Health Policy

MedPAC Got It Wrong (pt 3)


This is the third part of former Kaiser Permanente CEO George Halvorson’s critique of Medpac’s new analysis of Medicare Advantage. Part 1 is here. Part 2 is here. Eventually I’ll be doing a summary article about all the back and forth about what Medicare Advantage really costs!-Matthew Holt

Risk status and RAF

What is on the MedPac radar screen and what keeps their attention and what actually takes up several long portions of the annual report this year is the other factor that changes the payment levels to the plans — the risk status of their enrollees.

The capitation levels that are paid to the plans are affected very directly by the health status levels of the actual enrollees.

Risk levels for the members set and change the payment levels for the plans. The very first capitation programs didn’t factor in relative risk status for the members, and it was possible for some care sites to make major profits on capitation just by enrolling healthier than average people and by being paid an average cost level for each area for the people they enrolled.

That initial payment process has evolved very intentionally into having diagnosis-based cost factors that attempt to link the health status of the members and a fair payment level for the plans. The plans identify for the risk filing process the diagnosis levels for the members and their payment levels as plans are directly affected by the risk levels they report for their members.

People have had some concern about whether some parts of that coding process have been done badly, incorrectly or with purely avaricious intent.

There have been significant levels of concern expressed about whether the plans might be able and willing to produce and present inaccurate and distorted information in the process. That alarm was triggered in part by the fact that some of the plans made getting that information into their annual filings a high priority and some were more successful than others in that process.

It is good to have accurate diagnosis information.

We actually should as a nation and a health care macro system want to see an expansion of our data base and our medical records on basic levels of diagnostic information.

As a nation and as a macro care system we should definitely want to have full diagnosis information for each patient. Care can be better when caregivers have the right diagnosis for all of their patients.

How CMS  has changed Risk Adjustment

CMS just did a brilliant thing and completely eliminated the filing system and process for risk coding and data.

The CMS Hierarchical Conditions Categories Risk Adjustment Model was just killed. CMS just took the system that has created the vast majority of concerns and churn about the issues of coding intensity and shut it down.

It no longer is a factor for any risk scores. CMS will still look at the relative risk levels of patients but will get that information completely from patient encounter filings and direct patient information and not from any plan filings or reports.

An entire industry of organizations working to enhance risk scores just became obsolete and irrelevant.

That was a wonderful thing to do because we can now focus on having the best care and the right payment levels for people where we have no doubt about the diagnosis and levels of care.

That makes 20 pages of this year’s MedPac report completely irrelevant to any parts of the new risk payment processes.

MedPac knew that was happening — and a much better report would have looked at the new issues that are created by that flow of data about diagnosis and risk.

All of the concern about whether nurses gather accurate information about diagnosis in home visits is no longer a coding concern and it is now, very appropriately, a care delivery issue and a care improvement concern.

No caregivers under the old model actually invented diagnosis. That accusation was made in fairly public ways a number of times. It was also fake news. For a very good reason. The medical records are considered almost sacred records by the caregivers in actual care sites, and the likelihood of caregivers putting fake information into that sacred, highly functional and carefully protected workspace simply to distort risk codes for health plans is actually very low in the real world.

MedPac said in one critical part of this year’s report that the discernment process on diagnosis only flowed in one direction and MedPac pointed out that the process didn’t provide for the nurses who were doing the health assessments in the homes to change the initial medical record if the nurse discovered in the home that the patient was not actually diabetic.

That particular concern about the one-sided directional flow of the coding process by MedPac does not understand that the likelihood of a nurse in a home changing the medical diagnosis of a patient that had been done by another caregiver is pretty close to zero because nurses aren’t allowed or in any position to change that information in a medical record and we would actually would not want that change in our medical records to be allowed or to happen.

Some people will still be concerned that as long as plans are paid more for having patients with certain diagnoses, the number of patients with those diagnoses might explode. It probably will expand, but there is a very good reason why it won’t explode.

The biological reality is that we humans have a limited number of diagnoses to discern and once they are discerned for each of us, that particular data burst is done and it won’t grow much from that point on.

