In the telling of MedPAC (the Medicare Payment Advisory Commission, a Congressional agency), the premise governing the government's creation and support of Medicare Advantage, and predecessor private Medicare programs, has always functioned as a kind of slipknot.
The premise is that by managing care efficiently, private insurers will serve Medicare enrollees more cheaply than does FFS Medicare while improving (or at least not worsening) outcomes -- and also pass a good portion of the savings to enrollees in the form of reduced out-of-pocket costs and additional services.
MA plans do serve enrollees more cheaply, in the sense that they pay out less to providers per enrollee. They also pass about $2,000 in annual savings to members in the form of reduced out-of-pocket costs and added coverage, such as (variously limited) dental, hearing and vision care.
What they don't do -- ever -- is save the government (i.e., taxpayers) money, at least not in an absolute sense. By MedPAC's estimate, covering enrollees through MA currently costs about 104% of the cost of covering the same enrollees through traditional fee-for-service (FFS) Medicare.
Whether the value of the extra coverage and services MA plans offer exceeds the value of the government's excess payments remains opaque (as does whether the plans improve outcomes for enrollees). MedPAC considers the failure to collect sufficient data to make that assessment one of the program's persistent failures. But the overpayments tilt the market in favor of MA, as the plans do pass on much of the excess payment generated by benchmarks that exceed the cost of FFS Medicare to enrollees in one form or another.
For decades, the federal government has played cat-and-mouse with MA insurers: overpayments rise (e.g., via the George W. Bush administration/Republican Congress of 2003/4) and fall (e.g., via the Obama administration and Democratic Congress of 2009/10), but never go to zero. The payment reforms that Congress and/or CMS manage periodically manage to muster are necessary but never sufficient.
MedPAC keeps hammering, and sometimes its recommendations are followed to a degree, but its role is purely advisory. The "MedPAC with teeth" nominally established by the ACA -- the Independent Payment Advisory Board (IPAB) -- was relentlessly vilified by Republicans as "government-controlled health care," similarly resisted by healthcare providers, never staffed, and ultimately repealed. Putting enough pressure on the private plan program to enable it to produce excess benefits without cannibalizing the public program appears to be beyond the capacity of U.S. government.
Here is how MedPAC's 2022 report to Congress frames the complaint:
However, these efficiencies [realized by MA plans] are shared exclusively by the companies sponsoring MA plans and MA enrollees, in the form of extra benefits. The taxpayers and FFS Medicare beneficiaries who help fund the MA program (nearly all Medicare beneficiaries with Part B coverage pay a Part B premium, although for some that payment is made by a state Medicaid agency) do not realize any savings from MA plan efficiencies. Instead, Medicare spends 4 percent more for MA enrollees than it would spend if those enrollees remained in FFS Medicare. The MA program has been expected to reduce Medicare spending since its inception—under the original incorporation of private plans in Medicare in 1985, payments to private plans were set at 95 percent of FFS payments—but private plans in the aggregate have never produced savings for Medicare, due to policies governing payment rates to MA plans that the Commission has found to be deeply flawed (Ch. 12, p. 411).
MedPAC's 2021 report proposed several tweaks to MA's complex payment formula -- to benchmarks, rebates, and quality bonuses, capped by a 2% overall discount rate for all plans. But according to the 2022 report, almost all of the 4% excess in payments to MA plans (above what CMS would pay for the same enrollees in FFS Medicare) comes from the way MA plans game CMS's risk adjustment system, which compensates plans for serving enrollees with higher than average "risk scores," i.e., expected medical needs. Risk adjustment is necessary to deter plans from cherry-picking healthier-than-average enrollees. But as administered in the MA program, it's been a major cash cow for MA insurers.
According to the 2022 MedPAC report, MA plans' "coding intensity" (inflating of risk scores relative to what they would be were their enrollees in FFS Medicare) accounts for 3.6 percentage points of excess payment -- i.e., essentially all MA overpayment.
Plan payments—In 2022, plan payments remain higher than FFS spending levels. Total Medicare payments to MA plans (including rebates that finance extra benefits) average an estimated 104 percent of FFS spending, similar to the percentage of FFS spending in 2021. The 2022 estimate incorporates about 3.6 percentage points of uncorrected coding intensity. Relative to FFS spending for Part A and Part B benefits, quality bonuses in MA account for 3 percentage points of MA payments. Using plan bid data for 2022, and ignoring the impact of coding intensity, we estimate that MA payments are 100 percent of FFS spending (pp 412-413).
According to MedPAC's analysis, MA risk scores were 9.5% above FFS risk scores in 2020 (the report also cites an outside analysis (Kronick and Chua, 2021) that pegs the excess at 20%). U.S. lawmakers and administrators anticipate (and so implicitly accept) greater coding intensity from MA plans and, in accordance with a statutory requirement, CMS cuts the plans' risk scores across the board. Since 2018, the annual haircut has been 5.9 percentage points, though CMS has the authority to increase the percentage. In 2022, that "coding intensity adjustment" left the 3.6% excess payment cited above.
