The Centers for Medicare & Medicaid Services (CMS) issued a proposed rule last week that will change the way it audits Medicare Advantage plans–and the new policy may result in significant financial penalties for insurers. The agency wants to extrapolate data generated from Risk Adjustment Data Validation (RADV) audits dating back to 2011 without the use of a fee-for-service (FFS) adjuster to offset the error rate. RISE looks at the changes in the proposed rule.

Editor's note: This article was updated on Nov. 15, 2018 to include insights about the 2012 regulations and comments from Kevin Mowll, executive director of RISE.

CMS intends to overhaul the Medicare Advantage RADV audit process, a change which could result in insurers paying back millions of dollars to the federal government.

The agency outlined the changes in a proposed rule, announcing that it would extrapolate data in RADV contract-level audits dating back to 2011 and it would not apply a FFS adjuster to audit findings.

The clawback of improper payments would save the government $1 billion in 2020. CMS estimates that it would save $381 million each year in subsequent years because it would not make improper payments to insurers. The savings pile up quickly. The agency said that by implementing the changes to the audit process it can save as much as $4.5 billion over the next 10 years.

How audits work now

Each year CMS selects 30 plans for a RADV audit to confirm that the documented diagnoses in the patient’s medical record match the Hierarchical Condition Category codes that insurers submit to Medicare for payment.

But the data from the RADV audits is not extrapolated. Instead, if the audit reveals that in some cases the Medicare Advantage insurer exaggerated the patient’s condition and received higher payments than allowed, the plan is only responsible for paying back overpayments for the cases identified in the audit. In February 2012, CMS said it would conduct a RADV-like study on FFS payments to determine if the agency should apply a FFS adjuster to account for any effect of erroneous diagnosis codes in the data used to calibrate the Medicare Advantage risk adjustment model.

Why CMS wants to change the rules

A 2017 Government Accountability Office (GAO) report found that the federal government made more than $16 billion in improper payments to private Medicare Advantage health plans in 2016. The GAO investigation revealed that RADV audits were not targeted to contracts with the highest potential for improper payments.

In the proposed rule, CMS said it would audit up to 201 enrollees in any given plan and then extrapolate the audit results across the plan’s entire Medicare Advantage population. The agency also announced its plans to not adopt a FFS adjuster based on the results of an analysis, which found that diagnosis errors in Medicare data don’t lead to payment errors in the Medicare Advantage program when the FFS adjustment methodology is applied.

However, even if the study had found that diagnosis errors in claim data do lead to systematic payment errors in the Medicare Advantage program, CMS said it no longer believes that a RADV-specific payment adjustment is appropriate. That’s because errors in specific Medicare Advantage contract audits can produce both overpayment and underpayment, the net result of which washes out in the aggregate.  If anything, CMS says, there is a slight error in favor of the Medicare Advantage plans. 

For this reason, CMS said it’s not appropriate to correct any systematic payment error in the Medicare Advantage program through a payment adjustment that was only applied to audited contracts. “Doing so would introduce inequities between audited and unaudited plans, by only correcting the payments made to audited plans,” CMS said in the proposed rule.

This is not to say that the RADV audit extrapolation is based upon achieving total perfection, according to Kevin Mowll, executive director of RISE. Mowll refers back to the Notice of Final Payment Error Methodology that CMS published on February 24, 2012. The reference notes that the calculation allows for an error rate that only exceeds three standard deviations from the mean. In 2011, CMS states that the mean error rate was 11%. That provides a lot of leeway when you consider three standard deviations on top of the 11%, Mowll said. The FFS adjuster was contemplated as a further buffer if it was found to skew payment levels, which this recent study says it does not do. Here is the quote regarding the methodology from 2012:

"To derive the payment error estimate for each MA contract, the annual payment error for each sampled enrollee will be multiplied by the enrollee’s sampling weight (computed for each stratum [h] during the sampling phase as Nh/nh, where N represents the number of enrollees in the RADV-eligible population and n is the number of enrollees sampled). The weighted enrollee annual payment error will be summed across all enrollees in the sample to determine an estimated payment error for the MA contract (the “point estimate”). A 99 percent confidence interval (CI) will then be calculated for the estimated payment error for each audited MA contract."

But the agency’s rationale conflicts with a recent district court decision to grant UnitedHealth's motion for summary judgment and vacate the 2014 overpayment rule. The court determined that the overpayment rule, which required Medicare Advantage insurers to return overpayments within 60 days of identifying them, imposed a stricter standard on Medicare Advantage carriers than on CMS itself when it pays Medicare benefits.

CMS acknowledged the decision in the proposed rule, but said the ruling was made based on the administrative record before the court and didn’t include the results of the FFS adjuster study. The agency said it was reviewing the court decision and considering its response.

The proposed rule’s impact on Medicare Advantage plans

Matt Eyles, president and CEO of America’s Health Insurance Plans (AHIP), said in a statement that the organization is concerned that the CMS proposal reverses a long-standing position – held by both the agency and other stakeholders – that the FFS adjuster is legally and actuarially required. He said AHIP is performing a comprehensive assessment of how CMS’ proposal to modify how they conduct RADV audits without a FFS adjuster would impact beneficiaries and program stability.

“We are further concerned about significant methodological issues, which according to a recent study by the Wakely Consulting Group, would cause highly random and arbitrary results. CMS’ retroactive application of this proposed rule would exacerbate these problems,” he said.

Mowll says that perhaps a bigger concern to Medicare Advantage Organizations than this FFS adjuster ruling is the recent CMS announcement regarding relaxation of redocumentation requirements for chronic conditions under Original Medicare. This move, while notable for its intent to reduce administrative burden on primary care practitioners, augers in the opposite direction of robust and comprehensive coding and documentation under risk adjustment models. The result is likely to be that Medicare Advantage plans will be exposed to thinner documentation practices that present greater challenges under the threat of RADV audits and the extrapolation penalties, he says. However, Mowll says, that assumes that CMS does not clarify and accept less rigorous chart documentation under those specific circumstances addressed by its relaxed rules for FFS charting. 

CMS will accept feedback on the proposed rule through Dec. 31. Comments can be submitted electronically through its e-regulation website here.