Healing soles: a microbiology-driven electronic health record-algorithm and order set to decrease antipseudomonal use in diabetic foot infections, a retrospective, observational, quasi-experimental study

• DOI: 10.1017/ash.2025.59
• Publication Date: 2025-03-27T10:46:27Z
• Authors: Antoinette Marie ...; Naida Koura-Mola; Terrence McSweeney; Hongkai Bao; Mei Chang; Kelsie Cowman; Priya Nori; Yi Guo

Abstract

Background Antipseudomonal antibiotics are commonly prescribed for diabetic foot infections (DFI) at our institution despite a low local prevalence of Pseudomonas aeruginosa. A multidisciplinary team implemented a DFI electronic health record (EHR)-embedded treatment algorithm and order set.

Methods This multi-center, quasi-experimental study evaluated adults on antibiotics admitted for DFI to vascular surgery or medical units pre- and post-implementation of an EHR-embedded treatment algorithm and order set. Exclusion criteria: duplicate patients, concomitant infection, transfer from an outside hospital. Primary endpoint: antipseudomonal use among included patients (DOT/1000 DFI days present). Secondary outcomes: empiric antipseudomonal use, length of stay, 30-day readmission, mortality, amputation, and Clostridioides difficile infection. Patient demographics, diagnostics, treatments, and outcomes were evaluated.

Results Two hundred ten patients were included with 70 patients included in each group. The post-algorithm group had lower antipseudomonal DOT/1000 DFI days present compared to the pre-intervention group (360 vs 503, P < 0.001). The post-order set group had the lowest antipseudomonal use (347 vs 503, P < 0.001). Empiric antipseudomonal use decreased from 85.7% pre-intervention to 72% post-algorithm and 68.5% post-order set. Collectively, 30-day mortality was < 5%. Amputation during and within 30 days of hospitalization was similar in the pre-intervention (48.6%), post-algorithm (30%), and post-order set (41.4%) groups. Methicillin-susceptible Staphylococcus aureus and Streptococcus spp. were most frequently isolated. Wound cultures were not collected in 24.3%, 22.9%, and 40% of the pre-intervention, post-algorithm, and post-order set group.

Conclusions EHR-embedded clinical decision-making tools reduce antipseudomonal use for DFI treatment without increasing 30-day mortality or amputation.