Audit of empirical antibiotic therapy for sepsis and the impact of early multidisciplinary consultation on patient outcomes

Link to article at PubMed

Int J Antimicrob Agents. 2021 Jun 20:106379. doi: 10.1016/j.ijantimicag.2021.106379. Online ahead of print.

ABSTRACT

OBJECTIVES: To perform an audit of empirical antibiotic therapy (EAT) of sepsis at the emergency department and to analyse the impact of an antimicrobial stewardship (AMS) programme on process and patient outcomes.

PATIENTS AND METHODS: A prospective single centre cohort study including patients with SOFA score ≥2 from whom blood cultures were taken, was conducted between February 2019 and April 2020. EAT assessment was done using eight applicable quality indicators for responsible antibiotic use in inpatients (IQIs). Patient outcomes were hospital length-of-stay (LOS), ICU admission and ICU LOS, and in-hospital mortality.

RESULTS: The audit included 900 sepsis episodes of 803 patients. Full guideline adherence regarding choice and dosing was 45.9%; adherence regarding choice alone was 68.1%. EAT was active against all likely pathogens in 665/787 (84.5%) episodes. In the guideline non-adherent group, choice of EAT was inappropriate in 122/251 (48.6%) episodes. Changes within three days occurred in 335/900 (37.2%) episodes. Treating physicians changed administration route more often, whereas microbiological/ID/AMS consultant advice resulted in de-escalation and discontinuation (p = .000). Guideline-adherent choice was associated with significantly shorter LOS (6 (4-11) vs. 8 (5-15) days), and fewer ICU admissions (54 (10.1%) vs. 45 (17.9%)). Full adherence was associated with significantly lower mortality (23 (6.4%) vs. 48 (11.3%)) and shorter LOS (6 (4-10) vs. 8 (5-14) days).

CONCLUSION: Five global quality indicators of EAT were measurable in routine clinical practice. Full adherence to guidelines was only moderate. Adherence to guidelines was associated with better patient outcomes.

PMID:34161787 | DOI:10.1016/j.ijantimicag.2021.106379

Leave a Reply

Your email address will not be published.