fbpx
Select Page

Industry Insights

Artificial Intelligence: Payers vs Medical Billers Round One

By: Mick Raich, former President, RCM Consulting, Lighthouse Lab Services

There are some interesting changes in the medical billing world surrounding artificial intelligence (AI). There is an AI war going on with payers and billers in the revenue cycle management (RCM) arena and the winner will be the one who invests in AI quickest and maintains their edge in technology. We are just beginning to see this technology utilized, but several firms are already in the game. 

In essence, billers are using algorithms to review front-end demographics to ensure a clean claim is submitted. 

 

Utilizing new tools

Healthcare providers and billers are using data automation tools like Robotic Process Automation (RPA) to verify demographics and ensure a clean claim is submitted. They are also using these tools to find the accurate data needed to work claim denials and gain some leverage with payers. In fact, it has been reported that automation drove $122 billion in efficiency savings for the U.S. healthcare system in 2020[1], yet the same report highlighted that, “opportunities for automation remain.”

One firm that deploys RPA in the development of their software platform is FrontRunner Healthcare.

John Donnelly, CEO, Frontrunner HC

“Leveraging RPA and other data automation technology, we are helping labs, physician groups, and health systems better manage their rev cycle,” says John “JD” Donnelly, FrontRunnerHC’s Founder and CEO. “Sonora Quest, for example, achieved significant improvement in their DSO, cash collections, at-risk AR, correspondence and postage costs and most importantly, their patient satisfaction.

 

 

“Our software instantaneously gets the patients’ information, and then cross-checks it and fixes it if needed — all in real-time — and then automatically posts the correct data back into the client’s existing system if they’d like. While our software can be applied anywhere in the client’s process, we encourage use at the front end to catch the problems early and avoid the need for rework later.”

While we have seen work queue programs for years sorting claims within billers’ software systems (think of firms like OnBase), we are now seeing more add-on AI to the RCM process. This machine learning is being used on both the demographic side and the denial side.   

 

The difficulty with denials

The denial process is difficult for billers due to the volume and timing needed. There are huge volumes of denials which need to worked, both front-end edit denials from your clearinghouse and back-end denials from the payers. Furthermore, these claims must be appealed within a certain time window, or the revenue is gone.

The key, therefore, is to work the claims which are most likely to be paid in the quickest amount of time. Imagine if you could sort your denials and know with relative certainty which claims have a higher likelihood of getting paid. You could then allocate your human resources to work the highest probability claims.

If you can sort your claims and denials to the highest likelihood of payment, then you can really streamline your revenue cycle management (RCM) process. Keep in mind that every claim has to be billed, but not every claim has to be worked diligently. We can see a review of this in the dunning statement process. This is the process of sending paper claims to self-pay patients. Many billers sort and truncate their paper claims by billable amount and zip code. You don’t have to send three statements if you know it’s unlikely the patient won’t pay the co-pay or deductible.

Vitali Khvatkov, Chief Information Officer of our own Lighthouse Lab Services, has unique experience when it comes to these trends. Khvatkov has developed an AI program, RCM Spotlight, which providers can use to help sort claims correctly.

Viltali Khvatkov, CIO, Lighthouse Lab Services

“Medical billing is well suited application field for computer intelligence because this is a finding the needle in the haystack task, it involves repeatable processing of a lot of repeatable data looking for small irregularities and trying to predict payment outcome and outcome based on a number of details,” Khvatkov says.

 

 

For Khvatkov, the most significant promise AI presents to the RCM process is through Natural Language Processing (NLP). Transformers, the latest generation of NLP models pioneered by Google in 2018, can model and understand the meaning of human language.

“That means AI engines automatically read and summarize payer reimbursement policies, translate the diagnosis into ICD-10 codes, find answers to billing questions and answer patient questions about their bill,” Khvatkov says. “The potential of AI for medical billing is tremendous and we are working to realize this potential for our clients.”

 

Looking ahead

So, what is the end game in the AI RCM war? First, we cannot go back from this point. Firms that do not adapt AI technology will in time become dinosaurs and fade from competition. Those who adapt and improvise will likely win.

Second, the insurance industry will not back off this thinking as they have shareholders to please.

Third, whoever finds a way to work with Amazon or Google type firms will have an edge. Leading firms act as industry leaders. Finally, ten years from now, every firm will use AI to handle claims. With the declining payment rates and the current employment market, this is inevitable. We may even eventually see offshore server farms fighting each other for payment. Stay tuned.

Interested in learning more about RCM Spotlight or improving your overall revenue cycle management? Contact us today for a free consultation.

 

Sharing is caring!