Because increased risk scores result in higher payments to Medicare Advantage organizations, some stakeholders have alleged that the risk adjustment payment methodology encourages providers and plans to inappropriately upcode or inflate patients’ health risks to receive greater compensation. More specifically, critics allege that the methodology creates an incentive to meticulously document patients’ conditions to produce a higher risk score that will in turn produce higher payments, which has been dubbed “coding intensity.”³ As a result, there is often a marked difference in the risk scores of Medicare Advantage and traditional Medicare beneficiaries. For example, in 2020, the Medicare Payment Advisory Commission (MedPAC) estimated that Medicare Advantage beneficiaries’ risk scores were approximately 9.5% higher than similarly situated traditional Medicare beneficiaries. Consequently, in MedPAC’s estimation, the risk-adjusted payment structure led to roughly $12 billion in overpayments to Medicare Advantage plans.⁴

In its March 2023 Comment Letter⁵ to CMS regarding the Advance Notice, MedPAC supported policy changes outlined in the Advance Notice to reduce coding differentials between Medicare Advantage and traditional Medicare beneficiaries’ risk scores. Specifically, MedPAC approved of the proposal to eliminate or constrain the coefficients of certain HCCs that CMS had identified as most subject to intentional or unintentional discretionary coding variation or inappropriate coding by health plans and providers. MedPAC stated, “[w]hen diagnostic discretion, intentional or unintentional, leads to large differences in the coding rates between [Medicare Advantage] and [traditional Medicare] . . . it diminishes the accuracy of risk-adjusted payments . . . and increases the excess payments that [Medicare Advantage] plans receive due to higher coding intensity.”⁶

Common risk adjustment allegations

I. Upcoding and the disparity in diagnosis-based risk scores between Medicare Advantage and fee-for-service Medicare beneficiaries

Researchers have conducted studies on the disparity in risk scoring between traditional Medicare and Medicare Advantage beneficiaries, finding that Medicare Advantage beneficiaries’ risk scores invariably are higher than similarly situated traditional Medicare beneficiaries. For example, in February 2017, Richard Kronick determined that the average risk score for Medicare Advantage beneficiaries had risen by approximately 1.5% each year as compared to the average risk score of a traditional Medicare beneficiary in the previous 10 years. Kronick attributed these changes to identification of additional diagnoses (i.e., coding intensity) and not changes in Medicare Advantage beneficiaries’ actual health. He hypothesized that the growth rate of coding intensity could decrease over time, reaching 0.75% per year in 2026; however, he noted that this hypothesis was limited by a lack of evidence of deceleration of the growth rate at the time of publication.⁷ In November 2021, Kronick and another researcher, F. Michael Chua, conducted a similar review and concluded that coding intensity, or upcoding, increased average Medicare Advantage risk scores by approximately 2% from 2017-2019.⁸

In its March 2022 Report to Congress (MedPAC Report), MedPAC reiterated similar claims to those above, noting that “documenting additional diagnosis codes raises [beneficiaries’] risk scores, generating two distinct benefits for [Medicare Advantage] plans: (1) . . . boost[ing] the monthly payment amount a plan receives, and (2) . . . increas[ing] the rebate amount a plan uses to provide extra benefits to enrollees, thereby giving plans . . . a competitive advantage over other plans.”⁹ In turn, plans that engage in greater coding intensity gain unfair competitive advantages over other plans. MedPAC highlighted several methods that plans and providers have employed to facilitate information access and documentation of diagnoses in the MedPAC Report, including:

• Providers grant plans access to their electronic medical records, and, in exchange, through capitated arrangements, the plans pay the providers a risk-adjusted sum for each enrollee, which passes the financial incentive to upcode onto the physicians.
• Through chart reviews and pay-for-coding programs, plans provide physicians with a patient assessment, including diagnosis codes, which the plans have identified through chart reviews; ask the provider to verify the diagnoses; and if verified, include them in future claims.¹⁰

As described in more detail below, in recent years, chart reviews have been heavily litigated on a variety of claims, including that such reviews are used to identify new diagnoses but not to verify the accuracy of previously submitted diagnoses, even in cases when the Medicare Advantage plan knows that the diagnoses are not supported by documentation or episodes of care.¹¹

II. Vertical integration and increased risk of upcoding

Vertical integration has become increasingly pervasive in many aspects of the health care system, including integration of providers and health systems, health insurers and pharmacy benefit managers (PBMs), and Medicare Advantage plans and other related businesses. A March 2023 article from the Brookings Institution noted that health plan spending directed to related businesses that are owned by the health plans’ parent companies amounted to roughly 65% of Medicare Advantage spending in 2022.¹² Further, the authors found that gross and risk-adjusted spending by Medicare Advantage plans that are owned by parent companies, which own related businesses, increased 11% and 6.9%, respectively, from 2016-2019.

