Our findings revealed that the small/middle-sized hospitals had more frequent interventions regarding broad-spectrum antimicrobial use within seven days in inpatient settings. In a previous Japanese study, large hospitals tended to have more ASP and the reason was thought to be a greater number of full-time equivalent pharmacists compared to small/middle-sized hospitals.14 We think that one of the reasons for the inconsistency between the results of our study and those of the previous study is participants. Although the participants of previous study were certified members of the Japanese College of Infection Control Doctors in nationwide, those in our study were highly motivated medial staffs who participated in hospital epidemiology workshops during July 2018. The reason small/middle-sized hospitals intervened more frequently in our study is that small hospitals may be more likely to change their practices if they have highly motivated personnel. Another study reported that a barrier to ASP implementation in small hospitals was a lack of dedicated IDS to support the ASP.3 In our study, small/middle-sized hospitals tended to have fewer IDS, but had more frequent interventions compared to large hospitals. In small/middle-sized hospitals, it may be that pharmacists play a key role in AST, especially regarding interventions for antimicrobial use, despite the lack of IDS. We hypothesize that in small/middle-sized hospitals, it is difficult for an IDS to dedicate an adequate amount of time for AST activities because of their many tasks in Japan. However, pharmacists have more time to participate in AST and thus, are more often designated for AST activities compared to their IDS counterparts. This situation may suggest that pharmacists have more initiative for AST activities in small/middle-sized hospitals than in large hospitals. This may lead to prompt actions for intervention in cases of inappropriate antimicrobial use. In fact, it has been reported that pharmacists play a central role in AST.20 Regarding interventions within seven days when using carbapenem, piperacillin/tazobactam, and intravenous quinolone, small/middle-sized hospitals seemed to be more involved. On the other hand, the number of hospitals that had intervention protocols within seven days of using of 3rd and 4th generation cephalosporin was small in both small/middle-sized hospitals (1/21 [4.8%], 6/21 [28.6%]) and large hospitals (0/18 [0%], 1/18 [5.6%]). AMS for 3rd generation cephalosporins was thought to be related to the decrease of ESBL-producing pathogens.19 Because JANIS reported that the proportion of Escherichia coli resistance to cefotaxime increased from 23.3–27.5% between the years 2014–2018 in Japan,21 prompt intervention protocols for not only 4th generation cephalosporins, but also 3rd generation cephalosporins should be emphasized.
Our study showed that most hospitals, regardless of size, appeared to provide few interventions regarding antifungal use in inpatient settings (intervention within 7 d: 3 [14.3%], 0 [0%], intervention within 28 d: 7 [33.3%], 3 [16.7%]). The reason for this is thought to be due to a lack of human resource power and the low incidence of fungal infections compared to bacterial infections in inpatient settings.21 However, the importance of AFS for both patient benefit and cost-saving was well reported in a previous study.22,23 Based on the national database of health insurance claims and specific health checkups in Japan, although the cost of antimicrobials decreased from 163 billion dollars in 2013 to 121 billion dollars in 2017, the cost of antifungals increased from 34 billion dollars in 2013 to 50 billion dollars in the same year.24 A previous observational study in Japan pointed out that AFS was important to reduce the cost of antifungal use.25 In terms of not only patient benefit, but also the cost savings, we think that the importance of appropriate use of antifungal agents is paramount.
There were several limitations to our study. First, this study was based on a self-reported questionnaire, so there may be differences between the results of our study and actual hospital activity regarding AMS and AFS. Second, we have the possibility of respondent bias due to a low proportion of respondents in that a respondent facility which also participated in the hospital-epidemiology workshop might be a motivated cohort. The results of our study had the possibility of being overestimated, compared to one that directly observed all targeted institutions.