Inappropriate antibiotic prescribing has decreased in recent years, but 50 percent of antibiotic prescriptions are still unnecessary. Patients and physicians contribute to this problem. Samore and colleagues designed a randomized trial to educate physicians on appropriate prescribing for acute respiratory tract infection. The control arm received only a community intervention, whereas the intervention arm also received a direct intervention using a clinical decision support system (CDSS).

Twelve rural communities of various sizes in Idaho and Utah were selected and randomly assigned to the community intervention or to the community plus CDSS intervention. An additional six communities were included for reference. The community intervention consisted of media campaigns, distribution of educational materials, and distribution of self-management guides to individual patients. For the CDSS intervention, primary care physicians were recruited through physician contacts, hospitals, and continuing medical education courses. Participants were given the choice of two paper-based tools (i.e., a chart documentation tool and a graphic f low chart) and one personal digital assistant (PDA) electronic-based tool. Each physician was asked to use the chosen tool on at least 200 consecutive patients with acute respiratory tract infection. Antimicrobial prescribing patterns were determined through detailed chart review and retail pharmacy volume. Diagnoses were divided into categories of those never indicated, sometimes indicated, and always indicated according to appropriateness of antibiotic use.

Of the CDSS arm, 45 percent of participating physicians attended at least one continuing medical education session, and 71 percent used the decision-support tools; about one half used the PDA tool, one fourth used a paper tool, and one fourth used a mix of PDA and paper tools.

Before the intervention, prescribing rates ranged from 26 to 198 prescriptions per 100 person-years and were similar across study arms. In the first year of the study, prescribing rates did not change significantly in either arm. In the second year, prescribing rates decreased by 10 percent in the CDSS arm, whereas they increased 1 percent in the community intervention arm and increased 6 percent in nonstudy communities. The chart review revealed that the top diagnoses for respiratory infections remained unchanged over three years. In the never-indicated category, there was an 11 percent reduction of antimicrobial prescribing in the CDSS arm and a 2 percent reduction in the community-intervention-only group. There were no statistically significant prescribing reductions in the other categories. Prescriptions for macrolides declined correspondingly in the CDSS group. Finally, the greater the number of case-based algorithms used, the greater the decline in antimicrobial prescribing.

These results suggest that repetitive use of a handheld decision-support tool reduces inappropriate antibiotic prescribing. Further studies are recommended to determine whether these interventions are sufficient to lower the incidence of antibiotic resistance.