AI is gaining ground in the revenue cycle faster than the strategies needed to govern and deploy it, a recent study from UPMC and KLAS Research found. ‘
“Interviewed leaders said the biggest barriers to executing an AI strategy are not a lack of interest or perceived value. Instead, many organizations struggle to create the capacity, structure, and discipline needed to implement AI responsibly,” the report stated.
That spurs the thought that maybe it’s time for healthcare revenue cycle executives to consider the questions they aren’t asking but should be.
Three hospital officials with experience in revenue cycle AI responded to the following: “What is one question healthcare revenue cycle leaders should be asking about AI?”
What are our measures for success?
“How will we know six or 12 months after implementation that an AI solution is still delivering the operational, financial and workforce outcomes we expected, and who is accountable for proving it?” said Caris K. Newton, CRCR, MS, CSM, director, revenue cycle strategic partnerships and automation, revenue cycle shared services for Corewell Health, Grand Rapids, Michigan. “Ultimately, the measure of success is not how many AI solutions we deploy. It is whether we can clearly demonstrate, over time, that they are improving outcomes, reducing burden and creating measurable value for the organization.”
Are we leading with the business case for AI?
“AI can create the greatest measurable impact across the revenue cycle when it is applied to areas where there are significant volumes, repetitive work, avoidable revenue leakage and opportunities to improve decision-making,” said Sheila Augustine, FHFMA, MHA, executive director of revenue cycle operations for Nebraska Medicine in Omaha.. “Key opportunities include denial prevention and management, coding and documentation, charge capture, claims optimization, accounts receivable prioritization and automating manual administrative processes.
“To ensure we are investing in the right opportunities, revenue cycle leaders should start with the business problem — not the technology,” she said. “Each AI opportunity should have a clearly defined baseline, measurable financial or operational outcome and a way to demonstrate ROI. This includes evaluating whether AI can reduce denials, accelerate cash, improve accuracy, reduce cost to collect, increase staff capacity or improve the patient experience.
“Ultimately, the question should not be, ‘Where can we use AI?’ It should be, ‘Where can AI create a measurable, sustainable improvement in our revenue cycle that we could not achieve as effectively through traditional processes alone?’” she said.
How do we become an AI-enabled and human-centered team?
“It will be harder than ever when weighing what work should be automated and [performed] utilizing new tech tools and what work needs to remain for human judgment,” said Nikki Harper, division chair, revenue cycle – analytics automation and diversified revenue for Mayo Clinic, whose office is based in St. Louis. “And between each are the decisions on how we as leaders ensure that AI is governed and reviewed continually to keep data integrity and decision-making sound within our organizations.
“AI is an enabler to help us create more capacity than ever before and should not be treated as a total solution that replaces; it instead enhances our capabilities,” she said. “AI integration is vital to our strategic and operational models; however, how and when we use it versus having a human-in-the-loop approach is key.”
What UPMC and KLAS researchers learned
The AI study, conducted by the Center for Connected Medicine at UPMC and KLAS Research, found that AI adoption has moved so quickly in healthcare that important questions about governance, testing, validation, data quality and AI strategy may not be getting enough attention.
Among the 27 healthcare leaders interviewed, 59% (16) said their AI strategy is still developing. Another 30% (8) described their AI strategy as “established,” while just one respondent representing almost 4% of the total said their organization’s AI strategy is advanced and another was “not sure.” The leaders spanned health systems, hospitals, clinics and ambulatory care organizations.
