Development of antibiotic metrics for hospitalists via multi-institutional modified Delphi survey

Background

Closing the gap between evidence-supported antibiotic use and real-world prescribing among clinicians is vital for curbing excessive antibiotic use, which fosters antimicrobial resistance and exposes patients to antimicrobial side effects. Providing prescribing information via scorecard improves clinician adherence to quality metrics.

Objective

We aimed to delineate actionable, relevant antimicrobial prescribing metrics extractable from the electronic health record in an automated way.

Design

We used a modified Delphi consensus-building approach.

Settings and Participants

Our study entailed two iterations of an electronic survey disseminated to hospital medicine physicians at 10 academic medical centers nationwide.

Main Outcomes and Measures

Main outcomes comprised consensus metrics describing the quality of antibiotic prescribing to hospital medicine physicians.

Results

Twenty-eight participants from 10 United States institutions completed the first survey version containing 38 measures. Sixteen respondents completed the second survey, which contained 37 metrics. Sixteen metrics, which were modified based on qualitative survey feedback, met criteria for inclusion in the final scorecard. Metrics considered most relevant by hospitalists focused on the appropriate de-escalation of antimicrobial therapy, selection of guideline-concordant antibiotics, and appropriate duration of treatment for common infectious syndromes. Next steps involve prioritization and implementation of these metrics based on quality gaps at our institution, focus groups exploring impressions of clinicians who receive a scorecard, and analysis of antibiotic prescribing patterns before and after metric implementation. Other institutions may be able to implement metrics from this scorecard based on their own quality gaps to provide hospitalists with automated feedback related to antibiotic prescribing.