Proposed Rule2021-28317
Patient Protection and Affordable Care Act; HHS Notice of Benefit and Payment Parameters for 2023
Primary source
Metadata and text below are from the Federal Register, a public-domain U.S. government work. Always verify the official published version before relying on it for any legal matter.
Published
January 5, 2022
Issuing agencies
Health and Human Services Department
Abstract
This proposed rule includes proposed payment parameters and provisions related to the risk adjustment and risk adjustment data validation programs, as well as proposed 2023 user fee rates for issuers offering qualified health plans (QHPs) through federally- facilitated Exchanges and State-based Exchanges on the Federal platform. This proposed rule also proposes requirements related to prohibiting discrimination based on sexual orientation and gender identity; guaranteed availability; the offering of QHP standardized options through Exchanges on the Federal platform; requirements for agents, brokers, web-brokers, and issuers assisting consumers with enrollment through Exchanges that use the Federal platform; verification standards related to employer sponsored coverage; Exchange eligibility determinations during a benefit year; special enrollment period verification; cost-sharing requirements; Essential Health Benefits (EHBs); Actuarial Value (AV); QHP issuer quality improvement strategies; accounting for quality improvement activity (QIA) expenses and provider incentives for medical loss ratio (MLR) reporting and rebate calculation purposes; re-enrollment, and requirements related to a new State Exchange improper payment measurement program. This proposed rule also seeks comment on how HHS can advance health equity through QHP certification standards and otherwise in the individual and group health insurance markets, and how HHS might address plan choice overload in the Exchanges.
Full Text
<html>
<head>
<title>Federal Register, Volume 87 Issue 3 (Wednesday, January 5, 2022)</title>
</head>
<body><pre>
[Federal Register Volume 87, Number 3 (Wednesday, January 5, 2022)]
[Proposed Rules]
[Pages 584-728]
From the Federal Register Online via the Government Publishing Office [<a href="http://www.gpo.gov">www.gpo.gov</a>]
[FR Doc No: 2021-28317]
[[Page 583]]
Vol. 87
Wednesday,
No. 3
January 5, 2022
Part III
Department of Health and Human Services
-----------------------------------------------------------------------
45 CFR Parts 144, 147, 153, et al.
Patient Protection and Affordable Care Act; HHS Notice of Benefit and
Payment Parameters for 2023; Proposed Rule
Federal Register / Vol. 87 , No. 3 / Wednesday, January 5, 2022 /
Proposed Rules
[[Page 584]]
-----------------------------------------------------------------------
DEPARTMENT OF HEALTH AND HUMAN SERVICES
45 CFR Parts 144, 147, 153, 155, 156 and 158
[CMS-9911-P]
RIN 0938-AU65
Patient Protection and Affordable Care Act; HHS Notice of Benefit
and Payment Parameters for 2023
AGENCY: Centers for Medicare & Medicaid Services (CMS), HHS.
ACTION: Proposed rule.
-----------------------------------------------------------------------
SUMMARY: This proposed rule includes proposed payment parameters and
provisions related to the risk adjustment and risk adjustment data
validation programs, as well as proposed 2023 user fee rates for
issuers offering qualified health plans (QHPs) through federally-
facilitated Exchanges and State-based Exchanges on the Federal
platform. This proposed rule also proposes requirements related to
prohibiting discrimination based on sexual orientation and gender
identity; guaranteed availability; the offering of QHP standardized
options through Exchanges on the Federal platform; requirements for
agents, brokers, web-brokers, and issuers assisting consumers with
enrollment through Exchanges that use the Federal platform;
verification standards related to employer sponsored coverage; Exchange
eligibility determinations during a benefit year; special enrollment
period verification; cost-sharing requirements; Essential Health
Benefits (EHBs); Actuarial Value (AV); QHP issuer quality improvement
strategies; accounting for quality improvement activity (QIA) expenses
and provider incentives for medical loss ratio (MLR) reporting and
rebate calculation purposes; re-enrollment, and requirements related to
a new State Exchange improper payment measurement program. This
proposed rule also seeks comment on how HHS can advance health equity
through QHP certification standards and otherwise in the individual and
group health insurance markets, and how HHS might address plan choice
overload in the Exchanges.
DATES: To be assured consideration, comments must be received at one of
the addresses provided below, no later than 5 p.m. on January 27, 2022.
ADDRESSES: In commenting, please refer to file code CMS-9911-P.
You may submit comments in one of three ways (please choose only
one of the ways listed):
1. Electronically. You may submit electronic comments on this
regulation to <a href="http://www.regulations.gov">http://www.regulations.gov</a>. Follow the ``Submit a
comment'' instructions.
2. By regular mail. You may mail written comments to the following
address ONLY: Centers for Medicare & Medicaid Services, Department of
Health and Human Services, Attention: CMS-9911-P, P.O. Box 8016,
Baltimore, MD 21244-8016.
Please allow sufficient time for mailed comments to be received
before the close of the comment period.
3. By express or overnight mail. You may send written comments to
the following address ONLY: Centers for Medicare & Medicaid Services,
Department of Health and Human Services, Attention: CMS-9911-P, Mail
Stop C4-26-05, 7500 Security Boulevard, Baltimore, MD 21244-1850.
For information on viewing public comments, see the beginning of
the SUPPLEMENTARY INFORMATION section.
FOR FURTHER INFORMATION CONTACT:
Jeff Wu, (301) 492-4305, Rogelyn McLean, (301) 492-4229, Grace
Bristol, (410) 786-8437, Sara Rosta, (301) 492-4223, or Kaye Wells,
(301) 492-4301, for general information.
Cam Moultrie Clemmons, (206) 615-2338, or Anthony Galace, (301)
492-4400, for matters related to past-due premiums.
