Introduction
Antimicrobial resistance (AMR) represents a critical global public health crisis.Reference Charani, McKee and Ahmad1,Reference Charani and Holmes2 In response, Antimicrobial Stewardship (AMS) programs have become essential interventions to mitigate AMR progression and optimize therapeutic outcomes.Reference MacDougall and Polk3 The evolving role of pharmacists—transitioning from traditional drug dispensing to active engagement in patient-specific pharmacotherapy—has been increasingly recognized in AMS frameworks.Reference Toklu and Hussain4 International guidelines now explicitly endorse pharmacists as pivotal contributors to AMS, with organizations such as the Infectious Diseases Society of America (IDSA) and the Centers for Disease Control and Prevention (CDC) advocating for pharmacist leadership or co-leadership in these initiatives to enhance clinical outcomes.Reference Ponto5–7 Similarly, the Royal Pharmaceutical Society highlights pharmacists’ expertise in delivering evidence-based antimicrobial optimization strategies.8
Clinical benefits of pharmacist-integrated AMS are well-documented.Reference Alexis, Niccolò and Quentin9–Reference Nampoothiri, Mbamalu and Mendelson12 The multinational EUROBACT-2 cohort study demonstrated that pharmacist consultations reduced 28-day mortality rates in ICU patients with hospital-acquired bloodstream infections (adjusted OR: 0.62; 95% CI: 0.41–0.93).Reference Alexis, Niccolò and Quentin9 While China’s three-tier antimicrobial classification management system (established in 2004) has effectively standardized the clinical use of restricted-use antimicrobials (classified as Reserve group antibiotics under the WHO AWaRe framework), 13–15 the implementation of the Volume-Based Procurement (VBP) policy in 2018 introduced new challenges for managing these agents.16 The VBP policy utilizes a “volume-for-price” negotiation model to substantially reduce medication procurement costs (average price reduction: 53%).17 However, this price signaling mechanism may inadvertently create a “low-price preference” effect. Following the inclusion of restricted-use antimicrobials in procurement catalogs, their relative cost advantages incentivize clinicians to preferentially prescribe these agents—particularly broad-spectrum antibiotics—even when economic considerations conflict with evidence-based clinical judgment in infection management. Such conflicts may promote defensive medical practices, ultimately lowering the therapeutic threshold for restricted antimicrobials and fostering an anomalous “low-price, high-volume” consumption pattern.Reference Wang, Yang and Xu18 Compounding this issue, studies reveal that hospitals may relax prescription restrictions on restricted antimicrobials to fulfill procurement agreements, resulting in a 19%–28% increase in defined daily doses (DDDs) of carbapenems—agents requiring stringent stewardship.Reference Yang, Chen and Ke19
This study innovatively developed a pharmacist-led Antimicrobial Stewardship–Prospective Audit and Feedback System (AMS-APAS), establishing a dual real-time prescription review mechanism involving clinical pharmacists and infectious disease specialists for restricted-use antimicrobials. Our research aims to evaluate the impact of this pharmacist-driven AMS-APAS intervention on antimicrobial utilization patterns and antimicrobial resistance trends under VBP policy constraints, thereby proposing a novel strategy to reconcile cost containment with rational antimicrobial use.
Methods
Study design and setting
A retrospective quasi-experimental interrupted time-series analysis was conducted at Tongji Hospital, a 5,000-bed tertiary academic center in Wuhan, China, from January 2022 to December 2024. The intervention—prospective audit of special-restricted antimicrobials by clinical pharmacists and infectious disease specialists—was implemented in January 2023. The intervention covered all adult inpatient departments across the hospital. Due to systemwide upgrade in 2021, standardized data were available from 2022 onward. Clinical utilization data for restricted-use antimicrobials were extracted from the Hospital Information System (HIS) and Clinical Rational Drug Use System, spanning two periods: January–December 2022 (preimplementation phase) and January 2023–December 2024 (postimplementation phase). The data set encompassed Antimicrobial use metrics, Infection control parameters, Pathogen culture submission compliance, and Process indicators.