In their March 2020 article in the Journal of Political Economy, Michael Geruso and Timothy Layton claim that vertical integration of plans and providers, such as provider-owned plans, increases the risk of upcoding. Geruso and Layton found that beneficiaries of fully vertically integrated plans have 16% higher risk scores as compared to similarly situated traditional Medicare beneficiaries.¹³ Geruso and Layton state that this “suggests that the cost of aligning physician incentives with insurer objectives may be significantly lower in vertically integrated firms [and] may facilitate the gaming of health insurance payment systems.”¹⁴ Geruso and Layton highlighted that vertical integration may allow the plan to exert greater control over a provider’s coding practices, especially in instances where a provider’s compensation is tied to risk scores produced by that provider’s coding and documentation practices.¹⁵

Such incentives have been subject to litigation in recent years. For example, in United States v. Anthem, Inc., No. 10-CV-2593 (ALC) (S.D.N.Y. July 2, 2020), the government alleged that Anthem was party to profit-sharing agreements with physicians pursuant to which Anthem shared a percentage of its risk adjustment payments with the contracted providers, creating a financial incentive for providers to upcode patients’ diagnoses as well as the severity of those diagnoses. Similarly, in United States ex rel. Osinek v. Kaiser Permanente, No. 3:13-CV-03891-EMC (N.D. Cal. Oct. 7, 2021) (the Kaiser Complaint-in-Intervention and, together with the Amended Complaint, the Kaiser Matter),¹⁶ Kaiser Permanente employed several different methods of financial incentives to providers if and when patients’ risk scores increased or remained stable, including arrangements through which capitated providers received (1) a percentage of additional revenue due to increased premiums, (2) flat capitated and fee-for-service providers received bonuses, and (3) incentives and coding competitions related to the approval of a certain number of diagnosis codes or high-risk patient scores.

III. Use of retrospective record reviews and health risk assessments and consequential upcoding

Medical record reviews and health risk assessments have been identified as areas of particular concern from the perspectives of the Department of Health and Human Services Office of the Inspector General (OIG) and MedPAC. In MedPAC’s estimation, in 2017, health risk assessments and chart reviews were responsible for generating 4.6% of total payments to Medicare Advantage plans¹⁷ and 64% of Medicare Advantage coding intensity.¹⁸ Similarly, the OIG determined that plans and providers may use chart reviews as a way to bypass the face-to-face requirement and inflate risk-adjusted payments.¹⁹ From a review of 2016 Medicare Advantage encounter data, OIG found that:

• Plans used chart reviews more than 99% of the time as a method of adding, rather than deleting, diagnoses.
• Diagnoses reported only on chart reviews, and not in patients’ records of services rendered, totaled approximately $6.7 billion in risk-adjusted payments for 2017.
• Roughly $2.7 billion in risk-adjusted payments were linked to chart review diagnoses rather than services rendered and diagnoses made during a face-to-face visit.
• Although only in a limited number of cases, nearly 50% of the Medicare Advantage plans received payments pursuant to chart reviews in which the plan had no record of providing services to the beneficiary in 2016.

Use of these practices has been questioned by a number of courts in recent years in light of the perceived incentives created by the risk adjustment payment model. The government alleged in United States ex rel. Ormsby v. Sutter Health, 444 F. Supp. 3d 1010, 1023 (N.D. Cal. 2020) that Sutter Health and its affiliated Palo Alto Medical Foundation violated the False Claims Act through several methods they employed to maximize the number of risk-adjusting diagnosis codes reported to CMS. Such methods included: (i) preparation of annual visit reminders/forms for physicians to send to patients to encourage patients to schedule visits intended to capture diagnosis codes; (ii) supplying providers with lists of common risk-adjusting diagnosis codes (i.e., “cheat sheets”) to the exclusion of others; (iii) prompts in patients’ electronic medical records and prepopulated lists of diagnosis codes, which made it easier for providers to select and approve diagnosis codes that were preselected by the plan; and (iv) daily alerts that included a list of HCC codes that had not yet been captured during the year for the patient. In the Kaiser Matter, the court questioned Kaiser Permanente’s chart review practices, which appeared to focus exclusively on high-value diagnosis codes or codes that lead to risk adjustment payments and be dedicated to looking for documentation of diagnosis codes for new conditions rather than reviewing for accuracy or repetition. Further, Kaiser Permanente employed an addendum process that providers were encouraged to use to add diagnosis codes to patients’ medical records retrospectively, and in some cases, a year or more after a patient visit.

Implications for Medicare Advantage risk adjustment

As outlined above, chart reviews and health risk assessments were the primary focus of the No UPCODE Act in light of the perceived risks that these activities present for coding intensity and overpayments. Notably, the No UPCODE Act incorporates two of the three recommendations proposed in the MedPAC Report: (i) use of two years of diagnostic data in calculating the risk-adjusted payment rates; and (ii) removal of diagnoses collected from a chart review or health risk assessment when calculating risk adjustments.²⁰ As common themes raised by MedPAC and legislators alike, we expect that such recommendations will be a focus of discussions as the No UPCODE Act moves through the legislative process.

Nevertheless, the focus on these methods and the risk-adjusted payment methodology raises some important questions. Particularly, are these fair criticisms of the payment methodology or are plans and providers working within the confines of the system that was created for them? And, is the issue really upcoding in the Medicare Advantage program or are the coding practices in traditional Medicare less stringent?

