BEYOND THE NUMBERS
Private Medicare Advantage (MA) insurers have long received overpayments from Medicare due to their “upcoding,” which makes MA enrollees appear less healthy than they are — and a recent court decision could worsen the problem. As policymakers look to strengthen Medicare, eliminating upcoding should be high on their agenda.
While two-thirds of Medicare enrollees participate in the traditional, federally run Medicare program, a third receive their benefits through MA plans. Medicare pays MA plans a fixed amount per enrollee that is “risk-adjusted” to reflect each enrollee’s health status, as measured by the health conditions with which he or she has been diagnosed. That gives MA plans a financial incentive to ensure that their providers record all possible diagnoses: more diagnoses lead to higher risk scores and, in turn, higher payments. In contrast, traditional Medicare pays many providers (particularly doctors) based on the procedures they perform, so they have no incentive to report more diagnoses.
More and more research documents the impact of upcoding. The Medicare Payment Advisory Commission (MedPAC) concludes that risk scores are about 8 percent higher, on average, for MA enrollees than for similar beneficiaries in traditional Medicare, although the gap varies considerably from plan to plan. Federal law requires the Centers for Medicare & Medicaid Services (CMS) to adjust MA risk scores across the board to make them more consistent with scores for traditional enrollees. But even after this “coding intensity” adjustment, upcoding adds an average of about 2 percent to MA payments, MedPAC says. Some plans likely receive much larger payment increases.
Other researchers estimate similar or higher rates of upcoding. In a National Bureau of Economic Research working paper, economists Michael Geruso and Timothy Layton find that MA enrollees generate 6 to 16 percent higher risk scores than they would in traditional Medicare. In a recent paper in Health Services Research, Paul Jacobs and Richard Kronick conclude that upcoding could be as high as 13 percent. Kronick estimates that coding intensity could add $200 billion to Medicare spending over ten years.
One step to reduce upcoding would be to exclude diagnoses gathered in health risk assessments (HRAs) from risk score calculations. MA plans increasingly provide health assessments of their enrollees; for example, a nurse may visit a patient’s home to conduct a physical exam. Some plans use these assessments more to gather diagnoses that raise enrollees’ risk scores than to improve care. MedPAC has recommended excluding diagnoses obtained solely from HRAs, but CMS hasn’t implemented the recommendation.
MedPAC also recommends that CMS develop and apply a coding intensity adjustment that fully accounts for the remaining differences in coding between MA plans and traditional Medicare. To improve equity across MA plans, CMS could apply different adjustments to plans with low, medium, or high coding intensity. Currently, it uses only the minimum adjustment that the law requires.
A recent U.S. district court decision may open the way to additional upcoding. Overturning a 2014 CMS regulation on reporting and returning MA overpayments, the court found that CMS was treating MA plans and traditional Medicare inconsistently because it audits MA diagnoses, which affect payments, but doesn’t audit diagnoses in traditional Medicare’s Part B. If sustained, the decision will facilitate upcoding by reducing CMS’s ability to recover overpayments, and it could undermine the entire risk-adjustment process. CMS now must issue a new overpayment rule, and further regulatory and legislative changes also may be needed to address upcoding effectively.