Medicare Advantage UpCoding Has Been Eliminated by CMS Effective 2022

By GEORGE HALVORSON

Medicare Advantage now enrolls almost exactly half the people enrolled in Medicare — and has both significant fans and hardline opponents in the health care policy circles who disagree about its performance.

The biggest attack point that comes from the critics deals with the issues of coding accuracy by the plans. The payment model for the program is capitation — and that capitation is based on the average cost of fee-for-service Medicare in every county. The people who designed the model believed that the country should use the average cost of fee-for-service Medicare in every county as the baseline number and should have the plans paid less than that average Medicare cost going forward every year in their capitation cash flow.

Medicare fee-for-service has a strict and consistent payment level based on a list of approved Medicare services — and they add up the cost of those services in every county and let the plans bid a lower number than that fee for service Medicare cost, if the plans believe they can offer all of the basic benefits and possibly add more benefits and additional services for that amount.

The fee-for-service Medicare cash flow and costs in each county tend to be very stable over time, with a continuous and steady increase in the actual functional cost for taking care of those fee-for-service patients for each year that they receive care. That total cost of fee-for-service Medicare care is a visible and clear baseline number that we can use each year with confidence and knowledge that it is what we are spending now on those Medicare members in the counties.

The direct capitation amount that is then paid to each of the plans is based on the age, gender, health status, and diagnosis profile of the Medicare Advantage members who enroll in the plans. The plans have been reporting those patient profile numbers to the government through the Risk Adjustment Processing System (RAPS) over a couple of decades to set up their payment levels and to create the monthly cash flow for each plan.

That’s where the upcoding accusations relative to the plans arise.

The plans get paid more if the patients have more expensive diagnoses — so the plans have had a strong and direct incentive to make sure that every diabetic enrollee is recorded and reported as being diabetic for their RAPS filings.

They also have a strong incentive to be sure that every congestive heart failure patient has their diagnosis recorded in their RAPS report.

The Medicare Advantage critics say that the plans inflate and tend to over report and even upcode the higher impact diagnosis for their members. There have been multiple media stories about plans doing things like sending nurses into patients’ homes to collect diabetic diagnosis rather than having those deliver care to those patients.

The plans do often have nurses going into patients’ homes, and it is definitely accurate to say that one of the things that the nurses do is make sure that their diabetic patients are all recorded on the RAPS system as being diabetic.

One plan owned its own pharmacies, and that plan was accused of looking at the pharmacy records to identify patients with insulin prescriptions who had not been listed on the RAPS profile as being diabetic. There were a few patients who met that category in the year that the study was done, and that plan did end up with a slightly higher number of diabetics than the RAPS report for that plan would have shown for that year if the pharmacy files and records had not been reviewed for those patients.

There have been a number of both incidents and processes over the years where plan activity that might have increased the diagnosis levels has happened and the plan critics looking at those activities had a consensus conclusion that the total negative impact of that upcoding approach might be inflating the payment levels of the plans. Some critics say that number now approximates 9%.

The critics have said that, even though the Medicare Advantage plans consistently bid 12–20% lower amounts than the fee-for-service average cost per county cost, that consistent 9% upcoding basically offsets those gains. The critics that what appears to be Medicare Advantage’s lower costs in every county isn’t really lower costs because of the RAPS system upcoding.

A number of health care policy people and some of the academics who have looked at those issues have accepted that upcoding argument and some have even accepted that upcoding number. A number of observers have written various solutions to those approaches which sometimes include eliminating Medicare Advantage as a program, because some have said that the business model of the plans is to upcode not to deliver care.

Medicare Advantage’s supporters point to the fact that plan bids are significantly below the average fee-for-service cost in every county. They also argue that the quality agenda for the plans has reached the level where many care sites across the country publicize their achievements in the Medicare Advantage quality program. Those supporters and advocates say that it would be a major mistake for millions of patients to kill Medicare Advantage.

CMS decided two years ago to directly end the argument about upcoding. They decided to replace the highly contentious old debate about the risk levels of the plans with functional and direct data about the members. More importantly, CMS very wisely, intentionally, and explicitly decided to make upcoding impossible. They made that effective with the 2022 data flow.

It’s done and in place.

CMS actually killed RAPS. They completely eliminated the coding system. Plans can’t upcode anything now, because with the 2022 filing process, they actually can’t code anything.