Allison Yadsko, (410) 786-1740, John Barfield, (301) 492-4433, or
Jacqueline Wilson, (301) 492-4286 for matters related to risk
adjustment or risk adjustment data validation (HHS-RADV).
Aaron Franz, (410) 786- 8027, or John Barfield, (301) 492-4433, for
matters related to federally-facilitated Exchange (FFE) and State-based
Exchange on the Federal platform (SBE-FP) user fees.
Nora Simmons, (410) 786-1981, for matters related to advance
payment of the premium tax credit (APTC) proration.
Aaron Franz, (410) 786- 8027, or Hi'ilei Haru, 301-492-4363, for
matters related to cost-sharing reduction reconciliation.
Josh Van Drei, (410) 786-1659, for matters related to actuarial
value (AV).
Becca Bucchieri, (301) 492-4341, for matters related to essential
health benefit (EHB)-benchmark plans and defrayal of state-required
benefits.
Marisa Beatley, (301) 492-4307, for matters related to employer
sponsored coverage verification.
Susan Kalmus, (301) 492-4275, for matters related to agent, broker,
and web-broker guidelines. Dena Nelson, 240-401-3535, or Carly Rhyne,
301-492-4188, for matters related to income calculation for eligibility
for advance payments of premium tax credits.
Katherine Bentley, (301) 492-5209, or Ariel Kennedy, (301) 492-
4306, for matters related to special enrollment period verification.
Leigha Basini, (301) 492-4380, for matters related to
nondiscrimination based on sexual orientation and gender identity; and
EHB nondiscrimination.
Christina Whitefield, (301) 492-4172, for matters related to the
medical loss ratio (MLR) program.
Nidhi Singh Shah, (301) 492-5110, for matters related to quality
improvement strategy standards for Exchanges.
Erika Ourisman, (301) 492-4170, for matters related to downstream
and delegated entities.
Nikolas Berkobien, (301) 492-4400, or Leigha Basini, (301) 492-4380
for matters related to standardized options.
Erika Melman, (301) 492-4348, Deborah Hunter, (443) 386-3651, or
Whitney Allen, (667) 290-8748, for matters related to network adequacy
and essential community providers.
Linus Bicker, (803) 931-6185, for matters related to State Exchange
improper payment measurement.
Phuong Van, (202) 570-5594, for matters related to advancing health
equity through qualified health plans (QHPs).
Angelica Torres-Reid, (410) 786-1721, and Robert Yates, (301) 492-
5151, for matters related to State Exchange general program integrity
and oversight requirements.
Zarah Ghiasuddin, (301) 492-4308, for matters related to re-
enrollment in the Exchanges.
SUPPLEMENTARY INFORMATION: Inspection of Public Comments: All comments
received before the close of the comment period are available for
viewing by the public, including any personally identifiable or
confidential business information that is included in a comment. We
post comments received before the close of the comment period on the
following website as soon as possible after they have been received:
<a href="http://www.regulations.gov">http://www.regulations.gov</a>. Follow the search instructions on that
website to view public comments. CMS will not post on <a href="http://Regulations.gov">Regulations.gov</a>
public comments that make threats to individuals or institutions or
suggest that the individual will take actions to harm the individual.
CMS continues to encourage individuals not to submit duplicative
comments. We will post acceptable comments from multiple unique
commenters even if the content is identical or nearly identical to
other comments.
[[Page 585]]
Table of Contents
I. Executive Summary
II. Background
A. Legislative and Regulatory Overview
B. Stakeholder Consultation and Input
C. Structure of Proposed Rule
III. Provisions of the Proposed HHS Notice of Benefit and Payment
Parameters for 2023
A. Part 144--Requirements Relating to Health Insurance Coverage
B. Part 147--Health Insurance Reform Requirements for the Group
and Individual Health Insurance Markets
C. Part 153--Standards Related to Reinsurance, Risk Corridors,
and Risk Adjustment
D. Part 155--Exchange Establishment Standards and Other Related
Standards Under the Affordable Care Act
E. Part 156--Health Insurance Issuer Standards Under the
Affordable Care Act, Including Standards Related to Exchanges
F. Part 158--Issuer Use of Premium Revenue: Reporting and Rebate
Requirements
G. Solicitation of Comments Regarding Health Equity and
Qualified Health Plans
IV. Collection of Information Requirements
A. Wage Estimates
B. ICRs Regarding State Flexibility for Risk Adjustment (Sec.
153.320)
C. ICRs Regarding Distributed Data and Risk Adjustment Data
Submission Requirements (Sec. Sec. 153.610, 153.700, and 153.710)
D. ICRs Regarding Ability of States To Permit Agents and Brokers
and Web-brokers To Assist Qualified Individuals, Qualified
Employers, or Qualified Employees Enrolling in QHPs (Sec. 155.220)
E. ICRs Regarding Verification of Eligibility for Special
Enrollment Periods (Sec. 155.420)
F. ICRs Regarding General Program Integrity and Oversight
Requirements (Sec. 155.1200)
G. ICRs Regarding State Exchange Improper Payment Measurement
program (Sec. Sec. 155.1500-155.1540)
H. ICRs Regarding State Selection of EHB-Benchmark Plan for Plan
Years Beginning on or After January 1, 2020 (Sec. 156.111)
I. ICR Regarding Differential Display of Standardized Options on
the websites of Web-Brokers (Sec. 155.220) and QHP Issuers (Sec.
