Korean Standardized Antimicrobial Administration Ratio in Long-Term Care Hospitals: A Benchmark Tool

doi:10.21203/rs.3.rs-8091805/v1

Download PDF

Research Article

Korean Standardized Antimicrobial Administration Ratio in Long-Term Care Hospitals: A Benchmark Tool

https://doi.org/10.21203/rs.3.rs-8091805/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Antimicrobial stewardship in long-term care hospitals (LTCHs) is increasingly important ; however, standardized benchmarking tools for antimicrobial use in these settings are lacking. The Korean Standardized Antimicrobial Administration Ratio for long-term care hospitals (K-SAAR–LTCHs) was developed to benchmark antimicrobial use in these facilities.

Methods

We used 2021 National Health Insurance claims data from 1,432 LTCHs (232,826 inpatients) in Korea. Data were split into training (80%) and test (20%) sets for model validation. Main outcome measure was measured in defined daily doses per 1,000 patient-days (DPD). Predicted antimicrobial use was estimated using a negative binomial regression model, selected by the Akaike Information Criterion (AIC) for best model fit. Predictive performance was evaluated using mean absolute percentage error (MAPE) and coefficient of determination (R²). The standardized antimicrobial administration ratio (SAAR) was calculated as observed-to-predicted antimicrobial use for each LTCH.

Results

Antimicrobial use under fee-for-service payment was substantially higher than under fixed-rate payment (1,199.3 vs 82.6 DPD). Higher antibiotic use was associated with male sex, severe illness, and indwelling catheter use, whereas greater pharmacist staffing was linked to lower use. Mean SAAR values for pneumonia and septicemia were approximately 1.0, indicating overall antimicrobial use near predicted levels. However, 19.6% of hospitals exceeded expected use for pneumonia and 5.2% for septicemia. Model performance was robust (MAPE 1.2–2.6; R² up to 0.93).

Conclusions

The K-SAAR–LTCHs model enables risk-adjusted benchmarking of antimicrobial use in LTCHs. This model identifies institutional variability and supports data-driven antimicrobial stewardship, offering an evidence-based foundation for national surveillance and quality improvement in Korea.

long-term care hospitals

antimicrobial

defined daily doses

Korea

Due to the rapidly accelerating pace of population aging, the importance of long-term-care hospitals (LTCHs) is increasing worldwide [1, 2]. LTCHs specialise in managing chronic diseases and providing sustained care for older adults [1, 3, 4]. As of 2025, 20.3% of Korea’s population is aged 65 years or older indicating that the country has reached super-aged status. Approximately 1,600 LTCHs are in operation nationwide, out of about 37,500 hospitals and clinics [5].

As the number of older adults requiring long-term care continues to grow, healthcare–associated infections (HAIs) have emerged as a major cause of morbidity and mortality among residents of LTCHs [6]. According to a systematic review of 31 studies, the overall pooled point prevalence of HAIs in LTCFs was 3.5–3.9%, with urinary tract infections (38.9%), respiratory tract infections (33.6%), and skin or soft-tissue infections (23.7%) being the most frequently reported [7]. To manage these infections, antimicrobials are commonly prescribed among LTCH patients [6, 8].

The US Centers for Disease Control and Prevention (CDC) recommend that hospitals implement antibiotic stewardship programs and have introduced the standardized antimicrobial administration ratio (SAAR) as a tool for evaluating and managing antimicrobial use in hospitals [9, 10]. This metric assesses the appropriateness of antimicrobial use by comparing the actual amount administered with the predicted amount.

In Korea, several studies have reported that antimicrobial use in LTCHs is higher than in acute care hospitals [11–13], and that high and inappropriate antibiotic use contributes to the emergence and spread of antimicrobial resistance [14], Although promoting appropriate antibiotic use in LTCHs is a major challenge for both patient safety and public health [8, 15]. broad-spectrum antimicrobials are administered more frequently in some LTCHs than in others, potentially serving as a major driver of antimicrobial resistance [16]. In this context, the Korean Standardized Antimicrobial Administration Ratio (K-SAAR) model has been developed and is currently being implemented in acute care hospitals [17].

However, despite its success in acute care settings, applying the SAAR framework to specialized institutions such as LTCHs remains challenging. LTCHs differ substantially from acute care hospitals in patient characteristics, care objectives, and reimbursement structures. Therefore, it is essential to develop an antimicrobial administration evaluation model that reflects the unique characteristics of LTCHs. This study aimed to adapt the Korean Standardized Antimicrobial Administration Ratio for long-term care hospitals (K-SAAR–LTCHs) framework to the LTCH setting to establish a new system for evaluating antimicrobial administration.

