This study demonstrates the impact of an antimicrobial stewardship program (ASP) and the associated changes in antimicrobial use patterns on susceptibilities of specific Gram negative organisms (Pseudomonas aeruginosa, Escherichia coli, Enterobacter cloacae, Klebsiella pneumonia) via time series analysis from 2006 to 2014.
Antimicrobial use converted to defined daily dose (DDD) per 1000 patient days and bacterial susceptibilities were incorporated into a time series analysis, with a structural break at January 2009 (via Chow Test), corresponding to the implementation of the ASP. Seasonal Autoregressive Regressive Moving Average (ARIMA) to adjust for seasonal changes, lag was determined via X-13-ARIMA-SEATS. Transfer functions (ARMIX) were used to compare the use of antimicrobials on each bug-drug combination.
Antimicrobial consumption in DDD/1000 patient days decreased from 288.76 (β= 1.08) pre-ASP to 236.90 (β= -2.97) after implementation of ASP (p=<0.001, figure 1). Antibiotics which significantly decreased from pre-ASP to ASP included ciprofloxacin (p=<0.001), piperacillin/tazobactam (p=0.05), and tobramycin (p=<0.001). Amikacin, cefepime, ceftazidime, ceftriaxone, and imipenem demonstrated stable usage.
Decreased piperacillin/tazobactam usage resulted in improved susceptibilities of P. aeruginosa (figure 2), stable E. coli and E. cloacae sensitivities, and worsening K. pneumonia susceptibilities. Decreased ciprofloxacin usage correlated with improved E. coli and P. aeruginosa susceptibilities, with stable E. cloacae and K. pneumonia susceptibilities. Decreased tobramycin usage correlated with improving susceptibilities of K. pneumoniae, stable E. coli and E. cloacae sensitivities, but worsening P. aeruginosa susceptibilities.
The ASP had a favorable impact on stabilizing or improving susceptibilities. Total antibiotic consumption decreased significantly, demonstrating that reduced utilization of antibiotics including piperacillin/tazobactam has not applied pressure on other broad spectrum agents. Time series models can be utilized to illustrate the impact of optimal antimicrobial use on susceptibilities.
M. Ott, None
K. Mergenhagen, None