156.265)
J. ICRs Regarding Network Adequacy and Essential Community
Providers (Sec. Sec. 156.230 and 156.235)
K. ICRs Regarding Payment for Cost-Sharing Reductions (Sec.
156.430)
L. ICRs Regarding Quality Improvement Strategy (Sec. 156.1130)
M. ICRs Regarding Medical Loss Ratio (Sec. Sec. 158.140,
158.150, 158.170)
O. Summary of Annual Burden Estimates for Proposed Requirements
P. Submission of PRA-related Comments
V. Response to Comments
VI. Regulatory Impact Analysis
A. Statement of Need
B. Overall Impact
C. Impact Estimates of the Payment Notice Provisions and
Accounting Table
D. Regulatory Alternatives Considered
E. Regulatory Flexibility Act
F. Unfunded Mandates
G. Federalism
I. Executive Summary
American Health Benefit Exchanges, or ``Exchanges,'' are entities
established under the Patient Protection and Affordable Care Act (ACA)
\1\ through which qualified individuals and qualified employers can
purchase health insurance coverage in qualified health plans (QHPs).
Many individuals who enroll in QHPs through individual market Exchanges
are eligible to receive a premium tax credit (PTC) to reduce their
costs for health insurance premiums and to receive reductions in
required cost-sharing payments to reduce out-of-pocket expenses for
health care services. The ACA also established the risk adjustment
program, which transfers funds from issuers that attract lower-than-
average risk populations to issuers that attract higher-than-average
risk populations to reduce incentives for issuers to avoid higher-risk
enrollees.
---------------------------------------------------------------------------
\1\ The Patient Protection and Affordable Care Act (Pub. L. 111-
148) was enacted on March 23, 2010. The Healthcare and Education
Reconciliation Act of 2010 (Pub. L. 111-152), which amended and
revised several provisions of the Patient Protection and Affordable
Care Act, was enacted on March 30, 2010. In this rulemaking, the two
statutes are referred to collectively as the ``Patient Protection
and Affordable Care Act'', ``Affordable Care Act'', or ``ACA.''
---------------------------------------------------------------------------
In previous rulemakings, we established provisions and parameters
to implement many ACA requirements and programs. In this proposed rule,
we propose to amend some of these provisions and parameters, with a
focus on maintaining a stable regulatory environment. These proposed
changes are intended to provide issuers with greater predictability for
upcoming plan years (PYs), while simultaneously enhancing the role of
states in these programs. The proposals would provide states with
additional flexibilities, reduce unnecessary regulatory burdens on
stakeholders, empower consumers, ensure program integrity, and improve
affordability.
On January 20, 2021, the President issued an Executive Order which
stated the Administration's policy on preventing and combating
discrimination on the basis of gender identity and sexual
orientation.\2\ This Executive Order instructed the Secretary of Health
and Human Services (Secretary of HHS, or HHS Secretary) to review all
existing regulations, guidance documents, and other agency actions to
determine whether they are consistent with the aforementioned policy,
and to consider whether to suspend, revise, or rescind any agency
actions that are inconsistent with it. In consideration of this
Executive Order, and as a result of our review of certain regulations,
we propose to amend HHS regulations such that Exchanges, issuers, and
agents and brokers are prohibited from discriminating based on sexual
orientation and gender identity. The provisions in this proposed rule
reflect the aspects of the Executive Order 13988 and aligns with the
HHS' Notice, released on May 10, 2021, that HHS interprets and enforces
section 1557's and Title IX's prohibition on discrimination on the
basis of sex to include: (1) Discrimination on the basis of sexual
orientation; and (2) discrimination on the basis of gender identity,
based on the Supreme Court's decision in Bostock v. Clayton County.\3\
---------------------------------------------------------------------------
\2\ Executive Order 13988 on Preventing and Combating
Discrimination on the Basis of Gender Identity or Sexual
Orientation, January 20, 2021, see 86 FR 7023.
\3\ U.S. Dep't of Health & Hum. Servs., Notification of
Interpretation and Enforcement of Section 1557 of the Affordable
Care Act and Title IX of the Education Amendments of 1972, 86 FR
27984 (May 25, 2021). Also see, Bostock v. Clayton County, 140 S.
Ct. 1731 (2020). <a href="https://www.supremecourt.gov/opinions/19pdf/17-1618_hfci.pdf">https://www.supremecourt.gov/opinions/19pdf/17-1618_hfci.pdf</a>.
---------------------------------------------------------------------------
Risk adjustment continues to be a core program in the individual,
small group, and merged markets both on and off Exchanges, and we
propose recalibrated parameters for the HHS-operated risk adjustment
methodology. We published a technical paper, the 2021 HHS-Operated Risk
Adjustment Technical Paper on Possible Model Changes \4\ in October
2021, and sought comment on potential updates to the risk adjustment
models. Consistent with the model changes discussed in the October 2021
Risk Adjustment (RA) Technical Paper, in this rule, we propose the
following three updates to the HHS risk adjustment models beginning
with the 2023 benefit year: (1) Adding a two-stage weighted approach to
the adult and child models; (2) removing the current severity illness
factors from the adult models and adding an interacted hierarchical
condition category (HCC) count model specification to the adult and
child models; and (3) replacing the current enrollment duration factors
in the adult models with HCC-contingent enrollment duration factors.
These proposals are intended to improve prediction in the adult and
child risk adjustment models for the lowest-risk enrollees, the
highest-risk enrollees, and partial-year enrollees, whose plan
liabilities are underpredicted in the
[[Page 586]]
current models. We also propose to recalibrate the 2023 benefit year
risk adjustment models using the 2017, 2018, and 2019 enrollee-level
External Data Gathering Environment (EDGE) data. We further propose to
continue applying a market pricing adjustment to the plan liability
associated with Hepatitis C drugs in the risk adjustment models,
consistent with the approach adopted beginning with the 2020 models. We
discuss our consideration of the targeted removal of the mapping of
hydroxychloroquine sulfate to Immune Suppressants and Immunomodulators
(RXC 09) in the 2018 and 2019 benefit year enrollee-level EDGE data
used for the 2023 benefit year model recalibration,\5\ as well as the
targeted removal of Descovy[supreg] from mapping to Anti-HIV Agents
(RXC 01) in all three benefit years' enrollee-level EDGE datasets used
for the 2023 benefit year model recalibration. We also propose for the
2024 benefit year and beyond to recalibrate the adult models using the
final, fourth quarter (Q4) RXC mapping document that was applicable for
each benefit year of data that is included in the current year's model
recalibration. We propose to begin to use this approach for
recalibration of the 2023 adult risk adjustment models, with the
exception of the 2017 enrollee-level EDGE data year, for which we
propose to use the most recent RXC mapping document that was available
when we first processed the 2017 enrollee-level EDGE data (that is, Q2
2018).
