How to fix Medicare Advantage: Restoring the Payor-Provider Firewall

Why many “value based” contracts are code for “Risk Adjustment profiteering”

Authors: Vivek Garipalli & Andrew Toy

TL;DR

Risk adjusted revenue in Medicare in and of itself makes sense — sicker patients are expected to incur more healthcare services and should have more dollars allocated to their anticipated care to account for this likelihood. Risk adjustment works best when there is a firewall between those projecting, receiving and allocating the dollars (payors / plans) and those making the clinical diagnosis decisions that determine risk calculations (physicians / providers). Due to government program requirements that are unclear and ambiguous, this firewall is breaking down, resulting in a mis-aligment of incentives that drive up costs to the Federal Government and in turn taxpayers. And to make things even worse, CMS program requirements do not take into account risk adjusted revenue in provider-plan agreements, effectively undermining enforcement of the Medicare 85% Medical-Loss-Ratio minimum and ultimately hurting consumers and taxpayers.

CMS must address the breakdown of the payor-provider firewall to ensure the viability and credibility of the Medicare Advantage program. To do so, CMS must introduce clear program requirements that firmly reestablish the independence of providers’ clinical decision-making.

Risk Adjustment in Medicare Advantage is intended to address health equity and fairness

The spirit of the Medicare Advantage risk adjustment program is that the government is allocating a fixed number of dollars (known as a capitated payment) for the expected care of Medicare-eligibles, with a higher-fixed payment for sicker beneficiaries to cover their higher anticipated healthcare costs.

This personalized risk adjustment program is designed to prevent plans from purposefully recruiting healthier patients and avoiding those with a heavy disease burden (and therefore higher projected costs), which often correlates with socioeconomic, ethnic and geographic factors. If Medicare used a generic capitation payment set at the same level for all patients, then the government would be allocating unnecessary dollars to care for healthier patients, and simultaneously fewer dollars to care for sicker patients.

So, Medicare Advantage provides more dollars for the anticipated costs of treating sicker patients, and it does this by allocating those dollars against the clinical conditions diagnosed and documented by a physician for a given patient. The key thing is this: the program is designed so that the entity that receives higher revenue from risk adjustment (i.e., the health plan) cannot document these conditions themselves. Rather, a healthcare provider must do so. In this way, there is an implicit firewall between the risk economics of Medicare Advantage, and the clinicians that are diagnosing and documenting a patient’s health status. When this works well, the program results in the allocation of more Medicare dollars to sicker and underserved populations as compared to healthier, typically affluent populations.

How have things broken down?

Unfortunately, in the last decade, ambiguous and unclear program guidance has contributed to breakdowns in the firewall between providers and plans. While plans are not allowed to pay for diagnoses (ie, a plan cannot make direct payments tied to increasing the number of diagnoses captured), breakdowns occur when:

1. The plan IS the health system. This is the most straightforward way to break down the firewall — just be the same company financially. Even if they are technically separate legal entities, the ultimate economics are unified. This is what we see as a natural consequence of verticalized health systems. A verticalized healthcare system by nature serves as both the plan receiving Medicare Advantage payments from the government and the provider documenting patient health conditions. This kind of corporate structure allows for incentive alignment that may encourage disproportionate focus on risk adjustment coding by clinicians whose primary job is to manage patient care. Indeed, some physicians may even fear losing their jobs if they don’t aggressively code.
2. The plan and the provider have a “value based contract.” This is a more subtle, but an even more prevalent way to break the risk adjustment firewall. In this case, the provider group will receive a fixed payment to manage the Medicare patient from the plan, and this is structured as a percentage of the overall premium. While ostensibly this places incentives for the provider to control costs — they are now on the hook for medical expenses — this has the effect of also financially rewarding the provider if the overall Medicare payment from the government increases. So once again, the incentives and rewards are now in place that encourage providers to pursue revenue through risk adjustment. Even if the pursued revenue meets program requirements, the incentive to pursue that revenue has the potential to distract providers from patient care.
3. The plan has a tightly managed narrow network that selects for providers that code well. Medicare Advantage plans are commonly HMOs, which are those plans that allow patients to select from only a narrow pool of providers. One reason for the prevalence of this model may be tied to the ability to select for higher coding intensity. While in this case, economics from the plan might not be directly passed through to providers, the reward/incentive for the provider to code diagnoses in a manner that provides optimal financial benefit to the plan is continued participation in the HMO network. Faced with the potential of being dropped from the network, the provider must again focus on onerous coding requirements, rather than patient care.
4. Arrangements permit excess profits. As part of the Affordable Care Act, there is a minimum requirement that 85% of premiums be spent by plans on direct or indirect medical care offered by providers. Many capitation agreements now dictate that the physician/provider organization will receive 85% (or more of the premiums) for spending on “care.” The provider group can then direct those dollars as it sees fit or choose to simply keep it as profit, arguably evading the 85% requirement imposed on plans. We’d prefer to see this amount under the 85% sent back to the Federal Government.

