Mastermind participant Christina Slemp of Community Health Systems explains how RCM leaders need to balance the potential of AI with the damage it can do if not managed properly.
AI is fast becoming an important tool in revenue cycle management and finance operations. But healthcare leaders need to understand the risks associated with a new technology.
For starters, anyone using AI needs to know the difference between AI and automated solutions. AI can generate information by studying patterns or inputs, while automated solutions require some manual programming to complete specific tasks.
“An AI agent can have what we call a ‘hallucination’ and produce something that’s wrong,” says Christina Slemp, vice president of revenue cycle for Community Health Systems and a participant in the HealthLeaders Mastermind program for AI in RCM and finance operations. “Typically, when you have [an automated solution], it’s that the process breaks, not that it’s producing the wrong data.”
CHS leverages vendor solutions alongside those developed in-house. The health system is currently testing an in-house AI solution to write appeals letters for denials.
“There’s a very specific niche in revenue cycle on the back end to be able to write an appeal and understand what the payers are looking for and how to combat that,” Slemp says. “We’re working on AI technology and agents that will go in, pull the clinical [information needed], and say ‘You should pay us. Here’s the clinical reason.’”
Healthcare leaders should also understand how their staff may react to the introduction of AI tools and be prepared for some pushback. One common misconception—at least at this stage of AI development—is that the technology is replacing people.According to Slemp, AI has helped CHS fill in the gaps due to workforce challenges, completing tasks that the health system might have problems doing because of a lack of staff.
AI tools are also being used to improve efficiency. Where it previously took staff over an hour to write an inpatient appeal, depending on length of stay and the number of records they had to review, the AI solution will allow them to become clinical reviewers alleviating the amount of time needed to sift through records and decreasing that appeal writing time by over 50 percent.
AI is still fairly new and the technology isn’t advanced enough to run unsupervised, nor should it be. CHS staff review all AI output, Slemp says. And while they’re not seeing many mistakes, there are instances where the tool needs more data to produce an acceptable result.
As staff identify where more information is needed, that information is fed back into the tool to improve its knowledge.
The increase in AI use also comes with concerns about cybersecurity. One way to potentially mitigate cybersecurity risks, Slemp says, is by developing AI solutions in-house and keeping all sensitive information within the organization. That, however, could increase costs.
“Once we got past the finance barrier of saying here’s the risk/reward in it, and then got to the next level of that,” Slemp says. “It’s the cybersecurity and the compliance piece that really has to be reviewed closely to ensure that there is no risk and cause delays."
The HealthLeaders Mastermind series is an exclusive series of calls and events with healthcare executives. This Virtual Nursing Mastermind series features ideas, solutions, and insights on excelling your virtual nursing program. Please join the community at our LinkedIn page.
To inquire about participating in an upcoming Mastermind series or attending a HealthLeaders Exchange event, email us at exchange@healthleadersmedia.com
Jasmyne Ray is the revenue cycle editor at HealthLeaders.
KEY TAKEAWAYS
One of the biggest risks with AI is not being able to differeniate it from automation.
Lack of transparency and communication with staff prior to the implementation of an AI solution can negatively affect its success.
Developing solutions in-house, thus keeping all sensitive informtion within the organization, is one way to mitigate cybersecurity risks.