Comparison of UTI antibiograms stratified by ED patient  disposition

ObjectiveInstitutional antibiograms guide Emergency Department (ED) clinicians' empiric antibiotic selection. For this study, we created and compared antibiograms of ED patients stratified by disposition (admitted or discharged).MethodsWe conducted a cross-sectional study at two hospitals for 2014, comparing antibiograms limited to Escherichia coli urinary tract infections. Study-Specific Antibiograms, created for the study, excluded polymicrobial samples and multiple cultures from the same patient. Study-Specific Antibiograms were arranged by patient disposition: admitted (IP-Only) vs discharged from the ED (ED-Only). Antibiogram data were presented as average antibiotic sensitivities with 95% confidence intervals and demographic data as medians with interquartile ranges. Sensitivities between Study-Specific Antibiograms were compared by Fisher's Exact Test, alpha = 0.05, 2 tails.ResultsFor Hospital A, 13 antibiotics were compared between Study-Specific ED-Only (n = 313) vs IP-Only (n = 244). We found that sensitivities to all four antibiotics appropriate for empiric outpatient therapy by Infectious Disease Society of America guidelines were significantly (p < 0.0001) higher in the ED-Only compared to IP-Only groups: ciprofloxacin 80% (76-90%) vs 60% (53-69%), levofloxacin 81% (77-91%) vs 63% (57-72%), nitrofurantoin 75% (70-84%) vs 51% (44-58%), and trimethoprim/sulfamethoxazole 73% (68-82%) vs 58% (52-67%). For Hospital B, 14 antibiotics were compared between Study-Specific ED-Only (n = 256) and IP-Only (n = 168). Two out of the five appropriate empiric outpatient antibiotics had significantly (p < 0.0001) higher sensitivities for ED-Only compared to IP-Only: ciprofloxacin 87% (83-91%) vs 71% (64-78%) and levofloxacin 86% (82-91%) vs 71% (65-78%).ConclusionsWe found higher antibiotic sensitivities in ED-Only than the IP-Only Study-Specific Antibiograms. Our Study-Specific Antibiograms offer an alternative guide for antibiotic selection in the ED.