Intervention components
The implemented interventions comprised multiple complementary components designed to strengthen antimicrobial stewardship within the healthcare facility. These components are outlined as follows:
1. Prospective audit system
Prior to the intervention, clinical pharmacists participated in ward rounds but did not conduct real-time prescription reviews. The cornerstone of the intervention was a web-based audit system that seamlessly integrated with the hospital’s electronic health record system. This platform facilitated real-time submission and review of prescriptions, ensuring prompt scrutiny and decision-making. A particular focus was placed on 32 special-restricted antimicrobials (provided in Supplementary Table 1), including meropenem and tigecycline. For these agents, a stringent dual approval process was instituted, requiring endorsement from both a clinical pharmacist and an infectious disease physician. The system required the prescriber to complete the application form. This application included indicators associated with the patient’s infection, a summary of the patient’s conditions, previous and current use of antimicrobials, and the reason for the application. The system incorporated automated alerts to flag any prescriptions that lacked microbial culture confirmation or deviated from the guideline-recommended dosages, thus enhancing the accuracy and appropriateness of antibiotic use. The workflow for prescription review and the form of application are depicted in Figure 1 and Supplementary Figure 1.
Figure 1. Flowchart of the prospective audit process.
2. Standardized audit framework
The ASP team included pharmacists, infectious disease physicians, and microbiologists. Pharmacists led prescription reviews; ID physicians advised complex cases. Prescriptions were reviewed within 24 hours of submission, covering both empirical and definitive phases, using national and institutional guidelines. Nine full-time clinical pharmacists conducted reviews, supported by 56 hours of structured training. Audit criteria integrate the WHO AWaRe classification 14 and China’s Guidelines for Clinical Antimicrobial Use,13 establishing a multidimensional evaluation system: ① Appropriateness of Indication: Assessed by infection site, severity grading, and pathogen evidence level. ② Regimen Optimization: Adjusted based on creatinine clearance, Child-Pugh score, and drug interaction databases. ③ Upon identifying inappropriate prescriptions, pharmacists contacted prescribers directly and documented intervention outcomes.
3. Capacity-buildingprogram
A tiered training system was implemented: ① Foundational: Monthly AMS-VBP policy workshops with quarterly antimicrobial resistance surveillance updates. ② Professional: Pharmacist-focused advanced training (56 cumulative hours) on PK/PD modeling and MIC interpretation. ③ Practical: Multidisciplinary case discussions for simulated prescription decision-making (1–2 cases weekly).
4. Quality improvement cycle
A PDCA (Plan-Do-Check-Act) closed-loop management system was constructed: ① Monitoring: Monthly prescription quality reports with departmental rankings and error analysis. ② Feedback: Quarterly cross-departmental benchmarking reports and AMS performance scoring. ③ Improvement: Root-cause analysis for deviations in restricted antimicrobial use.
Data collection
Three data streams were extracted from the hospital information system: ① Medication Metrics: Antimicrobial Use Density (AUD, DDD/100 bed-days), prescription appropriateness rate, VBP drug expenditure ratio. ② Infection Control Parameters: Hospital-acquired multidrug-resistant organism rates, including carbapenem-resistant Klebsiella pneumoniae (CRKP), Acinetobacter baumannii (CRAB), Methicillin-Resistant Staphylococcus aureus (MRSA), and Pseudomonas aeruginosa (CRPA). ③ Process Indicators: Prescription review timeliness, microbiological culture submission rate, audit compliance rate. Primary outcomes: prescription appropriateness, AUD; Secondary outcomes: resistance incidence, costs. Audit compliance = % prescriptions reviewed within 24h.
Statistical analysis
Interrupted time series analysis (ITS) was employed to evaluate intervention effects. Monthly data from January 2022 to December 2024, with January 2023 as the intervention point. Segmented regression analysis using R 4.2.1 to assess significance of level (β2) and trend (β3) changes. Model robustness verified via Mann-Kendall trend testing. Stratified comparisons for restricted antimicrobials (Fisher’s exact test). Continuous variables were tested for normality using Shapiro-Wilk tests. Nonparametric data are presented as median (IQR), with between-group comparisons analyzed via Mann-Whitney U tests. All analyses were conducted using SPSS 26.0 and R 4.2.1, with two-tailed α = .05.
Results
Prescription rationality and antimicrobial utilization
The intervention significantly optimized prescription quality for restricted-use antimicrobials. ITS evaluated monthly proportion of appropriate prescriptions. Postimplementation, the prescription appropriateness approval rate increased from a baseline of 80.24% ± 3.21% to 93.83% ± 2.75% (P < .001), demonstrating a sustained upward trend (β = .43, P = .002). Overall antimicrobial utilization decreased by 4.08% (.49 ± .03 vs .47 ± .02, P = .032), while the proportion of restricted antimicrobial use declined by 12.84% (6.31% ± 1.12% vs 5.50% ± .89%, P = .004). Detailed results are presented in Table 1.
