Application of a Novel Desirability of Outcome Ranking Endpoint Incorporating a Framework for Resistance Assessment and Microbiologic Evaluation in Antibiotic Trials: A Proof-of-Concept in a Post Hoc Analysis of the BALANCE Trial

Abstract Background Desirability of Outcome Ranking (DOOR) with Response Adjusted for Duration of Antibiotic Risk (RADAR) has been suggested as an outcome for antibiotic duration trials. An endpoint incorporating a measure of antimicrobial resistance may better reflect trade-offs between different antibiotic strategies. Methods We constructed a novel endpoint, DOOR-FRAME (Framework for Resistance Assessment and Microbiologic Evaluation), and applied it as a proof-of-concept in a post hoc analysis of the BALANCE trial comparing 7- versus 14-day antibiotic duration for uncomplicated non-Staphylococcus aureus bloodstream infection (BSI). The DOOR comprised mortality, readmission, and reinfection, while the FRAME component was a quantitative composite measure of incident Clostridioides difficile infection (CDI) and positive cultures for antibiotic-resistant organisms (AROs). We included BALANCE patients in Ontario who consented to data linkage to provincial health administrative databases. We calculated the probabilistic index (PI), the probability of a randomly selected patient in the 7-day group having a better outcome at 180 days. Results Two hundred seventy-one participants were included. The DOOR distributions were similar between the 7- and 14-day groups, with PI of 46.8% (95% CI 40.4%–53.2%) using DOOR alone. Nine (6.9%) and 12 (8.5%) patients in the 7 and 14-day group respectively had ARO/CDI within 180 days. Probabilistic index was 53.6% (95% CI 46.7%–60.4%) with DOOR-RADAR and 47.4% (95% CI 40.9%–53.8%) with DOOR-FRAME. Conclusions There were no significant differences in DOOR outcomes at 180 days between patients with uncomplicated BSI treated with 7 versus 14-day antibiotic duration. Incorporating a novel measure of antimicrobial resistance did not change interpretation of the trial.