Comparison of Hospitalwide and Custom Antibiograms for Clinical Isolates of Pseudomonas aeruginosa

Abstract

Background Hospital antibiograms, which are commonly used to determine empiric antibiotic therapy and as a tool in stewardship in a given institution, are open to bias when combining susceptibility results from various sources, hospital locations, and patient groups. Methods We assessed such differences, using Pseudomonas aeruginosa as a test case, with susceptibility data from 2008 through 2010 in our institution. Each year's data were analyzed separately. A variety of specific or subcategorical antibiograms were compared with each other as well as with versions including all tested isolates and those with results from inpatients and outpatients only. Statistical significance was determined at the .01 level using either chi-square or Fisher exact test, and clinical significance was defined as ≥10 percentage points. Results A variety of clinically significant differences were found that illustrated important differences within the intensive care unit environment and based on population, specifically adult versus pediatric. Concordance between statistically significant and clinically significant differences was poor. Conclusion These results corroborate and extend previous similar observations and point to the potential importance of subanalyses in preparing the annual hospital antibiogram.