"The creation of a cognitive machine-enabled approach to identify which COPD patients are on a trajectory toward an avoidable inpatient admission within 30 days addresses a significant need with regard to care management in a complex patient population. By identifying which patients need additional assistance and who is at risk of an acute exacerbation, Geisinger is able to better intervene on those patients with a significantly elevated risk with regard to inpatient admissions."
In October 2017, a unique collaboration between Geisinger, GSK, the COPD Foundation, and Jvion launched with the goal of improving the lives of chronic obstructive pulmonary disease (COPD) patients. Following an initial, successful pilot and validation of Cognitive Machine performance, Geisinger has elected to move forward with the integration of AI as a key asset in driving down rates of complications and hospitalizations for COPD patients served by the system’s largest medical center.
COPD encompasses a group of progressive lung diseases that gradually increase breathlessness and wheezing. The disease -- which includes emphysema, chronic bronchitis, refractory (non-reversible) asthma, and some forms of bronchiectasis -- impacts more than 30 million people in the U.S. and is the nation’s third leading cause of death. According to the COPD Foundation, COPD patients account for the highest rate of avoidable inpatient stays and the second highest rate of preventable ED visits. And this trend could reach epidemic levels within the next 15 years as COPD related in-patient days are estimated to increase by 185% by 2030.
To help prevent avoidable admissions for COPD patients and associated deterioration, Geisinger is applying Jvion’s Cognitive Machine to identify at risk patients and action the most effective interventions that will keep a patient healthy and out of the hospital. These insights target individuals who are in the outpatient setting and whose trajectory toward an avoidable inpatient stay can be impacted through individualized clinical intervention.
Cognitive Machine Application
Current approaches to identifying COPD patients at risk of an avoidable inpatient admission stratify the target population without accounting for the clinical differences driving the risk. This blanket approach has a significant negative impact for COPD patients at risk of an acute exacerbation.
Acute exacerbations of COPD (AECOPD) are a leading cause of COPD patient deterioration. These incidences also have a significant financial impact on the healthcare system. According to a recent study released in the International Journal of Chronic Obstructive Pulmonary Disease, AECOPD-related costs are estimated to be around $4069/year per patient. These costs increase with exacerbation frequency, severity, and the presence of comorbidity.
The true incidence of AECOPD is difficult to determine because of underreporting. Approximately 50% of AECOPD are not reported by patients. But AECOPD drives 2.4% of acute hospitalizations. And acute exacerbations have an overall mortality rate of 11.6%, which increases up to 37% in patients with repeat admissions.
Jvion’s Cognitive Machine helps Geisinger more precisely identify patients at risk of an AECOPD within the next 30 days. For each patient, the clinical, behavioral, and socioeconomic factors contributing to his or her risk are identified by the machine. Using the Eigen Sphere technology at the heart of the machine’s engine, the solution extrapolates which interventions will result in the best outcome while also ensuring patient engagement. Each output is actionable, highly personalized, and patient-specific to drive to the greatest likelihood of improved health. This capability to addresses the shortcomings of existing predictive or machine learning models by enabling a more granular and personalized view into each patient’s risk of an AECOPD event.
Current Cognitive Machine performance is enabling the identification of COPD patients who are at a 30x increased risk of an acute exacerbation. By reducing treatment failure and the possibility of a relapse of AECOPD, Geisinger is working to reduce the disease burden and improve the management of COPD patients. Preliminary results are demonstrating a 50% reduction in avoidable admissions for COPD patients when Cognitive Machine recommendations are actioned.