There are few areas of medicine as complex and compelling as delving into the intricacies of the human mind. The very uniqueness that creates individuality means there are rarely ‘one-size-fits-all’ solutions to multifaceted physical and behavioral health issues.
However, the team behind a newly launched national company believes they have crafted a way to inform clinical decisions and improve outcomes on a highly individualized basis while simultaneously lowering costs. Faros Healthcare, LLC – a spinout of Centerstone Research Institute and Indiana University Research & Technology Corporation (IURTC) – has developed a clinical tool that combines advanced predictive analytics with a patent-pending artificial intelligence (AI) platform.
The need for more sophisticated means to effectively treat the whole person has become increasingly evident as the industry moves toward value-based healthcare. “Behavioral health, in general, is really exploding,” noted Tom Doub, PhD, CEO of Centerstone Research Institute, one of the nation’s largest not-for-profit providers of community-based behavioral health and addiction services. “The rest of healthcare is beginning to realize, based on data and their own practice experience, that behavioral health is a very important part of achieving good health outcomes.”
When depression or other behavioral health conditions are layered on top of chronic diseases such as diabetes or heart disease, Doub continued, the cost to care for that patient is two to three times higher than if the patient was dealing solely with the physical condition. And studies over the last two decades have consistently shown an increased prevalence of depressive disorders or other psychiatric conditions in the presence of chronic illness.
“I think all of healthcare is really converging on not separating the body and the mind, and that goes along with the science,” Doub said of the rise in integrated care.
Jim Stefansic, PhD, a biomedical engineer who serves as president of Faros, said finding the optimal route to treat complex physical and behavioral issues is part of the company’s core mission. Casey Bennett, PhD, the company’s co-founder and chief scientific officer, invented the analytics tool while working with Doub at Centerstsone Research Institute as a data architect and research fellow during graduate school at Indiana University’s School of Informatics and Computer Science.
“We wanted to see if we could essentially develop a smart algorithm,” Doub said. “It’s not just one decision you have to make in healthcare, it’s many decisions; and the better you make those series of decisions, the better the outcome for patients.”
With data and demographics from more than 6,700 Centerstone patients with a clinical diagnosis of major clinical depression, of which between 65-70 percent also had a chronic physical co-morbid condition, Bennett and IU Assistant Professor Kris Hauser showed the efficacy of applying the augmented intelligence platform. Using 500 randomly selected patients for simulations, the team was able to utilize AI to improve outcomes by nearly 40 percent compared to baseline while simultaneously lowering the cost of care by about 50 percent. The results were published in Artificial Intelligence in Medicine in January 2013,
Stefansic explained Faros AI & Analytics Platform, which is co-owned by Centerstone and IURTC, personalizes the approach to care through the use of machine learning. Not only does the software make recommendations on the optimal course of treatment for complex conditions through data analytics from an initial set of parameters and markers, but it has the ability to learn over time and suggest adjustments to the protocol based on patient outcomes or changes in parameters while also calculating treatment costs.
The cloud-based platform can update in real time to provide point-of-care notifications to providers, who can then factor the new cost and outcomes data into the decision-making process to determine whether or not a treatment plan should be modified. Stefansic stressed it was equally critical that providers both feel confident in the results and be able to access the information as part of their natural workflow.
“Moving both sides of the costs-of-care value equation is essential in transforming our healthcare system, and we’re incredibly excited to bring this power to providers,” Stefansic said of the platform that integrates with existing EHR and population health software.
“What’s great about our technology,” he continued, “is it’s perfectly suited to treat complex health conditions, and behavioral health fits right in that wheelhouse.” He added the technology is not limited to those with a behavioral health diagnoses but also could be deployed for patients with any number of variables and co-morbid conditions that might complicate a treatment plan.
Stefansic noted there are ‘big tent’ considerations when treating for a set of conditions. However, he continued, when you drill down into the patient population under that big tent, there are many variables. For example, he said, a 65-year-old, widowed diabetic who lives alone in the country probably faces different challenges … and potentially different outcomes … than a 45-year-old diabetic mother of two in the inner city.
Doub likened the additional layer of machine learning on top of the predictive analytics component to the refinements over time in way-finding technology. “It’s like Waze vs. MapQuest,” he explained. “A decade ago, MapQuest told us the straightest route between two points, but it didn’t allow for variables like road closures or a traffic accident.” Similarly, he continued, both of the women in Stefansic’s example were trying to arrive at the same destination of optimal health, but their journeys would likely look different so the most effective approach to treating them might also vary.
Stefansic added, “We trust the clinician knows how to get from point A to point B, but sometimes you still use your GPS because you don’t know what the conditions will be like. We just want to give them more tools.”
RELATED LINKS:
Centerstone Research Institute
Artificial Intelligence Framework for Simulating Clinical Decision-Making (2013)
Multimedia Option:
Personalized Medicine: Applying Artificial Intelligence to Mental Healthcare
