P-912. Correlation between Novel Antibiotic Stewardship Metrics and Antibiotic Use Practices for Lower Respiratory and Urinary Tract Infection across 24 VA Hospital Systems

TL;DR

This study evaluates new antibiotic use metrics to see if they reflect better antibiotic practices in treating infections at 24 VA hospitals.

Contribution

The study introduces two new antibiotic stewardship metrics (RSR-DOT and RSR-DASC) and evaluates their validity in measuring appropriate antibiotic use.

Findings

Lower RSR-DASC was significantly associated with better antibiotic selection and appropriate duration for lower respiratory tract infections.

RSR-DOT did not show significant associations with antibiotic appropriateness.

For UTIs, neither metric showed significant associations with antibiotic use practices.

Abstract

Various metrics have been proposed to measure antibiotic consumption in inpatient settings, but little is known about how these metrics correlate with appropriateness of antibiotic usage (i.e. construct validity). We developed two metrics (risk-standardized ratio: RSR), one based on Days of Therapy (RSR-DOT) and another based on Days of Antimicrobial Spectrum Coverage (RSR-DASC). We aimed to evaluate whether hospital performance on these metrics is associated with appropriateness of antibiotic selection and duration.Table 1:Assessments of Diagnosis, Antibiotic Selection on Day 3, and Antibiotic Duration for LRTI and UTIFigure 1:Application of Inclusion and Exclusion Criteria

Assessments of Diagnosis, Antibiotic Selection on Day 3, and Antibiotic Duration for LRTI and UTI

Application of Inclusion and Exclusion Criteria

Using data from October 2020 to September 2021, we constructed RSR-DOT and RSR-DASC for 118 acute-care hospitals in the Veterans Health Administration and selected 24 with lower or higher RSRs for DOT and DASC. For each hospital, an Infectious Disease physician, blinded to the hospital’s performance, reviewed 10 cases each of lower respiratory tract infections (LRI) and urinary tract infections (UTIs) to assess the appropriateness of antibiotic selection on day 3 (6-level ordinal outcomes) and appropriateness of antibiotic duration. Associations with RSR metrics were assessed using ordinal logistic regression (antibiotic selection) and logistic regression (antibiotic duration).Figure 2:Pie Charts representing Frequency of LRTI and UTI Diagnosis, Antibiotic Selection on Day 3, and Duration AssessmentTable 2.Association between Hospital Performance on the risk-standardized metrics and antibiotic appropriateness for LRTI and UTI

Pie Charts representing Frequency of LRTI and UTI Diagnosis, Antibiotic Selection on Day 3, and Duration Assessment

Association between Hospital Performance on the risk-standardized metrics and antibiotic appropriateness for LRTI and UTI

The final cohort included 240 cases for both LRI and UTI diagnoses after applying exclusion criteria (Figure 1 and Table 1). Lower RSR-DASC was significantly associated with better antibiotic selection at day 3 and appropriate duration for LRI, while RSR-DOT did not show significant associations (Table 2). For UTIs, we did not observe significant associations between antibiotic selection at day 3 or appropriate duration with either metric. When cases of LRI and UTI were combined, lower RSR-DASC was significantly associated with better antibiotic selection at day 3.

Hospital evaluations by RSR-DASC were associated with appropriateness of antibiotic selection at day 3 and the antibiotic duration in LRI, while RSR-DOT did not show any significant associations, suggesting higher construct validity with RSR-DASC. The lack of associations with both metrics in UTI might be due to fewer intra-hospital differences in antibiotic selection when microbiologic data (e.g. urine cultures) are available and high frequency of asymptomatic bacteriuria treatment.

All Authors: No reported disclosures

Figures