CMS still needs data to make the payments to plans, so they now get all of their data from direct and individual patient encounter reports. That’s a far superior system. They get both the procedures done for each patient and the medical record for each encounter, and each patient is now the source of the diagnosis. They get the data now in a steady flow of patient information about the encounters rather than getting the risk levels and the diagnosis for the patients in a RAPS summary report that was filed a year or more after the care happens.

That was a very good thing for CMS to do. It is much better information. They now know exactly who is enrolled in every plan. They know their patterns of care and their diagnoses, and they have actual information about their care outcomes that can be derived from that data flow.

CMS’ actuaries have been continuously refining and improving their formulas for projecting future costs. They also have the ability to use the most current data combined with the most current actuarial wisdom to figure out how much to pay the plans to make sure the payment process is accurate as the patient profile for making payments.

Fee-for-service Medicare for low-income people delivers very expensive and often bad care. Fee-for-service Medicare care sites for low-income people have twice as many people going blind and three times as many people with amputations as the Medicare Advantage special needs plan patients who enroll many of those patients now.

The average cost of an amputation for fee-for-service Medicare is slightly over $100,000. Our low-income Medicare patients are 10 times more likely to lose a limb to that procedure. The capitated Medicare Advantage plans know that 90 percent of the amputations are caused by foot ulcers. They also know that you can reduce the foot ulcers by more than 40 percent with dry socks and clean feet for their patients — and the Medicare Advantage plans all do exactly that with high levels of success.

The nurses that the plans have going into homes aren’t there, as the critics said, to “harvest diagnosis” for upcoding. They are there for amputation reduction and they also cut the diabetic blindness levels for patients by more than half. Just by doing that more than 90 percent of patients in Medicare Advantage five-star plans now manage their blood sugar.

The critics who continue to attack the Medicare Advantage plans for having quality goals included in the Five-Star quality program relating to diabetes clearly aren’t thinking of patients when they make their attacks on that part of the Medicare Advantage quality agenda.

Bad care is expensive.

The capitation target amounts for each county include all of that bad care, and that’s why the plans can bid at about eighty to 90 percent of the average local fee-for-service Medicare costs in their capitation in every county and still make both surpluses and profits for their care.

When CMS killed the RAPS system entirely for 2022 that created the opportunity to actually get the right risk-level data on all of the Medicare Advantage patients. CMS used that more complete and more accurate data to set the capitation levels for 2023 for the plans. It turns out that the 9% upcoding and overcharge level that many critics were citing so visibly in those critical attacks was a very wrong number. The actual risk and cost number for real patients was actually 8.5% under-coding when comparing the old risk levels to the actual diagnosis and to the current risk levels for the plans.

Look at the actual CMS 2023 decision, which increased the payments 8.5%. Those numbers are hiding in plain sight and they are based on actual data.

The upcoding accusations that have crippled so much of the Medicare Advantage discussions should now die a permanent death with the new system, because even if a plan somehow now wanted to upcode, there’s no coding left to do. It’s not possible to do any coding when the coding system has completely gone.

It’s time to look at actual plan performance and to make decisions based on actual outcomes, and to trust and understand that the plan discounts on the capitation levels are all direct savings for Medicare. Because the plan bids and actual payments are below the average cost of Medicare fee-for-service in every county, those are real dollars being saved every month for all of the people enrolled in Medicare Advantage.

Five million of the Medicare Advantage members have dual eligibility for Medicaid and Medicare. Those very low-income people have high health care needs and very few resources that help them get through the day without the enhanced dental, vision, and hearing benefits from by the plans. Those benefits were enabled by the surpluses created by having fewer people going blind in the plans.

The Medicare Advantage critics who continue to accuse the plans of upcoding have never pretended or attempted to explain how it’s possible to upcode that set of Medicaid eligible and high need patients.

CMS did a very good thing to take that issue completely off the table.

Let’s build on what we have in teeing up continuous improvement in some of those areas of care and let’s make the long-term costs at Medicare far lower than any one has projected just by making and maintaining those major improvements in care and having better care cost significantly less because it’s more effective at so many levels.

George Halvorson is Chair and CEO of the Institute for InterGroup Understanding and was CEO of Kaiser Permanente from 2002-14.