AI and Machine Learning Only Work if You Do
By Brian Robertson
Brian Robertson is CEO of VisiQuate of Santa Rosa, CA.
Do AI and ML represent a game-changing opportunity for revenue cycle management? Absolutely. An annual research report by EMC and IDC indicates that the digital universe will contain 44 trillion gigabytes of data next year, with nearly a third of that data collected and stored by the healthcare industry, according to a Ponemon Institute study.
Within this vast ocean of data, AI and machine learning are well equipped to act as the precision sonar to detect and solve business problems using advanced data-driven methods. Indeed, AI and ML are part of today’s buzzwords du jour, but few now question that it will play a role. We must now advance the conversation: how to get going and laser in on value.
Let’s rewind to a time not long ago when we couldn’t blink without seeing a plethora of white papers on big data. They seemed to all contain the same message: “Big data has the potential to be a game-changer.” As the CEO of a company in the data analytics arena, we sometimes struggled with how to best communicate the power of big data to our clients. Our ultimate answer was to focus less on the intelligentsia and more on “get stuff done” (GSD) thinking.
Using AI and ML as an accelerator
First on deck? Don’t get too caught up in the hype cycle. From a pure technology standpoint, it’s just not that hard. One of the benefits of back-office operations, as opposed to clinical departments, is easy access and availability of structured data.
The harder part? Prioritizing business problems where a return on analytics (ROA) could deliver big value. Back to the ocean. Don’t boil it! Invest more time with your team thinking through what you’re trying to accomplish and what can deliver ROA/ROI.
Let’s take something like denial management. AI and ML can help speed up the discovery of problems that are both acute and systemic.
First, resolve what’s in front of you. Then go upstream where the real potential is. If you’re fixing the same problems repeatedly, solve that problem at its core.
Consider a physician dictation issue where some dictate with great attention to detail the complete services and care provided during a complex surgical case. Coders love that because they rely on substantive information to correctly code. That’s in contrast to physicians with less attention to detail, where denials and/or lost revenue is impacted a la the old adage, “If it wasn’t documented, it wasn’t done.”
Automating variability by physician can help you better solve problems upstream. Maybe it’s a system glitch where a bill editor is not set up correctly. Inaccurate or incomplete payer edits often repeat month after month. Deeper trending insights can automate the illustration of consistent anomalies.
This is where AI and ML become a competitive advantage, particularly when you stay focused on business value vs. the glitter of new tech. Start narrow and allow the algorithms do some of the heavy lifting.
- Purely repetitive process automation. Take a binary process and drive automation via robotic process automation (RPA) tools and methods.
- Enhance user or consumer experience. Chatbots can deliver an exceptional user experience. Why not leverage voice automation and have your chatbot send you the daily cash report for your commute home? Or a report showing slow-paying payers? Or the bad debt forecast?
- Deep data mining. Use anomaly detection on historical claim data to empower upstream decision-making. Leverage ML to see what’s going on with the patterns. Let the decision-making power get smarter every day.
Done right, AI and ML will improve yield, increase velocity, and optimize FTE impact.
Final tips to those looking at AI and ML for back-end optimization
- Fail fast so you don’t lose precious time over-analyzing.
- Avoid the technology hype and focus more on business problems the technology can help enhance or catalyze.
- Train your staff. FTEs in repetitive roles will become obsolete — it’s just the reality of our future. We as leaders have a moral responsibility to train our talent. As Gartner often advocates, create learning pathways to enable your staff to become capable citizen data scientists. Give them a meaningful shot at surviving in the long-term.
- Lastly, pick three business problems. Go narrow and deep. but as deep as you possibly can. Then it’s time to grab a shovel and get after it.