Phelps Hospital has launched Dascena’s InSight algorithm house wide. InSight is a machine learning platform built for accurate sepsis onset predictions using only vital signs in the hospital’s EHR, greatly improving the timing of diagnosis and treatment.
Phelps Hospital operates on the fundamental values of community and patient service. It has achieved the highest level NICHE status (Exemplary) for care of older adults and has recently joined the ranks of the international “Baby Friendly Hospitals” distinction. This 238 bed hospital overlooking the Hudson river in Sleepy Hollow, New York, is also known for its excellence in cancer care and perpetual pursuit of new diagnostics and treatments to provide its community the best service possible.
In line with that goal, Phelps Hospital leadership decided to pursue improvements to sepsis-related outcomes, leading them to implement InSight house-wide. Before the use of InSight, Phelps clinicians used manual screening and the standard SIRS criteria to detect and treat sepsis patients. With the use of InSight, clinicians are given automated, advanced notifications of sepsis allowing for earlier and more successful treatments.
Supported by Dascena’s onboarding team, staff at Phelps Health were trained in the use of InSight, and unit specific risk escalation systems were established.
“We are excited to help improve sepsis-related outcomes at Phelps Hospital, and expect InSight to increase the already high level of care in PH by integrating with the hospital system and enabling early and accurate sepsis diagnosis,” said Ritankar Das, CEO of Dascena.
Dascena's website displays impressive numbers for the InSight technology. The stats, generated via use in over 75,000 patients include: 39.5% reduction in mortality, 32.3% reduction in length of stay and 22.7% reduction in readmissions.
The company also has two additional platforms, AutoTriage and Previse. AutoTriage predicts decompensation and helps physicians determine the best care setting. Previse predicts acute kidney injury. Dascena claims the platform can predict acute kidney injury more than a full day before patients meet the clinical criteria for diagnosis.