How can we fix this?

The spirit and intention of risk adjustment remains a good one: it allows the government to allocate fixed amounts of dollars for the anticipated care of Medicare patients, thereby capping and controlling costs while still fairly compensating organizations who spend their time caring for sicker patients, including those in underserved socioeconomic or ethnic groups. Any “one size fits all” approach would risk disrupting this balance by underfunding the disease management of sicker groups.

The fix for Medicare Advantage risk revenue is to clarify the program requirements so as to avoid the pitfalls highlighted above and restore the firewall between providers and plans.

There are multiple ways to do this but some approaches include:

Within Medicare Advantage:

• Prevent acquisitions of providers by plans. It’s important to prevent verticalization that removes the firewall between providers and plans, which prioritizes the financial well-being of the plan above other stakeholders and drives physicians into narrow networks, limiting the availability of care for the general population.
• Ban full capitation. Prevent entities from entering into fully capitated agreements between plans and providers for managing MA populations. To be clear, sharing of value generated by savings — where society, the plan, the provider and the patient all win — is generally fine. The problem comes when capitation delivers revenue to the healthcare provider as an actual or implied reward for coding intensity.
• Better tailor the Medicare Risk Adjustment coding intensity factor. In the near-term, administrators should audit and tighten protocols for risk adjusted payments for lives attributed to fully capitated practices, narrow network HMOs and health plan-employed physicians, as these are the places most likely to have the highest prevalence of Risk Adjustment coding intensity.

This is a crisis and must be addressed to ensure the viability and credibility of the Medicare Advantage program.

Part 2: How to Fix Medicare Advantage Q&A

Yes, we take a fundamentally different approach to Medicare Advantage, and we embrace that difference like a badge of honor. When we describe our strategy to people in the federal government, provider groups, and other health insurance companies, we get a lot of the same questions. We’ve collected the most common questions and taken the opportunity to answer them below. We believe creating a dialogue to pressure test existing beliefs is essential to creating positive change.

Q. Accurate and complete documentation of clinical conditions is important; why shouldn’t doctors be given incentives to document those conditions accurately? Clover is being naive and doesn’t understand this basic point.

To be clear, we absolutely believe in physicians being armed with data and accurate diagnostic truth so they can deliver better care to their patients. What we don’t believe in is financially incentivizing physicians to document conditions. We believe that the laws and regulations governing federal healthcare programs should be strengthened to support this focus as the government has done in other contexts, such as with the Anti-Kickback Statute and Stark law prohibitions on referrals. For example, the Anti-Kickback Statute (AKS) makes it a criminal offense to knowingly and willfully offer, pay, solicit, or receive any remuneration to induce referrals of items or services reimbursable by a federal healthcare program. The law thus reflects the principle that — even if a service is medically necessary, a referring doctor should not get paid for that referral because financial incentive should be separated from the clinical action. Yes, it’s a physician’s job to document patients’ clinical conditions accurately, but getting paid more or less based on the volume or intensity of conditions documented should not be permitted, and taints the credibility of the risk adjustment system. We believe that the federal government should set clear guardrails around such incentives.

Q: Why would it be wrong for doctors to be paid for making diagnoses? Are you saying doctors are overcoding, i.e., they’re documenting inaccurate diagnoses that don’t actually exist?

It’s a self-serving response to say, “hey, nothing inaccurate happening here, please move on.” The reality is, just because something is documented accurately doesn’t mean we should reward that via higher costs to the system. Bottom line, paying based on diagnoses creates higher costs to Medicare, as evidenced by increasing coding intensity in Medicare Advantage. The focus of a revenue-minded organization tends to go to where it’s easiest to generate a profit. Organizations that run health plans know the easiest way to generate a profit is risk adjustment, so their incentive structures focus behavior on revenue.

Here’s a rhetorical question. What’s harder for a physician:

1) Documenting a patient’s diagnoses accurately? Or…

2) Ensuring a patient from a challenged socioeconomic background is adherent to the right clinical protocols?

