/>iOn March 25, 2025, U.S. Senators Bill Cassidy, M.S. (R-LA) and Jeff Merkley (D-OR) introduced the No Unreasonable Payments, Coding, or Diagnoses for the Elderly (No UPCODE) Act (the “Bill”).
According to Senator Cassidy’s press release, the Bill aims to improve how Medicare Advantage plans evaluate patients’ health risks, reduce overpayments for care, and save taxpayers money by removing incentives to overcharge Medicare. If passed, this Bill would have a tremendous impact on plans, vendors, and risk-bearing provider groups relative to Medicare Advantage (“MA”).
Background
Traditional Medicare (Parts A and B) reimburses health care providers based on the cost of services already rendered (known as “Fee-for-Service” or “FFS”). Conversely, MA functions as a prospective payment model, whereby Medicare Advantage Organizations (“MAOs”) contract with the Centers for Medicare & Medicaid Services (“CMS”) to administer and insure their respective member population.
MAOs are paid what amounts to be a fixed amount per member per month that reflects the projected health care costs of a given MAO’s members. MA payments are adjusted by CMS using a risk adjustment (“RA”) model, which weighs the relative health care financial risks of each member based on (1) his or her health status and (2) demographic information. Broadly and logically speaking, sicker members and/or those with more conditions are at risk for higher medical spend. Accordingly, MAOs are paid more under the RA model to address those risks.
This model is not without its criticism on both sides of the table. Health plans generally feel like changes to the payment model have a significant impact typically in the negative direction. For example, the most recent update to the V28 model has gutted funding for a number of cost-driving conditions. And budget hawks take the opposite view, i.e., that MA is, to some extent, overfunded and subject to overcoding and abuse. Nonetheless, the aim of the risk adjustment model (and MA/value-based care generally) is to ensure that MAOs and other MA entities are not deterred from caring for sicker patients.[1]
Senator Cassidy’s Bill purports to address concerns regarding perceived incentives that could lead to exaggerated health conditions aimed at securing improper payments.
Breakdown of the No UPCODE Act
If passed, the Bill would amend title XVII of the Social Security Act, starting in 2026, and implement the following changes to the MA program:
- The Secretary of the U.S. Department of Health & Human Services (“HHS”), and under its umbrella, CMS, would use two years of patient diagnostic data in its RA model, instead of the existing one-year model;
- Patient diagnoses identified through “chart reviews” and “health risk assessments” (“HRAs”) would be excluded entirely from the RA model; and
- The Secretary of HHS would be required to assess the impact of coding differences between MAOs and traditional Medicare providers on risk scores, and to publicly report the findings, ensuring adjustments account for unrecognized coding pattern differences.
Implications on the Medicare Advantage Industry:
This is not the first time this Bill has been introduced in Congress, as Senators Cassidy and Merkley proposed an essentially identical bill in late March 2023 that did not lead to further congressional action.
The three components of this Bill would have a sizeable impact on the MA industry, as described in greater detail below.
Using Two Years of Diagnostic Data
The Bill requires CMS to use two years, rather than one year, of risk-adjusted data when determining payment. This facet of the Bill represents potential good news for MAOs, which spend time and money each year ensuring that chronic conditions are coded and submitted by a provider annually. The MA risk adjustment model, from a financial standpoint, currently “cures” all risk-adjusted diseases at the stroke of midnight each December 31, requiring annual recapture of diseases—even those that are chronic and/or do not resolve.
As discussed below, however, this may be inseparable from the other facets of the Bill, which render the upside of two years versus one year of data dubious. At a high level, requiring this change would appear to positively financially affect MAOs and MA value-based entities (“VBEs”), as their members/patients would now only need to have a Hierarchical Condition Category (“HCC”) attributed to them in just one of the prior two years to predict costs and trigger payment under the new RA model.[2]
The unintended consequence might be a reduction in patient care. The annual requirement of reconfirming diagnoses in many ways forces intervention by plans and their contractors. These take the form of in-home assessments, various gap closure interventions, and similar activities to effectuate a qualifying encounter with a licensed provider. As such, this change might have the unintended consequence of less care provided to MA beneficiaries, as MAOs and value-based entities may be disincentivized from identifying diagnoses and documenting them in the medical record on a yearly basis, as currently required, which in part incentivizes at least these annual provider encounters. This is especially true if in-home assessments are effectively hobbled by this Bill.
Notably, in its June 2020 Report to Congress, the Medicare Payment Advisory Commission (“MedPAC”) concluded from its study that using two years of diagnosis data to determine beneficiaries’ conditions produces payment adjustments that are about as accurate as using one year of diagnosis data, though it produces larger underpayments for those with high levels of Medicare spending than using one year of diagnosis data, which MedPAC attributes to lower coefficients for HCCs in the two-year RA model and high prediction-year (base year) spending.[3]
This means that not only did MedPAC determine that the two-year RA model would not substantively improve the predictive accuracy of the existing one-year RA model, but it would also likely result in greater underpayments to MAOs (and indirectly, VBEs) serving higher-risk member populations. MAOs and VBEs would naturally be reluctant to continue to serve those high-risk populations if they are not adequately compensated. This could lead to a reduction in members and potential increased utilization and government spending under traditional Medicare.
The 21st Century Cures Act in 2016 permitted CMS to use at least two years of diagnostic data, but CMS has yet to do so in nearly ten years. Doing so now would have a profound (and perhaps even unascertainable at this time) impact on the MA industry.
