Baltimore-based Johns Hopkins Medicine deployed an algorithm alert system that analyzes patients' EHR data to help identify individuals at-risk of sepsis.
The health system currently uses the targeted, real-time early warning system — or TREWS — at Howard County General Hospital in Columbia, Md., and Bethesda, Md.-based Suburban Hospital and will integrate it with the EHR system at Johns Hopkins Hospital, Johns Hopkins Bayview Medical Center in Baltimore and Washington, D.C.-based Sibley Memorial Hospital later this year.
If TREWS detects a patient is at risk of developing sepsis, the system will alert clinicians through the EHR and suggests the provider take measures such as prescribing antibiotics or requesting blood cultures.
Johns Hopkins computer science graduate student Katharine Henry, who helped develop TREWS, said the system makes fewer false-positive alerts than other sepsis detection systems because it analyzes more data points from the patients' EHRs than other systems do, according to the report. She added that one of every two to three TREWS alerts is correct, compared to one in 10 accurate alerts from other systems.