Data Source and Study Population

This study utilized data obtained from the National Health Insurance (NHI) claims data from Health Insurance Review & Assessment Service (HIRA) for the period spanning Januaryto December 2021. South Korea has a single-payer NHI system, which provides coverage for all citizens (51.8 million) and reimburses health care providers The medical institutions submit claims data to the HIRA for reimbursement for fee-for-service (FFS). The NHI claims database includes patient demographic characteristics, disease diagnosis codes, international non-proprietary names of prescribed medications, daily dosages, and treatment durations.[18, 19] This database NHI claims data and resource utilization data from LCTHs, and long-term care benefit appropriateness evaluation results. Long-term care hospitals in Korea operate under 2 reimbursement systems: fixed-rate (per diem) and fee-for-service (FFS). Because prescribing patterns may differ by payment type, antimicrobial use was analyzed separately for each system.

In 2021, a total of 429,348 patients admitted to 1,614 long-term care hospitals were identified. After excluding newly opened or closed institutions (n = 177) and patients without antimicrobial prescriptions (n = 191,166), 238,182 patients from 1,434 hospitals remained. Following the exclusion of those without fixed-rate or fee-for-service claims, the final analysis included 232,826 patients from 1,432 hospitals (Fig. 1).

Study Antibiotics and Measures

In this study, antibiotics were defined as J01 (antibacterials for systemic use) of the Anatomic Therapeutic Chemical (ATC) classification of the World Health Organization Collaborating Center (WHOCC).[20]Each drug has been registered in the NHI reimbursement list scheme and has manufacturer, individual product numbers, active ingredient codes, dose, and formulation. We mapped ATC classification using main active ingredient name and route of admistration, and classified the antibiotics into the ATC-3 level, which refers to the pharmacological subgroup.

The outcome measure was consumption presented as the numbers of Defined Daily Dose (DDDs)/1,000 patients/day (DPD), which is calculated as number of DDDs × 1000 / total patients / 365. The WHO defines the DDD of a single formulation based on the typical daily dose of an adult (70 kg), and since biological efficacy is different for each formulation, DDDs of the same drug differ according to the formulation. DDDs provide a fixed unit a measurement independent of price, package size, and strength, enabling to assess trends in drug consumption [21].

Admission episodes per patient unit were defined to facilitate calculation of the length of hospital stay, based on the benefit type provided by each long-term care hospital. The benefit types of FFS included the most severe medical condition, severe medical condition, moderate medical condition, mild medical condition, elective admission, pneumonia, septicemia, and isolation ward admission. Supplementary Table 1 summarizes the clinical factors that may influence antimicrobial administration in long-term care hospitals. Variables were excluded if missing values accounted for at least 10% of the data or if multicollinearity was identified among variables.

Development of Standardized Antimicrobial Administration Model

The standardized antimicrobial administration ratio (SAAR) serves as a metric for comparing predicted versus actual antimicrobial use in long-term care hospitals, while accounting for institutional characteristics. A SAAR value greater than 1 indicates that actual antimicrobial use exceeds the predicted amount, whereas a value less than 1 indicates lower-than-predicted use. This ratio is valuable for evaluating the appropriateness of antimicrobial use relative to the specific features of long-term care hospitals.

$$\:\varvec{S}\varvec{A}\varvec{A}\varvec{R}=\:\frac{\mathbf{A}\mathbf{c}\mathbf{t}\mathbf{u}\mathbf{a}\mathbf{l}\:\mathbf{a}\mathbf{m}\mathbf{o}\mathbf{u}\mathbf{n}\mathbf{t}\:\mathbf{a}\mathbf{d}\mathbf{m}\mathbf{i}\mathbf{n}\mathbf{i}\mathbf{s}\mathbf{t}\mathbf{e}\mathbf{r}\mathbf{e}\mathbf{d}}{\mathbf{P}\mathbf{r}\mathbf{e}\mathbf{d}\mathbf{i}\mathbf{c}\mathbf{t}\mathbf{e}\mathbf{d}\:\mathbf{a}\mathbf{m}\mathbf{o}\mathbf{u}\mathbf{n}\mathbf{t}\:\mathbf{a}\mathbf{d}\mathbf{m}\mathbf{i}\mathbf{n}\mathbf{i}\mathbf{s}\mathbf{t}\mathbf{e}\mathbf{r}\mathbf{e}\mathbf{d}}$$

A model was developed to standardize various factors influencing antimicrobial administration, tailored to the characteristics of long-term care hospitals. A negative binomial regression model was employed to predict the standardized amount of antimicrobial administration. The negative binomial regression model was selected because antimicrobial use data showed overdispersion, which could not be adequately addressed by Poisson regression. Model selection was based on the Akaike Information Criterion (AIC) to ensure the best balance between goodness of fit and parsimony. Predictive accuracy was evaluated using the mean absolute percentage error (MAPE), which provides an interpretable measure of prediction error across antibiotic categories.