When the new coding process kicked into gear, auditors looked at Medicare Fraud issues and some auditors estimated that basic Medicare billing fraud runs about 6 percent of fee-for-service Medicare and those first auditors also estimated that 6 percent of coding fraud probably also existed for MA .

The billing level fraud will probably continue — because it is so hard to stop — but the 6 percent matching fraud on coding has just ended.

When that report was written, the auditors said that the only sure way of getting rid of any coding fraud would be to build the data flow into the actual encounter system. The people who run CMS clearly read their own audit and then just did exactly that. They should be able to score a zero in that category for this year’s filing — and that is not insignificant progress.

Well done, CMS.

The current MedPac report has not figured that out and the report is still deeply concerned about possible coding fraud. They can probably relax and the fraud issue should look very different for next year’s report.

So the 6 percent coding fraud number and the 9 percent risk inflation number are now both irrelevant and both should be off the table for the next report.

So what are the real costs of MA ?

MA costs less.

MA costs a lot less. When the plans did their current bids, they ended up averaging 86 percent of the fee-for-service cost averages. They offer much better benefits, much better care, much better patient support, and they cost members nearly $1600 each in lower direct out-of-pocket costs.

The plans are much more effective users of the Medicare dollar than fee-for-service Medicare, and the plans generate surpluses for the members from that lower capitation level that this MedPac report says average more than $300 a month.

This report acknowledges the record levels of surpluses. Each of those surpluses represents MA costing less than standard Medicare because they come from the capitation and they are absolutely a better use of the Medicare dollar than fee-for-service Medicare gets from spending that same dollar.

Plans have created a record number of surpluses each year and they all come from spending fewer dollars per person than standard Medicare. Fee-for-service Medicare is a very poor purchaser of care — and that fee-for-service Medicare program now has its average member each spending more than $5000 in out of pocket costs each year because it is an inferior way to buy care.

When we look at the lower amount we are spending on capitation in all of those counties, it is easy to see that MA costs less. We can use that money going forward as our pathway to bringing Medicare spending to much better levels as a country.

We don’t need actuarial estimates of relative values of some kind to measure the actual impact of the costs. We can do just like anyone else buying coverage does and measure the difference in actual dollars against the alternative costs of the competing program and we can look at how much the actual dollars go up from year to year for each county in the country.

That’s the number we need to bring under control and we can do it by setting the capitation at the right level to achieve that goal.

The average per capita costs of fee-for-service Medicare are a very legitimate measure of the cost of the program and we know from that data that MA costs less. We need to build on that success.

People who don’t want to accept that reality can do various creative approaches to distort the costs of each program by inventing pseudo risk adjusted surrogate expense levels for the comparative calculations, but Medicare itself actually spends real dollars and not pseudo surrogate dollars and we know from looking at the data outlines in this year’s report that Medicare spends about 10 percent less per member right now on MA than it spends on fee-for-service Medicare in all of those counties.

If you pull the target expansion counties and that program out of the MA payment flow and calculations, the average cost in the other counties ends up significantly less than 90 percent of the actual average cost of care for fee-for-service Medicare in those counties and that is measured in real dollars.

Those numbers are all in this year’s report — hiding in plain sight.

Low income Americans in MA

All MA members are not the same. The biggest gap in the MedPac MA report this year is to miss and not report the extremely important and extremely relevant fact that two out of three low-income Americans have now joined MA. More than half of all African American Medicare members and more than 60% of Hispanic members are now in MA plans. 

The reason for this is much better benefits for MA plans. When the average net worth level of the Hispanic MA member is actually $12,000 — then the fact that the MA plans are language proficient and actually offer vision benefits and hearing benefits and dental benefits is hugely important because it’s functionally the only way that some of those members are ever going to be able to see and to hear.

The MedPac staff who are advising the Commission and the Congress with this report continue their tradition of leaving people entirely out of their report and entirely out of their priorities and they also believe and continue to say that those eyeglasses somehow indirectly distort and somehow increase the overall costs of Medicare because they exist and therefore they must cost Medicare money in some way and they don’t want Medicare to pay for anything that wasn’t in the original set of Medicare expenses.

