Discussion

The present study described trends of three distinct antimicrobial consumption metrics at each participating hospital. Moreover, structured interviews with narrative comments provided further insight into the current status, progressions and challenges relevant to ASP activities. Most hospitals faced insufficient staffing of clinical pharmacists dedicated to ASP activities, resulted in difficulties in promoting various ASPs. Consequently, although ASPs were formally established at all participating hospitals, the structured interviews revealed that many ASPs may not be fully operational. Moreover, the findings suggested that the presence of ASPs did not consistently translate into ideal antimicrobial prescribing practice as reflected by changes in antimicrobial consumption metrics. The present study underscores the importance of ongoing evaluation and continuously monitoring ASP activities.

As seen in Table 1, despite the presence of ASP, most participating hospitals had not implemented concurrent, broad range of ASPs interventions to optimize hospitalwide antimicrobial prescribing; many participating hospitals have implemented PPRF for broad-spectrum antimicrobials (Hospitals 1, 2, 5, 8, 9, and 10) and ASP interventions for prolonged (eg, >2 wk) antimicrobial use (Hospitals 2, 3, 5, 6, 7, 8, 9, and 10) whereas only 2 institutions (Hospitals 6, and 8) implemented preauthorization for broad-spectrum antimicrobials. Moreover, very few institutions have implemented beta-lactam allergy de-labeling and optimizing duration of syndromic-specific antimicrobial therapy. These findings suggest that many hospitals have faced challenges in advancing ASP activities beyond initial PPRF, which was widely adopted during the early phase of antimicrobial stewardship era. ^{Reference Honda, Goto, Uehara and Takamatsu9} Our interviews also revealed a universal shortage of pharmacists engaged in ASP. The finding was consistent with a previous review which demonstrated both a shortage of clinical pharmacists and labor-intensive pharmacist workloads are considered major barriers to promoting ASPs. ^{Reference Jantarathaneewat, Camins and Apisarnthanarak10} Limited opportunities for acquiring advanced ASP-related knowledge and experiences were reported (Hospitals 1, 3, 5, 6, 8, 9, and 10). Notably, a similar phenomenon has been observed even in other high-income countries; a national survey in the United States revealed that various ASP activities were performed by pharmacists without sufficient training in infectious diseases (IDs). ^{Reference Dionne, Wagner, Chastain, Rosenthal, Mahoney and Bland11} Moreover, the concept of beta-lactam de-labeling might be underappreciated because of limited availability of allergy-immunology specialists in Japan. Advancing successful ASP requires an ID physician-pharmacist team effort, which leads to a valuable opportunity to gain hands-on experiences in ASP pharmacists under the guidance and support of ID physicians. ^{Reference Barlam, Childs and Zieminski12}

Another challenge against promoting ASP included an inadequate EMR infrastructure to efficiently identify the target population of ASP (Hospitals 2, 5, and 8). Investment in hospital EMR systems to facilitate efficient and effective ASP may be particularly critical for institutions facing ASP clinical pharmacist shortages. While integration of EMR-based ASP tools such as antimicrobial stop orders, clinical decision support system, and prescription alert has progressed, ^{Reference Forrest, Van Schooneveld, Kullar, Schulz, Duong and Postelnick13–Reference Dutcher, Yeager and Gitelman15} the emergence of artificial intelligence-augmented antimicrobial stewardship strategies has a potential to substantially enhance ASP interventions. ^{Reference Marra, Langford, Nori and Bearman16}

According to the changes in trends of antimicrobial consumption metrics, hospitals with a larger number of beds (Hospitals 1, 2, 5, and 6) demonstrated an uptrend of DASC per 1,000 PD and DOT per 1,000 PD over time, despite ASP implementation. This pattern may reflect an increasing proportion of patients infected with multidrug-resistant organisms, who are more likely to be treated in large tertiary care centers. Consistent with interpretation, the national antimicrobial resistance (AMR) surveillance in Japan reported rising trends in AMR during the study period, especially among Enterobacterales. ¹⁷ Moreover, insufficient interventions focusing on hospitalwide antimicrobial streamlining process and optimization of treatment duration might have contributed to these observed trends. In contrast, Hospitals 7–9, representing acute care hospitals with smaller number of hospital beds (community hospitals) demonstrated decreasing trends in both DASC per 1,000 PD and DOT per 1,000 PD. However, absolute baseline values of DASC and DOT at the beginning of the study period were higher than those observed in larger tertiary care centers. Moreover, an uptrend of DASC/DOT in these hospitals was observed. These findings potentially reflected delayed ASP implementation and less intensive antimicrobial streamlining process.

Among all participating hospitals, Hospital 10 appears to show the most favorable changes in antimicrobial consumption metrics. The observed trends may be attributable to strongly implemented PPRF, which likely promoted antimicrobial streamlining, resulting in a significant decrease in DASC (β = −3.05 [95% CI −4.85, −1.26], P = .001), a slight increase in DOT (β = 0.08 [95% CI −0.15, 0.31], P = .47) and a significant decline in DASC/DOT (β = −0.01 [95% CI −0.015, −0.011], P < .001). In the structured interviews, the clinical pharmacist at Hospital 10 reported high effectiveness of PPRF targeting broad-spectrum antimicrobials: “We are very confident that our PPRF has worked well” (Hospital 10-pharmacist). Similarly, Hospital 6 demonstrated a decreasing trend in DASC/DOT after August 2022, corresponding with the introduction of carbapenem preauthorization: ”Preauthorization for carbapenems was implemented since a carbapenem shortage occurred, and it has been effective in maintaining low carbapenem consumption” (Hospital 6-pharmacist). These findings would be a clear example of measurable changes in antimicrobial streamlining process and limited broad-spectrum antimicrobial following a successful hospital ASP intervention. ^{Reference Kouyama, Uehara and Takamatsu18}

The present study has some limitations. Because participating hospitals were selected from specific regions without random selection, selection bias may be present, and the findings may not be representative of hospitals nationwide. Since we did not examine antimicrobial streamlining process at individual patient level, the interpretation of changes in these metrics may not reflect actual antimicrobial prescribing practices, especially a lack of adjustment of case-mix and changes in the incidence of antimicrobial resistance at each institution. We did not assess local antimicrobial susceptibility patterns for specific pathogens at the individual hospital level, which may have influenced antimicrobial selection. The findings from the structured interviews may contain interviewer bias, participant bias, and confirmation bias. Since the baseline antimicrobial consumption in each hospital varied, the magnitude of ASP intervention effect may be higher in hospitals with high baseline antimicrobial consumption (high DASC and DOT) or less intensive antimicrobial streamlining process (low DASC/DOT) at the beginning of the study period (eg, baseline effect). Lastly, this study did not account for the impact of the COVID-19 pandemic on antimicrobial practice, which may have substantially influenced inpatient antimicrobial consumption, particularly increased antimicrobial use associated with healthcare-associated infections during this period. ^{Reference Weiner-Lastinger, Pattabiraman and Konnor19}

In conclusion, the present study described long-term trends in inpatient antimicrobial consumptions using DASC per 1,000 PD, DOT per 1,000 PD and DASC/DOT across Japanese hospitals over an 8-year period. Although ASP were present at all institutions, their scope and ASP implementation varied widely, and the establishment of ASP alone did not result in meaningful changes in antimicrobial consumption or streamlining indicators. Structured interviews highlighted persistent barriers to ASP advancement, including shortages of clinical pharmacists, limited educational opportunities, and inadequate infrastructure, particularly EMR integration. Strengthening hospital ASPs will require systematic identification of existing deficiencies and targeted investments in ASP personnel, education, and technological resources.