Data from medical services provided in LTCHs were categorized by claim type. Eighty percent of this data was assigned to model training (training set), and an optimized prediction model was developed using k-fold cross-validation. The remaining 20% of the data served as a test set to evaluate the model’s predictive accuracy (Supplementary Fig. 1). Institutional characteristics influencing antimicrobial administration were analyzed, and the most suitable model was selected using stepwise regression. The final model was chosen based on the Akaike information criterion. Furthermore, the coefficient of determination was used to assess the predictive power of the model developed from the training set.

Two standardized antimicrobial administration models for fee-for-service claims were developed, specifically targeting pneumonia and septicemia. Additional models were constructed for broad-spectrum antimicrobials, primarily used for hospital-acquired infections, and for carbapenem antimicrobials, which are commonly prescribed for both pneumonia and septicemia.

Distribution of Antimicrobial use in Long-term Care Hospitals

When the total amount of antimicrobials administered was calculated according to claim type and length of hospital stay, antimicrobial use under the fee-for-service system was substantially higher than that under the fixed-rate system, at 1,199.3 DPD and 82.6 DPD, respectively. Among fixed-rate claim types, the highest density of antimicrobial administration per hospital day was observed in cases involving the most severe medical conditions (262.6 DPD), followed by severe medical conditions (144.3 DPD), moderate medical conditions (57.3 DPD), mild medical conditions (43.7 DPD), and elective admissions (39.6 DPD) (Table 1).

Table 1

Antimicrobial Administration by Antimicrobial Claim Type (Denominator: Sum of the Length of Hospital Stay for Each Benefit Type)
Category	Claim type	Total amount administered	Total number of patients	Total length of hospital stay	Amount used per 1,000 days of hospital stay
Total		9,896,412.0	390,986	74,996,480	132.0
Fixed rate	Total	5,970,003.7	382,445	72,275,626	82.6
	Most severe medical conditions	278,065.4	9,277	1,058,850	262.6
	Severe medical conditions	3,222,743.3	141,251	22,332,887	144.3
	Moderate medical conditions	1,505,188.7	171,814	26,260,938	57.3
	Mild medical conditions	720,299.1	101,685	16,465,483	43.7
	Elective admission group	243,707.2	65,526	6,157,468	39.6
Fee-for-service	Total	3,926,408.4	94,494	3,273,996	1,199.3
	Pneumonia	1,878,672.9	45,959	1,242,303	1,512.3
	Septicemia	1,730,000.8	48,663	1,278,289	1,353.4

Factors Influencing Antimicrobial Use in LTCHs

Table 2 summarizes the determinants of antibiotic utilization under both fixed-sum payment and fee-for-service systems. At the patient level, male sex was associated with higher odds of antibiotic prescribing (OR, 1.04; 95% CI, 1.03–1.06). Patients with severe or very severe medical conditions also had increased odds (OR, 1.06; 95% CI, 1.04–1.03), as did those with indwelling catheter–related urinary tract infections (OR, 1.04; 95% CI, 1.02–1.06). At the institutional level, facilities with more than 20 doctors (OR, 1.07; 95% CI, 1.06–1.19) and more than 40 nurses (OR, 1.13; 95% CI, 1.10–1.15) showed higher antibiotic use, whereas the presence of ≥ 2 pharmacists was inversely associated (OR, 0.94; 95% CI, 0.91–0.97).

