New study cast doubts on how much MA plans lower Medicare spending, but is the methodology flawed?

Research published by the Kaiser Family Foundation (KFF) made headlines this month for calling into question whether Medicare is overpaying Medicare Advantage (MA) plans by billions of dollars each year. The study found that even after risk adjustment, beneficiaries who switch to MA plans spent less on care and used fewer services in the year prior to signing up for their MA plans than members who stay in traditional Medicare plans. The research raises questions about how much MA plans lower spending. But Kevin Mowll, executive director of The RISE Association, believes the data researchers used to make the comparisons is flawed.

The study adjusted for health risk and compared average traditional Medicare spending and use of services in 2015 among members who switched to MA plans in 2016 with those who remained in traditional Medicare that year. Researchers said they adjusted Medicare spending values for health conditions and other factors with a model like the Centers for Medicare & Medicaid Services Hierarchical Condition Codes (HCC) Risk Adjustment Model that it used to adjust payments to MA plans.

They found that that people who switched from traditional Medicare to MA in 2016 spent $1,253 less in 2015, on average, than beneficiaries who remained in traditional Medicare, after adjusting for health risk. The pattern held true even among members with specific health conditions,including asthma, breast or prostate cancer, and diabetes. 

“The study findings suggest that the current method of setting payments to Medicare Advantage plans based on spending for people in traditional Medicare may systematically overestimate expected costs of Medicare Advantage enrollees,” KFF said in a press release. “Adjusting payments to reflect Medicare Advantage enrollees’ prior use of health services could potentially lower total Medicare spending by billions of dollars annually.”

The findings have major implications for Medicare spending in the future if CMS sets payments to more accurately reflect expected costs. MA plans cover nearly a third of all Medicare beneficiaries, and enrollment is expected to increase to more than 30 million people within the next decade.

Mowll said that the study is important and raises critical questions that require the attention of all stakeholders. Indeed, if the MA program is ultimately more expensive than Original Medicare, the next question stakeholders must ask is whether it is worth the cost. But before they can come to that conclusion, Mowll said there are three other factors to consider:

• Is the methodology of the research study solid, and does it adequately make its case for the conclusions it suggests? 
• If there is a higher relative price tag on MA, how can the industry evaluate the relative value of MA versus Original Medicare to come to a decision? 
• The study does not address the cost savings to Medicare that are counted every year when all plans file their annual bids to continue with updated products in the next year.  These are certified results showing that, on average, weighted by membership, the government realizes approximately 8 percent savings (see graph on the bid process below).

Is the methodology of the research study solid?

The proposition in the study is that, through an apples-to-apples comparison, it appears that less costly beneficiaries migrate towards enrollment in MA plans when compared to those that remain on Original Medicare, according to Mowll. The methodology attempts to balance the research by even-handedly using the HCCs that arise through the capture of diagnosis codes (ICD-10-CM) from the respective claims submitted.  Thus, researchers assume that the risk adjustment factors (RAFs) resulting from these two populations are equivalent. That is, to illustrate, that Patient A on Original Medicare with a RAF score derived from a specific set of ICD-10 codes would have the same burden of illness as Patient B that is enrolled in a MA plan and shares all the identical ICD-10 diagnoses. Mowll believes that is a faulty assumption and is the source of a disconnect in the study.

“There is no way to know the true burden of illness of either population under the study,” he said. “That is because we learned in the MA business that diagnostic coding and documentation practice varies greatly between Original Medicare and MA.”

Mowll asserts that fee-for-service (FFS) based practices notoriously under-code and understate the burdens of illness in four ways:

1. They report only a limited number of diagnosis because their revenues depend almost exclusively on procedure codes, not diagnoses
2. They understate the severity and complexity of illness, instead opting for simplified and understated diagnoses, despite the ICD-10 guidelines requirement to code to the highest level of specificity
3. They are not constrained by the calendar year “clock” of risk adjustment that resets the effective diagnoses to zero every January
4. They do not report all co-morbid conditions reviewed in a visit, and instead focus on the presenting complaints or reasons for the visit

If one population ends up costing more than the other over a one-year period, it stands to reason that the burden of illness reported was understated to begin with, according to Mowll. The risk adjustment methodology has been shown to be highly predictive, so he said it is difficult to understand why it would not work if comprehensive and accurate diagnoses were captured.

“Therefore, the only logical conclusion is that the data capture was poor and inadequate,” Mowll said. “The only way to remedy this study’s flaw is to go the next step and conduct chart audits of all the patients on both sides to correct the RAF scores and run the analysis again with corrected risk adjustment calculations and projections.”

