Improving empiric antibiotic selection using computer decision support

Abstract

Background: Physicians frequently need to start antibiotic therapy before the results of bacterial cultures and antibiotic susceptibility tests are available. We developed and evaluated a computerized antibiotic consultant to assist physicians in the selection of appropriate empiric antibiotics.

Methods: We used a two-stage random-selection study to compare antibiotics suggested by the antibiotic consultant with 482 associated antibiotic susceptibility results and the concurrent antibiotics ordered by physicians. The antibiotics ordered by randomized physicians were then compared between crossover periods of antibiotic consultant use.

Results: The antibiotic consultant suggested an antibiotic regimen to which all isolated pathogens were shown to be susceptible for 453 (94%) of 482 culture results, while physicians ordered an antibiotic regimen to which all isolated pathogens were susceptible for 369 culture results (77%) (P < .001). The physicians who prescribed antibiotics to which all pathogens were susceptible did so a mean of 21 hours after the culture specimens were collected. Physicians ordered appropriate antibiotics within 12 hours of the culture collection significantly more often when they had use of the antibiotic consultant than during the period before use (P < .035). Moreover, 88% of the physicians stated they would recommend the program to other physicians, 85% said the program improved their antibiotic selection, and 81% said they felt use of the program improved patient care.

Conclusions: Information from computer-based medical records can be used to help improve physicians' selection of empiric antibiotics for infections.