As everyone in healthcare knows, the COVID-19 pandemic turned out to be devastating not only in terms of the number of lives lost and the illness that it caused, as well as the pressure on clinicians and other healthcare workers; it also created an unanticipated rearrangement of the job recruitment nationwide.
As senior leaders in hospitals, clinics, and health systems nationwide sent workers home to work remotely whenever possible, because of the intensive danger of infections in the first months of the pandemic, an unexpected thing happened: suddenly, health systems were competing for non-clinician workers not just with local competitor organizations, but nationwide—and even across other industries. That became particularly true in such areas as revenue cycle management (RCM), where already a certain percentage of staff were working remotely. Very quickly, the job recruitment market in RCM became almost entirely nationalized, meaning that RCM leaders at a hospital in Buffalo were now competing for workers with hospitals in Cleveland, Atlanta, Albuquerque, and beyond.
At the same time, hospital, medical group and health system leaders were already moving forward to advance their leveraging of robotic process automation (RPA) tools, automating many of the most routine and non-conceptual tasks in the revenue cycle management area, while beginning to dip their collective big toe into the swimming pool of actual machine learning (ML) and artificial intelligence (AI). So what’s been happening since then? Has the leveraging of automation begun to enhance operations and balance out long-term staffing challenges?
The pandemic has definitely nationalized the recruiting market. It created a larger pool of competition, but at the same time, also a larger pool of opportunity. So, availability of resources, competition for that availability. The pool of available resources has significantly improved, based on the nationalization and the willingness to shift employees into a remote or a hybrid environment. The challenge is that everybody sees the same thing and is pursuing the same goals. So how you go about this and how quickly you can move, matter a lot.
“Automation is absolutely turning out to be a contributor to how you can mitigate some of your staffing challenges; in fact, it factors in, in a couple of different ways,” says Jim Akimchuk, a managing director at the Naperville, Ill.-based Impact Advisors consulting firm. “There is a level of what we do that is repetitive and easily automated, based on cause-effect and action-reaction types of scenarios. Where it’s really helped in that space has been in the area of more closely directing our revenue cycle folks to a focused area. If I throw 1,000 marbles in with one yellow one, it’s harder to find that one marble, than if 200.” Automation, he says, is advancing such work by exception, which is still early on in its development.
And the staffing shortages continue to be very real, says Sue McBride, principal in the Advisory Services division at the Charlotte, N.C.-based Premier Inc., and the association’s revenue cycle practice lead. “COVID really set off alarms around staffing shortages,” the Pittsburgh, Pa-based McBride affirms. “It allowed people to work remotely; and that bigger pool of people affected recruitment. So, for the first time in revenue cycle, we’ve had clients tell us that they’re having a difficult time attracting people; they’re competing against retail or fast food; that was surprising. So it became more of a focus to get automation in there.”
There is a nuance in all this, McBride notes: while on one hand, salaries are rising for all sorts of positions, including coders, utilization management nurses, and call center workers, she says, “It’s good and bad, because while there’s competition to attract those people, but it does also open up the talent pool.”
In any case, Craig McKasson, Premier’s senior vice president and chief financial officer, notes that “What’s interesting is that the pendulum swinging a little. That nationalization is happening in healthcare and also in other industries. But health system leaders are starting to want to bring people back on-site. So it will be interesting to see what happens. Tech companies used to say, everyone will work virtually; now, they’re creating hybrid systems.”
Those working in the trenches are experiencing the challenges first-hand, among them Amy Trogdon, revenue cycle director at Integris Health, an eight-hospital integrated health system based in Oklahoma City. While on one hand, Trogdon reports, “We have not had an issue with staffing of our remote workforce; we’ve actually seen a decrease in our turnover rate; we have very little turnover among our remote workers. With our on-site staff, we are seeing staffing challenges. First, just finding people who want to work in healthcare is an issue. We have competitors outside of the healthcare industry that can afford to pay more, with benefits. So trying to retain high-quality, knowledgeable individuals is difficult. We’ve put together a good retention plan. And we’ve cut turnover in our remote workforce by 50 percent.”
Paul Shorrosh, CEO of AccuReg, a Mobile, Ala.-based solutions provider in the revenue cycle area, says, that, “Nationwide, I think the field is still in crisis. There’s so much competition for these employees, and health systems are not doing well financially; profit margins are low. So it’s hard to attract and retain good-quality employees. It’s true across health systems; it’s definitely a crisis, we hear it everywhere we go. And the answer to that is technology: automate everything you can; and we’re so data-intensive in healthcare that there’s a lot of opportunity there. And second is digital; allow your employees to focus on high-value work.”
