Three questions with Jason Burke: Connecting the dots across the revenue cycle with AI

I sat down with Jason Burke, vice president of revenue cycle solutions at Solventum, to talk about shifting the revenue cycle closer to the point of care, taking a layered approach to autonomous coding, preventing potential denials at the source and predicting the future of healthcare analytics.

What are the biggest challenges revenue cycle leaders are facing, and how can they set up their teams for success in the current environment?

A lot of organizations got through COVID-19 and felt like healthcare would return to the way it was before. In reality, only some organizations have, and they’re likely the for-profit health systems; those that have not are most likely not-for-profit. Struggling with costs and margins isn’t new to healthcare, but in my 24 years in the industry I have not seen this level of urgency to address these issues. Organizations are making drastic decisions, with some eliminating service lines or even closing hospitals or clinics to reduce costs. Healthcare organizations are also challenged to deliver quality care at the same standards they were at before COVID-19. Improving care quality while cutting costs is a very difficult thing to do.

So, what can revenue cycle leaders do in this environment? The first thing is having the right people who are trained to deliver at the top of their license in their area of expertise. The second is making sure information is accurate from documentation to reimbursement. Artificial intelligence (AI) and automation can play key roles here, both in terms of helping staff make the most of their time and supporting accuracy. It is critical to ensure that these technologies are carefully vetted, and workflows are adjusted to take full advantage of generative AI’s capabilities. It is also extremely important to vet security standards and risks associated with the technology.

I think leaders will need to shift their mindset around how the revenue cycle is managed and where tasks get done. The revenue cycle is quickly moving as close as possible to the point of care because documentation can be done more accurately and efficiently when patients are present than after they leave. 

How can automation help shift the revenue cycle closer to the point of care?

Automation can speed up administrative processes, enabling tasks to happen sooner. For instance, ambient intelligence makes clinical documentation a byproduct of the patient-physician interaction, and computer-assisted physician documentation (CAPD) prompts clinicians to proactively address issues of accuracy or completeness that could otherwise lead to a denial down the line.

When it comes to traditional revenue cycle tasks, the innovation everyone seems to be looking for is autonomous coding. But coding with AI alone is not going to produce the most accurate codes because large language and generative AI models mimic what coders do – and there's always some variation or inconsistencies between coders. That’s why, as we’ve built our own large language AI model for autonomous coding, we’ve taken a layered approach. We start with AI, add coding guideline rules to create compliant code sets, layer on rules unique to the organization and top it off with payer rules. Together, those pieces allow us to keep honing the right codes and enable automation with confidence, accuracy and compliance.

Denials management is another topic on many healthcare leaders’ minds. I’ve heard from CFOs and industry organizations that the only way to resolve denials in a meaningful way is to proactively address them within workflows. That’s why I’m really excited about our new revenue integrity solution – which we announced earlier today – because it’s essentially a denials avoidance program. It helps identify what types of denials are occurring and where they’re happening in the organization, and then uses predictive modeling to generate a score based on the likelihood, severity and potential reason for denial. Based on these scores, organizations can build prioritization workflows to prompt clinical documentation integrity (CDI) specialists, coders and clinicians to take action to handle potential issues before the claim is submitted. For the people doing these processes, preventing denials becomes part of their everyday workflow.

Implementing automation throughout the revenue cycle presents opportunities to connect the dots between payment and operational execution. For instance, if we predict a denial will occur for a code set that has been automated, we can build edits into the coding, CDI, and utilization and care management workflow to solve the issue permanently. Predictive systems can work hand in hand to help identify and address problems. We’re collaborating with other companies that have technologies that complement our strategies and workflows to provide a complete automated solution.

Prior to leading the revenue cycle solutions team, you led the data informatics and strategic business development teams. How does that background influence how you think about the applications of predictive analytics across healthcare, and what do you see as the next frontier?

My business partner experience really opened my eyes to the opportunities of leveraging the strengths of other companies and not always having a “build it yourself” mentality. There are times when working with other companies allows you to fast track product delivery and results.  We don’t always have to build everything from the ground up, but can look at what’s available in the market and decide when to collaborate with someone to move things along. In fact, we've done that with our autonomous coding and revenue integrity solutions – we’ve combined our own AI models and technology with the functionality and expertise of third parties. Collaborations can be challenging to navigate because there are expectations on both sides, but when done strategically, it can help us stay ahead in the market and meet customers’ needs more quickly.

As for the next frontier of analytics, I think we’ll soon see an evolution from the traditional approach of running reports and looking back at historical data, to proactive analytics where actionable insights are delivered in near real time to drive immediate change. For instance, providing physicians or nurses with data and insights at the point of care based on past history and benchmarks can help them treat patients in the best way possible. AI models and the speed with which we can process data will allow people to make faster, better decisions across the continuum of care – and that’s where healthcare needs to go.

Jason Burke is the vice president of revenue cycle solutions at Solventum.

Angela Haile-Selassie is a marketing communications specialist at Solventum.