Reducing Hospital-Acquired Infections with Epic’s Sepsis Predictive Analytics Models
5 min read
Background
Incisive Consultants began working with Southeast Health in 2019 when they began implementing Epic’s clinical applications across the enterprise. Prior to Southeast Health’s implementation of Epic, the only tools available to identify patients for being potentially septic or at risk of becoming septic required manual documentation from the nursing staff. This was not very efficient since nursing had to manually document an assessment for the patient. This led to less accurate results due to varied clinical judgment and a limited number of factors that were taken into consideration when generating a score for the patient.
“On any given day, 1 out of every 31 hospital patients has at least one healthcare acquired infection.”
Implementation of Epic’s Sepsis Predictive Analytics Models
With the implementation of Epic, Southeast Health took advantage of implementing the 5 Early Detection of Sepsis Predictive Models. These automated models take data points from the patient’s chart and use them to calculate a risk score for the patient. Using this score, we have configured alerts for the clinical staff when the patient is at risk.
Barriers to Implementation
One obstacle that we encountered with this implementation was working with clinician to mutually understand the criteria available for the different predictive model alerts. We addressed this issue with the creation of educational Quick Start Guides and establishing workgroups to optimize the criteria for the various alerts.
Since the workgroups creation, we have successfully adjusted several aspects of the model resulting in keeping our alerts beneficial and reducing alert fatigue for our clinicians.
Predictive Models in Action
By utilizing the sepsis predictive models, clinicians are able to quickly and easily identify their patients that are at risk for sepsis the moment they log into the system. The efficiency of the models allows clinicians to provide appropriate interventions faster leading to less infections and better patient care.