At the 8th Annual Becker's Health IT + Digital Health + RCM Meeting, Andrew Lockhart, CEO of Fathom, led an executive roundtable discussion with hospital and health system leaders about the challenges and opportunities of automating medical coding with AI technology.
He described the current state of medical coding as slow, expensive, and inconsistent and highlighted how COVID-19 exacerbated issues like workforce shortages. While many healthcare leaders view AI as a solution to these long-standing challenges, some have questions around implementation approaches, accuracy levels, and how new processes or guidelines could affect reimbursement.
Here are three takeaways from the discussion.
Organizations must be grounded in their goals before embracing new tech
Healthcare is in an era of technological innovation. During the roundtable discussion, Lockhart shared data points that highlight healthcare's shift toward AI: nearly 60% of health system leaders already have an AI adoption strategy, or will have one established in the next 12 months, and about 40% of leaders view revenue cycle management as the number one priority for health IT investment, according to Bain & Co. and KLAS Research.
But tight budgets, staff shortages, and a pressing need for efficiency mean hospital and health system leaders need to demonstrate strong returns and determine the effects of automation on patient outcomes and revenue to justify technology investments.
Roundtable participants expressed the need for clear metrics to measure the success of new tools like automation and AI, and emphasized the importance of continued support from technology partners. Lockhart said Fathom's most successful hospital and health system partners have established goals and mechanisms for measurement that are flexible enough to embrace new technology.
"I think whether it's coding or any other AI tooling, really understanding what your goals are, what success looks like, and how you're going to measure it is ultimately how you're going to get a successful vendor partnership and show value," said Lockhart.
RCM workforce trends are posing challenges to innovation + improvements
Even before the pandemic, RCM teams were short-staffed, and many medical coders were nearing retirement age.
"We've had some clients where, just over the course of technology deployment, they've had enough people retire that they're having to pay people bonuses not to retire," Lockhart said.
Fast forward to today, and leaders are still grappling with persistent workforce shortages (this year, 63% cited revenue cycle staffing shortages) and are training newer, less experienced medical coders. Roundtable participants discussed how health systems have to consider the demographics of their staff because getting younger people into RCM can be difficult.
"Another challenge for coding staffing is, if you think about the areas where coders would learn and 'cut their teeth' — those are all widely automated today," Lockhart said. "There's not a lot of radiology coding to pick up, and most health systems in the next year or two are not going to have those simple cases that are good spots to learn."
Lockhart advised budding medical coders to get into a position where they're able to work with new technology like AI and automation. This makes their jobs more interesting and empowers them to work at the top of their credentials on the most challenging cases, such as troubleshooting unclear documentation, educating physicians, and working denials.
Evaluating ROI + the impact of automation requires a multifaceted approach
Integrating new technology is challenging on multiple levels, and participants identified three key ROI domains most important to them: better cash flow, expense reduction, and accelerated claims processing.
"Ultimately, what you want to move toward is a tool that's fully autonomous — you're not touching it, it's got a high level of accuracy, and it's highly scalable," Lockhart said.
He acknowledged not all organizations see volume fluctuations, but especially for those that do, it's critical to have reliable technology in place that can adapt to changes and produce clean claims faster. Determining how quickly cash will come in because of quick claims turnarounds is instrumental to demonstrating ROI.
Regarding the impact of automation, several leaders raised questions around trust and confidence in the technology due to general unfamiliarity with AI, as well as how it could affect reimbursement. To address these valid concerns, Lockhart shared his approach to validation: "Increasingly, what we've been doing with our clients is we'll code a month's worth of volume, overlay our results with theirs, and then give them back the results blinded and let their team pick which results they prefer. This allows them to extrapolate respective accuracies. Our clients have found they're surprised to see that the AI system outperforms their teams."
AI coding developments signal a promising future
Hospital and health systems are prioritizing technology investments, especially in RCM. While traditional medical coding sees slow turnarounds, high costs, inadequate accuracy, and too many inconsistencies, automation has emerged as the way forward. To accelerate transformation and address stubborn issues in staffing, costs, and inefficiencies, healthcare leaders need clear, intentional goals as they leverage AI-supported tools in the revenue cycle.