Hospitals, Health IT

How one hospital used Ayasdi’s AI to improve clinical outcomes

St. Augustine, Florida-based Flagler Hospital used Ayasdi’s artificial intelligence technology to take on its clinical variation problem, specifically within the realms of pneumonia and sepsis.

Flagler Hospital, a 335-bed hospital in St. Augustine, Florida, had a problem.

“It’s the same story that every hospital in the country has faced in the last 50 years: trying to tackle clinical variation,” Dr. Michael Sanders, the chief medical informatics officer of the hospital, said in a recent phone interview.  “We tried to do the best we could in tackling this problem of clinical variation.”

But none of its tactics for addressing the issue were yielding great results. It was time for a change.

About a year and a half ago, Sanders stumbled across an article on how Mercy in St. Louis partnered with a company called Ayasdi to work on improving clinical outcomes. Though he had never heard of the Silicon Valley business, Sanders gave the team a call and found they had an insightful approach to solving the issue.

Ayasdi’s AI-powered clinical variation management tool constructs possible care process models for acute and non-acute conditions. It can pull data from a variety of sources (such as the EMR), find patterns and discern the best treatment pathway for certain patient populations.

By the end of February, Flagler launched a pilot of Ayasdi’s technology. Its specific goal was to focus on pneumonia.

“It was a good opportunity to learn the tool and get our arms around the process and what we would need to do,” Sanders said.

The hospital went on to complete the pneumonia pilot and sign a contract with the California company. By the time it was signed, Flagler had moved on to tackling sepsis.

By implementing Ayasdi’s tool, Flagler Hospital has realized various positive outcomes. In the pneumonia realm, it’s saved money, reduced the length of stay and reduced readmission rates. It has also lowered the risk of dying from sepsis, Sanders noted.

Looking ahead, the St. Augustine hospital wants to use Ayasdi’s approach to get to work on a variety of other care paths, including diabetes, alcohol abuse, drug abuse, cardiac bypass, COPD, stroke, GI problems and pulmonary embolism.

When Sanders initially began exploring Ayasdi’s technique, he worried about whether a smaller hospital like Flagler could handle it. The company’s solution seemed more tailored to a large health system.

But the results prove otherwise, and the hospital plans to continue its work with Ayasdi.

“We are excited to support Flagler as it applies machine learning to the challenges of clinical variation,” Ayasdi COO Gary Hagmueller said in a statement. “In particular, we are pleased to see our technology get adopted by a community hospital, albeit one of the leading community hospitals in the country.”

Photo: Madmaxer, Getty Images