---------------------------------------------------------------------------
\4\ Available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\5\ The same concern was not present for the 2016 or 2017
enrollee-level EDGE data because hydroxychloroquine was not included
in the crosswalk until 2018.
---------------------------------------------------------------------------
Additionally, we propose to repeal the ability of states to request
a reduction in risk adjustment state transfers starting with the 2024
benefit year, while proposing to provide an exception for states that
previously requested a reduction to transfers under Sec. 153.320(d).
In addition, we solicit comments on the requests from Alabama to reduce
risk adjustment state transfers for the 2023 benefit year in the
individual (including the catastrophic and non-catastrophic risk pools)
and small group markets.
We also propose the 2023 benefit year risk adjustment user fee for
states where HHS operates the risk adjustment program. We also propose
to collect and extract five new data elements including ZIP code, race,
ethnicity, individual coverage health reimbursement arrangement (ICHRA)
indicator, and a subsidy indicator as part of the required risk
adjustment data that issuers must make accessible to HHS in states
where HHS is operating the risk adjustment program. We also propose to
extract three new data elements issuers already provide to HHS as part
of the required risk adjustment data submissions (plan ID, rating area,
and subscriber indicator) and to expand the permitted uses of the risk
adjustment data and reports. Finally, we propose that whenever HHS
recoups high-cost risk pool funds as a result of audits of risk
adjustment covered plans, actionable discrepancies, or successful
appeals, the recouped funds would be used to reduce high-cost risk pool
charges for that national high-cost risk pool for the next applicable
benefit year for which high-cost risk pool payments have not already
been calculated.
We propose further refinements to the HHS-RADV error estimation
methodology beginning with the 2021 benefit year to (1) extend the
application of Super HCCs (which are currently based on the coefficient
estimation groups defined in the applicable benefit year's ``Additional
Adult Variables'' Table of the ``Do It Yourself (DIY)'' software (Table
6 in the 2021 Benefit Year DIY Software), which is published on the
CCIIO website) \6\ from their current application only in the sorting
step that assigns HCCs to failure rate groups to broader application
throughout the HHS-RADV error rate calculation process, (2) specify
that Super HCCs will be defined separately according to the age group
model to which an enrollee is subject, and (3) constrain to zero any
failure rate group outlier with a negative failure rate, regardless of
whether the outlier issuer has a negative or positive error rate.
---------------------------------------------------------------------------
\6\ <a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance</a>. The August 3, 2021 version of the 2021 DIY Software Tables
is available at <a href="https://www.cms.gov/files/document/cy2021-diy-tables-07092021.xlsx">https://www.cms.gov/files/document/cy2021-diy-tables-07092021.xlsx</a>.
---------------------------------------------------------------------------
As we do every year in the HHS notice of benefit and payment
parameters, we propose updated parameters applicable in the individual
and small group markets. We propose the PY 2023 user fee rates for
issuers offering plans through the Exchanges using the Federal
platform. We propose maintaining the Federal-facilitated Exchange (FFE)
and State-based Exchange on the Federal platform (SBE-FP) user fees at
the current PY 2022 rates, 2.75 and 2.25 percent of total monthly
premiums, respectively, in order to preserve and ensure that the FFEs
and Federal platform have sufficient funding to cover the cost of all
special benefits provided to FFE and SBE-FP issuers during PY 2023. We
also note that HHS will issue the 2023 benefit year premium adjustment
percentage index and related payment parameters in guidance, consistent
with the policy finalized in part 2 of the 2022 Payment Notice.
We also propose to require all Exchanges to prorate premiums and
advance payments of the premium tax credit (APTC) when administering
APTC for enrollees enrolled in a particular policy for less than the
full coverage month, including when the enrollee is enrolled in
multiple policies within a month, each lasting less than the full
coverage month.
We are proposing changes to clarify that the cost-sharing reduction
(CSR) data submission process is mandatory only for those issuers that
received CSR payments from HHS for any part of the benefit year, and
voluntary for other issuers. We propose a technical correction to the
definition of large group market in Sec. 144.103 to delete the
concluding phrase ``unless otherwise provided under state law.''
We propose new display requirements for web-broker non-Exchange
websites, including requirements related to QHP comparative information
and standardized disclaimer language; a prohibition on displaying QHP
advertisements or otherwise providing favored or preferred display of
QHPs based on compensation agents, brokers, or web-brokers receive from
QHP issuers; and a requirement to prominently display a clear
explanation of the rationale for explicit QHP recommendations and the
methodology for the default display of QHPs on web-broker non-Exchange
websites to better inform and protect consumers using such websites.
We propose a number of policies to address certain agent, broker,
and web-broker practices. These policies would be added as part of the
FFE standards of conduct codified at Sec. 155.220(j)(2), improving
CMS's ability to enforce existing responsibilities agents, brokers, and
web-brokers utilizing the Exchange are required to adhere to without
substantially burdening other agents, brokers, and web-brokers, while
also providing more detail about specific business practices that are
prohibited. We believe the proposed new regulatory text would protect
consumers, ensure the efficient operation of the Exchange, minimize the
risk of future tax discrepancies, reduce unauthorized enrollments in
Exchange coverage, and provide a stronger basis for CMS to take
enforcement action against agents, brokers, and web-brokers for
violations of these requirements.