Traditional Medicare is a fee-for-service system, so it does not use capitated payments and risk adjustment to account for patients’ individual health status. While diagnosis codes are collected in traditional Medicare, they are of lesser importance in the clinical documentation, claims submission, and reimbursement processes. Moreover, while coding intensity may be a product of the Medicare Advantage risk adjustment payment system, it is not necessarily a fair criticism to use this practice as evidence of providers’ and plans’ wrongdoing for several reasons. First, the payment system requires documentation of diagnoses as a condition for reimbursement. Therefore, it is not unreasonable for plans and providers to develop tools to capture these diagnoses to ensure that patient records reflect the risks the plans are undertaking. Second, these diagnoses are not fabricated. Beneficiaries are living with these conditions, and plans and providers should be appropriately compensated for the care that these patients require to manage their existing conditions. In sum, in addition to taking a second look at the coding practices in the Medicare Advantage program, it may also be time to address the accuracy and depth of coding practices in fee-for-service Medicare.

If passed, the No UPCODE Act will dramatically alter plans and providers’ current practices with respect to documentation of diagnoses. Plans and providers will necessarily need to be more judicious in their coding practices, and so will be unlikely to receive the same types of risk-adjusted payments. As a result, patients may experience a reduction in care due to a loss of the financial incentive for plans and providers to identify, document, and provide services for all of a patient’s conditions. Providers may reduce the amount of care provided to reduce the impact on the provider’s and plan’s profit margins. However, assuming that a provider/plan did not unnecessarily document a patient’s conditions, and they have satisfied all of Medicare’s conditions for payment, there should be limited impact on the reimbursement. Passage of the No UPCODE Act could also lead plans to implement more stringent approval criteria for their beneficiaries as a cost-saving measure, thereby resulting in loss of coverage or lack of coverage options for sicker or more medically fragile patients.

As noted above, the No UPCODE Act was introduced on March 28, 2023 as Senate Bill 1002. It was read twice in the Senate and then referred to the Committee on Finance. As of April 19, 2023, the No UPCODE Act remains in committee and no further action has been taken. Assuming it is referred out of committee, the bill will need to pass a vote in the Senate, pass a vote in the House, and be signed into law by President Biden.²¹ Should the No UPCODE Act become law, providers and plans may contend with more stringent documentation and coding practices either as a direct result of the No UPCODE Act or as a result of plans’ responses to new obligations imposed by CMS pursuant to the provisions of the No UPCODE Act.

The authors will continue to track developments regarding the No UPCODE Act and the impact of policy decisions on Medicare Advantage reimbursement and associated practical considerations. Please reach out to the authors or other health care attorneys at Reed Smith if you have any questions about the No UPCODE Act or have other Medicare Advantage reimbursement concerns.

1. See No Unreasonable Payments, Coding, or Diagnoses for the Elderly Act, available via the PDF (last accessed Apr. 10, 2023).
2. 42 U.S.C. § 1395w-23(a)(1)(C)(ii).
3. See MedPAC Medicare Payment Policy Report to Congress (Mar. 2022), available via the PDF (last accessed Apr. 10, 2023).
4. See Commonwealthfund.org; see also MedPAC January 2022 Public Meeting Transcript (Jan. 14, 2022) at p. 90, available via the PDF (last accessed Apr. 10, 2023).
5. See generally, MedPAC Comment on CMS Advance Notice of Methodological Changes for CY 2024 for Medicare Advantage Capitation Rates and Part C and D Payment Policies (Mar. 1, 2023), available via the PDF (last accessed Apr. 10, 2023).
6. Id. at p. 5.
7. See Projected Coding Intensity In Medicare Advantage Could Increase Medicare Spending By $200 Billion Over Ten Years (Feb. 2017), available at healthaffairs.org (last accessed Apr. 10, 2023).
8. See Industry-Wide and Sponsor-Specific Estimates of Medicare Advantage Coding Intensity (Nov. 17, 2021), available online (last accessed Apr. 10, 2023).
9. MedPAC Report, supra note 3, at p. 433.
10. Id. at p. 437.
11. Id. at p. 438.
12. See Medicare Advantage Spending, Medical Loss Ratios, and Related Businesses: An Initial Investigation (Mar. 24, 2023), available at brookings.edu (last accessed Apr. 11, 2023).
13. See Upcoding: Evidence from Medicare on Squishy Risk Adjustment (Mar. 2020) at p. 989, available online at University of Chicago Journals (last accessed Apr. 10, 2023).
14. Id. at p. 989.
15. Id. at p. 1009-1010.
16. See also United States ex rel. Osinek v. Kaiser Permanente, No. 3:13-CV-03891-EMC (N.D. Cal. Oct. 25, 2021) (the Amended Complaint).
17. MedPAC Comment, supra note 5, at p. 19.
18. MedPAC Report, supra note 3, at p. 437.
19. See Billions in Estimated Medicare Advantage Payments From Chart Reviews Raise Concerns (Dec. 2019), available at Office for Inspector General (last accessed Apr. 10, 2023).
20. MedPAC Report, supra note 3, at p. 442.
21. To determine the current status of the No UPCODE Act, visit congress.gov.

In-depth 2023-096