Medicare Advantage audits reveal pervasive overcharges

Introduction

More than three dozen just-released audits reveal how some private Medicare plans overcharged the government for the majority of elderly patients they treated, often by overstating the severity of medical conditions like diabetes and depression.

The Center for Public Integrity recently obtained 37 Medicare Advantage plan audits from federal health officials through a Freedom of Information Act lawsuit. The audits have never before been made public, and though they reveal overpayments from 2007—money that has since been paid back— many plans are still appealing the findings.

Medicare Advantage is a privately-run option to standard Medicare which has been growing in popularity and now enrolls more than 17 million seniors. In 2014, Medicare paid the health plans more than $160 billion.

But there’s growing controversy over the accuracy of billings based on a formula called a risk score, which is designed to pay Medicare Advantage plans higher rates for sicker patients and less for people in good health. In a series of articles published in 2014, the Center for Public Integrity reported that overspending tied to inflated risk scores has cost taxpayers tens of billions of dollars in recent years.

In May, a Government Accountability Office report called for “fundamental improvements” to curb excess charges linked to faulty risk scores. In addition, at least half a dozen health-industry insiders have filed whistleblower lawsuits that accuse Medicare Advantage insurers of manipulating risk scores to boost profits.

The Centers for Medicare and Medicaid Services audits show that all but two of the 37 health plans audited for 2007 were overpaid – typically several hundred thousand dollars too much for the sample of 201 patients examined at each plan. Among the insurers charging the government too much: five Humana, Inc. health plans, three UnitedHealth Care Group plans and four Wellpoint, Inc. plans. None had any comment.

The high rate of overpayments for many diseases could signal millions in losses to the government, since many of the plans enroll thousands of people.

Among other findings:

• Auditors on average could confirm just 60 percent of more than 20,000 medical conditions plans were paid to treat. The confirmation rates were much lower for some conditions, such as diabetes with serious complications, depression and some forms of cancer.
• Overpayments triggered by unsupported medical diagnoses at the 37 plans audited topped $10,000 per patient for more than 150 patients. The health plans overcharged the government by $2,000 or more for at least 3,500 people in the 2007 sample group.
• The health plans overall were three times as likely to charge Medicare too much than too little for some of the 70 medical conditions examined as part of the audits. Two of the 37 health plans – Group Health Cooperative in Washington State and a Kaiser Foundation Health Plan in California – had no net overpayments.

Michael Geruso, an assistant economics professor at the University of Texas at Austin, said aggressive coding practices have had a “huge impact on taxpayer spending” for the Medicare Advantage program.

Geruso, co-author of a study on Medicare Advantage billing, noted that error rates revealed in the audits suggest many overcharges have escaped scrutiny. “Clearly, there’s room for more auditing,” he said.

Diabetes “without (medical) complications” was the most common disease code reported by the health plans and auditors typically validated the payments in three of four cases.

But extra payments made to health plans which claimed some diabetic patients also had complications of the disease, such as eye or kidney problems, were reduced or invalidated in nearly half the cases, sometimes more — meaning that auditors found insufficient evidence these complications actually existed. Some of these reductions are still being disputed by the plans.

Several other disease categories triggering larger payments, including “major depressive bipolar and paranoid disorders” and “drug/alcohol dependence” also were rejected as unfounded by auditors nearly as often as they were confirmed at many plans.

Not every medical condition impacted the bottom line and in more than 200 cases auditors said patients justified higher fees that plans charged because of the severity of their illnesses. But auditors were between three and four times more likely to slash payments than raise them for many medical conditions.

The audits provide new evidence of how federal officials have struggled to stamp out inflated coding, which is known in health policy circles as “upcoding.”

None of the plans faced closer scrutiny following the audits, no matter the size of the overpayment. The 2007 audits, which collected a total of $12 million in overpayments, are the only ones CMS has completed since officials adopted risk scores in 2004 at the behest of Congress. In some cases, health plans are still appealing the results, nine years later.

The special Medicare Advantage audits are called Risk Adjustment Data Validation, or RADV. RADV audits involved a lengthy and secretive process in which medical records are selected from a sample of 201 patients enrolled in a health plan. Auditors review the medical files to confirm that the diseases billed for, and their severity, are properly documented. When they are not, CMS cuts or reduces the payment. Some plans have disputed nearly every pay cut while others chose not to contest most of the findings.

Yet the paucity of these audits, and their tendency to drag on for years unresolved, brought a stern rebuke from the GAO, the watchdog of Congress, in its May report.

The GAO criticized the Medicare agency for not expanding the scope of the audits as required by a provision of the Affordable Care Act in 2010. It also said that CMS had failed to zero in on health plans with histories of charging too much, or whether some plans have persistently exaggerated the severity of certain illnesses in order to jack up their fees.“We think that CMS has a lot of work to do,” James Cosgrove, who heads the GAO’s health care division, said in an interview shortly after the report’s release.

In response to written questions from the Center for Public Integrity, CMS officials agreed that some medical conditions are overbilled more often than others. But they said the results were “not conclusive given that the audit samples were not designed to produce statistically valid results” of overcharges for each disease.

Asked why overcharges are so much more common than underpayments, CMS said health plans have an “incentive to submit diagnosis codes to CMS and, as a result, are less likely to under-report these diagnoses.” Officials said unconfirmed diagnoses could be caused by “incorrect” coding by the health plan’s doctors or “incorrect diagnostic data” submitted to the government by the health plan.

Agency officials said they expect to complete 30 audits of 2011 billings by specific plans this year, but declined to say how much they expected to collect in overpayments. In the past, officials have put the figure at as much as $370 million.

That level of auditing would encompass about five percent of Medicare Advantage contracts for 2011. CMS has asserted that its goal is to audit every plan annually, but officials declined to say whether that’s likely to occur. However, they said they plan to expand the auditing program in 2017.

Medicare Advantage plans have challenged the validity of the audits from the start. Clare Krusing, a spokeswoman for America’s Health Insurance plans, an industry trade group, said that CMS doesn’t allow health plans to submit “additional data” such as drug prescriptions that could verify patients have the diseases claimed. The group also has called the RADV audit process “not yet stable and reliable.”

Expanding Medicare Advantage audits also “could disrupt the care being provided by plans that are working hard to meet the needs of their enrollees,” the trade group says.