Case Study

How Flagler Hospital Fought Pneumonia with AI and CVR

20 Minute Read

Key Takeaways

The challenge

Hospital-acquired conditions (HACs) like pneumonia are difficult to detect and costly to treat. It’s common for data-driven organizations to address this challenge by using data to identify patients at risk of deterioration and then to implement a standard treatment plan. This retroactive approach is starving systems of valuable resources because it targets patients at risk of deterioration instead of preventing deterioration in the first place.

The organization

Flagler Hospital is a private not-for-profit community hospital in St. Augustine, Florida. The 335-bed acute care center has 400 physicians on staff.

The approach

Flagler took a unique approach, using artificial intelligence (AI) to proactively identify care paths that prevent HACs. Partnering with Symphony AyasdiAI, a machine intelligence software company, helped Flagler take advantage of the readily available data that sits in electronic health records (EHRs). Together, Flagler and Symphony AyasdiAI implemented a data-driven solution using AI to identify the best care paths, develop new order sets, and reduce care variation. Flagler started with pneumonia and has since moved on to cover 11 other conditions.

The result

Flagler advanced beyond the pneumonia pilot stage to successfully implement standardized pneumonia order sets. Within the first few months of implementation, length of stay (LOS), mortality, and costs decreased.

 

The Approach

How Flagler prevented hospital-acquired pneumonia with AI and standardized order sets

Instead of using AI to detect patients with HACs, Flagler used AI to proactively generate care paths that protect patients from developing these conditions. Flagler’s IT and physician teams then worked together to match the care paths to a standardized order set. A successful pneumonia pilot showed improved care outcomes and decreased costs, encouraging Flagler to permanently implement standardized order sets for pneumonia.

 

The Four Elements

After piloting the AI application, Flagler identified four elements that helped them successfully implement AI-generated standardized order sets:

  • Element

    Pitch executives on the benefits of success and failure

    Read More Collapse
  • Element

    Solve problems alongside your vendor

    Read More Collapse
  • Element

    Learn from your own data

    Read More Collapse
  • Element

    Encourage physicians to learn from their peers

    Read More Collapse
 

Results

Standardized pneumonia order sets reduced LOS, mortality, and costs

During a two-week pilot, the AI solution helped Flagler determine the best order set for pneumonia. Within nine months of implementation, Flagler decreased LOS, rate of admission, and mortality rate, and saved $1.06 million.

1.5

Days reduction in LOS

0.8%

Rate of admission, down from 8%

0%

Mortality rate, down from 4%

$1.06M

     Saved

The same solution is currently being applied to 11 other conditions including sepsis, chronic obstructive pulmonary disease (COPD), coronary artery bypass grafting (CABG), and hip and knee replacements. Flagler credits the AI solution for saving 54 lives during the first year of operation.

X
Cookies help us improve your website experience. By using our website, you agree to our use of cookies.