Insurers profit from Medicare Advantage's incentive to add coding that boosts reimbursement

The CMS pays Medicare Advantage organizations a per member, per month fee to care for seniors enrolled in the plans. Since the 1980s, that payment has been adjusted based on demographic information such as age and sex. But in 2004, the government implemented a new risk-adjustment model that tweaked payment based on demographic information and the health conditions of each member.

Each Medicare Advantage beneficiary is assigned a risk score based in part on their health diagnoses. Insurers are supposed to be paid more for sicker members with higher risk scores because they use more resources to care for them, while plans receive a lower payment for healthier-than-average members with lower risk scores. The payment system is designed to prevent health insurers from cherry-picking only healthy members and it has worked to that effect.

At the same time the system created strong incentives for health insurers to report as many diagnosis codes as possible because that can lead to higher risk scores and higher payments from the CMS. The annual payment to an Advantage plan for an 84-year-old male patient with diabetes without complications who is not eligible for Medicaid would be $6,765, according to the Medicare Payment Advisory Commission. But tack on a diagnostic code for vascular disease and that same diabetic patient's Medicare payment would jump to $9,796.

It's important to note that the CMS requires diagnosis codes to be backed up by the patient's medical record and result from a hospital stay or face-to-face clinician visit. There have been allegations that some health insurers haven't followed this rule.

The traditional Medicare fee-for-service program is paid differently, and there is no incentive for providers in that program to code all possible diagnoses. Because that incentive is present in Advantage plans, the average risk score for Advantage enrollees is about 8% higher than the scores for similar Medicare fee-for-service beneficiaries, according to MedPAC's latest report, despite strong evidence that Advantage members are not sicker than fee-for-service patients.

The result is that payments to Medicare Advantage were 2% to 3% higher in 2016 than they would have been if those same patients were treated under fee-for-service Medicare, according to MedPAC. With that extra money, Advantage plans can offer benefits to their patients that fee-for-service beneficiaries don't have access to, such as vision care and health club memberships. Payments to Medicare Advantage plans totaled $210 billion in 2017 to manage the care for 19 million seniors.

The CMS has taken steps to adjust for the higher rate of coding in Advantage to create more parity in payment with traditional Medicare, but experts like Kronick argue the adjustment isn't enough. Powerful insurance industry lobbies and voter reaction to Medicare payment cuts are one reason the CMS has not done more, he said.

When the CMS began risk-adjusting Medicare Advantage payments, a cottage industry of third-party vendors formed to help insurers rake in more Medicare dollars by reviewing patient medical charts to find diagnoses that physicians didn't code. Some also send clinicians to patients' homes to perform health risk assessments to find additional diagnoses they can submit to the CMS to increase their payments.

Vendors with advanced analytics platforms that can comb through patient data and identify missing codes in claims often promise insurers high returns for using their risk-adjustment products. Verscend Technologies says on its website it helped Advantage plans secure $279 million in risk-adjustment payments in 2015 by deploying its “best-in-class” medical coders to find undocumented patient conditions. Thirty health plans, including Anthem, contract with Verscend for risk-adjustment. And business is good: Verscend closed last week on a $4.9 billion deal to acquire analytics company Cotiviti.

A case study published by Optum, which is owned by UnitedHealth Group, said it helped one payer identify a single provider whose patient medical records showed $1 million worth of medical codes for a year, but who didn't send a claim for them. It also helped the payer identify a capitated provider with $250,000 in claims, but who did not submit the diagnosis data. Optum and Verscend declined to be interviewed for this story.

Another vendor called SCIO Health Analytics said its clients see an average increase to their patients' risk scores of 18% in the first year of using its risk-adjustment services and 56% in the second year, translating to higher Medicare payments for the health plan. In one case study, SCIO said it helped a Texas-based Advantage plan with more than 65,000 enrollees increase its revenue potential by $24 million.

SCIO said its goal is to increase the accuracy of its clients' diagnostic data through prospective and retrospective chart reviews and provider education. It analyzes patient data to look for gaps in documentation or diagnostic coding patterns that don't make sense, and it fixes mistakes.

Donna DuLong, a senior consultant with SCIO, said the team often subtracts as many diagnostic codes from the claims as it adds. The CMS requires diagnosis codes submitted for risk-adjustment to be present in the patient's medical record.

“Our focus is accuracy of documentation. The coding team is looking at the documentation and for this service, you either have the support (in the medical record) or you don't,” DuLong said. Accurate documentation of patients' medical conditions also helps ensure they are benefiting from care-management programs and other services that the payers provide, she said.

Most insurers are probably not out to inflate patient illnesses through their chart review programs, sources said. But DuLong acknowledged there is potential that plans will game the system by fraudulently making patients look sicker on paper to increase risk-adjustment revenue, though she said those bad actors are few and far between (and not her clients).

The Justice Department has some possible suspects. Last month it told a federal court that Anthem wasn't fully complying with a civil investigative demand to offer testimony about its chart-review program and asked the court to force the insurer to comply. In one document filed with the court, Assistant U.S. Attorney Li Yu said Anthem has already admitted in response to the investigation that it does not attempt to identify which medical codes that providers submit on claims are not supported in the medical record.

“Instead of using the chart review process to identify both additional codes to submit and corrections to be made, Anthem executed its chart review program solely to generate additional diagnosis codes to report so that it could receive more payments from CMS,” the court filing stated.

The federal government similarly argued that UnitedHealthcare for many years used its chart review program to find additional codes to send to the CMS for payment but didn't bother to delete codes it knew were false and that would have saved Medicare a lot of money. One of the whistle-blower lawsuits against UnitedHealth was dismissed; the other case was pared down but is still alive.

Mary Inman, an attorney at law firm Constantine Cannon who brought one of the whistle-blower cases against UnitedHealthcare, said that the False Claims Act cases against Advantage insurers are the tip of the iceberg.

“This is clearly an industrywide practice that lots of managed-care organizations are involved in,” Inman said.

She also noted there have yet to be Medicare Advantage fraud lawsuits against the vendors that orchestrate the chart-review programs. Yet Verscend is mentioned throughout the Justice Department's case against Anthem, as Ingenix—now named Optum—is in the whistle-blower lawsuit against UnitedHealthcare.

“That may be the common thread: Vendors will often work for multiple plans and be unscrupulous and encourage them to do this kind of risk-adjustment work; to go out and comb the records. There will be cases against the vendors,” she predicted.