Antibiotic Diversity Index: A novel metric to assess antibiotic variation among hospitalized children

Abstract

Background: Despite nationally endorsed treatment guidelines and stewardship programs, variation and deviation from evidence-based antibiotic prescribing occur, contributing to inappropriate use and medication-related adverse events. Measures of antibiotic prescribing variability can aid in quantifying this problem but are not adequate.

Objective: The objective of this study is to develop a standardized metric to quantify antibiotic prescribing variability (diversity) within and across children's hospitals, and to examine its association with outcomes.

Methods: We performed a cross-sectional study of empiric antibiotic exposure among children hospitalized during 2017-2019 with one of 15 common pediatric infections using the Pediatric Health Information System database. Encounters for children with complex chronic conditions, transfers in, and birth hospitalizations were excluded. Using the Shannon-Weiner entropy index, we quantified antibiotic diversity for each infection type using the d-measure of diversity. Generalized linear mixed-effects models were used to examine the association between hospital-level antibiotic diversity and risk-adjusted length of stay and costs.

Results: A total of 79,515 hospitalizations for common pediatric infections were included. Antibiotic diversity varied within and across hospitals. Infections with low mean antibiotic diversity included appendicitis (mean diversity [mDiv] = 4.9, SD = 2.5) and deep neck space infections (mDiv = 5.9, SD = 1.9). Infections with high mean antibiotic diversity included pneumonia (mDiv = 23.4, SD = 5.6) and septicemia/bacteremia (mDiv = 28.5, SD = 12.1). There was no statistically significant association between hospital-level antibiotic diversity and risk-adjusted LOS or costs.

Conclusions: We developed and applied a novel metric to quantify diversity in antibiotic prescribing that permits comparisons across hospitals and can be leveraged to identify high-priority areas for local and national stewardship interventions.

FIGURE 1

Consort diagram.

FIGURE 2

Heat map demonstrating antibiotic diversity within and across hospitals. Within our heat map, each row represents an individual infection, whereas each column represents an individual hospital. Each cell represents an individual hospital’s infection-specific Antibiotic Diversity Index (ADI). Hospitals were arranged left to right in the heat map based on an individual hospital’s overall mean ADI across all infections. Color was applied independently to each infection type and to the overall hospital-specific mean ADI. The red color represents increased diversity, and the green color represents decreased diversity.