According to the National Health Care Anti-Fraud Association, $3.6 trillion – or roughly $11,000 per person – was spent in the U.S. in 2018 on healthcare services. Up to 10 percent of U.S. healthcare claims are fraudulent, totaling an estimated $240 billion. Estimates are that only 5-10 percent of healthcare fraud, waste and abuse (FWA) is intercepted.

No discussion about healthcare can overlook the fact that trillions of dollars of valuable treatments and medications save lives every day. But in such a large industry, it’s reasonable to expect there will be those who take advantage. Beth Griffin, Mastercard’s Vice President, Healthcare, Product Development and Innovation, noted in a recent PYMNTS interview about healthcare fraud that a staggering amount of that money is buying nothing at all. Instead, she noted, about $240 billion is going out the door to fraud, waste and abuse (FWA).

Mastercard is no stranger to healthcare; they’ve been a leader in the payments side of the business for over 20 years. But when they started looking more closely at healthcare, they recognized a strong need for Mastercard’s innovative artificial intelligence (AI) for fraud prevention.

“We store 18 petabytes of sensitive data – there are just huge amounts of data that we’re dealing with every day,” she said. “And we detect and defend against attacks on our network all day, every day. So as we looked for areas to focus on in healthcare, fraud, waste and abuse are areas where we have some really strong competencies, especially around using predictive analytics,” Griffin said, noting Mastercard’s 2017 acquisition of AI firm Brighterion.

Closing the gaps to healthcare fraud

Healthcare billings and claims are extremely complex, and open opportunities for fraudulent activities, Griffin said. A provider sends a claim file to the insurance company payer who edits and adjudicates the claim, then sends the file and payment back to the provider.

A Special Investigations Unit (SIU) reviews the file to look for fraudulent or suspicious activities after the payment is made. These highly skilled investigators are looking for unusual charges, wasteful practices, more expensive procedure codes, excess numbers of claims by one provider and more. Sometimes they uncover complex schemes that involve a number of providers, in which case it can take years to put a case together.

Trying to recover the money paid is a Herculean task, usually resulting in only a five to 10 percent reimbursement.

Complementing healthcare payments with AI

The Mastercard team decided to leverage its advanced AI to secure healthcare payments. Its workflow moves smoothly for legitimate interactions while locking out  bad ones. Just as Mastercard AI is trained for the payments industry, it can also be trained on historical healthcare data to prevent FWA, Griffin says.

“We take those trained models that know how to detect fraud and include some third-party sources, such as exclusion lists that give the names of providers that aren’t supposed to practice medicine anymore,” she said.

Early trials proved that using AI in the claims process helped SIUs become more operationally efficient. Current efficiencies show a reduction in false positives by 10 to 20 times, enabling investigators to focus on actual fraud cases.

“So, we’re stopping the fraud in its tracks or we’re pending those claims to take a closer look at them. And we’re allowing the investigative units to get those alerts that provide them with highly suspicious activity so they’re not wasting time.”

Early results in fraud, waste and abuse prevention

An early result surprised the Mastercard team while they were training models for a new deployment, confirming the platform’s self-learning was spot on.

Upon upload, the system immediately flagged a treatment for a patient receiving genetic and molecular testing. “This was a woman who was a cancer patient and had gone through the typical testing and activities, and was ultimately at hospice. She was end of life state, but it was at that point that the providers started to put through genetic and molecular test billings,” Griffin said, noting this testing would have been done early on in the diagnosis to help determine treatment.

“It was totally out of place,” Griffin said. “The fact that the technology was able to pick up the time frame, the dollar amounts, the kind of procedures that should have been processed, and we didn’t even feed it any data to learn that, I was really excited to see.”

With an estimated $240 billion of healthcare costs attributed to FWA, it’s easy to see how important intervention is to controlling costs. Healthcare providers, insurers and most of all patients deserve to know the system is serving those who need it. By tapping into its advanced AI solutions, Mastercard is gearing up to transform that system.

Learn more about Mastercard Healthcare Solutions and its use of advanced AI to combat healthcare fraud, waste and abuse.