Comparison of Two Pharmacy Practice Models for Pharmacokinetic Management of Vancomycin at an Academic Medical Center

Purpose: National pharmacy practice model (PPM) advancements include pharmacokinetic (PK) programs.  The optimal PPM for vancomycin PK management is not known. This retrospective cohort review compared two progressive PK models at an academic medical center.

Methods: New PK services were implemented in 2 phases as part of an institutional comprehensive PPM change.  Phase 1 (May 2009 to April 2010) included universal monitoring by pharmacists with recommendations made to prescribers (business hours, 7 days per week). Phase 2 (November 2010 to October 2011) expanded coverage to 24/7 and provided optional pharmacist-managed consults. Consults included comprehensive medication therapy management and progress note documentation.  All adult inpatients receiving intravenous vancomycin were retrospectively reviewed.  Surgical prophylaxis, duration < 72 hours, and initiation prior to admission were excluded.

Results: Patient characteristics and indications were similar in both phases. Phase 2 had greater proportion of courses with initial therapeutic trough concentrations (45.3% vs. 27.1%, p=0.02), higher initial trough values (16.4 mcg/mL vs. 10.5 mcg/mL, p=0.03), greater proportion of therapeutic concentrations (77.3% vs. 59.5%, p=0.013), and reduced hospital length of stay (12.0 days vs. 14.5 days, p=0.02). Both phases had similar time to therapeutic trough concentrations (4 days vs. 5 days, p=0.76), incidence of vancomycin-associated nephrotoxicity (8.8% vs. 11.2%, p=0.40) and all-cause mortality (11.0% vs. 11.1%, p=0.98). In phase 2, patients with consults had greater proportion of initial therapeutic trough concentrations (51.4% vs. 39.4%, p=0.30), higher initial trough values (17.1 mcg/mL vs. 13.5 mcg/mL, p=0.64), and greater proportion of therapeutic concentrations (86.5% vs. 65.8%, p=0.02).

Conclusion: Metamorphosis of PPM afforded expansion to 24/7 PK services with comprehensive pharmacist consults which improved vancomycin management.  These results may help define optimal models as recommended by the Pharmacy Practice Model Initiative.  Further study is needed to assess the impact of these models on patient outcomes and in other practice settings.