Skip to main content

3 Reasons an AI Beta Test Is Good for Revenue Cycle

Analysis  |  By Alexandra Wilson Pecci  
   August 11, 2020

Owensboro Health Regional Hospital has found financial success as a beta client for an AI tool that prioritizes patient encounters.

If a patient hospitalized for simple pneumonia has documented symptoms and treatments that point to a different diagnosis, a clinical documentation improvement (CDI) specialist can help sleuth out the truth.

Clues like a consistently high fever, antibiotic treatments, and chest x-ray results might point instead to a bacterial pneumonia, prompting the CDI specialist to query the physician and clarify the diagnosis.

This is an important step for ensuring both proper payment and an accurate patient record.

"We're able to bill a more appropriate, higher-based DRG," says Sharon Cooper, manager of clinical documentation and appeals at Owensboro Health Regional Hospital, in Owensboro, Kentucky. "We want to be reimbursed for the resources that we used for that patient, and we want the chart and the documentation to be accurate, correct, and complete. We want any other physician looking at that record somewhere in the future to see that that patient had a history of bacterial pneumonia, not just simple pneumonia."

But the process of finding such inconsistencies in the first place is arduous and inefficient if CDI specialists must comb through each record looking for them. 

That's why when Owensboro Health Regional Hospital had the opportunity to become a beta client for a tool that uses AI to prioritize patient encounters, based on which ones would have the largest CDI impact—both clinical and financial—they jumped onboard.

"We wanted something where technology would work for us instead of us manually having to do everything," Cooper says. The hospital sought a tool that would "triage" which cases should be reviewed first, as well as a tool that would interface with Epic.  

The tool

That tool is CDE One from Nuance, a cloud-based documentation guidance software that "uses technology and pulls the information, so we don't have to look through every single document in the record looking for specific things," she says.

Once those high-impact cases are prioritized, CDI specialists can use their clinical expertise to determine whether they should query physicians based on what they see in the documentation.

Using the tool has also streamlined the query process, allowing users to develop the query within the program and put it directly into Epic in a progress note for the physician, showing up in physicians' in-basket in real time for further clarification or to specify the diagnosis.

From there, the physician reviews, comments on, and adds an additional diagnosis, if applicable. The completed query with the physician's response is in the record forever.

"We wanted to be transparent," she says. "We want the documentation to support the diagnosis that we're asking them for. We wanted to be really transparent and leave the query as a permanent part of the record."

Being a beta client has helped Owensboro Health Regional Hospital realize significant cost and CDI benefits by allowing technology to help maximize CDI specialists' clinical expertise.

Here are three reasons being a beta client was good for Owensboro Health Regional Hospital.

1. Users can shape the tool

Cooper says being a beta client allowed the hospital to tailor the tool to its needs.

"We wanted a hand in how the program operates: What's best for us, what doesn't work for us, what does work," she says. "Being a beta client was really important to us."

She says the hospital worked with Nuance to make the tool more user-friendly and sensitive in terms of diagnoses and descriptions.

Owensboro Health Regional Hospital also changed other things about the tool such as the wording on certain aspects of the tool and reducing the number of clicks needed to perform certain tasks.

The hospital can also set parameters for which cases and terms are triaged.

"Whatever is important for our facility, that's what they set the triage at," Cooper says.

To make changes to the tool, beta users note anything that could be easier or more user friendly. Those notes and requests are passed along to Nuance.

2. There can be early success before the contract

Owensboro Health Regional Hospital has been a beta user since January 2019. After the beta ends in October, it will enter a three-year contract for the tool at a reduced rate. But the hospital has already realized financial success.

Since the start of the beta, the hospital has recognized a 3.57% increase in case mix index (CMI), which resulted in $1,671,159 in additional reimbursement. A higher CMI means that there's a more complex case load and more use of resources.

Cooper says the hospital was "able to review the records that had the biggest impact … without having to look through every single record to see if there was an opportunity for impact or not," which in turn, resulted in increased productivity.

3. The hospital can think about expansion before the end of the beta test

Unlike some forms of automation, which can result in reducing the number of staffers needed, Cooper hopes to hire more CDI specialists as the hospital expands its use of the tool.

So far, staffing has remained the same, but the employees are more productive. They use the tool not only for Medicare and Medicaid records, but they also have added mortality reviews since the beginning of the beta.

"We were able to do [this] with our current staffing," Cooper says.

In the future, the hospital would like to expand the tool's work to all payers.

"Instead of just having our Medicare and Medicaid records having complete and accurate documentation, we would like for all of our accounts to have that," she says.

Alexandra Wilson Pecci is an editor for HealthLeaders.

Get the latest on healthcare leadership in your inbox.