Blood culture-guided de-escalation of empirical antimicrobial regimen for critical patients in an online antimicrobial stewardship programme.

A blood culture-guided review strategy was applied to a hospital-wide computerised antimicrobial approval system (HCAAS) at a medical centre in Taiwan. The study aimed to evaluate the impact of this deployment on prescribers' behaviours, antimicrobial consumption, antimicrobial expenditure and healthcare quality in adult intensive care units (ICUs). The HCAAS automatically identifies patients with positive blood cultures and notifies the pre-assigned infectious diseases (ID) physicians for an online second review of the current antimicrobial regimen. Patients from 16 adult ICUs were selected as a focus group. Descriptive analysis, McNemar's test, interrupted time-series analysis and univariate regression analysis were applied. The number of prescriptions assigned for second review increased from 304 in 2010 to 682 in 2012. The approval rate for the antimicrobial regimen in the second review exceeded 70%. In disapproved cases, prescribers accepted the recommendation from ID physicians in 66.1% of cases in the first year; the acceptance rate increased to 80.6% in 2012. Among the restricted antimicrobial agents, consumption gradients decreased for all eight drug classes. The overall antimicrobial expenditure gradient declined significantly following deployment of the second review strategy. The healthcare-associated infection rate continued to decrease over time, and the mortality and ICU re-admission rates remained stable after deployment. A blood culture-guided review of antimicrobial use based on clinical and microbiological evidence improves accuracy in choosing appropriate antimicrobial agents and encourages de-escalation. Consumption and expenditure gradients of antimicrobial agents decreased after the intervention, and healthcare quality was not compromised.