Table 2

Determinants of Antibiotic Utilization in Fixed-Sum Payment and Fee-for-Service Systems
Model		Factor	OR (95% CI)
Total antibiotics	Model1 (patient-level)
	Patient characteristics	Sex (male vs. female)	1.04 (1.03–1.06)
		Medical conditions (severe or worse vs. moderate)	1.06 (1.04–1.03)
		Standard score of urinary tract infection related to the indwelling catheter (yes vs. no)	1.04 (1.02–1.06)
	Institutional characteristics	Number of doctors (≥ 20)	1.07 (1.06–1.19)
		Number of nurses (≥ 40)	1.13 (1.10–1.15)
		Number of pharmacists (≥ 2)	0.94 (0.91–0.97)
	Model2 (claims-level)
	Patient characteristics	Sex (male vs. female)	1.02 (1.01–1.04)
	Patient characteristics	Medical conditions (more than severe vs. moderate)	1.03 (1.02–1.05)
	Institutional characteristics	Number of doctors (more than 20)	1.05 (1.04–1.07)
		Number of nurses (more than 40)	0.91 (0.90–0.92)
		Number of pharmacists (more than 2)	0.84 (0.83–0.86)
		Number of beds (more than 200)	1.06 (1.05–1.07)
		Standard score of local community return rates (Grade ≥ 3 vs. Grade < 3)	0.97 (0.96–0.98)
		Prescription prevalence of psychotropic agents (Grade ≥ 3 vs. Grade < 3)	1.05 (1.03–1.06)
Broad spectrum antibiotics	Model1 (patient-level)
	Patient characteristics	Insertion of an indwelling catheter (yes vs. no)	1.05 (1.02–1.07)
	Patient characteristics	Oxygen therapy (yes vs. no)	0.95 (0.93–0.97)
	Institutional characteristics	Number of doctors (more than 20)	0.95 (0.93–0.98)
		Number of nurses (more than 40)	1.16 (1.14–1.18)
		Number of pharmacists (more than 2)	1.24 (1.20–1.28)
	Model2 (claims-level)
	Patient characteristics	Insertion of an indwelling catheter (yes vs. no)	0.95 (0.93–0.97)
	Patient characteristics	Oxygen therapy (yes vs. no)	0.94 (0.92–0.96)
	Institutional characteristics	Number of doctors (more than 20)	0.96 (0.94–0.98)
		Number of nurses (more than 40)	1.13 (1.11–1.16)
		Number of pharmacists (more than 2)	1.17 (1.13–1.20)
Carbapenem	Model1 (patient-level)
	Patient characteristics	Sex (male vs. female)	0.95 (0.92–0.97)
	Patient characteristics	Oxygen therapy administered (yes vs. no)	0.95 (0.91–0.98)
	Institutional characteristics	Number of doctors (more than 20)	1.14 (1.10–1.19)
		Number of nurses (more than 40)	1.13 (1.06–1.19)
		Number of pharmacists (more than 2)	0.85 (0.79–0.97)
		Number of beds (more than 200)	0.91 (0.87–0.94)
		Standard score of local community return rates (Grade ≥ 3 vs. Grade < 3)	0.87 (0.85–0.90)
	Model2 (claims-level)
	Patient characteristics	Sex (male vs. female)	1.06 (1.03–1.08)
		Medical conditions (More than severe vs. moderate)	1.15 (1.10–1.21)
		Standard score of urinary tract infection related to the indwelling catheter (yes vs. no)	1.05 (1.02–1.08)
		)ECI Score (≥ 2 vs. <2)	0.93 (0.91–0.95)
		Long-term hospitalization (yes vs. no)	1.19 (1.16–1.23)
	Institutional characteristics	Number of doctors (more than 20)	1.17 (1.13–1.20)
		Number of nurses (more than 40)	1.16 (1.11–1.21)
		Number of beds (more than 200)	0.90 (0.87–0.92)
		Standard score of local community return rates (Grade ≥ 3 vs. Grade < 3)	0.95 (0.92–0.97)
		Prescription prevalence of psychotropic agents (Grade ≥ 3 vs. Grade < 3)	0.91 (0.88–0.94)

At the claim level, male sex (OR, 1.02; 95% CI, 1.01–1.04) and severe medical conditions (OR, 1.03; 95% CI, 1.02–1.05) remained significant predictors. Larger institutions with more than 200 beds (OR, 1.06; 95% CI, 1.05–1.07) and higher psychotropic prescribing prevalence (OR, 1.05; 95% CI, 1.03–1.06) were associated with greater antibiotic use, whereas having ≥ 40 nurses (OR, 0.91; 95% CI, 0.90–0.92), ≥ 2 pharmacists (OR, 0.84; 95% CI, 0.83–0.86), and higher community return rates (OR, 0.97; 95% CI, 0.96–0.98) were inversely associated.

For broad-spectrum antibiotics, at the patient level, indwelling catheter insertion increased prescribing (OR, 1.05; 95% CI, 1.02–1.07), whereas oxygen therapy reduced it (OR, 0.95; 95% CI, 0.93–0.97). At the institutional level, hospitals with more than 40 nurses (OR, 1.16; 95% CI, 1.14–1.18) and ≥ 2 pharmacists (OR, 1.24; 95% CI, 1.20–1.28) demonstrated greater use, while institutions with > 20 doctors showed lower use (OR, 0.95; 95% CI, 0.93–0.98). Claims-level findings confirmed similar patterns, with nurse and pharmacist counts positively associated and oxygen therapy negatively associated with broad-spectrum antibiotic use.

