How to Implement AI Platform to Fight Commercial Insurance Fraud

“How do we get in front of commercial insurance fraud issues as early in the scheme as possible?”

That is what Kurt Spears constantly asks himself as the vice president of financial investigations and provider review for Highmark Inc. And it is a question that haunts any anti-fraud department as new technologies and regulations open up more opportunities for commercial insurance fraud every day.

Fraud schemes in the US can consume as much as $300 billion of the nation’s healthcare spending, according to the National Health Care Anti-Fraud Association (NHCAA). 

But that is only the financial impact. Behind that figure are millions of patients, each of whom have been stripped of their personal security and sometimes undergone needless, unwanted, unapproved treatments. Over 2 million Americans have fallen victim to medical identity theft, the NHCAA estimates—and medical identity theft is merely one variant of the hundreds of fraud schemes in existence.

“The industry has historically been very dependent on rules-based analytic systems,” Spears explains. “It has been very focused on unusual trends and patterns, which is always going to be the case. And Highmark has always had that model in place as well.”

Like other healthcare payers, Highmark has relied on data analytics and health IT teams to detect and prevent fraud.

But Highmark did not settle comfortably into the industry’s anti-fraud model. 

The payer prides itself on being unique and ahead of the curve. It has solid evidence of that claim in its history, having established a children’s health insurance program that became foundational for the federal Children’s Health Insurance Program (CHIP) and a plan for seniors that prefigured modern Medicare.

Thus, it is no surprise that, even before introducing AI, Highmark’s fraud-fighting had a distinctive strategy.

Highmark’s approach to combatting commercial insurance fraud is multi-pronged, Spears explains. By multi-pronged, he means 15 different initiatives, various technologies, and multiple specialized teams.

Team members boast a blend of backgrounds—registered nurses, investigators, accountants, former law enforcement agents, clinical coders and programmers, and others.

“We take our findings and, if it's a clinical concern, we review it with our nurses and our MD. If we think it's a true fraud-related issue, we'll take it to some of our ex-undercover police officers,” Spears says.

All along the way, Highmark’s anti-fraud department works alongside the US Attorney's office, Drug Enforcement Administration (DEA), and Federal Bureau of Investigation (FBI).

But even Spears’s diverse team struggled to protect the payer’s over 4.5 million members across Pennsylvania, Delaware, and West Virginia against every health insurance fraud scheme.

Toward the end of 2019, Spears and Melissa Anderson, the executive vice president and chief audit and compliance officer of Highmark Health, decided to implement artificial intelligence to help catch fraud schemes earlier.

"We're trying to use the AI tool to add reasoning into the analytics that we use,” Spears says.

Whereas before a human agent with superior understanding of fraud schemes had to run through the data and search for repetitive, suspicious activities over a period of time, AI algorithms can identify patterns more quickly.

Highmark is one of the first payers to use AI to reinforce their anti-fraud strategies.

“Although just recently implemented, we’re already seeing positive results from our AI software,” Spears said in a press release. “The goal of AI is to adapt quickly to changing behavior and to help predict aberrances earlier than traditional tools that often rely on established rules to catch suspicious behavior.  We know it is much easier to stop these bad actors before the money goes out the door than pay and have to chase them.”

In 2019, Highmark’s anti-fraud team saved $260 million in prevented losses, recovered funds, and policy savings through its multi-pronged approach and the addition of AI. That is the highest savings it has seen in the past five years.

Since 2015, the payer has saved a total of $850 million due to the anti-fraud team’s efforts.

“My main goal is to help protect the health and safety of our customers, protect them from fraud, waste, and abuse, as well as being good stewards of their healthcare dollars,” Spears attests. “And this is just one more tool for us to try to get in front of issues before they get too far down the line.”

Payer executives can support their anti-fraud departments first and foremost by maintaining their companies’ focus on the consumer’s welfare, Spears advises. Secondly, they can support implementing models that identify risks.

“Melissa Anderson has been incredibly supportive of pretty much all the different initiatives that we've launched,” Spears emphasizes. “We're definitely aligned on the fact that we want to get in front of these issues quicker.”

In order to implement AI effectively, payers have to be willing to evolve and change, Spears adds. 

The mindset it takes to use AI against fraud is different than the traditional anti-fraud strategy in part because the problem is different. 

No longer are the teams fighting singular fraud schemes executed by a small set of criminals. Instead, payers are up against well-funded, organized schemes that take advantage of advanced technologies like telemedicine and new regulations including changes to the Affordable Care Act.

In light of the constantly metamorphosing landscape of commercial insurance fraud, payer executives now have to ask:

“Where are new healthcare dollars being spent and how can those systems or high-risk populations be exposed or taken advantage of?”

At the heart of this question is the consumer herself, reflecting the current trend of consumerism in healthcare.

In the face of national fraud schemes, the nation’s health plans have to band together.

“The health plans are going to have to continue to come together more and work together even more closely,” Spears concludes. “We’re going to continue to have to be aligned on anti-fraud strategies and methodologies to have the biggest impact we can.”