.79– 0.5713.626.49MidRejectAntibiotic cost per hospital day– 4.0003.076.26– 0.2503.926.75– 0.5713.506.39MidRejectCost of operating ASP (limited to personnel expenses)– 7.0001.715.400.2003.576.65– 1.6002.025.32ShortRejectCost savings from IV-to-oral conversion0.0004.026.87– 2.5002.926.08– 1.0003.786.66ShortReject Process indicators Appropriate preoperative prophylactic antibiotic use rate0.8004.657.690.3644.167.060.7144.577.54ShortAcceptTDM implementation and dose intervention rate0.8004.677.660.3643.986.910.5004.477.30ShortAcceptPresence of institutional antibiotic use guidelines0.5004.537.360.5004.537.360.5004.327.24ShortAcceptIntervention rate for prolonged antibiotic prescriptions0.5004.467.32– 1.0003.356.310.2003.936.85ShortRejectIntervention rate for combination (duplicate) therapy0.7144.337.33– 0.5713.436.350.3643.916.86ShortRejectConcordance of antibiotic use with organism identification0.8004.727.73− 7.0002.686.210.3644.157.08ShortReject Department-specific indicators (OR/Surgery appropriate rate of preoperative surgical antibiotic prophylaxis0.8005.826.630.7145.476.300.5004.856.37ShortAccept(OR/Surgery) Appropriate timing of postoperative antibiotic discontinuation0.8005.716.620.6155.106.240.5004.696.31ShortAccept(OR/Surgery) Appropriateness of prophylactic antibiotic selection0.8005.826.630.7145.576.320.5004.746.26MidAccept(ED) Appropriate empirical antibiotic selection for major infections(e.g., UTI, pneumonia)– 0.5714.545.68– 1.0003.274.95– 0.2504.575.76MidReject Other indicators Number of ASP training sessions conducted0.2005.096.350.8005.866.690.2004.706.19ShortRejectUse of ASP education indicators[e.g., (training hours × attendance)/staff]0.2004.796.320.7145.686.650.2004.576.09ShortRejectHealthcare provider satisfaction– 1.6003.965.59– 2.5003.645.02– 1.0004.475.64ShortRejectCore elements implementation rate0.5005.286.610.3644.776.230.6155.146.53ShortAcceptAbbreviations: ASP, Antimicrobial Stewardship program; CVR, Content validity ratio; CI, Confidence interval; DOT, Days of therapy; DDD, Defined daily dose; TDM, Therapeutic drug monitoring; ICU, Intensive care unit; ED, Emergency department; OR, Operating room; UTI, Urinary tract infection; WHO, World Health Organization; AWaRe, Access, watch, and reserve.Timeframe: Short = 3–4 years; Mid = 5–9 years; Long = ≥10 years; N/A = Not applicable.† C. difficile infection rate: Expected ASP Outcome CVR met the threshold (0.500); however, Feasibility CVR did not meet the acceptance criterion (CVR = −0.571). The Reject judgment was retained.‡ (ED) Appropriate empirical antibiotic selection: Not adopted primarily due to concerns regarding operational feasibility (systematic tracking not yet widely available in Korean hospitals), case heterogeneity, and reliance on microbiological confirmation data unavailable at the time of ED prescription. Outcome indicators related to antimicrobial stewardship programs in Korea: results of the third Delphi round Inappropriate antibiotic prescription rate (by syndrome or time point) Appropriateness of antibiotic duration (limited to representative syndromes) Unnecessary combination therapy (e.g., anti-anaerobic duplication) (ED) Appropriate empirical antibiotic selection for major infections (e.g., UTI, pneumonia) Use of ASP education indicators [e.g., (training hours × attendance)/staff] Abbreviations: ASP, Antimicrobial Stewardship program; CVR, Content validity ratio; CI, Confidence interval; DOT, Days of therapy; DDD, Defined daily dose; TDM, Therapeutic drug monitoring; ICU, Intensive care unit; ED, Emergency department; OR, Operating room; UTI, Urinary tract infection; WHO, World Health Organization; AWaRe, Access, watch, and reserve. Timeframe: Short = 3–4 years; Mid = 5–9 years; Long = ≥10 years; N/A = Not applicable. † C. difficile infection rate: Expected ASP Outcome CVR met the threshold (0.500); however, Feasibility CVR did not meet the acceptance criterion (CVR = −0.571). The Reject judgment was retained. ‡ (ED) Appropriate empirical antibiotic selection: Not adopted primarily due to concerns regarding operational feasibility (systematic tracking not yet widely available in Korean hospitals), case heterogeneity, and reliance on microbiological confirmation data unavailable at the time of ED prescription. Discussion This study represents the first national consensus in Korea on ASP performance indicators developed through a rigorous Delphi process in the context of a newly launched government-led reimbursement pilot. Through a systematic evidence review and structured expert consensus, we identified seven major domains and initially defined 59 indicators encompassing both short-term measures (antibiotic consumption and adherence to restricted antibiotic use) and medium-to long-term measures (resistance trends and clinical outcomes). Through the subsequent Delphi process, this list was iteratively refined to a final set of 13 core indicators comprising six antibiotic use metrics, three process-related indicators, three indicators targeting high-risk clinical settings, and one additional cross-cutting measure. Rather than replicating earlier single-center or drug-class–focused analyses, this framework was designed to enable nationwide benchmarking at a critical juncture when ASPs are formally integrated into national health policy. In Korea, antibiotic consumption remains among the highest compared with other high-income countries, and national strategies have increasingly prioritised ASPs [10]. Although many hospitals have independently implemented ASPs, these efforts have largely remained confined to individual institutions. Previous single-center studies have demonstrated reductions in antibiotic use and resistance through infectious disease specialist–led or stepwise ASP models, confirming the feasibility of institutional ASP implementation [12–14]. However, such experiences cannot substitute for a harmonized, evidence-based national framework. Therefore, the introduction of a reimbursement-based ASP pilot program in 2024 provided a timely opportunity to develop standardized indicators that could ensure consistent evaluation and support sustainable policy expansion. Notably, the