This study in the Journal of Emergency Medicine retrospectively reviewed patients with severe sepsis and septic shock who had positive blood cultures, and compared the antimicrobial sensitivities of the isolates with the initial ER therapy:
Effective antibiotic coverage was prescribed by emergency physicians in 82% (95% confidence interval [CI] .74–.88) of cases. Of the 25 patients who received ineffective antibiotics, the majority had infections caused by resistant Gram-negative organisms.
So the ER physicians missed the boat in 18% of the cases. This was an academic ER. One might expect a higher miss rate in community hospitals. The investigators then looked at what effect a commonly used readily available guideline would have had on the miss rate:
Health care-associated pneumonia guidelines were applied to all patients, regardless of the source of infection, and were 100% sensitive (95% CI .93–1) for selecting patients who had infections caused by highly resistant organisms.
Using HCAP guidelines may offer an advantage in antibiotic selection over the Surviving Sepsis guidelines because the latter make only general statements about administering sufficiently broad spectrum antibiotics. The HCAP guidelines, on the other hand, list specific regimens.
What's also unique about this method is that the antibiotic selection is based on the general level of microbiologic risk rather than the likely organ source of the infection which is the basis for more traditional recommendations. The presentation of many septic patients is undifferentiated and it has long been my feeling that basing selection on the likely organ source risks too narrow a spectrum in the very ill patients (severe sepsis and septic shock) who so often present to the ER.
Full text at Medscape.