456. The Effects of an Electronically Enhanced Antimicrobial Stewardship Program in Two Critical Care Locations at a 415 Bed Urban Hospital
Session: Poster Session: Hospital-acquired and Transplant Infections
Friday, October 30, 2009: 12:00 AM
Room: Poster Hall A
Background: An antimicrobial stewardship program (P) was started and staffed by an ID pharmacist in 2 targeted adult critical care locations at Bronx Lebanon Grand Concourse Hospital. The pharmacist responded to drug class mismatch, susceptibility mismatch, and IV to PO switch alerts from an electronic antimicrobial monitoring service. It has previously been shown that P improved the likelihood of fast (<12h) alert resolution of drug and susceptibility mismatch alerts. To further examine P’s effects, an expanded dataset was examined for targeted (TL) and non-targeted (NTL) locations.
Methods: Data from the MedMinedTM Health Outcomes Database were analyzed by multivariable regression techniques to estimate the effect of P on the likelihood of fast resolution and post-alert LOS. Controlling for alert type and location, pre-alert LOS, age, gender, day of week, and time from WP start, models were built to determine the time-dependent development of effects in TL and look for possible spillover effects in NTL. Pre-P (PrP) was from 6/07 - 2/08 and with-P (WP) was from 3/08 - 12/08.
Results: Of 1643 eligible alerts, 300 resolved rapidly PrP vs. 265 WP. In TL, the probability of fast resolution was 2.5 times greater WP than PrP (p=0.01), diminishing slightly with time after WP start (p=0.07); in NTL, this probability improved by nearly 10% per 100 days from WP start (p=0.05). Post-alert LOS decreased in TL by nearly 4 days per 100 days from WP start (p<0.01), but there was no improvement over time in NTL (p=0.46). There was no underlying hospital-wide, time-dependent change in fast resolution likelihood (p=0.5) and post-alert LOS (p=0.9).
Conclusion: For fast alert resolution likelihood in TL and NTL, P was associated with immediate, sustained and time-dependent improvements, respectively, suggesting that its effects spread hospital-wide over time. Post-alert LOS in TL decreased with time, indicating that P had an effect on this outcome from WP start.
Stephen Brossette, MD, PhD1, Emily Falk, MSPH1, Kyoung Sil Kim, PharmD2, Peter Lao, MS RPh2, Frank Palmieri, PhD, RPh2, Larry Schiller, RPh2, Prashant Tayshetye, MD, MPH1 and  F. E. Palmieri,
Bronx Lebanon Hospital Center Role(s): Employee, Received: Salary.
E. A. Falk,
Cardinal Health - MedMined Services Role(s): Employee, Received: Salary.
K. Kim,
Bronx Lebanon Hospital Center Role(s): Employee, Received: Salary.
P. Lao,
Bronx Lebanon Hospital Center Role(s): Employee, Received: Salary.
L. Schiller,
Bronx Lebanon Hospital Center Role(s): Employee, Received: Salary.
P. Tayshetye,
Cardinal Health - MedMined Services Role(s): Employee, Received: Salary.
S. Brossette,
Cardinal Health - MedMined Services Role(s): Employee, Received: Salary., (1)Cardinal Health - MedMined (TM) Services, Birmingham, AL, (2)Bronx Lebanon Hospital Center, Bronx, NY