Seema Verma, MPH
Administrator
Centers for Medicare & Medicaid Services
Department of Health and Human Services
200 Independence Avenue, SW
Washington, DC 20201

..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..

•
•

August 12, 2019

RE: Request for comments on Medicare and Medicaid Programs; Risk Adjustment Data
Validation [CMS-4185-N4]
Dear Ms. Verma:
The Medicare Payment Advisory Commission (MedPAC) welcomes the opportunity to comment
on the Centers for Medicare & Medicaid Services’ (CMS’s) “Medicare and Medicaid Programs;
Risk Adjustment Data Validation” request for additional comment published in the Federal
Register, vol. 84, no. 125, pages 30983 to 30984. We appreciate your staff’s work to implement
the risk adjustment data validation (RADV) provisions and specifically to evaluate whether an
adjustment for unsupported fee-for-service (FFS) diagnoses is needed. We commend the
transparency of your staff’s careful consideration of the issues and the substantial release of
additional data.
We submit the following comment in addition to our December 19, 2018, comment on the initial
proposed rule [CMS-4185-P] addressing Medicare Advantage (MA) RADV provisions.1
Request for public comment
The MA risk adjustment model uses FFS spending, demographic, and diagnostic data to calculate
a set of risk adjustment coefficients that represent the relative amount of FFS spending related to
different demographic factors and diagnoses. These factors are applied to MA enrollees’
demographic and diagnostic data to estimate a risk score—a standardized expected spending
amount—that adjusts CMS’s payment to the plan. CMS tracks beneficiary demographic data, but
MA plans submit diagnoses for their enrollees to CMS for risk adjustment. Because diagnoses are
a major factor in determining plan payments, RADV audits are necessary for ensuring payment
accuracy by verifying that diagnoses submitted by MA plans are documented in patients’ medical
records. Through the audits, plans are held accountable for submitting diagnostic data that does not
conform to program rules.
1

Medicare Payment Advisory Commission. 2018. MedPAC comment letter on policy and technical changes to the
Medicare Advantage program. December 19. http://medpac.gov/docs/default-source/commentletters/12192018_ma_medpac_comment_letter_v2_sec.pdf?sfvrsn=0.

 Seema Verma
Administrator
Page 2

In a 2012 RADV audit methods memo, CMS stated that after calculating any overpayment
recovery amount warranted by the RADV audits, the agency would apply an FFS adjuster to that
amount to account “for the fact that the documentation standard used in RADV audits to determine
a contract’s payment error (medical records) is different from the documentation standard used to
develop the Part C risk-adjustment model (FFS claims).” 2 The expectation may have been that
MA improper payment estimates would decrease if the calculation of risk adjustment coefficients
was based only on the FFS diagnoses with medical record support (rather than all FFS diagnoses).
However, as part of the initial proposed rule addressing MA RADV provisions [CMS-4185-P],
CMS released the results of its study demonstrating that unsupported FFS diagnoses introduce no
systematic bias on MA risk scores and thus proposed not to apply an FFS adjuster. CMS
subsequently duplicated the study and released results confirming the initial finding. 3
The Secretary now requests additional public comment about whether statutory text related to risk
adjustment for payments to MA plans mandates an FFS adjuster, prohibits an FFS adjuster, or
otherwise informs the proposal not to apply an FFS adjuster in the RADV extrapolated audit
methodology.4
Comment
Without weighing in on the legal question of the applicability of the statutory provision, we
support CMS’s conclusion that an FFS adjuster is unwarranted in determining overpayment
recovery amounts identified through RADV audits.
The Commission supports including private plans in the Medicare program because they allow
beneficiaries to choose between FFS Medicare and alternative delivery systems that private plans
can provide. Plans often have flexibility in payment methods, including the ability to negotiate
with individual providers; care-management techniques that fill potential gaps in care delivery
(e.g., programs focused on preventing avoidable hospital readmissions); and robust information
systems that can potentially provide timely feedback to providers. Plans also can reward
beneficiaries for seeking care from more efficient providers and give beneficiaries more
predictable cost sharing; one trade-off is that plans typically restrict the choice of providers.
By contrast, traditional FFS Medicare has lower administrative costs and offers beneficiaries an
unconstrained choice of health care providers, but it lacks incentives to coordinate care and is
limited in its ability to modify care delivery. Because private plans and traditional FFS Medicare
2

Centers for Medicare & Medicaid Services, Department of Health and Human Services. 2012. RADV methodology.
https://www.cms.gov/Research-Statistics-Data-and-Systems/Monitoring-Programs/recovery-audit-program-parts-cand-d/Other-Content-Types/RADV-Docs/RADV-Methodology.pdf.
3
Centers for Medicare & Medicaid Services, Department of Health and Human Services. 2019. Medicare Risk
Adjustment Data Validation Program: Resources. https://www.cms.gov/Research-Statistics-Data-andSystems/Monitoring-Programs/Medicare-Risk-Adjustment-Data-Validation-Program/Resources.html.
4
The request for public comment cites 42 U.S.C. 1395w-23, in particular (a)(1)(c) “which requires risk adjustment in
subclause (a)(1)(c)(i), mandates a downward adjustment of risk scores in subclause (a)(1)(c)(ii), and includes
provisions about risk adjustment for special needs individuals with chronic health conditions in subclause
(a)(1)(c)(iii).”

