Villages Health is using natural language tech to capture patient info, save money

Photo: Al David Sacks/Getty Images

Despite investments in analytics, key patient attributes are often buried in a facility's electronic health record.This was the challenge facing The Villages Health (TVH).

The 75-physician multi-specialty group was built at the request of The Villages active community living in central Florida, TVH has been aggressive in operating under value-based care models, including full-risk capitation since its inception in 2010. To support this model, TVH has used an EMR, invested in analytics, put together a coding team and conducted extensive training on documentation quality.

Clinical leadership believed key patient attributes were not fully captured as structured information but rather were buried in the EMR's text. 

According to TVH Chief Medical Officer Jeffrey Lowenkron, this negatively affects care efficiency, clinical efficacy and revenue capture. Important diagnoses weren't being managed, and the organization was missing opportunities to close quality care gaps.

"TVH takes full financial risk for a large segment of its patient panel, and has an inherent deep commitment to the overall health The Villages community writ large," said Lowenkron. "As such, TVH is committed to quality outcomes as determined by our payer partners, as well as ensuring long-term affordable access to our top-notch delivery system. 

"Achieving these objectives of high-quality, full documented and coded care is a lot of work for the care teams," he said. "We ask them to continually capture new revenue opportunities already evident in the chart, ensure HEDIS measures are addressed when they do, and more. We needed to make this easier for them."

Using Clinical Natural Language Understanding (cNLU), TVH was able to identify previously hard-to-access clinical attributes in at least 15% of its patient population. In an initial pilot, TVH provided care teams with more complete problem lists with conditions identified for evaluation and management. The system was able to generate $2.5 million in net new revenue as the clinical insights more fully captured patients' underlying disease burdens.

"The chart is read looking to answer very specific questions such as whether there were anatomic or functional findings in the chart of which we were unaware," said Lowenkron. "Equally important was to assure we did not ask team members to open charts and look when the opportunity was not there."

That early pilot resulted in a nine-fold return on investment, and that return is much larger now, although not explicitly calculated, said Lowenkron. 

"On the workflow side, we are saving team members a few minutes per chart for wellness visits," he said. "We are still learning and anticipate more gains as our collaboration continues."

Lowenkron will expand on TVH's journey on Tuesday, April 18 at 12:30 p.m. CT during "Natural Language Understanding for VBC: Uncover Revenue and Disease Burdens," at the McCormick Center, South Building, in Room S104 during the HIMSS23 annual conference in Chicago.
 

Twitter: @JELagasse
Email the writer: Jeff.Lagasse@himssmedia.com