How to match people and processes?
Richard L. Gundling, senior vice president in the professional practice at the Downers Grove, Ill.-based HFMA (Healthcare Financial Management Association), says that, in his view, the great opportunity/challenge of the moment is around “broader thinking about how to use the human staff, and matching it up with AI and robotic processes, so that purely transactional processes like payment processes can be automated. Meanwhile, patients are responsible for more of their healthcare costs, so there needs to be more advisement, as well as time and energy devoted to collections.”
In that regard, what are the smart CFOs and senior revenue cycle management leaders doing right now? “Among other things, they’re looking at regulatory finance and audits, looking at the No Surprises Act, and updating compliance processes, and training their staffs on these regulatory complexities,” Gundling says. “And they’re looking at the different payer rules, and at where they’re experiencing high claims denials rates.” Leaders need to be leveraging technology to get into the weeds, he says, and analyze such areas as “payer-provider coding disagreements and medical necessity judgments, and even prompting the clinician at the bedside—is there a comorbidity that needs to be documented? Or an alert telling the physician a claim will be denied unless something isn’t done?” Or they might need to be prompted to document the fact that they’ve done a psychosocial assessment on a patient.
The key point with regard to all the process improvement work that can be technology-facilitated, Gundling says, is that “Those changes require a lot of training; so you have to make sure you’re dedicating enough resources to staff compliance. You have to avoid penalties, that would be a cost; avoid public condemnation. And make sure you’re documenting comprehensively, so you get paid. That requires staff. You can have staff working on that instead of purely mechanical processes.”
Asked the same question, McBride says, “How are successful organizations leveraging automation? They’re leveraging certain types of solutions in the revenue integrity/coding space, for example: AI solutions that can look at massive numbers of accounts and immediately spit out the variance. Some solutions can machine-learn so you don’t keep making the same mistakes over and over. Meanwhile, in the billing, collections, and patient services area, automating cash posting, and pinging Medicare for claims status, we see automation increasing there; and recently, more places have moved towards digital self-scheduling and bill pay. With self-scheduling, fewer calls go to the call center, and so there’s greater adherence to appointments. In that context, digitizing the patient experience helps.” The key, she emphasizes, lies in examining processes that could be dramatically optimized through automation.
Indeed, she notes, technology can be used to enhance processes that might not be directly related to revenue cycle management, but which can impact finance: for example, the implementation of systems of automated reminders through text messages, emails, and phone calls, to ensure that patients show up for their medical appointments, thus ensuring that there are not major gaps that lead to sub-optimal billing because of unfulfilled appointments. Developing systems that encompass a broad range of processes across health systems will be important in that regard as well, she emphasizes.
Asked how RCM leaders should move forward process-wise to leverage technology to optimize staffing, HFMA’s Gundling advises, “Do an assessment to find out where the opportunities are. Different communities might have different needs. If you’re in a socioeconomically disadvantaged area, you might need more staff to help patients access Medicaid, or help them navigate the health insurance exchanges; in other areas, it might be around how to set up a payment plan, or how to navigate their insurance benefits.” Per that, asked which technologies are helping the most right now, he says that it is “the ones that handle a lot of the prior authorizations. The denials are still a problem, but the ones that can ensure that edits are going through and that all the appeals are done as expeditiously as possible. Helping patients through the preauthorizations, and making everything that can be, is automated. And then beyond preauthorizations, it helps to divert staff to work with the payers to eliminate categories that require preauthorizations. If 99 percent of the preauthorizations are approved, maybe we don’t need to do that.”
On the border between RPA and AI
Many in the industry are very excited by the potential of artificial intelligence and machine learning to dramatically improve processes in revenue cycle management; but inevitably, Impact Advisors’ Akimchuk says, RCM leaders are going to have to get there through stages. For one thing, he observes, “We haven’t even gotten to the point yet where the majority of RCM departments are truly working by exception. What’s happened is that we’ve been able to eliminate some widget-level activity by utilizing technology through editing and grouping activity to create a subset predictable in terms of its outcome, to modify workflows, so that they’re automating their systems as much as possible. Where automation has now expanded is in the ability to identify and interpret scenarios based on payer behavior, allowing us to preemptively respond without a human intervention.”