Table 1. Pre post comparison of the rationality and rate of Antimicrobial Usage (Intervention: 2023–2024 vs Pre intervention: 2022)
Antimicrobial use density (AUD)
The intervention achieved a 28.3% reduction in AUD, decreasing from 47.87 ± 5.12 to 34.25 ± 4.03 DDDs/100 bed-days (P < .001). The Emergency and Critical Care Department demonstrated the most pronounced reduction (158.9 ± 18.7 vs 126.2 ± 15.4 DDDs/100 bed-days, P < .001). Interrupted time series analysis revealed a 2.7-fold acceleration in the monthly AUD decline rate postintervention (β = -1.23, P = .008). Significant interdepartmental heterogeneity was observed (χ2 = 18.37, P = .002), with statistically meaningful reductions in Respiratory Medicine (-16.1%), Hematology (-17.9%), and Cardiothoracic Surgery (-16.2%) departments (P < .01 for all). Detailed data are presented in Figure 2.
Figure 2. Pre Post Comparison of Antimicrobial Drug Use Intensity (Intervention: 2023–2024 vs. Pre-Intervention: 2022).
High-risk antimicrobial classes and resistance patterns
Carbapenem use intensity decreased by 26.4% (5,234 ± 1,245 vs 3,876 ± 987 DDDs), with costs declining from $446,778.57 [$406,428.57–$509,571.43] to $300,642.86 [$268,.00–$362,.00] (P < .001). A strong positive correlation was observed between carbapenem use and CRKP isolation rates (R = .62, P < .05). Tigecycline DDD declined by 29% in DDDs (1,256 ± 284 vs 892 ± 167, P < .001), and costs decreased from $149,331.43 [$125,142.86–$176,285.71] to $98,214.29 [$76,.00–$120,714.29] (P < .001). A weak association with CRAB was noted (R = .38, P = .07). Antifungals DDD decreased by 26% (2,845 ± 567 vs 2,104 ± 432, P < .001), with total costs reduced from $304,928.57 [$268,.00–$362,.00] to $208,.00 [$176,285.71–$268,.00] (P < .001). Detailed results are shown in Tables 2 and 3.
Table 2. Pre post comparison of key Antimicrobial drug Usage cost metrics (Intervention: 2023–2024 vs Pre intervention: 2022)
Table 3. Correlation analysis of antimicrobial use and the incidence rate of multidrug-resistant infections
Positive correlations suggest potential link between usage and resistance trends.
Multidrug-resistant organism (MDRO) incidence
The total number of MDRO infections decreased from 15.2 ± 4.3 to 9.8 ± 2.7 (P < .05), with the overall incidence rate declining from .084 ± .012% to .062 ± .008% (P < .05). Significant reductions were observed for: MRSA: .017 ± .003% → .008 ± .002% (P < .05), CRAB: .036 ± .006% → .020 ± .004% (P < .05). No significant changes were detected for CRPA (.011 ± .002% → .010 ± .001%, P = .12) or CRKP (.026 ± .004% → .025 ± .003%, P = .08). Temporal trends in AUD and MDRO incidence rates are visualized in Figure 3.
Figure 3. Temporal Trends in Antimicrobial Use DDD and Incidence Rate of MDRO Infections (2022–2024).
Process metrics
Rejected prescriptions were revised or withdrawn upon pharmacist consultation. ‘Presubmission’ refers to documentation submitted before antimicrobial initiation. Pharmacists rejected 10.24% of prescriptions (10,340/100,937), primarily due to inappropriate indications (62%) or dosing errors (28%). Preantimicrobial therapy submission rates increased from 46.47 ± 7.8% to 58.27 ± 6.5% (P < .05).
Discussion
Optimized drug stewardship and health economic benefits
The pharmacist-led AMS model developed in this study significantly improved antimicrobial utilization. The appropriateness rate for restricted-use antimicrobial prescriptions increased by 13.6 percentage points (80.24% to 93.83%), outperforming comparable U.S. interventions by 51%.Reference Barlam, Ellen and Zieminski11,Reference Nampoothiri, Mbamalu and Mendelson12 This breakthrough is attributable to the dual-review mechanism’s systematic correction of empirical prescribing patterns: 62.3% of rejected prescriptions lacked appropriate indications, reflecting systemic overreliance on broad-spectrum antimicrobials in severe infection management.Reference Pulcini, Binda and Lamkang6 Notably, a 26% reduction in carbapenem use coincided with increased β-lactamase inhibitor combinations, validating the effectiveness of antimicrobial spectrum-narrowing strategies.
