What do Centene, CVS Health-Aetna and Humana all have in common? (Trying really hard not to start off with a “three health plans walk into a bar” joke…). Well, if you were at the 41st annual J.P. Morgan Healthcare Conference, you would know that all three are waiting to see what the upcoming final rule on Medicare Advantage risk adjustment data validation (known as RADV) audits will mean for the industry when it drops in February. And, wait, before you switch-off and think that this is a subject of intensive geekery, you should know that this topic impacts the foundational transformation of the healthcare industry from traditional fee-for-service reimbursement to value-based and risk-based reimbursement models, underlies the ability of the healthcare industry to undertake risk stratification to put healthcare resources where they are needed, and implicates billions and billions of dollars paid to health plans and providers. Estimates vary widely in the marketplace, from $12 billion (MedPac) to $200 billion (Richard Kronick, Health Affairs). That’s worth a look in my book. It’s also enough to cause the chief executive officers of these three plans to discuss the possibility of industry litigation against the federal government – which threat was effective in an earlier round of health plans versus CMS and delayed audits, enforcement actions and repayments under the Medicare Advantage risk adjustment program for years.
Risk adjustment is intended to reimburse health plans and providers whose Medicare Advantage members have a higher burden of illness, and therefore require more care, than the average Medicare Advantage population. Risk adjustment is not paid for in the Medicare fee-for-service program, as the health plan and providers are not at risk.
This higher burden of illness is determined by providers coding their patients for both current and past disease history (using the Hierarchical Condition Categories codes and Risk Adjustment Factor approach, or HCC-RAF), based upon the provider interaction and as documented in the medical record. Providers then share this coding with the health plans, which aggregate the data and submit it to CMS. CMS analyzes the data and then both (i) pays out additional lump sum payments in the year after date-of-service to compensate for the higher cost of care for the impacted patient populations that were coded; and (ii) uses the risk adjustment coding to set payments for that population going forward. This data, which is critical for identifying and adjusting appropriate CMS payments to health plans and providers, also is critical for effective risk stratification and patient management strategies by health plans and providers as they create treatment and management plans for chronic conditions. The HCC-RAF risk scores for a patient population will vary based upon the age, demographics, burden of illness and presence or absence of an effective coding program. The average HCC-RAF score is supposed to be a 1.0, but we see many patient populations that have not been managed or effectively coded in the past running below that, in the 0.7 – 0.9 range. A HCC-RAF score for many Medicare Advantage seniors, especially more elderly or frail, can go up into the 1.1 – 1.6 range, with a 1.3 – 1.4 score showing a well-managed, effectively coded population with a higher burden of illness, and scores above that typically going higher due to the patient being dually eligible or part of a high acuity cohort, such as a PACE patient (eligible for skilled nursing facility (SNF) care), for whom risk scores may be in the 2.0 to perhaps a 2.3 range.
Obviously though, there could be a strong financial incentive to be “creative” with the coding and to add codes retrospectively that may or may not be supported by the medical record. A whole industry sector has grown up around retrospective coding review, providing information to providers for HCC-RAF code “suspecting” or “prospecting” and encouraging physicians to code patients. And the regulators and enforcers have noticed and are pushing back, with the Department of Justice pushing hard over the years (although mostly unsuccessfully in their litigation that goes to trial) to require “two-way” coding be implemented by providers and health plans – not just adding codes upon review but also looking for and deleting unsupported codes that previously were noted. There is no statutory requirement currently for two-way coding to be used (a 2014 CMS proposed regulation was never put into effect), but it is considered to be best practice.
There’s also not a one-to-one correlation with adding more codes and receiving more dollars. Many codes are grouped, so adding a code in a similar group does not increase any payment. Additionally, CMS processes the data through their algorithm and adjusts as appropriate, so the inclusion of an extra code by a doctor does not directly result in more payment. And, interestingly, most doctors across the country have either not even heard of this program or are receiving a nominal sum (in the hundreds of dollars) for getting patients in for an annual wellness visit (AWV) and getting them properly coded. That’s not much of an incentive for overcoding. It is possible to have agreements between payors and providers to share the additional HCC-RAF payments received from CMS, and we see that commonly in certain parts of the country where provider organizations are taking downstream risk or are being delegated administrative functions from the health plans.
So, what could be in the upcoming RADV rule that has everyone waiting with baited breath? It’s three things:
- Retroactivity – for what period will CMS be auditing under the new rule? Will the rules be applied retroactively or only prospectively? If retroactively, then there are many dollars already received that are at stake and the need to potentially unearth thousands of patient charts in response to the RADV auditors.
