Two or three years ago, investors were telling startups that clinical coding would be fully automated by LLMs within a year – which didn’t quite pan out.
The healthcare industry’s revenue cycle has proven stubbornly resistant to simple technology solutions, but AI can still make real progress, according to Lee Kupferman, co-CEO of R1the innovation laboratory of.
AI has the most impact when it is used to handle large, simple work, freeing up humans for cases that actually require their expertise, Kupferman noted in an interview last week at HFMAannual conference in National Harbor, Maryland.
For example, a simple inpatient encounter where a patient presents for a known procedure with no complications is the kind of case where 50 coders would all arrive at the exact same answer — that’s where AI should just work, he said.
However, most AI models still struggle to handle complex situations that involve longer documentation and more varied payer rules. Kupferman thinks the current goal is to figure out how to route good work to AI and reserve human efforts for gray areas.
“You can benefit from [AI] tools across the entire revenue cycle, as long as you have the right guardrails in place and you’re honest about where it’s working well and where there’s still a way to go,” he noted.
Part of what makes the revenue cycle resilient to AI is that the healthcare payment system itself is so deeply fragmented, Kupferman added.
There are hundreds of vendors in the healthcare revenue cycle, but most of them sell point solutions that don’t communicate with each other, he pointed out.
Health system coding teams often operate in near-total isolation from the prior authorization team, meaning a denial that could have been caught up front triggers weeks of rework downstream, Kupferman explained.
He sees this fragmentation as one of the biggest obstacles to AI realizing its promises. Kupferman said these tools need to be connected for efficiencies to actually be realized.
The good news, however, is that the environment could change. Kupferman noted that while payers and providers have historically rejected the idea of a more collaborative, AI-driven revenue cycle, they are starting to show greater willingness to work together to improve the payment process.
“Everyone is violently in agreement about what the problem is – they’re just trying to figure out the best way to solve it,” he said.
Photo: Uchar, Getty Images