Medicare Advantage coding intensity could cost taxpayers $200 billion

Projected Coding Intensity In Medicare Advantage Could Increase Medicare Spending By $200 Billion Over Ten Years

By Richard Kronick
Health Affairs, February 2017


Over the past decade, the average risk score for Medicare Advantage (MA) enrollees has risen steadily relative to that for fee-for-service Medicare beneficiaries, by approximately 1.5 percent per year. The Centers for Medicare and Medicaid Services (CMS) uses patient demographic and diagnostic information to calculate a risk score for each beneficiary, and these risk scores are used to determine payment to MA plans. The increase in relative MA risk scores is largely the result of successful efforts by MA plans to identify additional diagnoses, also known as coding intensity, and not of changes in enrollees’ true health. In this article I estimate the effects of coding intensity on Medicare spending over the next decade. Under the moderately conservative assumption that coding intensity will decelerate, Medicare expenditures are expected to increase by approximately $200 billion. CMS has implemented a variety of strategies since 2010 that lessened the impact of coding intensity on Medicare spending; it has a variety of policy responses at its disposal to mitigate the impact going forward. The problem could be largely solved if CMS adjusted for coding intensity using the principle that MA beneficiaries are no healthier and no sicker than demographically similar fee-for-service Medicare beneficiaries, returning to the budget-neutrality approach that was introduced in 2004 and later abandoned.

From the Introduction

In the three decades that Medicare has been contracting with health maintenance organizations and other health plans, figuring out how to pay the plans accurately and fairly has posed a persistent challenge. Rapid growth in enrollment in Medicare Advantage (MA) plans has raised the stakes involved in getting the answer right: Over 30 percent of Medicare beneficiaries are enrolled in Medicare Advantage and account for an estimated $200 billion in Medicare payments in 2017, or approximately 1 percent of the gross domestic product.

House Speaker Paul Ryan’s proposal to convert Medicare into a premium support system raises the stakes even further. In such a system, beneficiaries might be required to pay a larger premium for traditional Medicare than for Medicare Advantage. It is critically important that any premium differential between traditional Medicare and Medicare Advantage under premium support reflect differences in efficiency, quality, and covered benefits, and not differences in risk selection or coding practices.

The inclusion of diagnoses in the payment system creates strong incentives for MA plans to report as many diagnoses as they can legitimately support—incentives that are not present in traditional Medicare. As a result, the average risk score for MA enrollees has increased by approximately 1.5 percent per year more than the average risk score of fee-for-service (FFS) Medicare beneficiaries over the past decade, with little indication that the relative morbidity of MA enrollees has actually increased. If this remarkably consistent rate of increase in relative Medicare Advantage risk continues over the next decade, Medicare payment will increase dramatically.

Despite the attention given to coding intensity by CMS and Congress over the past decade, unless there is a further policy response, Medicare will substantially overpay MA plans over the coming decade—likely to the tune of hundreds of billions of dollars.

From the Discussion

Political And Technical Obstacles

The political obstacles to a robust solution are formidable. Aggressive action to mitigate the effects of coding intensity would raise the ire of the Medicare Advantage industry and might well engender a lobbying effort involving the spectacle of buses circling the Capitol, with multitudes of MA enrollees imploring members of Congress to protect their benefits. The Medicare Advantage industry and enrollees would be angered by the prospect that aggressive response to coding intensity would increase premiums or reduce supplemental benefits enabled by the rebates that plans receive for bidding below their benchmarks.

In addition, aggressive action to mitigate the effects of coding intensity might raise concern among some policy makers about the stability and growth of the Medicare Advantage program.

The technical obstacle is determining how to accurately measure coding intensity. As discussed more fully in the Appendix, various reasonable approaches exist to measure the effects of coding intensity.


