Methods: We developed a prevalence-weighted, condition-specific, real-time antibiogram that aggregates and displays bacterial culture data from our electronic health record according to Clinical and Laboratory Standards Institute guidelines. Using a commercial data visualization platform (Qlikview) that can be refreshed periodically to incorporate new information, we compared susceptibility rates for selected pathogen-drug-patient combinations between the cumulative and stratified antibiograms using χ2 or Fisher’s exact test.
Results: The stratified antibiogram revealed important differences in antimicrobial susceptibility across epidemiologic subsets of patients. For example, the overall susceptibility of methicillin-resistant S. aureus (MRSA) to clindamycin was 78%, but was 91% from wound cultures obtained in the Emergency Department [c2 (1, N=3425) =74.7, p <0.01]. Similarly, the susceptibility of E. coli to cefazolin among all urine cultures was 87% compared with 67% in hospital-onset urinary tract infections [c2 (1, N=2754) =53.1, p <0.01]. The web-based antibiogram also displayed real time information regarding the most common organisms isolated, with important differences when stratified by hospital unit. For example, Staphylococcus aureus accounted for 21% of community-onset (<48 h from admission) bacteremia but only 6% of pathogens among oncology patients with bacteremia.
Conclusion: Stratified antibiograms reveal clinically important differences in susceptibility patterns that can be obscured by hospital-cumulative antibiograms. Further research is needed to determine if delivery of this information to front line clinicians can facilitate appropriate antibiotic selection.
T. A. Metjian, None
K. Ota, None
B. Desai, None
T. Zaoutis, Merck: Grant Investigator , Research grant
J. S. Gerber, None