Enabling New Models of Care: Applying Eigenspace to Stop Patient Readmissions -- Testimonials Title
This Central Florida Integrated Delivery Network (IDN) is the largest provider in the area and delivers health insurance for thousands within the community. As a stand-out innovator and leader in the adoption of value-based/at-risk models of care, the system is continuously looking for new and more effective ways of improving patient health outcomes.
When leadership turned their focus to reducing readmissions across all of the provider’s inpatient facilities, they knew they needed powerful predictive analytics. They wanted a solution that could pin point patients at risk of a readmission and provide the intervention effectiveness information that would help ensure a healthy discharge.
Jvion's solution delivers readmission predictions at the patient-level to help providers reduce readmission rates and drive improved health outcomes. The solution, which combines deep machine learning technology, clinical intelligence, and advanced statistics, uses the latest data science technique know as clinically relevant data pods to render the most accurate predictions available in the market.
Jvion's solution uses the data that is already on hand and can account for incomplete and inaccurate data elements because of the advanced data capabilities that drive the predictive engine. The resulting analytic power extends patient-level predictions to interventions and likelihood of patient engagement. Moreover, Jvion's solution is equipped for much more than readmissions. From nosocomial events to community health efforts, Jvion's solution is pre-seeded with dozens of use cases designed for all types of provider environments.
In the first month of results, 175 individuals were identified at high risk of readmissions across all facilities. With each readmissions costing the system an average of $11,200, the one-month savings adds up to a potential total savings of more than $1.9M.
It is important to note that these results were rendered without historical data information from the system. The clinically relevant data pods enabled accuracy levels three-time that of a LACE score even without initial tuning to the system’s patient population. And as the solution processes more of the system’s data, accuracy will continue to increase.