The impact of an infectious disease (ID) service on the optimal antibiotic
management of 103 patients with bloodstream infections, defined as bacterem
ia and systemic inflammatory response syndrome, was evaluated. The optimal
antibiotic management was defined according to the Sanford Guide to Antimic
robial Therapy (1996) or written internal guidelines. The judgment on optim
al antibiotic management was made at the time of reporting the positive blo
od culture results. Switching from a broad-spectrum to a narrow-spectrum ag
ent was carried out significantly more often by the ID service than by the
attending physicians (25 of 25 vs. 20 of 40; P<0.001). In patients without
empirical therapy, the ID service initiated optimal antimicrobial therapy s
ignificantly more often than physicians without training in infectious dise
ases (12 of 12 vs. 4 of 10, P = 0.0028). Three of 12 patients in whom the a
ttending physician misinterpreted the positive blood culture result needed
8 days to 4 months of additional hospitalization. In summary, patients for
whom an ID service was provided received appropriate treatment more often a
nd experienced significantly fewer complications.