Is it possible to evaluate the relative value of MA versus Original Medicare?

The industry has much greater insight into the value delivered with the year of treatment and care provided in the MA program, Mowll said. It can rely on the publicly transparent reporting and tracking supplied through the Medicare Stars program, such as:

1. HEDIS® measures, which reveal whether the members received appropriate and necessary services against national benchmarks as a proxy for the quality of care
2. The Health Outcomes Survey (HOS), which reports on specific health outcomes measures (such as 30-day hospital readmissions), that combine with the HEDIS metrics for a more comprehensive view of quality
3. The CAHPS study, which reflects the direct feedback of MA members as to their levels of satisfaction with the care received and services provided by the plan in terms of access, availability, and adequacy on a self-reported basis
4. Administrative feedback in the Stars program, which reports on the general citizenship levels of the plan in the eyes of the government, including elements such as grievances and resolutions, sales practices, and compliance factors  

The public transparency of the MA program is another way of gauging the value received for the money paid by the government for this program, according to Mowll.  “Such accountability and comparability with Original Medicare are not currently possible,” he said. “We know that bonuses are paid to MA organizations for their achievements and rankings using these metrics. However, with respect to 30-day hospital readmissions, the Original Medicare program sees nearly one in five patients coming back into the hospital for the same condition for which they were originally hospitalized. That generates a whole lot of unnecessary costs.”

Furthermore, he said, when the patients who use Original Medicare receive their health care from FFS physicians, no one tracks the information to determine whether they receive all the necessary care or whether they fall through the cracks as measured by HEDIS reporting. But it wouldn’t surprise Mowll if they don’t get the care they need. Without the same type of monitoring and metrics used for MA plans, it is understandable why they have unnecessary hospital readmissions, and fail to have coordinated care, all of which will cost a lot more in the end.

“There are hefty financial hurdles to seeking care due to the deductibles and cost-sharing under Original Medicare Part B (and Part A, as well).  Again, since the data is not reported or gathered, it is impossible to compare program performance.  Yet it stands to reason that people who are not receiving preventive and comprehensive routine care will end up with deferred needs that crop up later with more costly interventions when missed needs worsen and result in ambulance trips, emergency room visits, and emergency admissions. These alone could account for some of the higher costs observed in Original Medicare patients,” he said.

MA plans pack a lot of extra value into the benefits delivered to their members in comparison to Original Medicare, Mowll noted. These plans are not only accountable for the cost, quality, accessibility, and acceptability of health care, plus their fulfillment of their contractual obligations, they bring quantifiable value to their members through the benefits they cover, he said.

What about the cost savings indicated via the annual bid process?

The annual bid process requires plans to bid on how much it costs them to provide the same services as Original Medicare Parts A and B. Most MA plans bid below the benchmark, and four-star and five-star plans can recapture up to 70 percent of that amount. They must then spend 100 percent of that rebate on benefits or reducing out-of-pocket costs for members. Therefore, Mowll said, the government is nearly assured that the cost to pay MA plans is less than Original Medicare due to the competitive bidding process.  

Mowll pointed to the most recently published annual bid results from 2016 to show that MA plans come in under their cost benchmark in their annual bids. This data, which is taken from all plans, shows the national average. It doesn’t account for the rebates based upon Stars ratings, however. The rebates on average will be much lower, he said, thus ensuring that the total cost to the CMS budget (and the tax payers) is always less than FFS Medicare (Original Medicare).

A November 2018 KFF issue brief summaries the facts about expected benefits and out-of-pocket costs for MA plans. CMS provides an online Plan Finder for Medicare beneficiaries, and it includes the estimated out-of-pocket cost of every plan option in contrast with their cost for Original Medicare. Invariably, the government calculation of estimated out-of-pocket costs of MA plans show significant savings. However, CMS does not necessarily consider all the extra benefits packed into the plans, such as gym memberships, Mowll said.

“The value received for MA plans is extremely high. The level of accountability for overall performance is high. When considering whether we are overpaying for MA coverage, it has not necessarily been proven due to the flawed data used to make the comparisons,” he said. “Poor quality management may be to blame for higher costs under Original Medicare, not that the MA plans are overpaid. At worst, the cost comparison is ambiguous, but it may well be outweighed when taking the whole question of value into consideration. “

Since the annual retention and satisfaction levels among MA members is so high, it is likely that the 35 percent of all Medicare beneficiaries are largely convinced that the value proposition posed by their plans is highly favorable, according to Mowll. “In the end, we know that the members of such plans are receiving high quality care and their performance levels only continues to improve, as reported by the Medicare Stars program,” he said.