Under China’s VBP policy, this study achieved the first documented dual reduction in drug expenditures and resistance risks. Total antimicrobial costs decreased by 41.3% ($4.09 million to $2.40 million), with per-capita restricted antimicrobial expenditure declining 45.7% ($988.29 to $536.00), surpassing European AMS benchmarks (15–30% reductions).Reference Pulcini, Binda and Lamkang6,Reference Burrowes, Drainoni and Tjilos20 Three synergistic mechanisms drove these outcomes: First, the intelligent review system intercepted 10.24% of inappropriate prescriptions. Second, microbiological testing rates increased by 25.4%, enabling precision therapy. Third, prescribing behavior restructuring established a virtuous cycle of rational use, cost control, and resistance mitigation.
Antimicrobial use intensity modulation and resistance dynamics
The 28.3% reduction in AUD was primarily driven by workflow redesign in critical departments. The ICU achieved a 20.6% AUD decrease via sepsis de-escalation protocols, while Respiratory Medicine and Hematology implemented pathogen-directed therapy using rapid molecular diagnostics (25.4% increase in testing rates). Significant spatial heterogeneity (χ2 = 18.37, P = .002) underscores the need for department-specific AMS protocols.
The dose-response relationship between CRKP detection (R = .62) supports theories of horizontal resistance gene transfer via gut microbiome disruption.Reference Tamma, Heil and Justo23 Although CRKP reduction was statistically nonsignificant (P = .078), its delayed response aligns with ecological niche persistence mechanisms.Reference Cassini, Högberg and Plachouras24 Conversely, the weak tigecycline-CRAB association (R = .38, P = .072) may reflect subtherapeutic lung penetration—a pharmacokinetic limitation driving A. baumannii adaptation.Reference Yaghoubi, Zekiy and Krutova25 Crucially, empirical use restrictions reduced CRAB detection by 44.4%, providing direct evidence of AMS-driven resistance evolution control.
Pharmacist role transformation and practical implications
The dual-review model redefined antimicrobial governance. Compared to traditional single-pharmacist reviews, this system reduced care coordination errors by 28% (P = .003) through e-prescription prescreening and multidisciplinary consultation, matching Indian AMS efficacy with superior cost-effectiveness.Reference Vrinda, Mohamed and Oluchi28,Reference Marins, de Jesus and Holubar29 Continuous professional development programs addressed AMS workforce challenges in low-resource settings.
Three limitations warrant consideration: First, residual carbapenem overuse (3.2 DDDs/1,000 patient-days) highlights clinical demand-stewardship tensions. Second, process-oriented metrics may undervalue AMS’s long-term benefits.Reference Xia, Li and Yang30 Third, the single-center design limits generalizability, particularly to primary care settings.
Conclusions and perspectives
This intervention improved antimicrobial appropriateness and reduced AUD, with observed secondary benefits in resistance and cost reduction. This study demonstrates that the pharmacist-led AMS-APAS system effectively resolves the “cost-resistance” paradox under VBP policies. The integrated intervention model—combining intelligent auditing, capacity building, and quality improvement—reduced antimicrobial expenditures by 41.3% and hospital-acquired MDRO infections by 26.2%. This breakthrough offers global AMR containment a replicable protocol: policy-aligned institutional innovation can synergize pharmacoeconomic efficiency with microbiological safety.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ash.2025.10155.
Data availability statement
The data that support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to privacy and ethical constraints.
Acknowledgments
The authors sincerely acknowledge Ms. Xuan Li (M.S.) for her expert medical guidance throughout this study. Her specialized medical consultation significantly contributed to the scientific rigor of this work.
Financial support
This research was supported by Tongji Hospital Research Grant for Medical and Health Sciences (202406) and Merck Sharp and Dohme (MISP 102061).
Competing interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this study.
Ethical standard
This study was performed at Tongji Hospital, Wuhan, Hubei Province, China. This study was approved by the Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (Approval No.: TJ-JRB202308134) and registered with the Ethics Review Committee for Clinical Trials in China (Chinese Clinical Trial Registry; Approval No.: ChiCTR2300077449; Date of Registration: 2023–11–09). The requirement for informed consent was waived owing to the retrospective nature of the study.
Open access