For carbapenem antibiotics, at the patient level, male sex (OR, 0.95; 95% CI, 0.92–0.97) and oxygen therapy (OR, 0.95; 95% CI, 0.91–0.98) were inversely associated with use. Institutional characteristics such as > 20 doctors (OR, 1.14; 95% CI, 1.10–1.19) and > 40 nurses (OR, 1.13; 95% CI, 1.06–1.19) correlated with higher use, whereas ≥ 2 pharmacists (OR, 0.85; 95% CI, 0.79–0.97), larger bed capacity (> 200 beds; OR, 0.91; 95% CI, 0.87–0.94), and higher community return rates (OR, 0.87; 95% CI, 0.85–0.90) were associated with lower use. At the claim level, severe medical conditions (OR, 1.15; 95% CI, 1.10–1.21), indwelling catheter infections (OR, 1.05; 95% CI, 1.02–1.08), and long-term hospitalization (OR, 1.19; 95% CI, 1.16–1.23) increased carbapenem use, whereas higher ECI scores (OR, 0.93; 95% CI, 0.91–0.95), larger institutions (> 200 beds; OR, 0.90; 95% CI, 0.87–0.92), and higher psychotropic prescribing (OR, 0.91; 95% CI, 0.88–0.94) were inversely associated.

Comparison of Standardized Antimicrobial Administration Ratios in Long-Term Care Hospitals

As shown in Table 3, the SAAR for each long-term care hospital was calculated by comparing predicted and observed antimicrobial use. While overall antimicrobial use for pneumonia and septicemia approximated predicted levels (SAAR ≈ 1.0), pneumonia cases exhibited greater variability, with nearly one-fifth of institutions falling outside the expected range. In contrast, broad-spectrum and carbapenem antimicrobials demonstrated more stable patterns, although several hospitals consistently showed higher use.

Table 3

**Comparison of SAAR in Pneumonia and Sepsis**
		Amount of antimicrobials administered			Abnormal values
		Observed value (95% CI)	Predicted value (95% CI)	SAAR*	Total	Lower limit	Upper limit
Pneumonia	Total	129.0 (127.5, 130.5)	127.3 (126.1, 128.6)	1.0 (0.5, 1.9)	193(19.6%)	134(13.6)	59(6.0%)
	Broad-spectrum antibiotics primarily used for hospital-acquired infections	114.0 (112.6, 115.3)	119.0 (117.8, 120.2)	1.0 (0.99, 1.00)	153(17.1%)	148(16.4%)	5(0.6%)
	Carbapenem antimicrobials	47.0 (44.9, 49.1)	54.1 (52.6, 55.6)	1.0 (0.90, 1.01)	168(19.2%)	122(13.9%)	46(5.3%)
Septicemia	Total	111.2 (109.9, 112.4)	109.2 (108.2, 110.2)	1.0 (0.8, 1.1)	163(18.9)	118(13.7)	45(5.2%)
	Broad-spectrum antibiotics primarily used for hospital-acquired infections	103.2 (102.0, 104.5)	109.9 (108.7, 111.0)	0.9 (0.8, 1.1)	136(17.1%)	131(16.5%)	5(0.6%)
	Carbapenem antimicrobials	49.6 (48.7, 50.5)	56.4 (55.7, 57.2)	1.0 (0.8, 1.3)	92(15.5%)	74(12.5%)	18(3.0%)

For pneumonia, predicted antimicrobial use was 127.3 DPD (95% CI, 126.1–128.6) compared with an observed 129.0 DPD (95% CI, 127.5–130.5). The mean SAAR was 1.0 (95% CI, 0.5–1.9), indicating substantial variability across hospitals. A total of 193 institutions (19.6%) fell outside the 95% CI, with 89 (6.0%) exceeding the upper limit. For broad-spectrum antibiotics, the predicted mean was 119.0 DPD compared with an observed 114.0 DPD, yielding a SAAR of 1.0 (95% CI, 0.99–1.00). Variability was minimal, with only five hospitals (0.6%) above the upper limit. For carbapenems, predicted use was 54.1 DPD (95% CI, 52.6–55.6) versus an observed 47.0 DPD (95% CI, 44.9–49.1); the mean SAAR was 1.0 (95% CI, 0.90–1.01), and 46 hospitals (5.3%) exceeded the upper threshold.

For septicemia, predicted antimicrobial use was 109.2 DPD (95% CI, 108.2–110.2) compared with an observed 111.2 DPD (95% CI, 109.9–112.4), resulting in a SAAR of 1.0 (95% CI, 0.8–1.1). Forty-five hospitals (5.2%) exceeded the upper limit. Broad-spectrum antibiotic use was 103.2 DPD observed versus 109.9 DPD predicted, with a SAAR of 0.9 (95% CI, 0.8–1.1), and only five institutions (0.6%) were above the upper limit. For carbapenems, predicted use was 56.4 DPD (95% CI, 55.7–57.2) compared with an observed 49.6 DPD (95% CI, 48.7–50.5), yielding a SAAR of 1.0 (95% CI, 0.8–1.3); 18 hospitals (3.0%) were above the threshold.

