All healthcare organizations are cognizant of the problems presented by fraud, waste and abuse (FWA). Until recently, however, there had yet to be a solution that significantly addresses an issue that each year results in losses of at least $300 billion.

Enter artificial intelligence. A 2019 Optum survey of 500 U.S. healthcare leaders showed that 62 percent of their organizations had adopted AI strategies, compared to 33 percent the year before. These leaders believe AI will be capable of meeting a wide variety of challenges — authorization automation, personalized care and EHR management among them — but FWA detection is high on the list. Fully 43 percent of the leaders polled pointed to that as a potential benefit.

Also in 2019, the Centers for Medicare and Medicaid Services (CMS) adopted a new, five-pronged approach to mitigating FWA, with AI emerging as one of the organization’s primary countermeasures. 

The belief throughout the healthcare industry is that AI and machine learning are capable of tackling a variety of issues, including:

  • Intricate Criminal Schemes: These include ransomware attacks and credit card fraud, and are so prevalent that healthcare FWA dwarfs even securities fraud, which runs between $10 billion and $40 billion annually. The pandemic has, sadly, availed bad actors of further opportunity to exploit the system, but AI has shown itself to be capable of anticipating and quashing threats, and adapting to meet the challenges that lie ahead.
  • FWA System Inefficiencies: Keeping legacy systems up to date is a costly, labor-intensive task. AI, by contrast, is capable of providing real-time updates, as it learns patterns and adapts accordingly.
  • Procedural Code Errors: It is estimated that as many as eight of every 10 medical bills in the U.S. contain coding errors, and that in the case of bills that run over $10,000, these errors can cost patients as much as 13 percent of the total. Sometimes these errors are honest mistakes, sometimes the result of deliberate upcoding. But AI is capable of constantly updating National Correct Coding Initiative (NCCI) codes, thus minimizing the possibility of such eventualities.
  • Institutional Waste: The problem is epitomized by unnecessary testing following elective surgery — notably the 1.7 million cataract procedures performed each year, which make them the most common undergone by Medicare beneficiaries. While some of that testing falls under the category of fraud, some of it is performed out of excess caution or other factors. Either way, AI can detect unusual billing patterns.
  • Operational Efficiency: One of the great challenges in the healthcare sector is achieving data interoperability — essentially, ensuring that all systems can communicate with one another, providing physicians with real-time insights and ensuring the best outcomes. AI makes this possible by shattering silos and melding information.

As an example of the way AI can benefit an organization, consider the insurance provider Highmark Inc., which has developed technology that quickly adapts to changing behavior to help spot things that might otherwise be missed. A review of the company’s payment integrity programs, including its Financial Investigations and Provider Review (FIPR), showed that Highmark’s performance was superior to that of  the industry standard, while at the same time offering customers significant savings.

This technology is not unlike Amazon’s developing technology, which gives the company insights into how customers think and shop, yielding a better application or website producing better revenues. 

Whether during or after the pandemic, aggregating data to further improve this artificial intelligence is paramount because COVID-19 changed the way so much of business is done. What might have been considered a potentially fraudulent anomaly before is now status quo.

Customers will also stand to benefit from an insurance system that can more accurately and efficiently provide the right healthcare coverage at the right time. A system that doesn’t get bogged down by waste is one that can deliver better results for its customers.

At least one expert in the field cautions organizations to be wary of the hype surrounding AI — to understand what your company might need, and what vendors are offering. Further, business owners would do well to have a firm understanding of whether AI can enhance machine learning’s efficacy. And finally, it is essential that AI and machine learning are capable of adapting to changes in claiming behavior, since bad actors are forever changing their approach.

The bottom line is that AI has already made a considerable impact in the healthcare sector, particularly as it pertains to FWA. As AI becomes more prevalent (and as the challenges within the space become increasingly complex), it is expected that it will become an even bigger factor in the years ahead.