We propose revising our interpretation of the guaranteed
availability requirement to prohibit
[[Page 587]]
issuers from applying a premium payment to an individual's or
employer's past debt owed for coverage and refusing to effectuate
enrollment in new coverage. We believe this proposal would have a
positive impact on the risk pool by removing barriers to enrollment for
low-income individuals who lost prior coverage due to nonpayment of
premiums. In addition, this proposal would promote more equitable
access to health insurance coverage by ensuring that enrollment is not
delayed as a result of non-payment of past-due premiums to the same
issuer or control group, regardless of an individual's or employee's
status as an APTC recipient.
Stable and affordable Exchanges with healthy risk pools are
necessary for ensuring consumers maintain stable access to health
insurance options. In order to minimize the potential for adverse
selection in the Exchanges, we propose to allow Exchanges to conduct
risk-based employer sponsored coverage verification.
We propose to clarify that only those provider incentives and
bonuses that are tied to clearly defined, objectively measurable, and
well-documented clinical or quality improvement standards that apply to
providers may be included in incurred claims for MLR reporting and
rebate calculation purposes. We also propose to specify that only
expenses directly related to activities that improve health care
quality may be included as quality improvement activity (QIA) expenses
for MLR reporting and rebate calculation purposes.
In addition, we propose to make a technical amendment to remove a
reference to a provision that was vacated by the United States District
Court for the District of Maryland in City of Columbus, et al. v.
Cochran, 523 F. Supp. 3d 731 (D. Md. 2021), and thus deleted in part 2
of the 2022 Payment Notice final rule.
With regards to the essential health benefits (EHB), we propose an
evergreen deadline for EHB-benchmark plan applications by states, as
well as proposing to remove the ability for states to permit issuers to
substitute benefits between EHB categories. In addition, we propose
changed de minimis thresholds for the actuarial value (AV) for plans
subject to EHB requirements, as well as narrower de minimis thresholds
for individual market silver QHPs and income-based CSR plan variations.
We also propose to remove the state annual reporting requirement to
report state-required benefits in addition to the EHB to HHS. We
believe there may be ways to achieve compliance with the defrayal
policy without imposing the rigid submission requirements on states
that exist under the annual reporting requirement.
We propose policies to strengthen and clarify our network adequacy
standards, including expanding the provider specialty list for time and
distance standards and adding appointment wait time standards. For
plans with tiered networks, we propose that, to count toward the
issuer's satisfaction of the network adequacy and essential community
provider (ECP) standards, providers must be contracted within the
network tier that results in the lowest cost-sharing obligation. We
also propose to require issuers to submit information about whether
providers offer telehealth services. We propose to increase the ECP
threshold from 20 percent to 35 percent.
We also propose to amend the current regulation, which provides
that, notwithstanding any relationship or relationships a QHP issuer
may have with delegated or downstream entities, the QHP issuer
maintains responsibility for its compliance and the compliance of any
of its delegated or downstream entities with all applicable Federal
standards related to Exchanges. Specifically, HHS proposes adding a
requirement that all agreements between QHP issuers and their
downstream and delegated entities include language stating that any
Exchange authority, including State Exchanges, may demand and receive
records related to the QHP issuers' obligations and compliance with
applicable Federal standards related to Exchanges. We also propose
other amendments to extend the obligation to oversee compliance of
delegated and downstream entities to QHP issuers in all models of
Exchange. These proposals would hold QHP issuers in all models of
Exchange responsible for their downstream and delegated entities'
adherence to applicable Federal standards, and make their oversight
obligations, and the obligations of their downstream and delegated
entities, explicit. We also propose to amend the title of subpart D of
45 CFR part 156 from ``Standards for Qualified Health Plan Issuers on
Federally Facilitated Exchanges and State-Based Exchanges on the
Federal platform'' to ``Standards for Qualified Health Plan Issuers on
Specific Types of Exchanges'' to more accurately reflect the
applicability of the regulations within the subpart.
We solicit comments on incorporating the net premium, maximum out-
of-pocket (MOOP), deductible, and annual out-of-pocket costs (OOPC) of
a plan into the Exchange re-enrollment hierarchy as well as additional
criteria or mechanisms HHS could consider to ensure the Exchange
hierarchy for re-enrollment aligns with plan generosity and consumer
needs, such as, re-enrolling a current bronze QHP enrollee into an
available silver QHP with a lower net premium and higher plan
generosity offered by the same QHP issuer. We also propose to update
the quality improvement strategy (QIS) standards to require QHP issuers
to address health and health care disparities as a specific topic area
within their QIS beginning in 2023.
We also propose to require issuers of QHPs in FFEs and SBE-FPs to
offer through the Exchange standardized QHP options beginning in PY
2023.
Finally, we solicit comments regarding additional ways HHS could
incentivize QHP issuers to design plans that improve health equity and
health conditions in enrollees' environments, as well as how QHP
issuers could address other social determinants of health (SDOH)
outside of the QHP certification process.
II. Background
A. Legislative and Regulatory Overview
Title I of the Health Insurance Portability and Accountability Act
of 1996 (HIPAA) added a new title XXVII to the Public Health Service
Act (PHS Act) to establish various reforms to the group and individual
health insurance markets.
These provisions of the PHS Act were later augmented by other laws,
including the ACA. Subtitles A and C of title I of the ACA reorganized,
amended, and added to the provisions of part A of title XXVII of the
PHS Act relating to group health plans and health insurance issuers in
the group and individual markets. The term ``group health plan''
includes both insured and self-insured group health plans.\7\
---------------------------------------------------------------------------
\7\ The term ``group health plan'' is used in title XXVII of the
PHS Act and is distinct from the term ``health plan'' as used in
other provisions of title I of ACA. The term ``health plan'' does
not include self-insured group health plans.
