The New Revenue Cycle Management
The leaders of U.S. patient care organizations, whose incomes were dramatically fragilized by the shutdown of nearly all elective medical procedures during the first and second quarters of 2020, as the Centers for Medicare and Medicaid Services (CMS) worked to protect clinicians, staff members, and patients from infection during the early months of the COVID-19 pandemic, have been working their way back to financial stability since then. And it is in that context that revenue cycle management (RCM) has once again come into focus as a tool critical to ensuring viable operating margins for hospitals, medical groups, and health systems nationwide.
The good news? After decades of the advancing evolution of core RCM operations through such methods as business process automation (BPA)—also known as robotic process automation (RPA), which have been successfully incorporated into core RCM processes around managing all claims processes—a new phenomenon is emerging that could prove very helpful: the leveraging of machine learning- and artificial intelligence (AI)-based technologies to achieve what is being called predictive denials management. Essentially, this involves the development of algorithms based on data analytics, that can trigger interventions based on anticipated insurance claims denials.
“We’ve spent more than a decade using BPA/RPA technologies to automate what had largely been manual work, since payers pushed all the administrative activity onto portals,” says Jeffrey Porter, vice president of revenue cycle and chief revenue cycle officer at the 40-plus-facility, Pittsburgh-based UPMC health system. But within the next two years, Porter says, he and his team will be developing machine learning-based algorithms to support the denials management and claims management work of their team.
Paymon Farazi, chief product officer at the Eden Prairie, Minn.-based Optum, a data analytics company, agrees. “The business process services element is fairly far along; and simply building if/then rules into the claims process is one thing,” he says. “What’s more complicated is when you build in payer rules. You go to a payer and you ask them, ‘How do you deny claims?’ And either that payer shares proactively with the provider, or your team at the provider organization builds out machine-learning techniques to game out situations. That last type of activity is the leading edge, and is a huge leap.”
The percentage of patient care organizations even in the early stages of leveraging machine learning and AI tools to fuel predictive analytics in this way is still tiny, says Farazi’s colleague Doug Hires, who is COO of Optum’s provider market segment. “The intelligence component, we’ll find, is not very prevalent at all—maybe some of the more sophisticated and large healthcare delivery systems are using analytics themselves, but it’s not necessarily yet an automated component running as part of the daily system,” Hires says. “So it ends up being surveillance, intervention and correction, in cycles. So you can identify, for instance, a bulk of claims based on a denial code that you get, and then start doing data searches to try to find them, and then remediate.”
Meanwhile, says Farazi, “There’s a mirror-image process on the payer side. The payer is probably a little bit further ahead in terms of using advanced techniques to generate denials. But the bleeding edge is to share the outcome of that with the provider.” Only a few health plans have done so to date, he notes; but that doesn’t mean that more won’t be doing so in the near future.
Looking at the broader picture around all of this, “It’s really about efficiency,” says James McHugh, managing director of revenue cycle consulting at the Naperville, Ill.-based Impact Advisors consulting firm. “Revenue cycle is all about workforce efficiency now; the revenue cycle is an assembly line. Now, you have these great EHRs [electronic health records] in place; but even then, with all that you can do with those systems, it’s still about leveraging the system you have. And COVID has accelerated the workforce management piece, including driving efficiency and the work-by-exception process. And I don’t think these are new things, but there are more and more tools and technology available for us to work with.”
Beyond BPA/RPA, McHugh says, “A lot about truly advanced analytics, or AI, is aspirational, but will become reality, and involves denials management. The best way of handling denials is to prevent them from occurring in the first place. If we know that Health Plan X is always going to deny endoscopy for reason Y, using advanced analytics and AI, we can prevent those denials in advance by identifying what the cause would be; that is predictive denials management.” But, he says, fewer than 1 percent of U.S. patient care organizations are yet using predictive analytics or AI/machine learning.
Meanwhile, what are the keys to success with AI- and machine learning-based predictive denials management/predictive analytics for revenue cycle management? Industry leaders see two. “The foundation to this is a really good analytics platform,” says Impact Advisors’ McHugh. “If you don’t have a good analytics program, you’re going to struggle to do this. And it’s funny, because when RPA was first becoming hot, the IT people would say, that’s operational, you can build that, so RCM would take it on. So because of that, there wasn’t an enterprise-wide approach, and as a result, RPA evolved in silos. It’s really important to have a center of excellence for analytics and a center of excellence for business process automation. Those are two very different, distinct things. IT doesn’t do everything, but they’re a coordinating entity in this. And you need to prevent silos. Also, a lot of health systems are outsourcing too much to their technology vendors around this. You need the control.”
UPMC’s Porter says very firmly, “First, you have to invest in the people—you need actual data scientists, who have a high technical capability,” he emphasizes. “But those people are difficult to get, and are expensive.”
So, clearly, there are prerequisites involved: a strong data analytics foundation, and actual data scientists. And, everyone agrees, there’s no question that it will be a major challenge to acquire the data scientists with the expertise to make AI- and machine learning-based analytics work.
Once the AI- and machine learning-based tools are implemented, Porter says, he and his colleagues will be engaged not only in predictive denials management, but also in developing models to predict patient payment behavior. “We’re looking at our patient segmentation, to understand patient payment patterns and behaviors—who pays and how, for example, on the first bill, second bill, etc. We want to create a better patient experience around this.”
Clearly, there are challenges involved here; but, everyone agrees, this is one area in which the potential of technology to fuel a new kind of revenue cycle management is very real; and all agree that a significant number of patient care organizations will be engaged in the most advanced forms of RCM within the next two to three years.
