Impact of Two Different Antimicrobial Stewardship Methods on Frequency of Streamlining Antimicrobials Agents in Patients with Bacteremia

Background: The Antimicrobial Stewardship Program (ASP) at the Hospital of the University of Pennsylvania was initiated in 1993 and focused on prior authorization alone. In 2009, several broad-spectrum agents were de-restricted, including cefepime, vancomycin, and piperacillin/tazobactam, and these agents were monitored via prospective audit and feedback. The objective of this study was to compare rates of antimicrobial streamlining between these two ASP approaches.

Methods: We compared the frequency and time to streamlining PRE- and POST-ASP change using a retrospective cohort design. The PRE and POST periods were 5/18/08-5/31/09 and 7/1/09-6/30/11, respectively. Inclusion criteria were patients age ≥ 18 years with monomicrobial bacteremia with S. aureus, Enterococcus spp., or an aerobic gram-negative organism. Episodes were excluded if antimicrobials were discontinued within 48 hours of susceptibility results or if therapy was inappropriate based on predefined criteria. Streamlining was assessed by an algorithm developed by study investigators based on spectrum of activity of antimicrobial agents.

Results: With 875 patients meeting the inclusion criteria, 245 patients were excluded and 632 patients with 665 episodes of bacteremia were available for analysis. Baseline demographics were similar between groups. 62.9% (158/251) and 70.0% (290/414) of bacteremic episodes were eligible for streamlining in the PRE and POST groups, respectively. PRE-ASP change, 53.2% (84/158) of episodes were streamlined compared with 60.7% (176/290) in the POST group (p=0.123). Streamlining occurred within 24 hours of susceptibility results in 61.9% and 55.1% in the PRE and POST episodes, respectively (p=0.435). In a bivariate analysis, factors associated with a higher rate of streamlining include: absence of β-lactam allergy (p<0.001), gram-negative bacteremia (p<0.001), and admitting service (p=0.011).

Conclusion: Prospective audit and feedback led to a modest overall increase in streamlining compared to the prior model using prior authorization alone. The benefit of prospective audit may be enhanced when focusing on gram negative bacteremia and patients without β-lactam allergies.