Since 2016, MedPAC has recommended several measures to close the gap between MA and FFS risk scores. If fully adopted, MedPAC estimates, these measures would shave about 2 percentage points off of total MA payments, or about half the excess (again, plans are paid an estimated 104% of what CMS would pay were their enrollees in FFS Medicare). As summarized in the 2022 report:
The Congress should direct the Secretary to develop a risk-adjustment model that uses two years of FFS and MA diagnostic data and does not include diagnoses from health risk assessments from either FFS or MA, and then apply a coding adjustment that fully accounts for the remaining differences in coding between FFS Medicare and MA plans.
"Health risk assessments" (HRAs) are questionnaires deployed by the plans, designed to facilitate treatment plans but also used to tease out diagnoses that may not be reflected in any treatment but nonetheless inflate the enrollee's risk score. The 2022 MedPAC report states that 30% of conditions that MA plans documented through HRAs were not reflected in encounter data, i.e. in actual treatment.
Another means of inflating risk scores is via "chart reviews" -- retrospective analyses of an enrollee's medical records to identify diagnoses missed in treatment. These are often conducted by third parties that promise revenue enhancement. Citing a 2021 report by the HHS Inspector General on MA plans' inflation of risk scores, the 2022 MedPAC report flags HRAs and chart reviews as together responsible for 64% of MA "coding intensity" -- i.e., risk score inflation. MedPAC rather diffidently suggests, without outright recommending, that "eliminating chart reviews as a source of diagnoses for risk adjustment is consistent with the Commission's approach and would reduce the need for an across-the-board or tiered adjustment."
MedPAC's final recommendation to curb coding intensity is to use one of various calculation methods to increase the annual haircut (the "coding intensity adjustment") imposed on MA plans' risk scores. Noting that, according to the OIG report, just 10 MA plan parent companies accounted for 79% of inflated RA payments, MedPAC floats one method of calculating the haircut that would sort plans into tiers determined by their coding intensity and cut the higher tiers more severely.
A method to effectively eliminate excess risk adjustment payments to MA plans collectively was proposed in 2017 by Richard Kronick, coauthor of the 2021 paper finding huge overpayments cited above. Kronick proposed restoring a budget neutrality adjustment that was in place when CMS first started using encounter data in 2004.* Under this method, "the coding intensity adjustment would be calculated to ensure that aggregate payments to MA plans were equal to the amount that would have been paid using demographic risk adjustment under the adjusted average per capita cost method." Demographic risk adjustment uses basic demographic data -- age, sex, Medicaid enrollment (i.e. low income) and disability -- as opposed to scoring enrollees based on actual diagnoses. Using demographic risk adjustment as a benchmark to cut overall payments to MA plans "rests on a simple principle: namely, that MA enrollees are no healthier and no sicker than demographically similar FFS Medicare beneficiaries." The more fine-grained risk adjustment based on diagnoses would still serve to adjust payment among MA plans (presumably still rewarding the intensive coders) but would be revenue-neutral overall. Or rather, it would benefit MA plans only slightly, as Kronick cites evidence that their enrollees are in fact slightly healthier than demographically similar FFS enrollees.
Kronick's mid-range estimate in 2017 was that revenue-neutral risk adjustment would save the federal government $200 billion over 10 years. In the 2021 update cited by MedPAC, Kronick and Chua estimate that the proposed method would reduce spending by $600 billion from 2023-2031. Since 2017, coding intensity has...intensified, and MA enrollment growth has far exceeded the expectations cited in the 2017 paper.
By whatever method, CMS has so far refrained from increasing the annual coding intensity adjustment to a level that would come close to matching the excess in MA risk adjustment payments. Why? Jeannie Fuglesten Biniek, a senior Medicare policy analyst at the Kaiser Family Foundation, suggests that CMS may be worried about further penalizing MA plans that don't upcode. Recall that the OIG report found the likely excess in risk scoring to be heavily concentrated among a few insurers.
Political inhibitions are also plainly a factor. Kronick acknowledges that his proposal "cannot solve the political problem of creating support for a robust response to the problems created by differential coding in Medicare Advantage." He raises the specter of MA plans siccing busloads of indignant enrollees on Capitol Hill during Open Enrollment season, which happens to coincide with election season. That is the core reality ensuring that the Medicare Advantage slipknot will forever let slip the promise of better care for less money.
--
* Kronick recounts that when encounter data was first incorporated in risk scores in 2004, MA plans' demographic risk scores were below those of FFS enrollees, and CMS compensated by adding the 8% difference to the aggregate payment to MA plans. Congress, seeking to avoid that plus-up, required CMS to phase out that budget neutrality adjustment by 2010. "Ironically," Kronick writes, "the budget-neutrality adjustment was phased out just around the time that it would have started resulting in reductions in payment to MA plans."
All of your points have value, and you have many insights.
ReplyDeleteHere is another perspective:
I believe that as you say, MA purchasers are clearly not as sick as risk adjustment suggests; I would go a step further, and assert that MA purchasers may be healthier than those in traditional Medicare.
I take this from my own selling experience. MA purchasers were more affluent, more analytical, and more compliant with doctor's orders than the persons in traditional Medicare.