Excluding Diagnoses Identified from Chart Reviews and HRAs
Chart reviews, or so called “retrospective chart reviews” of medical records are utilized to help ensure member medical diagnoses—which are documented in the medical record, but for a variety of reasons are not submitted to health plans—are appropriately reflected in CMS data. Dropped or missing codes on claims data is exceptionally common. So common, in fact, that it birthed a private equity backed risk adjustment vendor industry. Providers typically focus on service level coding such as EM codes and CPT coding, not ICD diagnostic coding, and retrospective reviews are the last chance that plans have to ensure accurate data is submitted to CMS.
The prospective RA model, built to be an actuarial equivalent of FFS Medicare, requires MAOs to document all conditions to receive accurate payment. It stands to reason that MAOs (and VBEs) should utilize tools like retrospective chart review to capture valid conditions diagnosed and appropriately documented by a provider and submit member data that comprehensively reflects the risks the MAOs are assuming on behalf of CMS.
Should this Bill become law, plans and risk-bearing organizations that rely on retrospective review could be materially and negatively impacted. It would also likely serve as a death blow to a number of retrospective solutions vendors that provide such services.
HRAs differ significantly from chart review. HRAs are an established and longstanding practice in the MA care coordination model that are used to identify gaps in care, identify chronic conditions early, and prevent those conditions from worsening, causing comorbidities, or becoming more costly. These may occur in a patient’s home (often a vital component of comprehensive “In-Home Assessments” or “IHAs,” which are sometimes conflated with HRAs) or in the provider’s office. HRAs have evolved over the course of the MA program, as CMS has issued standards and best practices for conducting HRAs, and it continues to more broadly utilize other tools to ensure payment and data accuracy in MA (and in diagnosis codes identified via HRAs specifically), including through Risk Adjustment Data Validation (“RADV”) audits and HCC coefficient and model changes.
However, HRAs and IHAs have become the lightning rod for risk adjustment criticism by politicians, government agencies, and the media. Tapping into zeitgeist and perceived momentum, the Bill seeks to disqualify their use in the MA RA model. CMS has already taken a contrary view. CMS has, and continues to take, a balanced approach to these interventions and evaluated the appropriateness of using HRAs in MA. It has consistently supported their use, including in August 20, 2021 and September 5, 2024 letters from former CMS Administrator Chiquita Brooks-LaSure to the HHS Office of Inspector General (“OIG”) responding to OIG’s report that certain MAOs used HRAs to drive a disproportionate share of risk adjustment payments. Notably, in the 2024 CMS letter, CMS stated that it did not concur with OIG’s recommendation to “restrict the use of diagnoses reported only on in-home HRAs or chart reviews linked to in-home HRAs for risk-adjusted payments,” as it “allows MA organizations to use HRAs as a source of diagnoses used for the calculation of risk adjusted payments, as long as those diagnoses meet CMS’s criteria for risk adjustment eligibility.” To reinforce this point, CMS later addresses that it “will continue its efforts to conduct RADV audits to inform our understanding of the accuracy of these diagnoses.” CMS has also pointed out that despite OIG’s concern that diagnoses obtained from in-home HRAs may be inaccurate, OIG has “not conducted medical record reviews of the diagnoses that came from visits that may have contained an HRA and have not concluded that these diagnoses are not accurate.”
In addition to this support for HRAs in MA, CMS requires that Dual Eligible Special Needs Plans (“D-SNP”) use HRAs. If the Bill passes, it would likely cause confusion and potential misallocation of resources for MAOs with multiple plan products that have inconsistent legal requirements for HRAs. Additionally, similar to the above risks with excluding chart review-derived diagnosis codes from the RA model, if this Bill passes, MAOs and VBEs may forgo using HRAs altogether and instead initiate more basic encounters (like a less comprehensive PCP visit), which may lead to a deficiency in identifying and submitting diagnoses to CMS for payment under the RA model and therefore insufficient funds to manage the risk of the MAO’s member population.
Notably, this Bill charges the Secretary of HHS to “establish procedures to provide for the identification and verifications of diagnoses collected from chart reviews and health risk assessments,” but it does not give any other details, so it is unclear how these processes will be defined moving forward.
Adjustments in Coding Between FFS Medicare and Medicare Advantage
If passed, the Bill also requires CMS to calculate and publish the difference in coding growth between FFS Medicare and MA and set the adjustment to “fully account[] for the impact of coding pattern differences.” Currently, CMS applies a “minimum adjustment” of 5.9% for CY2025 pursuant to Section 1853(a)(1)(C)(ii)(III) of the Social Security Act, but this calculated adjustment would replace the minimum adjustment.
Importantly, CMS “may include [this coding intensity] adjustment on a plan or contract level,” meaning CMS could implement the adjustment differently among plans and contract years. This would create even more uncertainty for MA stakeholders.
Ultimately, given the wide-ranging implications this Bill will likely have on the MA industry and its members, MAO, VBEs, and other MA stakeholders should be monitoring this legislation and similar aims to materially change the MA program.
ENDNOTES
[1] It should also be noted that MAOs are bound to a percentage cap on profits, as Medical Loss Ratio (“MLR”) rules require MAOs to pay 85% of their revenue towards claims experience or quality improvement activities, with the remaining 15% allocated to various large administrative costs before profits are factored.
[2] For example, if projecting 2025 costs, DOS years 2023 and 2024 would be analyzed under the new RA model, and the MA beneficiary would need to have an HCC in just one of those years for the MAO to receive payment for that HCC.
[3] See Medicare Payment Advisory Commission (“MedPAC”) June 2020 Report to Congress. A separate, 2019 NIH study similarly found that incorporating additional years of clinical information into a Veterans Affairs prospective risk adjustment model did not result in a material increase in fit or predictive capability.