Among the examples he cites are “Things like eligibility mismatches, requests for information, payer pends on certain types of claims every time, where you can’t submit information until you get the request, because of HIPAA requirements. Historically, most of our intelligence has been based on anecdotal information or small subsets. Now, with technology, we’ve been able to look at large amounts of data, so you can identify the small number of situations that prompt denials. We can build process around that without having to involve human interaction. That’s the new space that we’re starting to get into.”
Is that the border between robotic process automation (RPA) and actual machine learning-based AI? “It’s actually two pieces of the same solution,” Akimchuk says. “If you think about AI as the intelligence to identify and capture and RPA as the ability to act on that intelligence, each has value, and combined, they have more value. The knowledge by itself gets you something, but not enough to drive towards a goal; and the action is only as good as the intelligence driving it. Traditional RPA has relied on a defined set of scenarios or rules: when A happens, you do B; when C happens, you do D. Machine learning/AI goes in and identifies those patterns. So instead of having a set of rules and outcomes, you can identify a set of hypotheses, and can program the RPA to act based on the insights you obtain. AI is a snipper: I may have hypotheses that I will use my machine learning to validate. Machine learning has the ability to take large pieces of data and establish its own hypotheses, based on identifying patterns. And you can then create process automation to react to that, or change your input before it happens.”
How far along are hospitals on using AI? “It’s early days,” Akimchuk says. “In healthcare right now, AI/machine learning is becoming prevalent in terms of diagnostics, such as in the mammogram/breast cancer space. AI identifies patterns and correlations, determines hypotheses, and then validates those hypotheses to determine veracity. Everything we’ve done has prioritized the clinical. The business side of what we do has always lagged; and that’s appropriate, frankly. We’re in the business of helping people. This use of AI and even earlier, with the beginning of RPA, hasn’t changed that construct. So we are very much at the beginning of how we can utilize this; the value proposition that most organizations are up against is simply, do I take my intellectual capital required to generate and create this, from the clinical to business solutions? And quite frankly, I’m less worried about whether we’ll get there and more focused on our best opportunities to get there early. If we can find and demonstrate the right use case, then it starts to get legs.”
A journey of 1,000 miles
Looking at this complex landscape, where do experts think most hospitals, medical groups, and health systems are right now on this journey of 1,000 miles? “So many hospitals are devoting more of their resources to clinical care and health IT, and we tend to undervalue revenue cycle management, because it’s seen as a cost center,” HFMA’s Gundling says. “But that prevents as good a consumer experience as with other types of experiences.”
How might the technology advance over the next few years? Impact Advisors’ Akimchuk says that, “As we go further down the path with better-defined technology, I believe that we’ll start to see more and more capsulized uses of this technology. Think about how data warehousing started—with a group of people trying to cobble together different tools into a solution. I believe that this is going to be similar to that; that it will take time for people to demonstrate value; and a technological approach to simplify this, whether that’s through visual programming or some other means.”
Meanwhile, asked whether revenue cycle leaders are moving forward with alacrity, Premier’s McBride says that “I think they want to do all of it; I think they’re very much interested in automation and such solutions; but budgetary issues, getting it into the cycle, and having IT resources available, are among the barriers right now, she says. “One client said, ‘Our IT department has two years’ worth of projects they still have to go live on.’”
Importantly, though, Premier’s McKasson observes, “The environment is becoming more difficult overall” in terms of the financial constraints under which RCM departments are operating, with the ongoing staffing shortages facing RCM leaders, and the realization of the need for hyper-efficiency. “It’s becoming harder overall, and they have to be creative,” he says. “Also, retention becomes a big thing. If these individuals are hourly employees, going down the street for an extra dollar per hour, they’re willing to leap away. And then after you’ve trained them, you’ve wasted that time. So they’re implementing retention bonuses as well.”
In this time of challenge and uncertainty what should senior health IT leaders be doing? “I would want them to partner with the revenue cycle leaders, especially around robotic process automation,” Premier’s McBride urges. “The revenue cycle people have a level of IT knowledge, but supporting bots and other transformative types of solutions—we need an IT partner to push that technology forward; it’s a joint effort. I would look to them to co-lead or help spearhead the effort. IT puts it in and supports it, but they don’t run it. The most successful implementations have been when they’ve been co-led efforts with revenue cycle and IT. It’s technology-enabled operations. Your operations need to change and adapt; and IT, they’re scanning what’s out there seeing what’s out there and looking at IT solutions; when they partner with each of the individual operational groups, it leads to a better outcome. It’s not just implementation, it’s optimization.”