indicators that achieved consensus were overwhelmingly structural and process-oriented, including antibiotic consumption metrics (DOT, DDD, and broad-spectrum reduction), intravenous-to-oral conversion, appropriateness of surgical prophylaxis, and implementation of core ASP elements. These domains are broadly consistent with established ASP indicator frameworks from other countries, which similarly prioritise antibiotic use monitoring, prescribing processes, and stewardship infrastructure [18, 19]. However, unlike some international frameworks, no clinical outcome, economic, or microbiological indicators met the predefined validity thresholds in our study. This distribution reflects a pattern similar to international experiences in which process measures consistently dominate ASP evaluation frameworks. For instance, O’Riordan et al. reported that 75% of 229 indicators across 16 studies were process-related measures, whereas only 1% were outcome indicators [18]. Similarly, the ASP indicator set established by Centers for Disease Control and Prevention is composed almost entirely of structure and process-related measures, with no routine requirements for clinical outcomes [19]. These differences likely reflect practical and contextual challenges in standardising indicators that are highly dependent on institutional and healthcare system characteristics. Clinical and microbiological outcome measurement is often hampered by variability in diagnostic criteria, laboratory capacity, surveillance practices, reporting standards, and patient case mix [20]. Consistent with these challenges, our Delphi panel found it difficult to reach consensus on CDI incidence as a clinical outcome because of the heterogeneity in diagnostic methods and surveillance practices across hospitals in Korea. Panel discussions revealed that many hospitals remain in the early stages of ASP implementation, and the wide variability in baseline resistance, patient populations, and stewardship interventions makes the derivation of standardized outcome indicators currently impractical. As ASP infrastructure and data systems advance, selected outcome measures may be incorporated, particularly in settings with standardized diagnostic practices and surveillance systems. However, as demonstrated in international frameworks, outcome indicators remain difficult to operationalize and are often included only conditionally or on a limited basis [21]. Thus, any future expansion of outcome metrics should be pursued cautiously and only when data comparability and methodological robustness are ensured. From a policy perspective, the indicators identified through the Delphi process—primarily antibiotic-use metrics, process-related indicators such as intravenous-to-oral conversion and surgical prophylaxis appropriateness, and structural elements aligned with core ASP activities—were deliberately selected for feasibility within existing hospital information systems, enabling near-term implementation. At the national level, these metrics can facilitate benchmarking, feedback mechanisms, and performance-based reimbursement [19, 21, 22]. Importantly, health authorities can use the indicator set to systematically identify areas requiring improvement, inform priority setting, and support evidence-based policy formulation, as ASP implementation expands nationwide. Although the framework was developed specifically for the Korean healthcare system, some of the principles identified through this consensus process may be informative for other settings facing similar implementation and resource challenges. Furthermore, defining a concise core indicator set for smaller communities or long-term care hospitals is essential to ensure equitable participation in future national programs. Successful implementation will require further operational refinement of several indicators that have reached consensus but lacked standardized definitions [21, 22]. In particular, metrics such as broad-spectrum antibiotic reduction and intravenous-to-oral conversion, require clearer specification of target drug classes and eligibility criteria to ensure consistent measurement across hospitals. Early assessments may focus on overall changes in antibiotic use; however, long-term approaches should stratify indicators by disease category, clinical department, or patient risk group to enhance interpretability and actionability. This study had two main limitations. First, although our consensus was grounded in systematic evidence and expert opinion, empirical validation of indicator reliability and responsiveness remains necessary [9, 18]. Second, although we conducted a structured literature review to identify candidate indicators, some relevant indicators may not have been captured because of limitations in search strategy, database coverage, or reporting in the published literature. In addition, the final indicator set may have been influenced by the composition of the expert panel and the operational context of Korean hospitals. The limited inclusion of clinical outcome indicators likely reflects both the early stage of ASP implementation in Korea and the methodological challenges of standardizing outcome attribution across heterogeneous settings. Future research should incorporate these dimensions to assess the long-term impact of ASP performance and establish the association between process measures and patient outcomes. Conclusions This study provides Korea’s first national consensus on ASP performance metrics at a critical policy juncture. As the first nationally scaled effort to standardize ASP evaluation, the proposed framework addresses the urgent needs identified by the 2024 government-led reimbursement pilot program. Although tailored to the Korean healthcare system, these indicators may serve as valuable references for comparable healthcare environments with limited resources. Adoption of these standardized metrics can enhance ASP accountability, guide equitable reimbursement, support the maturation of institutional ASPs, and foster international collaboration in evaluating ASP effectiveness. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Material 1.