4048 — In Data We Trust? Comparison of Electronic versus Manual Medical Use Evaluation for Hospitalized Veterans with Pneumonia
Lead/Presenter: Barbara Jones, COIN - Salt Lake City
All Authors: Jones B (VA SLC Health System IDEAS Center)
Haroldsen C (VA SLC Health System, IDEAS Center)
Madaras-Kelly K (Boise VA Medical Center)
Goetz MB (VA Greater Los Angeles Healthcare System)
Ying J (University of Utah Division of Epidemiology)
Greene T (University of Utah Division of Epidemiology)
Cunningham F (VA Pharmacy Benefits Management MedSAFE)
Neuhauser M (VA Pharmacy Benefits Management MedSAFE)
Samore M (VA SLC Health System, IDEAS Center)
VA Pharmacy Benefits Management (PBM) employs Medication Utilization Evaluation (MUE) to assess the quality of prescribing practices in VHA. Pharmacists review records using a standardized chart abstraction system. An MUE of pneumonia in 2013 at 30 sites revealed that therapy duration far exceeded guideline recommendations. The goals of this study were to: 1)develop and validate a MUE that uses electronic data extraction rather than clinician review; 2)apply the electronic MUE to all VA hospitalizations for pneumonia to examine national trends in antimicrobial use.
Using the existing MUE for pneumonia, we examined consistency between electronic and manual MUE for illness severity, antibiotic use, clinical outcomes, and daily clinical stability by calculating simple agreement and 2-way kappa statistics. We measured consistency of facility-level ranking between the electronic and manual extraction of data with the spearman correlation coefficient. We then applied the electronic MUE to all VA hospitalizations for pneumonia from 2008-2013 to examine national trends in antimicrobial choice and duration.
Among 2,004 hospitalizations, electronic and manual MUE demonstrated high agreement for initial severity measures (agreement 86-98%, kappa 0.5-0.82), initial antibiotic choice (agreement 89-100%, kappa 0.70-0.94), and ranking of facilities by empiric antibiotic use rates (anti-MRSA rho 0.97, p < 0.001; anti-pseudomonal rho 0.93, p < 0.001) and duration (rho 0.73, p < 0.001) but lower correlation for days to clinical stability (rho 0.50, p = 0.006) or excessive duration (rho 0.56, p = 0.005). Among 128,757 hospitalizations for pneumonia occurring in the VA system during 2008-2013, initial use of anti-pseudomonal coverage increased from 32% to 44%, anti-MRSA coverage increased from 26% to 39%, and median days of therapy decreased slightly from 10 [IQR 7-14] to 9 [IQR 7-12].
Electronic MUE correlated tightly with manual MUE for evaluating illness severity, antibiotic choice, and duration of therapy in pneumonia. National trends demonstrated a consistent increase in initial anti-MRSA and anti-pseudomonal coverage, despite evidence suggesting against this practice.
Validation of an electronic MUE allows for replicable, scalable tracking of national patterns of antimicrobial prescribing. This will enable us to the examine practice patterns and the impact of interventions to improve the quality of antimicrobial prescribing.