Artificial intelligence is all the rage right now — and several health system revenue cycle management leaders are integrating the technology into their workflows.
Becker's reached out to seven RCM executives to get their thoughts on AI:
Anne Robertucci. Vice President of Clinical Revenue Cycle of Prisma Health (Greenville, S.C.): We have moved to use autonomous coding with our emergency department population and seen great results. I think that AI will help simplify many of the more complex RCM processes that, at the moment, are still very time-consuming and resource-intensive. We've seen things like robotic process automation come in and help automate simple RCM tasks, but AI must be capable of understanding clinical language in medical records to help with more complex processes.
At Prisma, we've been exploring the use of AI in the form of autonomous medical coding. We're currently working with a company that has developed sophisticated AI technology which can fully understand the clinical information in medical records and accurately assign medical charge codes in a matter of seconds. As you can imagine, this has had a huge impact on our medical coding processes by speeding up payment cycles, improving coding quality, and reducing costs. All in all, our success with autonomous coding makes me pretty excited about the impact that AI will have in RCM down the line.
Laura Calkins. Vice President of RCM of Presbyterian Healthcare Services (Albuquerque, N.M.): AI and robotic process automation can be valuable tools for healthcare revenue cycle teams to lower administrative costs. These technologies can be used to optimize net revenue and improve the overall patient financial experience. For example, Presbyterian uses "claim statusing," where a bot pings insurance carriers regarding outstanding insurance claims. This automated process allows employees to focus on higher-value work and use their time more effectively.
AI-based tools do require careful consideration, monitoring and assessment to avoid potential operational issues, but they can play an important role in efficiently managing costs and time.
Michael Mercurio. Vice President of Physician Revenue Services of Mass General Brigham (Boston): I think AI is going to have an enormous impact on healthcare RCM. In fact, AI is not only already impacting healthcare (the startups in this space are almost too numerous to count), but it most definitely will have a transformative impact on RCM in the coming months and years. The two areas where I think the most direct impact will be felt are in the coding/billing space and in predictive analytics.
As we have seen across our own RCM teams at Mass General Brigham, AI is primarily automating repetitive and time-consuming tasks involved in both coding and billing; in some cases, it is even doing the work of more experienced coders. AI's benefits are evident in the speed and accuracy with which CPTs, diagnosis codes and modifiers are generated, thereby relieving the tedious burden our healthcare system imposes on our most expensive and scarce resources: providers. The result for us has been a far more efficient revenue cycle, a shorter time to payment, fewer payer denials, and a reduced cost to collect (in addition to a reduced burden for our providers, coders and billers).
There is also tremendous opportunity for predictive analytics in this space. Given AI tools can consume vast amounts of data to extract features and identify patterns, there is a huge opportunity to leverage AI to help us focus on the work for which we are most likely going to get paid. We have to be careful that we don't get caught in a catch-22 in this scenario, but the power of AI in this space is undeniable.
Olaf Faeskorn. Vice President of Revenue Cycle of South Georgia Medical Center (Valdosta): We have conversations here at South Georgia Medical Center, with our IT folks involved to really understand the security aspect of this. How can we work with this in the future without putting our data at risk?
There's clearly awareness around this in the industry and here locally, on a personal level, I look at these things certainly through a skeptical lens as what kind of damage this can do to our safety and security. On the other hand, we got to embrace it to some degree because it's not going away. We're actively actually looking into this right now. I have, in the past, looked at automation and, to some degree, limited machine learning in the coding realm.
Some of the bigger players have toyed around with it or have done it in the past. At the local level, some vendors have helped with infusion building automation. It is a good target because it is a rules-based system. You can build automation around it based on the rather complex rules, but they can be captured. I think with AI we will see more dynamic systems that are more truly looking at different scenarios and preferences and the actual clinical data behind it.
We're actively pursuing a coding AI solution to test it out and really see what it is capable of. Does it fit our needs and what does it really mean for us? We're going to be running a limited trial to see if this is something that we want to put into our regular workflows.
Paul Capello. Revenue Cycle Project Manager of Shriner's Children's Hospital (Tampa, Fla.): It's not a lot of people out there that have married AI to healthcare just yet. I know there are probably the Epics and other larger players of the world that have teams working on it. In order to make it make sense from a patient perspective, there are three criteria that I use: patient safety, financial return on investment and regulatory requirements. Healthcare is so regulatory requirement-heavy that AI may be able to help in this space.
Longer term, I think these players will come at it from a methodical type of approach. The analogy I use is electric vehicles in the automotive industry; everybody wants to get there, but right now, the cost is somewhat prohibitive to them. It is going to be something with healthcare, where AI may help in some capacity, but right now it is probably too cost-prohibitive and too unknown.
Seth Katz. Vice President of Health Information Management and Revenue Cycle of University Health KC (Kansas City, Mo.): Automation and machine learning holds the promise to help drive the front-, middle- and back-end units of the revenue cycle forward by taking over the mundane and repetitive tasks, allowing staff to work at the top of their skill set on what matters most. It has the potential to help with staffing shortages and allow organizations to be more nimble in adapting to new rules and regulations.
Though the technology isn't there yet, I have no doubt it'll continue to improve over the coming years, and it's something all current and future leaders need to become versed in and understand how the technology works and what use cases they could apply it to within their organization.
Todd Craghead. Vice President of Finance and Revenue Cycle of Intermountain Health (Salt Lake City): In my opinion, AI will continue to have a significant impact within the RCM space by automating more and more tasks that have traditionally been performed by the RCM workforce. I expect that this will reduce the workforce, reduce operating costs, and improve outcomes. Many providers/payers are already advancing their work together to leverage this technology in order to reduce friction for the patient/member.
Targeted workflows that were often performed by individuals are accelerating their move to AI, given that many of the decision trees can be replicated within this technology. Some of the more visible workflows impacted by AI include authorization workflows, payment posting/processing workflows, and claims status/adjudication. In addition, patient self-service payment tools continue to introduce increasing levels of artificial intelligence, which further reduces the need for patients to interact directly with staff and allows them to leverage technology in order to complete their interaction with the system.