Assessment of SAAR distributions using funnel plots (Supplementary Figs. 2 and 3) showed that most institutions fell within the 95% CI.

This study developed and validated a Korean version of the standardized antimicrobial administration ratio (K-SAAR–LTCHs) model tailored to long-term care hospitals. Using nationwide claims data, we identified patient-, institutional-, and claim-level determinants of antimicrobial use, compared observed and predicted consumption across antimicrobial categories, and demonstrated the feasibility of benchmarking antimicrobial use in LTCHs through a risk-adjusted model.

In this large-scale analysis including 232,826 inpatients across 1,432 Korean LTCHs, several key determinants of antimicrobial use were identified. At the patient level, male sex, severe medical conditions, and indwelling catheter–related infections were associated with higher antibiotic use. At the institutional level, a higher number of physicians and nurses was linked to increased antimicrobial consumption, whereas greater pharmacist staffing consistently reduced use. These findings reinforce the importance of multidisciplinary stewardship, particularly the role of pharmacists in optimizing antimicrobial therapy.

Although mean K-SAAR–LTCHs values for pneumonia and septicemia were close to 1.0, suggesting overall use near predicted levels, substantial interhospital variability persisted, with 19.6% and 5.2% of hospitals exceeding expected antimicrobial use, respectively. This variation highlights opportunities for targeted stewardship interventions in high-utilization facilities. Model performance was robust, with R² values up to 0.93 and MAPE ranging from 1.2 to 2.6, confirming the reliability of the K-SAAR–LTCHs model as a benchmarking tool.

The U.S. Centers for Disease Control and Prevention (CDC) introduced the SAAR metric for acute-care hospitals within the National Healthcare Safety Network (NHSN) [9, 10]. The K-SAAR–LTCHs expands this framework to a fundamentally different care environment. Unlike the U.S. SAAR, which focuses on short-stay acute-care facilities, the K-SAAR–LTCHs incorporates long-term admissions, fixed-rate reimbursement structures, and patient populations characterized by chronic conditions and functional dependency.

This model also differs from the previously developed Korean SAAR for acute-care hospitals [17]. The acute-care K-SAAR used data primarily from general and tertiary hospitals under fee-for-service reimbursement, emphasizing infection type and case-mix adjustment. In contrast, the K-SAAR–LTCHs integrates contextual factors unique to long-term care, such as prolonged hospitalization, fixed-rate payment mechanisms, patient cognitive and functional status, and institutional staffing levels. These features improve model fit and interpretability for the LTCH setting and demonstrate that the SAAR framework can be effectively adapted to diverse care environments. Notably, the inclusion of workforce-related variables, particularly pharmacist and nurse staffing, provides actionable insight for antimicrobial stewardship, which is less emphasized in both the U.S. and Korean acute-care models.

Our findings align with prior Korean studies reporting high and heterogeneous antimicrobial use in LTCHs [16, 17, 22]. and complement international evidence underscoring institutional culture, physician practice patterns, and resource allocation as major drivers of prescribing behavior [23, 24]. These converging findings suggest that antimicrobial stewardship in LTCHs must address not only clinical complexity but also organizational and behavioral determinants of prescribing. The K-SAAR–LTCHs distribution observed in this study parallels results from U.S. SAAR analyses, where outlier hospitals with excessive antimicrobial use were similarly identified [9, 10], supporting the generalizability of SAAR-based approaches to long-term care settings.

The K-SAAR–LTCHs model provides a scalable, data-driven foundation for antimicrobial stewardship and quality improvement within Korea’s long-term care sector. By enabling risk-adjusted comparisons across institutions, it can identify facilities with excessive antimicrobial use and support targeted feedback, education, and monitoring. These insights can guide national initiatives led by the HIRA and the Ministry of Health and Welfare to strengthen antimicrobial stewardship infrastructure in LTCHs. Incorporating SAAR-based indicators into regular institutional evaluations could further promote rational antibiotic prescribing, particularly in settings with limited infectious disease specialists.

Strengths and Limitations

The strengths of this study include its use of a comprehensive, nationwide dataset encompassing nearly all Korean LTCHs and its integration of patient-, institutional-, and system-level factors into a unified benchmarking model. This is the first study to adapt and validate the SAAR framework for long-term care institutions, which differ markedly from acute care hospitals in patient profiles, care objectives, and reimbursement structures. The revised model addresses limitations of the original SAAR, which was designed primarily for acute care settings. A standardized antimicrobial administration model tailored to LTCHs is now expected to provide critical data for evaluating prescribing appropriateness, refining antimicrobial guidelines, and ultimately reducing antimicrobial resistance in these facilities.