As originally conceived, the budget-neutrality adjustment was not designed as an adjustment for coding intensity, but it could serve that purpose well moving forward. It rests on a simple principle: namely, that MA enrollees are no healthier and no sicker than demographically similar FFS Medicare beneficiaries. If one accepts that principle, then it follows that aggregate payments to MA plans should be equal to the amount that would have been paid using demographic risk adjustment. As shown in this article, if CMS were to calculate the coding intensity adjustment using this principle, payments to MA plans would be approximately $200 billion less over the next decade than if the coding intensity adjustment remained at the statutory minimum of 5.91 percent.

Although there is evidence that the introduction of diagnostic risk adjustment using diagnoses from ambulatory care encounters in 2004 reduced favorable selection into MA plans, analyses by Joseph Newhouse and colleagues show that favorable risk selection into Medicare Advantage likely remains to some degree. More recent evidence provided by Pete Welch and me, and extended in Exhibit A-2 of the Appendix to this article, also suggests that MA enrollees are no sicker, and may well be healthier, than demographically similar FFS Medicare beneficiaries. Thus, the budget-neutrality method of calculating the coding intensity adjustment will result in payments to Medicare Advantage that are at least fair, if perhaps still slightly too generous.


The analysis presented in this article cannot solve the political problem of creating support for a robust response to the problems created by differential coding in Medicare Advantage. I hope, however, that this work will create the foundation for a solution to the technical problem by fostering a discussion of how best to measure and adjust for differential coding between Medicare Advantage and fee-for-service Medicare. Solving this problem is an important prerequisite to the establishment of a stable and equitable future for the current Medicare Advantage program, and even more important if Congress were to convert Medicare into a premium support system.



By Don McCanne, M.D.

Some long-time readers of these messages may be bored by yet another article confirming that the private Medicare Advantage plans have been successful in being paid significantly more than we are spending for comparable patients in the traditional fee-for-service Medicare program, but it is crucial to understand that this is a major part of the strategy to privatize the Medicare program, and we taxpayers are being cheated as they pull off their scheme.

This year we will see intensified efforts to convert to a premium support program (vouchers) which will enrich private insurers as costs are transferred onto the backs of Medicare beneficiaries, with the goal of privatizing the entire Medicare program. Insurers win, patients lose, and the libertarian ideologues snicker in the background.

There are two methods by which the private insurers are overpaid. They engage in favorable selection - using devious methods to enroll healthier patients - thus spending less than they would if their enrollees were of average health. They also have gamed risk adjustment - qualifying for extra payments by submitting claims indicating that their enrollees are sicker than they actually are.

About one-third of these extra payments are used to enhance benefits, thus enticing seniors to enroll in their plans instead of the traditional Medicare program. You can imagine how attractive lower deductibles could be to individuals who just turned 65 and are used to dealing with private plans. By marketing heavily to them they further improve their favorable selection since those just 65 years old are the healthiest and least expensive members of the Medicare population.

And that other two-thirds of the extra payments? They keep that to help pay for their high administrative costs that characterize the unique waste in the U.S. health care financing system.

But why should Medicare Advantage patients receive more benefits than those remaining in the traditional program? If the same extra payments were used to reduce deductibles and coinsurance in the traditional program then there would be no reason to select the private plans, especially since there are no network restrictions such as those instituted by the private plans. More choice, at no extra cost.

You say you don’t want to increase taxes to pay for better benefits in the traditional Medicare program? Then, for the sake of fairness, at least decrease the overpayments to the private Medicare Advantage plans to the same level that we are paying for the traditional program. Of course, because of their administrative excesses, they could never compete, and they would just go away (just as they were doing under the Medicare + Choice program that was discontinued when the insurers proved that it was impossible for them to provide higher quality at a lower cost).

The study author, Richard Kronick, suggests that we switch to a budget neutrality method of correcting the overpayments. That might reduce the problem of differential coding for risk adjustment, but we could still face the deceptive, opaque methods the insurers use to gain favorable selection.

Regular readers know what the solution is. Throw out the private insurers, improve the traditional Medicare program, and then use it to cover everyone - greater value while ensuring health care access for all.