Artificial intelligence and automation have started making their mark in the U.S., and the trend is expected to continue. A recent report from global accounting and consulting firm PwC forecasts 38 percent of existing American jobs could be affected by automation from robotics and AI by the early 2030s.
Within healthcare, revenue cycle management is especially susceptible to automation, as certain tasks are relatively transactional, according to Brian Sanderson, managing principal of the Crowe Horwath healthcare services group.
Here, Mr. Sanderson provides four thoughts on how automation and AI could transform RCM.
1. Cost. A significant part of RCM costs is labor. By reducing labor costs through automation and AI, hospitals and health systems may drive down their cost to collect, says Mr. Sanderson. Based on the number of revenue cycle positions that could potentially be performed by AI and automation, Crowe Horwath predicts the cost to collect at organizations will decrease 25 percent to 50 percent over the next five to 10 years.
2. Reliability. Mr. Sanderson believes automation and AI will also bring more confidence and quality control to RCM. He notes once hospitals and health systems program automation, that automation repeats at the same level of quality each time. "When you look at the revenue cycle with people performing the same function who actually sit next to each other, you see a lot of volatility in performance, accuracy, productivity," he says. "When you invoke data science, assuming it's programmed correctly, you should have confidence something will be performed consistently across a population of accounts. So for our internal audit folks in hospitals and health systems, reliability is a big component."
3. Identify issues, see trends in real-time. Automation and AI will provide greater transparency and visibility on a real-time basis to exceptions that are occurring in the revenue cycle,such as denials and credits, according to Mr. Sanderson. RCM workers have traditionally been tasked with identifying patterns that create revenue leakage. But Mr. Sanderson notes this method relies on people to recognize trends themselves — a process that happens over time.
Automation and AI, though, may identify those patterns on a real-time basis, and creates exceptions to what is the standard protocol. "You'll get that visibility very quickly," he says. Rather than waiting until the end of the month or the end of the quarter to see issues that need to be resolved, automation and AI could identify issues within 24 hours. "It could be a payer changed some reimbursement methodology. It could be that within organization somebody tripped up some mapping of codes. But rather than wait until revenue is lost, you'll find it on a real-time basis because you rely on that data to teach the machine how to do work," says Mr. Sanderson.
4. RCM leaders are hesitant. Although potential exists for automation and AI to transform RCM, Mr. Sanderson acknowledges there generally is hesitancy from RCM leaders. He says the biggest thing he hears from RCM leaders is complete unfamiliarity. "It's a concept you need to understand, then understand how to work and practice and then trust," he says. "If you don't really understand how it works, then it won't be put in place." Mr. Sanderson says RCM leaders also show reticence to displace jobs and face challenges with how to aggregate data in a way that allows a data scientist to create trends. "The clinical side is ahead of the revenue side on this because they can have access to a lot of clinical data that have outcomes already embedded in them, but it is coming to revenue cycle," says Mr. Sanderson.