“Deep Sepsis” Licensed to Cohere Med
Cohere Med licenses technology from Duke to drive adoption of early Sepsis detection using AI
PRESS RELEASE: 3 JULY 2019
Sepsis strikes more than a million Americans every year and 15 to 30 percent of those affected die. Caused by an overwhelming immune response to infection, sepsis rates have steadily been on the rise in the country. This is a major challenge in hospitals, where it is one of the leading causes of death. It is also a main reason why people are readmitted to the hospital. Sepsis occurs unpredictably and can progress rapidly. It often involves a prolonged stay in the intensive care unit and complex therapies with high costs. Sepsis as the most expensive condition treated in U.S. hospitals, costing nearly $24 billion in 2013.
Late last year, Duke Health developed and deployed an Artificial Intelligence (AI) system for early detection of sepsis. “Significant progress has been made since then to validate the accuracy of the model that we developed,” said Suresh Balu, Director of the Duke Institute for Health Innovation (DIHI). “With a deep learning model ingesting over 50,000 patient records and more than 32 million data points, we are able to identify patients at risk for developing sepsis with greater than 90% accuracy,” he added. Traditional scoring mechanisms such as NEWS, SIRS, and QSOFA usually start with high detectability from the time a patient presents to Emergency Department (ED) but their accuracy decrease over time making detectability of sepsis harder with a large number of false positives.
To scale the solution further across the globe, Cohere Med, a clinical analytics company based in the US and India has licensed technology from Duke University. Cohere Med’s CoMeT- Coherence of Medical Things® platform is built to deploy enterprise-class AI solutions for health systems in critical care. CoMeT further is expected to enhance the technology with real-time processing of events using internet of things (IoT) for high fidelity clinical data, interoperability standards such as Fast Healthcare Interoperability Resources (FHIR), electronic medical records (EMR) extensions to ease the integration of sepsis detection, and management into already installed information systems along with a host of other deep learning-based algorithms. “Identifying and predicting patient decompensation for critically ill is a key focus area as we bring in high fidelity information across systems together in real-time,” said Srikanth Muthya, CEO, Cohere Med.
For more information: www.cohere-med.com and contact@cohere-med.com
References:
https://www.nigms.nih.gov/education/pages/factsheet_sepsis.aspx#1
https://www.beckershospitalreview.com/quality/duke-university-hospital-to-roll-out-ai- system-for-sepsis.html