Discussion

In this interrupted time-series analysis using data in a nationwide claims database, issue of the Manual of Antimicrobial Stewardship had no significant impact on reducing antibiotic use for outpatients with ARTI or gastroenteritis during the 1-year period after the date of issue. Similar findings were noted for use of broad-spectrum antibiotics. The strengths of our study include the use of a quasi-experimental design and nationwide claims data covering primary care settings.

Issue of guidelines is a general and central approach to reducing inappropriate use of antibiotics. Nevertheless, controversy remains on the impact of such guidelines in clinical settings. ^{Reference Bhattacharyya and Kepnes15–Reference Hawker, Smith and Smith17} In the United States, for example, care patterns for medical treatment of acute adult sinusitis changed after guideline publication, with an increase in oral antibiotic prescription rates and an increase in the use of amoxicillin as the first-line antimicrobial agent. ^{Reference Bhattacharyya and Kepnes16} In the United Kingdom, implementation of national guidelines for primary care showed mixed results from 1995 to 2011. Antibiotic prescriptions for sore throats declined from 77% to 62%, but antibiotic prescriptions for otitis media increased from 77% to 85%. ^{Reference Hawker, Smith and Smith17}

The aforementioned studies used a before-and-after design to assess the impact of the issue of guidelines. However, a simple before-and-after design may be unable to fully address the causal impact of the guidelines because it cannot account for changes in trends between the pre-issue period and post-issue period. ^{Reference Cousens, Hargreaves and Bonell18} In this context, our study using an interrupted time-series analysis as a quasi-experimental design extended these previous studies and suggested that guideline publication may not be an immediately effective intervention to change the prescribing behavior of physicians.

Although we used Japanese data, we believe that our findings may be generalizable to any developed country and useful for an international readership because of the known international issues related to adopting guideline-issued changes in a timely manner. In addition, our findings can provide an important basis for further development of evidence-based interventions toward prevention of AMR.

Potential reasons for the nonsignificant impact of the issue of national guidelines are likely to be multifactorial. The most important reason may be that doctors were either unaware of the manual (42.2%) or did not use it if they were aware of the manual (43.7%), according to a previous survey of Japanese clinic doctors. ^{Reference Gu, Fujitomo and Soeda19} Qualitative studies showed that trying to meet patient expectations was a major factor for physicians and that pressure from patients, relatives, or carers was an uncomfortable influence on prescribing decisions. ^{Reference Gu, Fujitomo and Soeda19–Reference O’Connor, O’Doherty and O’Regan23} One of the studies also identified other possible factors, including limited time, poor doctor–patient communication, and diagnostic uncertainty. ^{Reference Fletcher-Lartey, Yee, Gaarslev and Khan22} Another factor may be direct marketing strategies by pharmaceutical industries toward prescribing physicians. ^{Reference Avorn and Solomon20} General practitioners can also prescribe antibiotics to gain financial incentives. ^{Reference Burkhard, Schmid and Wuthrich24}

The observed nonsignificant impact of the issue of national guidelines warrants further investigations to seek optimal and targeted strategies that can reduce antibiotic use. One potentially effective strategy is financial incentives. Indeed, recent studies in the United Kingdom showed that introduction of financial incentives successfully reduced all antibiotic prescriptions by 8.2% and broad-spectrum antibiotic prescriptions by 18.9% ^{Reference Balinskaite, Johnson and Holmes25} and reduced antibiotic prescription for uncomplicated ARTI by 3%. ^{Reference Bou-Antoun, Costelloe and Honeyford26} Multiple other studies have supported the notion that financial incentives may be an effective strategy for reducing inappropriate antibiotic use. ^{Reference Han, Chae and Kim27} The other potential strategies include (1) prescribing restrictions, (2) educating healthcare professionals, and (3) educating the general population. First, prescribing restrictions (eg, restrictions on antibiotic purchases without a prescription) had consistent evidence for effectiveness at reducing antibiotic consumption and prescription according to a recent systematic review on the impact of national interventions to promote responsible antibiotic use. ^{Reference Lim, Singh, Duong, Legido-Quigley, Hsu and Tam28} Second, in Japan, dissemination of knowledge on antimicrobial stewardship is needed. A recent survey of clinics in Japan reported that physicians in clinics were unaware of the manual (42.2%) and the national action plan (44.9%), but they were aware of the effect of appropriate antimicrobial use on the prevention of antimicrobial resistance (82.1%). ^{Reference Gu, Fujitomo and Soeda19} Thus, continuous education and recommendations to physicians is important. Third, according to a nationwide survey of the general population in Japan, ~30% of respondents preferred physicians who prescribed antibiotics for a cold. Furthermore, 80% of respondents believed that antibiotics were effective for colds and flu. ^{Reference Kamata, Tokuda, Gu, Ohmagari and Yanagihara29} In agreement with a recent study on the importance of educational interventions tailored to patient educational levels, ^{30,Reference Tsuzuki, Fujitsuka and Horiuchi31} further education for the general population should facilitate the appropriate use of antibiotics. As indicated by these studies, understanding the actual impact of individual strategies is critical for researchers and policy makers.

Our study has several limitations. First, because the database includes patient data from employee-based health insurance societies, little information is available for older patients, particularly those aged ≥65 years. However, the trends in practice patterns may not differ substantially across age categories. Second, after accounting for the dissemination period, the 1-year study period after issue of the manual may not be sufficient to evaluate its impact. Nevertheless, with the rapid increase in AMR-related clinical problems, ^{32–Reference de Kraker, Wolkewitz and Davey35} it is important to determine whether or not an intervention is immediately effective. ^{Reference Han, Chae and Kim27} Third, there was no control group and a basic interrupted time-series analysis design cannot exclude confounding by cointerventions or other events occurring around the time of the intervention. ^{Reference Lopez Bernal, Cummins and Gasparrini36} Nonetheless, the obvious linear trend without significant changes in the nationwide data may support the assumption that there were no impactful strategies during the study period.

In conclusion, despite the issue of the Manual of Antimicrobial Stewardship to promote optimal antimicrobial therapy, we detected no significant changes in trends of antibiotic use for outpatients with ARTI and gastroenteritis. The present study revealed that broad-spectrum antibiotics still accounted for ~80% of oral antibiotic prescriptions. Our findings indicate that the issue of national guidelines may not be an immediately effective intervention to change the prescribing behaviors of general practitioners, and they suggest the importance of further multifaceted strategies to promote optimal antimicrobial use.