Although the SAAR model serves as a valuable tool for assessing the appropriateness of antimicrobial use, this study identified several limitations. First, this analysis relied on claims-based data lacking microbiological results, infection severity, and clinical decision-making context, which limits evaluation of prescribing appropriateness at the patient level. Second, diagnostic coding variability may have led to misclassification of infection categories. Third, residual confounding due to unmeasured institutional factors such as stewardship culture or laboratory capacity cannot be excluded. Fourth, downstream outcomes—such as infection resolution, readmission, or mortality—were not examined. Finally, findings are specific to the Korean LTCH system and may not directly generalize to other health systems

Despite these limitations, the K-SAAR–LTCHs is the first nationally validated, risk-adjusted model for benchmarking antimicrobial use in long-term care hospitals. By extending both the U.S. CDC SAAR framework and the Korean acute-care K-SAAR to a new healthcare domain, this model highlights the feasibility of adapting standardized antimicrobial use metrics to population-specific and system-specific contexts. The K-SAAR–LTCHs provides actionable insights into prescribing behavior, identifies institutional variability, and supports evidence-based stewardship and policy development aimed at reducing antimicrobial resistance in Korea.

Ethical considerations

This study received approval from the Institutional Review Board (IRB) of the Health Insurance Review and Assessment Service (HIRA), under IRB No. 2024-015 -001. The IRB waived the requirement for written informed consent, as all datasets utilized were anonymous and had been de-identified.

Author Contribution

J.S and DS.K wrote the main manuscript text and J.S and J.C conducted the analysis.All authors reviewed the manuscript.

Data Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. De-identified participant-level data that support the findings of this study can be made available upon reasonable request and subject to approval by the institutional ethics committee and data-sharing agreements, where required.