---------------------------------------------------------------------------
Section 2702 of the PHS Act, as added by the ACA, establishes
requirements for guaranteed availability of coverage in the group and
individual markets.
Section 2718 of the PHS Act, as added by the ACA, generally
requires health insurance issuers to submit an annual MLR report to
HHS, and provide rebates to enrollees if the issuers do not achieve
specified MLR thresholds.
Section 2791 of the PHS Act defines several terms, including
``large group market''.
[[Page 588]]
Section 1301(a)(1)(B) of the ACA directs all issuers of QHPs to
cover the EHB package described in section 1302(a) of the ACA,
including coverage of the services described in section 1302(b) of the
ACA, adherence to the cost-sharing limits described in section 1302(c)
of the ACA, and meeting the AV levels established in section 1302(d) of
the ACA. Section 2707(a) of the PHS Act, which is effective for plan or
policy years beginning on or after January 1, 2014, extends the
requirement to cover the EHB package to non-grandfathered individual
and small group health insurance coverage, irrespective of whether such
coverage is offered through an Exchange. In addition, section 2707(b)
of the PHS Act directs non-grandfathered group health plans to ensure
that cost sharing under the plan does not exceed the limitations
described in sections 1302(c)(1) of the ACA.
Section 1302 of the ACA provides for the establishment of an EHB
package that includes coverage of EHBs (as defined by the Secretary of
HHS), cost-sharing limits, and AV requirements. The law directs that
EHBs be equal in scope to the benefits provided under a typical
employer plan, and that they cover at least the following 10 general
categories: Ambulatory patient services; emergency services;
hospitalization; maternity and newborn care; mental health and
substance use disorder services, including behavioral health treatment;
prescription drugs; rehabilitative and habilitative services and
devices; laboratory services; preventive and wellness services and
chronic disease management; and pediatric services, including oral and
vision care. Section 1302(d) of the ACA describes the various levels of
coverage based on their AV. Consistent with section 1302(d)(2)(A) of
the ACA, AV is calculated based on the provision of EHB to a standard
population. Section 1302(d)(3) of the ACA directs the Secretary of HHS
to develop guidelines that allow for de minimis variation in AV
calculations. Sections 1302(b)(4)(A) through (D) establish that the
Secretary must define EHB in a manner that: (1) Reflects appropriate
balance among the 10 categories; (2) is not designed in such a way as
to discriminate based on age, disability, or expected length of life;
(3) takes into account the health care needs of diverse segments of the
population; and (4) does not allow denials of EHBs based on age, life
expectancy, disability, degree of medical dependency, or quality of
life.
Section 1311(c) of the ACA provides the Secretary the authority to
issue regulations to establish criteria for the certification of QHPs.
Section 1311(c)(1)(B) of the ACA requires among the criteria for
certification that the Secretary must establish by regulation that QHPs
ensure a sufficient choice of providers. Section 1311(e)(1) of the ACA
grants the Exchange the authority to certify a health plan as a QHP if
the health plan meets the Secretary's requirements for certification
issued under section 1311(c) of the ACA, and the Exchange determines
that making the plan available through the Exchange is in the interests
of qualified individuals and qualified employers in the state. Section
1311(c)(6)(C) of the ACA establishes special enrollment periods and
section 1311(c)(6)(D) of the ACA establishes the monthly enrollment
period for Indians, as defined by section 4 of the Indian Health Care
Improvement Act.\8\
---------------------------------------------------------------------------
\8\ The Indian Health Care Improvement Act (IHCIA), the
cornerstone legal authority for the provision of health care to
American Indians and Alaska Natives, was made permanent when
President Obama signed the bill on March 23, 2010, as part of the
Patient Protection and Affordable Care Act.
---------------------------------------------------------------------------
Section 1311(c)(1)(E) of the ACA specifies that to be certified as
a QHP, each health plan must implement a QIS, which is described in
section 1311(g)(1) of the ACA. Section 1311(g)(1) of the ACA describes
this strategy as a payment structure that provides increased
reimbursement or other incentives to improve health outcomes of plan
enrollees, to prevent hospital readmissions, improve patient safety and
reduce medical errors, promote wellness and health, and reduce health
and health care disparities.
Section 1311(d)(3)(B) of the ACA permits a state, at its option, to
require QHPs to cover benefits in addition to EHB. This section also
requires a state to make payments, either to the individual enrollee or
to the issuer on behalf of the enrollee, to defray the cost of these
additional state-required benefits.
Section 1312(c) of the ACA generally requires a health insurance
issuer to consider all enrollees in all health plans (except
grandfathered health plans) offered by such issuer to be members of a
single risk pool for each of its individual and small group markets.
States have the option to merge the individual and small group market
risk pools under section 1312(c)(3) of the ACA.
Section 1312(e) of the ACA provides the Secretary with the
authority to establish procedures under which a state may allow agents
or brokers to (1) enroll qualified individuals and qualified employers
in qualified health plans offered through Exchanges and (2) assist
individuals in applying for PTC and CSRs for qualified health plans
sold through an Exchange.
Sections 1313 and 1321 of the ACA provide the Secretary with the
authority to oversee the financial integrity of State Exchanges, their
compliance with HHS standards, and the efficient and non-discriminatory
administration of State Exchange activities. Section 1313(a)(5)(A) of
the ACA provides the Secretary with the authority to implement any
measure or procedure that the Secretary determines is appropriate to
reduce fraud and abuse in the administration of the Exchanges. Section
1321 of the ACA provides for state flexibility in the operation and
enforcement of Exchanges and related requirements.
