Trial Summary

Study question and significance: Although fewer than 5% of non–critically ill hospitalized patients have an antibiotic-resistant infection, more than half receive extended-spectrum antibiotics. Tools to support judicious antibiotic prescribing are needed to curb the public health threat of antibiotic resistance.

Design and setting: Two cluster randomized trials in 92 hospitals compared routine hospital-based antibiotic stewardship programs with an enhanced approach that added a computerized physician order entry (CPOE) prompt to reduce unnecessary prescribing of extended-spectrum antibiotics in non–critically ill patients at low risk of infection by an antimicrobial-resistant pathogen. The study included adult patients treated with empiric antibiotics in non–intensive care units who were hospitalized with abdominal infections or skin or soft tissue infections.

Intervention and methods: In hospitals randomized to the intervention group, clinicians received CPOE prompts recommending standard-spectrum instead of extended-spectrum antibiotics when they ordered extended-spectrum antibiotics for patients with a less than 10% absolute risk of having an antibiotic-resistant infection. Clinicians also received education and feedback on their antibiotic prescribing practices.

Findings: There were 105,004 with abdominal infections and 60,725 patients with skin or soft tissue infections. The CPOE prompts led to 35% and 28% relative reductions in empiric extended-spectrum antibiotic days of therapy in the abdominal infections trial and the skin or soft tissue infections trial, respectively. These reductions occurred without an increase in transfers to an intensive care unit or length of stay.

Conclusions and relevance: A program that used automated prompts to deliver patient-specific risk information to clinicians led to significant reductions in the use of extended-spectrum antibiotics for hospitalized patients.

NIH Project Information