Barber SL, Ong O, Han ZA. Long-Term Care in Ageing Populations. World Health Organization; 2021.
Martinez-Lacoba R, Pardo-Garcia I, Escribano-Sotos F, Aging. Dependence, and Long-Term Care: A Systematic Review of Employment Creation. Inquiry. 2021;58:469580211062426. : 10.1177/00469580211062426. PubMed PMID: 34913376; PubMed Central PMCID: PMCPMC8695749.
Yun SY, Lim JY, Kim E, Oh J, Lee DH. Epidemiology and Clinical Characteristics of Older Patients Transferred from Long-Term-Care Hospitals (LTCHs) to Emergency Departments by a Comparison with Non-LTCHs in South Korea: A Cross-Sectional Observational Study. Int J Environ Res Public Health. 2022;19(14). Epub 20220721. 10.3390/ijerph19148879. PubMed PMID: 35886728; PubMed Central PMCID: PMCPMC9316191.
Prince MJ, Wu F, Guo Y, Gutierrez Robledo LM, O'Donnell M, Sullivan R, et al. The burden of disease in older people and implications for health policy and practice. Lancet. 2015;385(9967):549–62. 10.1016/S0140-6736(14)61347-7. Epub 20141106.
Statistics KN. Senior citizen statistics data in 2023. Daejeon: Korea National Statistics Office; 2024.
Ricchizzi E, Sasdelli E, Leucci AC, Fabbri E, Caselli L, Latour K, et al. Incidence of health-care-associated infections in long-term care facilities in nine European countries: a 12-month, prospective, longitudinal cohort study. Lancet Infect Dis. 2025. 10.1016/S1473-3099(25)00217-8.
Bennett N, Tanamas SK, James R, Ierano C, Malloy MJ, Watson E, et al. Healthcare-associated infections in long-term care facilities: a systematic review and meta-analysis of point prevalence studies. BMJ Public Health. 2024;2(1):e000504. 10.1136/bmjph-2023-000504. Epub 20240527.
Eze N, Cecchini M, Hsashiguchi TCO. Antimicrobial resistance in long-term care facilities. Paris: Organization for Economic Co-operation and Development; 2022.
Polk RE, Hohmann SF, Medvedev S, Ibrahim O. Benchmarking risk-adjusted adult antibacterial drug use in 70 US academic medical center hospitals. Clin Infect Dis. 2011;53(11):1100–10. 10.1093/cid/cir672. Epub 20111013.
van Santen KL, Edwards JR, Webb AK, Pollack LA, O'Leary E, Neuhauser MM, et al. The Standardized Antimicrobial Administration Ratio: A New Metric for Measuring and Comparing Antibiotic Use. Clin Infect diseases: official publication Infect Dis Soc Am. 2018;67(2):179–85. 10.1093/cid/ciy075. PubMed PMID: 29409000.
Park J, Han E, Lee SO, Kim DS. Antibiotic use in South Korea from 2007 to 2014: A health insurance database-generated time series analysis. PLoS ONE. 2017;12(5):e0177435. 10.1371/journal.pone.0177435. PubMed PMID: 28520761; PubMed Central PMCID: PMC5435228.
Chae J, Kim DS, Shin J, Kim YC, Ji SY, Kim Y, et al. Ten-year trends of antibiotics used for patients with pneumonia at long-term care hospitals in the Republic of Korea: An analysis based on national health insurance claims data. Heliyon. 2024;10(12):e33047. PubMed PMID: 39021963; PubMed Central PMCID: PMCPMC11252976.
Kim YC, Park JY, Kim B, Kim ES, Ga H, Myung R, et al. Prescriptions patterns and appropriateness of usage of antibiotics in non-teaching community hospitals in South Korea: a multicentre retrospective study. Antimicrob Resist Infect Control. 2022;11(1):40. 10.1186/s13756-022-01082-2. Epub 20220222.
Holmes AH, Moore LS, Sundsfjord A, Steinbakk M, Regmi S, Karkey A, et al. Understanding the mechanisms and drivers of antimicrobial resistance. Lancet. 2016;387(10014):176–87. 10.1016/S0140-6736(15)00473-0. Epub 2015/11/26.
WHO. Antimicrobial resistance. 2023.
Lee R, Park SY, Park JY, Kim B, Kim YC, Ga H, et al. Evaluating antimicrobial utilization in 20 Korean long-term care hospitals: a call to action for antimicrobial stewardship. J Hosp Infect. 2024;153:65–72. 005. PubMed PMID: 39181451.
Kim B, Ahn SV, Kim DS, Chae J, Jeong SJ, Uh Y, et al. Development of the Korean Standardized Antimicrobial Administration Ratio as a Tool for Benchmarking Antimicrobial Use in Each Hospital. J Korean Med Sci. 2022;37(24):e191. PubMed PMID: 35726144; PubMed Central PMCID: PMCPMC9247727.
Cho HJ, Chae J, Yoon SH, Kim DS. Factors related to polypharmacy and hyper-polypharmacy for the elderly: A nationwide cohort study using National Health Insurance data in South Korea. Clin Transl Sci. 2023;16(2):193–205. 10.1111/cts.13438. Epub 2022/11/20.
Park D, Lee H, Kim DS. J Gen Intern Med. 2022;37(10):2390–7. 10.1007/s11606-021-07165-x. Epub 2021/10/28. High-Cost Users of Prescription Drugs: National Health Insurance Data from South Korea.
Anatomical Therapeutic Chemical (ATC)/Defined Daily Dose (DDD) Index [Internet]. WHOCC. 2021. Available from: https://www.whocc.no/atc_ddd_index/
OECD Health Statistic. 2020 Deifnitions, Sources and Methods - Pharmaceutical Market [Internet]. Orgnization for Economic Co-operation and Development. 2020.
Park SY, Kim YC, Moon SM, Kim B, Lee R, Kim HB, et al. Developing a framework for regular and sustainable qualitative assessment of antibiotic use in Korean medical institutions: a Delphi study. Antimicrob Resist Infect Control. 2023;12(1):114. 10.1186/s13756-023-01319-8. Epub 20231018.
Lim CJ, Kwong MW, Stuart RL, Buising KL, Friedman ND, Bennett NJ, et al. Antibiotic prescribing practice in residential aged care facilities–health care providers' perspectives. Med J Australia. 2014;201(2):98–102. 10.5694/mja13.00102. PubMed PMID: 25045989.
Dyar OJ, Huttner B, Schouten J, Pulcini C, Esgap. What is antimicrobial stewardship? Clin Microbiol Infect. 2017;23(11):793–8. 10.1016/j.cmi.2017.08. Epub 20170904.

No competing interests reported.

SupplementaryTablesandfigures.docx

Download PDF

Reviewers agreed at journal
09 Jan, 2026
Reviewers agreed at journal
30 Dec, 2025
Reviewers invited by journal
30 Dec, 2025
Editor assigned by journal
13 Nov, 2025
Submission checks completed at journal
13 Nov, 2025
First submitted to journal
11 Nov, 2025

You are reading this latest preprint version

Korean Standardized Antimicrobial Administration Ratio in Long-Term Care Hospitals: A Benchmark Tool

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Data Source and Study Population

Study Antibiotics and Measures

Development of Standardized Antimicrobial Administration Model

Results

Distribution of Antimicrobial use in Long-term Care Hospitals

Factors Influencing Antimicrobial Use in LTCHs

Comparison of Standardized Antimicrobial Administration Ratios in Long-Term Care Hospitals

Discussion

Strengths and Limitations

Conclusions

Declarations

Ethical considerations

Author Contribution

Data Availability

References

Additional Declarations

Supplementary Files

Status:

Version 1