That is just plain wrong thinking. Coldhearted, callous, sad, short-sighted, and wrong.

The Commission should ask their staff after reading the report this year to change their priorities and to make helping people and improving team care and improving personal care for people a priority for the use of Medicare dollars instead of resenting and literally disparaging those efforts and those approaches that improve lives because those enhancements increase enrollment in the plans and the staff writing those pieces of the report clearly does not want that enrollment to increase.

Special Needs Plans

The most important work done by MA this year again is actually the Special Needs Plans, and the MedPac report team left them also entirely out of their report. We have a lot of people in this country who have been damaged for too many years in too many ways by social determinants of health (SDH) factors that have created inequities and disparities in both health and care. Many of these people are now eligible for both Medicare and Medicaid. The “dual eligibles” have some of the biggest health care needs in America and deserve our support and our best efforts on their behalf. It was a particular blessing that we managed to get nearly 4 million people into MA Special Needs Plan before Covid hit.

MA is helping in a major way right now — and that was not mentioned in the report.

CMS today clearly understands that situation. Very good and caring people at CMS and the Medicare program are working with the Special Needs Plans now to make a major difference in the lives of millions of people who very much need that difference to happen.

The MedPac report — continuing its multiyear practices of not placing any priorities on real or actual people or on any part of their functional lives — sadly managed to leave one of the best things we are doing in the country as a government to functionally, directly and individually remediate people damaged by social determinants of health completely out of their annual report.

MedPac exists to advise Congress. Most members of Congress care very much about people’s lives. We need whoever provides oversight to the group and that Commission from Congress to ask MedPac to add that priority and those actual impacts on people of MA to next year’s MA report.

More than 400 members of Congress already think that is a good idea. They signed this support letter for MA .

The 400 Members of Congress probably would not be happy with a MedPac recommendation that suggests the benefits for the lowest-income people should be reduced because the benefits are attractive to too many people and get too many people to enroll in those plans.

‘Tone deaf’ comes to mind as a level of priorities and a descriptor for this year’s MedPac MA report.

So this year’s report has flaws.

The Future

We are at the point in the history of American health care where we could actually be on the cusp of a golden age for care. Our medical science is exploding. Beautiful tools are emerging and are becoming available to us. Our ability to do better diagnosis and to do better care — with better connections between caregivers and between caregivers and the patients they serve — has never been as good as it is now and it is going to get even better going forward from now.

We have up to a third of the people who are dying in hospitals who have had wrong diagnosis along the way to their care. That should not be happening. We need better data, better outcomes, better connections, better diagnosis, and better care, and we need to spend less to get more because the science should make that entirely the path we are on.

We should be using artificial intelligence to build care agendas and treatments for people and to provide the very best care at the very best sites of care. Some care site programs today are actually providing fully connected in-home hospital care that has better outcomes and much lower costs than most traditional in-hospital care. That should be available broadly and soon because the people doing that work are beginning to do it to scale.

That is relevant to this report on the MedPac annual report on MA because MA should be a major asset in making that new world happen. Fee-for-service Medicare has crippled far too many care enhancements, but MA should be a platform for making that better future happen.

MedPac should help by learning now what that continuum of possibilities actually is and MedPac should become expert in their applicability of all of those resources to both MA and Medicare itself.

ACOs should be data enhanced. Plans should insist on and deliver continuously improving care.

The new tools will be entirely transformational. We should use the huge cash flow of Medicare to help make that happen and when we do that well, care will be better and much less expensive for everyone.

So MedPac — please move on and please make a major difference in the future delivery of care because we need continuously improving care and that will only happen if we do some smart things to make it happen.

And stop using cost numbers about MA and process warnings about possible cost distortions for plans and patients that that you know are not true.

We can’t afford politically skewed thought processes at this point in our history when we could be doing so much more and could actually cover everyone with better care for less money if we get this right.

Let’s set up an expectation of continuously improving care for our country and let’s use our Medicare cash flow to actually make it happen.

George Halvorson is former CEO of Kaiser Permanente