Impact of daily versus weekly service of infectious diseases consultation on hospital antimicrobial consumption: a retrospective study

Background To verify whether a daily service of Infectious Diseases consultation (ID-cons) is more effective than a weekly service in reducing antibiotic (ATB) consumption without worsening of clinical outcomes. Methods Two-year observational analysis of the ID-cons provided in a hospital setting in Milan, Italy. ID-cons resulted in: start-of-ATB; no-ATB; confirmation; modification-of-ATB. The impact of a weekly (September 1, 2016 - August 31, 2017 versus a daily (September 1, 2017 - September 30, 2018) service of ID-cons was evaluated in terms of: time-from-admission-to-first-ID-cons, type of ATB-intervention and number-of-ID-cons per 100 bed-days (bd). Primary outcomes: reduction of hospital ATB consumption overall and by department and classes expressed as Defined Daily Dose (DDD)/100bd (by Wilcoxon test for paired data). Secondary outcomes: overall and sepsis-related in-hospital annual mortality rates (as death/patient's admissions). Results Overall 2552 ID-cons in 1111 patients (mean, 2.3 ID-cons per patient) were performed (18.6% weekly vs 81.4% daily). No differences in patient characteristics were observed. In the daily-service, compared to the weekly-service, patients were seen by the ID-consultant earlier (time-from-admission-to-ID-cons: 6 days (IQR 2-13) vs 10 days (IQR 6-19), p < 0.001) and ATB was more often started by the ID-consultant (Start-of-ATB: 11.6% vs 8%, p = 0.02), rather than treating physicians. After switching to daily-service, the number-of-ID-cons increased from 0.4/100bd to 1.5/100bd (p = 0.01), with the greatest increase in the emergency department (1.5/100bd vs 6.7/100bd, p < 0.001). Total ATB consumption decreased from 64 to 60 DDD/100bd. As for the number-of-cons, the consumption of ATB decreased mainly in the emergency area. According to ATB classes, glycopeptides consumption was reduced from 3.1 to 2.1 DDD/100bd (p = 0.02) while carbapenem use decreased from 3.7 to 3.1 DDD/100bd (p = 0.07). No changes in overall mortality (5.2% vs 5.2%) and sepsis-related mortality (19.3% vs 20.9%; p = 0.7) were observed among the two time-period. Conclusions Daily-ID-cons resulted in a more comprehensive management of the infected patient by the ID-consultant, especially in the emergency area where we also observed the highest rate of reduction of ATB-usage. No change in mortality was observed.