CMS Considers AI and Value-Based Care Fraud Prevention Strategies

October 23, 2019 - CMS sent out two requests for information, one on program integrity issues regarding the fee-for-service to value-based care transition and the other on how to evolve data analytics and artificial intelligence to suppress fraudulent activities, CMS Administrator Seema Verma announced in a CMS blog post.

Medicare has tripled its membership since it first began in 1965 but its share of federal spending is ten times what it was then, Administrator Verma noted. As baby boomers graduate into Medicare coverage, opportunities for fraud only grow.

Aside from the program’s growth, the transition to value-based care has given further opportunity for fraud. If complexity enables fraud, then recent CMS rules and guidelines have only compounded Medicare’s vulnerability by introducing more intricate, advanced payment models and new systems, Administrator Verma admitted.

“More challenging cross-ownership issues have emerged, such as one corporate parent owning various providers and provider types,” explained Administrator Verma. “Increasingly complex webs of affiliations can allow unscrupulous providers to simply appear, disappear if they come under scrutiny, and then re-appear as ‘new’ entities.”

Thus, in light of the administration’s executive order to suppress Medicare fraud, the new CMS five-pillar program integrity strategy aims to revamp Medicare’s fraud prevention and response.

READ MORE: CMS Finalizes Rule to Crack Down on Medicare, Medicaid Fraud

The first pillar is to react to current fraudulent practices. To shut down “bad actors,” CMS collaborates with the OIG, the Department of Justice (DOJ), and Unified Program Integrity Contractors (UPIC).

OIG has been vocal on CMS Medicare fraud. In 2018, the agency reported that Medicare could save $367 million by auditing improper payments. CMS disputed OIG’s suggestion that physical therapy claims did not meet Medicare requirements and of notifying providers of overpayments through Medicare administrative contractors (MACs). However, CMS did agree with OIG’s suggestions for new monitoring systems for physical therapy claims and provider education.

Fraud prevention is the second pillar. This involves adapting systems after a potentially fraudulent scenario to ensure it will not reoccur. CMS also will work with law enforcement to create policies and processes that are immune to fraud.

The administrator gave a couple of examples, including enhancing the address validation process when sending out Medicare cards, a new process which Administrator Verma estimates saved billions. She also pointed to the Healthcare Fraud Prevention Partnership (HFPP), a voluntary collaboration between the public and private entities including private health insurance plans, employer organizations, and state agencies, numbering 144 members in total.

Third, CMS will protect against risks that are currently coming to light. For example, the agency has proposed a list of durable medical equipment, prosthetics, orthotics and supply (DMEPOS) products that are often used for telemarketing scams. The agency also instituted protections against home health fraud and says it is continuing to look for ways to regulate Medicare Advantage’s complex alternative payment models.

READ MORE: HHS Charges Dozens in $1.2B Telemedicine, DME Fraud Scheme

That last issue is particularly relevant as OIG recently found that CMS spent $160.8 million on drugs for Medicare Advantage that were also covered by hospices. It is an issue that has persisted since 2012, OIG said, and while OIG noted CMS corrective efforts, the results were lacking.

The fourth pillar acknowledges the administrative burden that anti-fraud preventive measures can place on providers and pledges to relieve that burden. CMS intends to accomplish this by educating providers so that they do not mistakenly commit activities that could be viewed as fraudulent.

Lastly, Administrator Verma expressed an interest in using artificial intelligence and machine learning to alleviate provider burden and combat fraud.

“Advanced analytics and artificial intelligence (AI) can perform rapid analysis and comparison of large scale claims data and medical records that could allow for more expeditious, seamless and accurate medical review, and ultimately, improved payment accuracy,” said Administrator Verma.

Administrator Verma foresees predictive analytics playing a major role in these technological advances. The agency already uses predictive analytics for its Fraud Prevention System and case management systems.

READ MORE: DOJ Recovers $2.5B in Healthcare Fraud, False Claims in 2018

Artificial intelligence and other advanced analytics has been a priority for private payers as well, which are exploring the technologies’ effectiveness for analyzing EHR data, remote patient monitoring systems, and more.

Administrator Verma said that Medicare improper payment rates have diminished but not disappeared, claims that Government Accountability Office (GAO) data seem to support.

Medicare has been on GAO’s list of government agencies at high risk for fraud and improper payments since 1990. Last year, GAO uncovered $90 billion in improper payments made from HHS to Medicaid and Medicare programs in 2017, $36.2 billion of which went to Medicare. Payments to Medicare represented 36.8 percent of the improper payments that year.

Though the number remains high, the agency can boast the lowest Medicare improper payment rate since 2010. In 2018, the Medicare improper payment rate dropped to 8.12 percent from a record-setting 9.51 percent the previous year. CMS attributed the achievement to decreased improper payments to home health agencies as well as skilled nursing facilities and false Durable Medical Equipment, Prosthetics, Orthotics, and Supplies (DMEPOS) payments.

A separate GAO report discovered that insufficient documentation was a primary reason for improper payments in 2017. Medicare suffered in total a 6.1 insufficient documentation rate in 2017.

The RFI comment period will be open until November 20, 2019.