Artificial intelligence (AI) is an exciting topic in healthcare – one which is often associated with the clinical or the research side. But what about the back office and the healthcare supply chain? Could artificial intelligence lead to significant improvements in healthcare operations and efficiency, saving billions of dollars each year? At the 2019 World Medical Innovation Forum, a group of healthcare executives formed a panel to discuss how they are using AI-based technologies to good effect.
The moderator and panelists included:
- Peter Markell, EVP, Administration, and Finance, CFO and Treasurer, Partners Health System (moderating)
- Kent Ivanoff, Co-founder, and CEO, VisitPay (panel)
- Ingeborg Garrison, CNO/VP of Strategic Advisory Services, Verge Health (panel)
- Mary Beth Remorenko, VP, Revenue Cycle Operations, Partners Health System (panel)
- Brian Robertson, CEO, VisiQuate
AI-based technologies cover a range of capabilities
AI can encompass a variety of capabilities from robotic process automation (RPA) to high-end analytics. Kent Ivanoff, CEO VisitPay explains: “There is a spectrum for AI. On one extreme, there is very narrow AI – essentially pre-programmed response tools.”
Ivanoff adds, “On the other end of the spectrum, there are uses that are more like Hal from 2001: A Space Odyssey – where a machine is learning and mimicking brain function using input from a lot of different sources, integrating the information, and deciding what to do based on reasoning derived through insight gained from the data. VisitPay operates further along the spectrum – towards machine learning.”
Intelligence-based automation helps reduce costs, save time, and prevent errors.
AI is a natural fit for the back office
The healthcare industry has been exploring the clinical use of AI, i.e., precision medicine, for some time. “It’s a tougher nut to crack because evidence-based medicine is a lot more amorphous,” says Brian Robertson, CEO, VisiQuate. However, because the business side of healthcare is so conducive to structured data, there are several areas where AI is making headway. Process automation, improving the patient and provider experience, and data analytics on the back-end are three everyday use cases where intelligence-based automation is helping reduce costs, save time, and prevent errors.
Do more with less
Administrative waste is a massive challenge for the healthcare industry. Repetitive and time-consuming tasks are easy targets for process automation. Robotic Process Automation focuses on routine and repetitive tasks that are expensive for the provider back office. Partners Health System (PHS) uses bots to help with accounts receivable follow-up to improve its efficiency. With the bot in production, the workflow has been reduced from eight to ten minutes per account, down to between 30 to 60 seconds.
PHS also deployed bot technology to validate licensure. Before the bot, the validation task was a full-time job. Now that headcount has been redeployed to focus on more complex, higher touch work. “In my revenue cycle business, approximately 70% of my budget is salary-related. For this reason, we are automating routine workflows and repetitive tasks,” explains Mary Beth Remorenko, VP, Revenue Cycle Operations, PHS. The staff engaged in these processes are being retrained and redeployed to roles which require higher levels of empathy, discretion, and personal engagement.
Mitigate risk, reduce errors, shorten claims processing times
Claims-processing errors create billions of dollars in unnecessary administrative costs, slow down payments to doctors, and frustrate patients. A lot of validation and quality review work happens in the back office. Eliminating errors occurring in the system and within transactions can optimize revenue, and identify additional revenue opportunities.
Eliminating errors can optimize revenue, and identify additional revenue opportunities.
Claims processing is an area where AI innovation has a significant impact at PHS. They’ve put bots to work handling follow up for Blue Cross denials. By mimicking the workflow of an end-user, the bot can go into the system and look to see if authorization is on file. If the authorization is there, then the data is collected; after logging into the payer website, the bot resubmits the information. In many instances, the once-denied claim is paid after being reprocessed.
Verge Health uses AI-based technologies to identify unreported events. Using a surveillance module that queries the EMR, Verge can identify undocumented or under-documented events, quantify the information, and then develop action plans that can sustainably mitigate risk, explains Ingeborg Garrison, CNO/VP of Strategic Advisory Services, Verge Health.
Improve the user experience
VisitPay mobilizes machine learning-based strategies in a way that enables health systems to break free from the one-size-fits-all approach to how they engage with the consumer. To address gaps in the revenue cycle area of healthcare operations, VisitPay designed a platform that uses analytics and machine learning to understand a patient’s financial needs, preferences, and propensity to pay. According to Ivanoff, VisitPay clients consistently generate a 70% increase in patient satisfaction, achieve 10X gains in efficiency, and experience an average 30% lift in revenue cycle yield.
With these kinds of benefits, there’s enormous potential for AI in the administrative and back-end processes of healthcare. But are there also risks?
Brian Robertson, CEO, VisiQuate, sees the potential for risk and waste if value engineering, business problem framing, and use case analysis are not tackled in tandem with the use of AI. “Think about painting. Sure it’s fun. But before you can paint, you have to do a lot of sanding and scraping,” says Robertson.
Don’t just jump to an AI conclusion. “AI is not the solution to a bad process,” notes Mary Beth Remorenko, VP, Revenue Cycle Operations, PHS. Have the workflows in a process been comprehensively reviewed against Lean Six Sigma principles? Has that process been fully optimized?
AI can go off the rails if you don’t properly train the machine. “You have to train that machine to use data and produce the outcomes you want. After you train it, verify that you’ve got the right outcomes. You can’t just rely on the machine to know that it’s arrived at the right answer,” explains Ivanoff.
Getting started with AI in the back office
If the heavy lifting gets done from the outset, AI and machine learning can amplify and accelerate what providers are doing in their business. “If humans haven’t done their job, then you can’t properly task a machine to do its job,” summarizes Peter Markell, EVP, Administration and Finance, CFO and Treasurer, PHS.
AI can amplify and accelerate what providers are doing in their business.
When thinking about where artificial intelligence can be used in your healthcare back office, the panelists at this event agreed that change must start with leadership support from the top.
- Get leadership excited about the potential advantages of AI and machine learning.
- Fund and invest in use case areas where the benefits of its application are visible.
- Evangelize and champion these use cases, building on their success to apply more significant levels of AI.
With so much potential for artificial intelligence to help the healthcare industry, the panel has many more interesting insights and experiences to share. To hear them all, watch the full-length video here.
Discover how VisitPay is using analytics and machine learning to improve the patient and provider experience.