Section 1321(a) of the ACA provides broad authority for the
Secretary to establish standards and regulations to implement the
statutory requirements related to Exchanges, QHPs and other components
of title I of the ACA, including such other requirements as the
Secretary determines appropriate. When operating an FFE under section
1321(c)(1) of the ACA, HHS has the authority under sections 1321(c)(1)
and 1311(d)(5)(A) of the ACA to collect and spend user fees. Office of
Management and Budget (OMB) Circular A-25 Revised establishes Federal
policy regarding user fees and specifies that a user charge will be
assessed against each identifiable recipient for special benefits
derived from federal activities beyond those received by the general
public.
Section 1321(d) of the ACA provides that nothing in title I of the
ACA must be construed to preempt any state law that does not prevent
the application of title I of the ACA. Section 1311(k) of the ACA
specifies that Exchanges may not establish rules that conflict with or
prevent the application of regulations issued by the Secretary.
Section 1343 of the ACA establishes a permanent risk adjustment
program to provide payments to health insurance issuers that attract
higher-than-average risk populations, such as those with chronic
conditions, funded by payments from those that attract lower-than-
average risk populations, thereby reducing incentives for issuers to
avoid higher-risk enrollees.
Section 1401(a) of the ACA amended the Internal Revenue Code (the
Code) to add Section 36B, which, among other things, requires that a
taxpayer reconcile APTC for a year of coverage with the amount of the
PTC the taxpayer is allowed for the year.
[[Page 589]]
Section 1402 of the ACA provides for, among other things,
reductions in cost sharing for EHB for qualified low- and moderate-
income enrollees in silver level qualified health plans offered through
the individual market Exchanges. This section also provides for
reductions in cost sharing for Indians enrolled in QHPs at any metal
level.
Section 1411(c) of the ACA requires the Secretary to submit certain
information provided by applicants under section 1411(b) of the ACA to
other federal officials for verification, including income and family
size information to the Secretary of the Treasury. Section 1411(d) of
the ACA provides that the Secretary must verify the accuracy of
information provided by applicants under section 1411(b) of the ACA for
which section 1411(c) does not prescribe a specific verification
procedure, in such manner as the Secretary determines appropriate.
Section 1411(f) of the ACA requires the Secretary, in consultation
with the Treasury and Homeland Security Department Secretaries and the
Commissioner of Social Security, to establish procedures for hearing
and making decisions governing appeals of Exchange eligibility
determinations. Section 1411(f)(1)(B) of the ACA requires the Secretary
to establish procedures to redetermine eligibility on a periodic basis,
in appropriate circumstances, including eligibility to purchase a QHP
through the Exchange and for APTC and CSRs.
Section 1411(g) of the ACA allows the use of applicant information
only for the limited purposes of, and to the extent necessary to,
ensure the efficient operation of the Exchange, including by verifying
eligibility to enroll through the Exchange and for APTC and CSRs, and
limits the disclosure of such information.
Section 1557 of the ACA applies certain long-standing civil rights
nondiscrimination requirements to ``any health program or activity, any
part of which is receiving Federal financial assistance, including
credits, subsidies, or contracts of insurance, or under any program or
activity that is administered by an Executive agency, or any entity
established under'' Title I of the ACA (or amendments). It did so by
referencing statutes that specify prohibited grounds of discrimination,
namely, race, color, national origin, sex, age, or disability, in an
array of federally funded and administered programs or activities.\9\
In addition, HHS has previously finalized rules unrelated to section
1557 of the ACA to address populations that have historically been
subject to discrimination.
---------------------------------------------------------------------------
\9\ 42 U.S.C. 18116.
---------------------------------------------------------------------------
Section 5000A of the Code, as added by section 1501(b) of the ACA,
requires individuals to have minimum essential coverage (MEC) for each
month, qualify for an exemption, or make an individual shared
responsibility payment. Under the Tax Cuts and Jobs Act, which was
enacted on December 22, 2017, the individual shared responsibility
payment is reduced to $0, effective for months beginning after December
31, 2018.\10\ Notwithstanding that reduction, certain exemptions are
still relevant to determine whether individuals age 30 and above
qualify to enroll in catastrophic coverage under Sec. Sec. 155.305(h)
and 156.155(a)(5).
---------------------------------------------------------------------------
\10\ Public Law 115-97, 131 Stat. 2054 (2017).
---------------------------------------------------------------------------
1. Premium Stabilization Programs
In the July 15, 2011 Federal Register (76 FR 41929), we published a
proposed rule outlining the framework for the premium stabilization
programs.\11\ We implemented the premium stabilization programs in a
final rule, published in the March 23, 2012 Federal Register (77 FR
17219) (Premium Stabilization Rule). In the December 7, 2012 Federal
Register (77 FR 73117), we published a proposed rule outlining the
benefit and payment parameters for the 2014 benefit year to expand the
provisions related to the premium stabilization programs and set forth
payment parameters in those programs (proposed 2014 Payment Notice). We
published the 2014 Payment Notice final rule in the March 11, 2013
Federal Register (78 FR 15409). In the June 19, 2013 Federal Register
(78 FR 37032), we proposed a modification to the HHS-operated
methodology related to community rating states. In the October 30, 2013
Federal Register (78 FR 65046), we finalized the proposed modification
to the HHS-operated methodology related to community rating states. We
published a correcting amendment to the 2014 Payment Notice final rule
in the November 6, 2013 Federal Register (78 FR 66653) to address how
an enrollee's age for the risk score calculation would be determined
under the HHS-operated risk adjustment methodology.
---------------------------------------------------------------------------
\11\ The term premium stabilization programs refers to the risk
adjustment, risk corridors, and reinsurance programs established by
the ACA. See 42 U.S.C. 18061, 18062, and 18063.