We believe number two is harder, but number one is more financially rewarding within the current incentive structures. That means more effort might be spent on comprehensive diagnosis “capture,” and less on making sure at-risk individuals are adherent to their care plans. This does not align with the best interests of patients, the government, or tax payers.

Q: If Clover is right, then the entire HMO model is ill-conceived, and we all know that cannot be right.

Well, the notion that HMO (narrow) Networks always promote physician quality as their primary measurement tool is false. Here’s an example of why.

One of the typical Medicare plan measures of primary care physician quality is risk-adjusted Hospital Admissions per 1000 patients. Risk-adjusted, by definition, takes into account coding quality. Unfortunately, this presents a bias against physicians who practice in areas where there is higher participation in Medicaid (i.e., lower income communities), as these physicians tend to receive lower reimbursements per patient, resulting in overworked practices seeing more patients per day, exacerbated further by local-area shortages of PCPs.

“Low-income patients and doctors felt they did not have enough time together. PCPs expressed regret that they could not spend more time with patients and acknowledged that the limited visit time got in the way of high-quality primary care. They attributed the problem to declining and insufficient Medicaid reimbursement and a decreased number of providers.” -Commonwealth Fund

With less time per visit, doctors have to focus on the most clinically urgent activities while administrative tasks like coding and documentation fall to the wayside. While this prioritization is essential for the patients’ well-being at that moment, it ultimately results in less accurate, and lower, risk-adjusted payments for these physicians. This then flows through to even lower revenue and penalization on quality measures like risk-adjusted Hospital Admissions per 1000 patients.

So once again, essentially, HMO networks that claim to filter physicians by “quality” end up biasing towards physicians in wealthier areas, with more socioeconomically advantaged panels. With these patients, they can then increase risk adjustment revenue while avoiding the difficulties of addressing social determinants of health. Put another way, they are effectively creating careful selection bias to include in their networks those physician groups who code diagnoses well and serve only certain populations. This does not promote quality or equity.

Q: Clover is saying physicians should not be rewarded for lowering total cost of care; that’s definitely wrong and short-sighted.

Incorrect. Physicians should absolutely be aligned around long-term total cost of care reduction, to the extent it does not lead to a rationalization of services that harms patients.

Given that coding is the single biggest driver of costs to the government in Medicare Advantage, we believe that physicians should not be incentivized by plans, whether directly or indirectly, around specious actions that create more payments from the government to those health plans.

Physicians should be rewarded for looking after sicker, underserved populations. Sadly, under the way the system is structured today, they’re actually financially punished for doing so.

Q: Physicians cannot be trusted, so we must pay them for “value,” and we as health plans must dictate how that’s defined.

Do you really believe that physicians don’t strive to make decisions that are in the best interest of their patients? Do you believe that reimbursement structures dictated by insurance companies is what makes them adhere to their Hippocratic oath?

Physicians are overworked and in dire need of more resources to support them in making the best decisions for patients. They know what they need to achieve this — better clinical software and meaningful face time with each patient. Current systems, unfortunately, demand that physicians spend increasingly more time in coding and administration on behalf of plans.

Q: I understand the point Clover is trying to make, but there are many full capitation and risk-sharing arrangements that exist with organizations that focus on disadvantaged populations.

There are organizations that generate significant margins via “accurate risk adjustment” with contracts that specifically ensure they are directly incentivized to “code accurately.”

Reiterating our core point — accurately diagnosing a condition is an imperative that helps drive better care decisions. What Clover stands against is tying physician incentives to accurate diagnosis, as the knock-on effects end up driving up the cost of Medicare.

Q: Everything Clover says about HMO networks is totally untrue. STAR ratings prove that HMO networks score the highest under STAR ratings.

You don’t need to look much further than the following study: https://jamanetwork.com/journals/jama-health-forum/fullarticle/2781100

This cross-sectional study of 1,578,564 Medicare Advantage enrollees found that simulated star ratings for persons with lower SES (socioeconomic status) and Black and Hispanic enrollees were substantially lower than ratings for those with higher SES and White enrollees in the same contract […] Contracts with higher Medicare star ratings have larger racial, ethnic, and socioeconomic disparities in quality. These findings indicate that the MA star ratings may need to be modified to explicitly consider and reward equity in care.

STAR ratings have evolved to unintentionally reward organizations that focus on wealthier and healthier populations. STAR ratings actually exclude race and ethnicity from their measurement adjustments and reward better coding, which penalizes lower-income populations.