br Results br Discussion Although the value

Results
Discussion Although the value of ID specialists in the USA has already been described in the literature, this study suggests that the same holds true for the Japanese healthcare environment [7], [17], [18], [19]. Valerio et al. reported that antifungal use in DDD decreased from 66.4 per 1000 patient-days to 54.8 per 1000 patient-days after implementation of an antifungal stewardship programme in a tertiary hospital in Spain [20]. These DDD values are relatively high and the change seems small compared with our results. In the present study, a change from 26.1 to 9.9 DDD per 1000 patient-days was observed between the first and last years of the study period. Possible reasons for the differences in results between our study and that by Valerio et al. are the variations in patient and hospital characteristics. Unlike the hospital in the study by Valerio et al., our hospital did not provide transplantation care, including bone marrow transplant. Another probable reason is the difference in the durations of the study periods. Valerio et al. only investigated up to two years post-implementation of their ASP, whereas our study investigated up to six years post-implementation of the intervention. However, two years may be sufficient for the study by Valerio et al., as the impact of their programme appeared to stabilize in the second year, whereas the impact observed in our study stabilized three years after implementation. The assignment of an ID specialist seemed to have a favourable impact on cost-effectiveness. The total expenditure on antifungals decreased by 73% from fiscal year 2006 to fiscal year 2015. This reduction amounted to ¥38 million per year, which is equivalent to about US$380,000 with a currency rate of US$1 = ¥100. The prudent use of antifungals has a strong economic impact, as antifungals are generally expensive drugs. In the Japanese healthcare market, there are almost no generic intravenous antifungals apart from fluconazole, which is not available in our hospital. Therefore, the decrease in expenditure directly reflects antifungal usage. Carling et al. revealed that a multi-disciplinary antibiotic management programme achieved annual cost savings of about US$250,000 in a 174-bed AGK 2 synthesis hospital [21]. Our study suggests that a similar cost-saving effect was achieved with reduction in antifungal use alone. This study had several limitations. First, it may be difficult to generalize the impact of the assignment of the ID specialist to other hospitals in terms of the level of reduction seen. This was the first time that our hospital had used an ID specialist. Therefore, the quality of ID management, including antifungal use, may have been below the desired standard prior to the arrival of the ID specialist. Different hospitals may have different margins for improvement based on the current standard of ID services provided; a similar effect would not be duplicated in a hospital that already meets the desired standard for ID management. Notably, the effect seen in the bundle implementation period disappeared in the long-term follow-up period. This might indicate that the educational impact of the ID specialist had pervaded the hospital within three years. However, we consider that this plateau in the effect is worth noting. Our study might suggest that an educational bundle, including ASP, has a certain limit in terms of the degree of reduction achievable, which provides important information about target levels of antifungal use. The Government of Japan released a national action plan to address antimicrobial resistance in April 2016 [22]. In the action plan, target levels were set in relation to the decreased use of certain antibiotics, but not with regard to any antifungals. Our study can thus be used as a reference to support the Government in setting target levels for antifungal use. Our study suggests that antifungal use is a potential performance indicator for ASPs and accompanying educational activities. Although the importance of educational activities, including ASPs, is well accepted, it is difficult to measure outcomes concerning antibacterial agents [10], [23], [24], [25]. Even if a reduction in a single class of antibacterial agents leads to lower resistance to that class, the unintended consequence of increased use of other classes and resistance to them may occur (‘squeezing-the-balloon’) [12]. Antifungals are less likely to be influenced by this phenomenon. Widely used classes of antifungals include the polyenes, echinocandins, and azoles. The armamentarium to combat fungal infections is limited, so it is not possible to forgo these antifungals during the treatment of fungal infections. Therefore, the total use of antifungals can be expected to reveal actual-use conditions without concerns related to the ‘squeezing-the-balloon’ phenomenon.