 Seema Verma
Administrator
Page 3
have structural aspects that appeal to different segments of the Medicare population, the
Commission’s long-standing position is that the Medicare program should provide a financially
neutral choice between private MA plans and traditional FFS Medicare. Medicare’s payment
systems, as well as monitoring and enforcement efforts, should ensure parity between the two
sectors and not unduly favor one component of the program over the other.
In assessing parity between FFS Medicare and MA, two diagnostic coding issues come into play.
First, MA plans have a greater financial incentive than providers in FFS Medicare to report all
diagnoses codes for their enrollees. For the past several years, we have reported on the impact of
these financial incentives and on factors that lead to greater coding intensity in MA (e.g., the use of
home health risk assessments, chart reviews, pay-for-coding programs, and electronic health
records to improve data sharing and to remind providers to identify diagnoses).We estimate that
such coding resulted in MA risk scores (and payments) that were on average 7 percent greater than
comparable FFS beneficiaries in 2017 without a commensurate difference in the characteristics of
the MA enrollee relative to an otherwise comparable FFS beneficiary. 5 This excess coding results
in payments to MA plans that are higher than warranted given the enrollees health status relative to
comparable beneficiaries in FFS. The Congress has partially addressed this issue by mandating a
coding intensity adjustment, and the Commission has recommended a set of policies that would
more fully and equitably address the issue.
The second coding issue that has the potential to undermine financial parity between MA and FFS
Medicare is the matter discussed in the current request for comment pertaining to medical record
documentation requirements in MA and FFS. While CMS requires medical record documentation
for diagnoses submitted by MA plans, it does not have a similar documentation requirement for
diagnoses in FFS claims, which are the source of diagnostic information used to calculate risk
adjustment coefficients in the MA payment system. If the FFS data used to calculate risk
adjustment coefficients include diagnoses that are not documented in a patient’s medical record,
use of those coefficients could introduce a bias into the system because the coefficients would be
either greater or smaller than if the documentation requirement were applied equally to FFS and
MA. This potential for bias has led some stakeholders to call for a counterbalancing adjustment,
the FFS adjuster which is the subject of CMS’s current solicitation.
We believe a guiding principle for the risk adjustment system is that one set of risk coefficients
should be used for all payments to MA plans (i.e., the same set of risk coefficients should
determine monthly payment amounts for all MA plans and should be used to determine
overpayment amounts through RADV audits). Applying one set of coefficients to determine
monthly plan payment amounts and a different set of coefficients (or an FFS adjuster based on a
different set of coefficients) to determine RADV overpayment amounts would introduce
inconsistency in payments to plans and would undermine financial neutrality between FFS
Medicare and MA. Therefore, we reiterate our comment from December 19, 2018, agreeing with
CMS’s conclusion that if unsupported FFS diagnoses caused bias, the impact would affect all MA
risk scores such that an appropriate remedy would require a modification to the entire risk
5

Medicare Payment Advisory Commission. 2019. Report to the Congress: Medicare payment policy. Chapter 13: The
Medicare Advantage program: Status report. Washington, DC: MedPAC. http://medpac.gov/docs/defaultsource/reports/mar19_medpac_ch13_sec.pdf?sfvrsn=0.

 Seema Verma
Administrator
Page 4
adjustment system, not an adjustment only for the contracts subject to RADV audits. An
adjustment only to MA contracts under RADV audits would fail to fully address the potential bias.
Although there is a potential for bias arising from differing documentation standards, CMS’s study
shows that bias due to including undocumented FFS diagnoses is negligible (i.e., causing a less
than 0.1 percent increase in MA risk scores in both iterations of the study). The negligible effect on
MA risk scores is likely due to the fact that (1) the risk adjustment model attributes Medicare
spending to the variables in the model so that all FFS spending is accounted for, and (2) the FFS
and MA populations are substantially similar with regard to the distribution of risk adjustment
coefficients.
Based on the results of CMS’s study, we conclude that CMS should not make an adjustment to the
entire risk adjustment system. Implementing an adjustment to remove unsupported FFS diagnoses
from the risk adjustment system in each year would require significant effort and would offer
negligible benefit. Because MA risk scores using coefficients with and without unsupported FFS
diagnoses are equivalent in aggregate, it is appropriate for the Secretary to require medical record
documentation of MA diagnoses and enforce program rules through extrapolated RADV audits.
Conclusion
The Commission values the ongoing cooperation and collaboration between CMS and our staff on
technical policy issues. We look forward to continuing this productive relationship. If you have
any questions, or require clarification of our comments, please feel free to contact James E.
Mathews, the Commission’s Executive Director, at 202-220-3700.
Sincerely,

Francis J. Crosson, M.D.
Chairman