VA collaboration: Algorithm helps identify underlying health risks

Can an algorithm help identify rising health risks in Veterans? The Veterans Health Administration Innovation Ecosystem (VHA IE) collaboration with CareCentra, a leading AI and Machine Learning company, is attempting to find out.

VA enrolled Veterans are provided some of the best care in the country. However, clinical care only impacts 20 percent of Veteran outcomes. Lifestyle behaviors, social, economic and environmental factors drive the other 80 percent.

A Veteran’s physical and mental health are far more impacted by factors outside of a clinical setting than within it. VHA IE realizes the importance of supporting Veterans beyond their local VA medical facility.

To address this, VHA IE is collaborating with CareCentra’s Unified Health Algorithm to detect disease, prevent progression, and reduce Veteran hospitalization. This algorithm identifies distant signals of disease and upstream risk to help Veterans address it in real-time, before they step into a VA medical center.

Closing the gap

“The biggest gap in health care today is the time when our patients are not interacting with the health care system,” explained Chief Health Informatics Officer and Physician Executive at the Ralph H. Johnson VA Medical Center Dr. Cole Zanetti. “Having the ability to seamlessly nudge our Veterans through behavioral decision support toward better health is going to allow us to close that gap.”

Zanetti also believes this algorithm will benefit Veterans with post-combat mental health issues.

“Poor mental health amplifies risk of other co-morbid conditions like hypertension, or diabetes,” he added. “Detecting depression risk from lifestyle behaviors and other non-clinical signals and connecting Veterans to ‘just-in-time’ intervention can reduce health risks for Veterans.”

Building the algorithm

CareCentra’s self-learning algorithm was tested through a randomized clinical trial presented at the Scientific Sessions of the American Heart Association in 2019. It combines behavior science, AI/Machine Learning, and social technologies to identify rising risks of conditions, even if undiagnosed. The algorithm then “nudges” patients toward small health actions to mitigate risk.

For example, if the algorithm picks up signals that indicate a Veteran could be heading toward a diagnosis of Type 2 Diabetes, it will suggest preventive micro-health-actions. These actions, linked to diet, exercise and stress management activities, can reduce risk and prevent the diagnosis.

The micro-actions that the algorithm suggests are personalized to the Veteran’s unique motivations and abilities on any given day.

Vasant Kumar, founder and CEO of CareCentra explained that “The algorithm learns what actions each patient is motivated and able to do, and [it] nudges them to do just that, no more.”

“Committed to helping Veterans live their best life.”

By taking these small actions, Veterans can move toward their larger goal of better health, showing measurable improvement.

With the algorithm’s universal application across chronic and acute conditions and delivery in a Prevention-as-a-Service model, VHA IE will identify the areas of greatest need within the VA population. The pilot program will be first launched across five states and will cover 10,000 Veterans.

“We’re committed to helping our Veterans live their best life and this collaboration is part of that commitment,” said Kit Teague, executive director of VHA IE.