Savvy AI Deployments Help Healthcare Providers Climb Reimbursement Learning Curve

In the darkest hours of the pandemic when stimulus money was being provided, COVID-19 tests and vaccinations were free to patients, but someone had to pay the bills. In most cases, it was the health system that delivered the services, and getting reimbursed is proving to be a bear.

This has much to do with the analog nature of healthcare records today and the labyrinthine nature of getting paid by the government, all of which need digital triage.

“There’s not an easy answer. I really feel for all the providers because it truly is a nightmare,” Experian Health Senior Vice President Rob Stucker told PYMNTS. “And it’s not limited to reimbursement for anything related to COVID vaccinations or tests. It’s across the board. Reimbursement has gotten much more challenging, especially over the last few years.”

His advice? Providers should first focus on areas where they have control — and can see the most dramatic improvements when trying to reduce the number of denied claims and get their reimbursements faster. The number one item on his list is eligibility checks, which are still done manually in many hospitals and physician’s groups, using external web sources that consume time and effort.

See also: Digital Solves Key Problems in the Face of Mounting Healthcare Collections

“The first question is, ‘Are you using automation to help your staff do their job?’” Stucker said. “Have you invested in a system that allows your staff to reach out to nearly every payer in a real-time environment to do that eligibility check in seconds, or are you expecting your staff to manually go to a payer’s website because it’s free?”

That’s where a March 2020 mindset persists, thinking anything COVID-related is somehow free.

“There’s a huge misconception. It’s not free. It’s costing you a lot of time and money,” Stucker told PYMNTS. “There’s also a much higher chance of something getting done inaccurately because of a manual process versus an automated process.”

The Trials of Denials

A major advocate of automation in healthcare, Stucker recited the litany of ways that the claims process can break down, and how automation coupled with data is curing this syndrome.

“Let’s assume as a provider, you are using automation. That’s great. That’s the first step,” he said. “But the next thing you really need to focus on is your staff [being] sufficiently trained to do everything they can do with that automation” to get a healthy reimbursements flow going.

Automation by itself can’t check the accuracy of data, leading to common problems like a claim being issued for someone named “Rob” instead of “Robert,” for example. If that person is a Medicaid member or has a Medicaid-managed plan, for example, it makes a big difference.

“Especially if they’re doing this manually, there’s a huge opportunity for error because [is the patient] really signed up for Medicaid, or is it a Medicaid-managed care system? If it’s put in as Medicaid and is truly a managed care plan, that claim’s going to get denied,” Stucker said.

He added that “a good system will have the claims piece that does the editing tied in to be able to see the electronic response, again, assuming you’re using automation, to see the electronic response that was run for eligibility for the patient to begin with.”

Stucker also said that a word of encouragement to those at the registration desk goes a long way toward improving this output, because the fewer mistakes, the more reimbursements.

Read more: Regulation Forces Healthcare Pivot to Patient Estimates

Partnering for Claim Editing and AI

Back to the fact that automation alone does not a digital transformation make, artificial intelligence (AI) makes the upgrade transformational — if it’s actual AI. Coupled with a code editing system that ensures proper treatment categorization, health systems can improve cash flow considerably by investing in systems that make sure claims are approved the first time.

“Regarding preventable denials, having a strong editing system [is crucial],” he said. “I’m talking about the third-party claims vendor that picks up the claim after it’s gone through your patient accounting system or practice management system. There’s going to be claim edits in both systems, but they’re typically not nearly as advanced as what you can get from a third-party vendor such as Experian and others.”

These advanced editing systems crosscheck a library of government and commercial payers against a set of predefined rules, which can put billing teams in a better position to align them with their own contracts and internal processes, he said.

Then there’s making the system “smart,” which is the AI part of the equation.

Stucker said, “The claims process is a good environment for artificial intelligence, because there is a very complicated data collection process, and you have to have a true system that can [handle] the algorithms if it’s actually going to be doing AI or machine learning.”

But for clarification, screen scraping from a payer’s website or scripting records into a patient accounting system from an electronic file is not AI. It’s just an electronic workaround.

“If a provider is hearing from a vendor, ‘We’re using the latest in AI technology,’ challenge them on it. Don’t just accept that their definition of AI is truly AI,” he said. “There is a lot of stuff being used for AI throughout the whole health industry.”