- Extrapolation – The RADV auditors used by CMS will not be able to look at every single encounter and patient chart. They instead will ask for a sample to be pulled and then, per the earlier HCC-RAF program (and dispute between the health plans and CMS), there will be an extrapolation done. So, if 300 charts are pulled, and there is an error rate of 15%, does that 15% then get applied to the total number of actual Medicare Advantage members attributed by that health plan on a pro rata basis? Put another way, does that then mean that 15% of the HCC-RAF payments for the Medicare Advantage local “H” contract then have to be repaid to CMS? Is extrapolation the right approach and how can it be done in a valid and statistically sound manner for the health plans. And…it’s even worse at the provider level. If we take the above example, then perhaps of the 300 charts pulled, maybe only 5 of them might be from a larger medical group in that geography under that “H” contract with CMS. So, is that medical group at risk for a 15% pay-back? Well, what if, of that 15%, the group had errors in 2 of their 5 charts that didn’t support the HCC-RAF coding. In that case, the group did not have a 15% error rate, but a 40% error rate. How much would they then owe back to the health plan of the repayment the health plan must make to CMS? You can see how this can get very large, very complex, very uncertain and very unfair quickly. It also could lead to massive litigation between the health plans and the providers, if health plans that are hurt by RADV audit results and forced to make a large repayment then look to recoup under their downstream provider agreements from their participating providers. Also, many plan-provider participation agreements today do not have an explicit extrapolation methodology in their contract forms, so that will then be open to discretion and dispute.
- Fee for Service Error Rate Adjustment – A core concept in Medicare Advantage is to balance the program with the original Medicare fee-for-service program (Medicare FFS). When auditing and enforcing Medicare FFS claims, providers who perform at less than the Medicare FFS error rate are typically seen as compliant, while those with a higher error rate than the CMS determined error rate are seen as potentially violative and of interest. That error rate changes and historically has ranged around 5% and somewhat higher, reaching about 7.5% for 2022 as the Medicare FFS Comprehensive Error Rate Testing result. CMS calibrates its HCC-RAF risk adjustment model and payments against the Medicare FFS data. For this upcoming rule, the questions raised are – will there continue to be this Medicare FFS adjuster and, if so, how will it be set? The health plans have argued that this approach introduces a systematic bias that causes underpayments to plans relative to Medicare FFS.
More to come in February when the new CMS rule drops….
The Winds of Change
As I was crowd-surfing the main staircase at the Westin St. Francis through the crashing wave of healthcare investors, I heard two guys behind me remembering when they first started attending decades ago that is was all three piece suits and uncomfortable shoes. Now, the winds of change had gotten rid of ties and allowed for sneakers. While I don’t mind losing the tie, in my mind, the more important change was the increasing number of women chief executive officers presenting. Sarah London of Centene, Roz Brewer of Walgreens Boots Alliance, Karen Lynch of CVS, and Kristen Peck of Zoetis, among others, all took the stage this year at the J.P. Morgan conference. That’s progress in our industry!
Any Leading Indicators for a 2023 Recession?
I was listening carefully to the health plan numbers announced for enrollment and what trends they were seeing as to growth in Medicaid members versus commercial and Marketplace enrollment. If we were to see an increase in Medicaid and a weakening of commercial or Marketplace numbers, that might be an indicator of employment losses or weakening consumer financial status, with the resulting loss or downgrading of healthcare insurance coverage/benefits. Thankfully, none of the health plans reporting shared troubling numbers. While there often is a lag between people moving from employers’ insurance or Marketplace plans to Medicaid when their employment or financial circumstances change, and therefore, the above cannot be dispositive, this at least correlates and supports the more generally available market indicators and does not markedly indicate a recessionary trend.
Food for Thought
Kristen Peck of Zoetis, one of the leading global animal health companies, shared an interesting note that I had not thought about. She suggested that if the world’s population were to grow by another 2 billion people by 2050, as forecasted, those additional people all will need sufficient protein for their diet. While non-animal protein continues to grow in the market as an additional source, Kristen shared that, as a percentage of total global protein production, it is relatively small compared to meat (animal and fish) production. Therefore, she was very positive on the future prospects for the animal health industry, as it will have an important role to play in supporting the animal meat sector. She also noted that Zoetis is also working on ESG and climate change initiatives in the animal health sector, as there is a significant opportunity there. Our firm has been growing its animal health and veterinary sector practice recently, and Kristen’s presentation was a good reminder of the way the world actually works. I will be thinking about this more as I sit down to dinner tonight…
On to Day 3 of the 41st Annual J.P. Morgan Healthcare Conference!