2026 Eastern States Conference

Title: 
Impact of a Bayesian clinical decision support tool on time to appropriate antimicrobial therapy in gram-negative bacteremia infections

Authors:
Jenna Mangan, Craig Worby, Rebecca Wang, Julie Ann Justo

Learning Objective:
Describe whether implementation of the clinical decision support tool pRxcision® reduces time to appropriate antibiotic therapy compared with standard clinical decision-making.

Background:
By integrating electronic health record inputs with Bayesian inference, this study will investigate whether pRxcision® can accelerate the identification and administration of appropriate antimicrobial therapy in gram-negative bacteremia.

Methods: 
This single-center observational cohort study includes adult patients hospitalized at Dartmouth-Hitchcock Medical Center with gram-negative bloodstream infections who received at least one dose of empiric antibiotics. Patients will be assigned to pre- or post- implementation groups based on whether they were admitted during the six months before or after the launch of the pRxcision® clinical decision support tool. Patients will be excluded if they have polymicrobial bacteremia, were discharged or deceased before final blood culture results, or experienced recurrent gram-negative bloodstream infections during the same hospitalization. Outcomes for both groups will be compared, with data collected before and after implementation to evaluate changes over time. The primary goal is to evaluate whether the Bayesian model used in pRxcision® improves time to appropriate antibiotic therapy compared to standard clinical practice in gram negative bloodstream infections.

Results : 
Data collection remains ongoing, as a number of operational challenges encountered during implementation have delayed the original timeline. The anticipated go-live date for pRxcision® at DHMC was initially February 2026; however, due to slower-than-expected development and various obstacles, it has been postponed to October 2026. While the implementation was originally estimated to require approximately 32.5 hours from the DHMC information technology (IT) team, the average time invested by DHMC IT resources to date has reached 170 hours. 

Conclusions: 
If successful, DHMC will be the first institution in the United States to integrate pRxcision® into the Epic® electronic health record. As a result of unforeseen implementation challenges within Epic® that remain unresolved, the initial project timeline has been delayed. Despite these setbacks, the use of pRxcision® has the potential to significantly improve outcomes for patients with gram-negative bacteremia. 

Self-Assessment Question:
All of the following are stages along the technical implementation cycle EXCEPT:
a.    Innovators
b.    Early Adopters
c.    Late Majority
d.    Stalwarts