---------------------------------------------------------------------------
In the December 2, 2013 Federal Register (78 FR 72321), we
published a proposed rule outlining the benefit and payment parameters
for the 2015 benefit year to expand the provisions related to the
premium stabilization programs, setting forth certain oversight
provisions and establishing the payment parameters in those programs
(proposed 2015 Payment Notice). We published the 2015 Payment Notice
final rule in the March 11, 2014 Federal Register (79 FR 13743). In the
May 27, 2014 Federal Register (79 FR 30240), the 2015 fiscal year
sequestration rate for the risk adjustment program was announced.
In the November 26, 2014 Federal Register (79 FR 70673), we
published a proposed rule outlining the benefit and payment parameters
for the 2016 benefit year to expand the provisions related to the
premium stabilization programs, setting forth certain oversight
provisions and establishing the payment parameters in those programs
(proposed 2016 Payment Notice). We published the 2016 Payment Notice
final rule in the February 27, 2015 Federal Register (80 FR 10749).
In the December 2, 2015 Federal Register (80 FR 75487), we
published a proposed rule outlining the benefit and payment parameters
for the 2017 benefit year to expand the provisions related to the
premium stabilization programs, setting forth certain oversight
provisions and establishing the payment parameters in those programs
(proposed 2017 Payment Notice). We published the 2017 Payment Notice
final rule in the March 8, 2016 Federal Register (81 FR 12203).
In the September 6, 2016 Federal Register (81 FR 61455), we
published a proposed rule outlining the benefit and payment parameters
for the 2018 benefit year and to further promote stable premiums in the
individual and small group markets. We proposed updates to the risk
adjustment methodology, new policies around the use of external data
for recalibration of our risk adjustment models, and amendments to the
HHS-RADV process (proposed 2018 Payment Notice). We published the 2018
Payment Notice final rule in the December 22, 2016 Federal Register (81
FR 94058).
In the November 2, 2017 Federal Register (82 FR 51042), we
published a proposed rule outlining the benefit and payment parameters
for the 2019 benefit year, and to further promote stable premiums in
the individual and small group markets. We proposed updates to the risk
adjustment methodology and amendments to the HHS-RADV process (proposed
2019 Payment Notice). We published the 2019 Payment Notice final rule
in the April 17, 2018 Federal Register (83 FR 16930). We published a
correction to the 2019 risk adjustment coefficients in the 2019 Payment
Notice final rule in the May 11, 2018 Federal Register (83 FR 21925).
On July 27,
[[Page 590]]
2018, consistent with 45 CFR 153.320(b)(1)(i), we updated the 2019
benefit year final risk adjustment model coefficients to reflect an
additional recalibration related to an update to the 2016 enrollee-
level External Data Gathering Environment (EDGE) dataset.\12\
---------------------------------------------------------------------------
\12\ ``Updated 2019 Benefit Year Final HHS Risk Adjustment Model
Coefficients.'' July 27, 2018. Available at <a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/2019-Updtd-Final-HHS-RA-Model-Coefficients.pdf">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/2019-Updtd-Final-HHS-RA-Model-Coefficients.pdf</a>.
---------------------------------------------------------------------------
In the July 30, 2018 Federal Register (83 FR 36456), we published a
final rule that adopted the 2017 benefit year risk adjustment
methodology as established in the final rules published in the March
23, 2012 (77 FR 17220 through 17252) and March 8, 2016 editions of the
Federal Register (81 FR 12204 through 12352). That final rule set forth
additional explanation of the rationale supporting use of statewide
average premium in the HHS-operated risk adjustment state payment
transfer formula for the 2017 benefit year, including the reasons why
the program is operated in a budget-neutral manner. That final rule
also permitted HHS to resume 2017 benefit year risk adjustment payments
and charges. HHS also provided guidance as to the operation of the HHS-
operated risk adjustment program for the 2017 benefit year in light of
publication of the final rule.\13\
---------------------------------------------------------------------------
\13\ ``Update on the HHS-operated Risk Adjustment Program for
the 2017 Benefit Year.'' July 27, 2018. Available at <a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/2017-RA-Final-Rule-Resumption-RAOps.pdf">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/2017-RA-Final-Rule-Resumption-RAOps.pdf</a>.
---------------------------------------------------------------------------
In the August 10, 2018 Federal Register (83 FR 39644), we published
a proposed rule seeking comment on adopting the 2018 benefit year risk
adjustment methodology in the final rules published in the March 23,
2012 (77 FR 17219) and in the December 22, 2016 editions of the Federal
Register (81 FR 94058). The proposed rule set forth additional
explanation of the rationale supporting use of statewide average
premium in the HHS-operated risk adjustment state payment transfer
formula for the 2018 benefit year, including the reasons why the
program is operated in a budget-neutral manner. In the December 10,
2018 Federal Register (83 FR 63419), we issued a final rule adopting
the 2018 benefit year HHS-operated risk adjustment methodology as
established in the final rules published in the March 23, 2012 (77 FR
17219) and the December 22, 2016 (81 FR 94058) editions of the Federal
Register. That final rule sets forth additional explanation of the
rationale supporting use of statewide average premium in the HHS-
operated risk adjustment state payment transfer formula for the 2018
benefit year, including the reasons why the program is operated in a
budget-neutral manner.
In the January 24, 2019 Federal Register (84 FR 227), we published
a proposed rule outlining updates to the calibration of the risk
adjustment methodology, the use of EDGE data for research purposes, and
updates to HHS-RADV audits. We published the 2020 Payment Notice final
rule in the April 25, 2019 Federal Register (84 FR 17454).
In the February 6, 2020 Federal Register (85 FR 7088), we published
a proposed rule that included updates to the risk adjustment models'
HCCs and a modification HHS-RADV error rate calculation methodology. We
published the 2021 Payment Notice final rule in the May 14, 2020
Federal Register (85 FR 29164).
In the June 2, 2020 Federal Register (85 FR 33595), we published a
proposed rule that proposed updates to various aspects of the HHS-RADV
methodologies and processes. We published a final rule titled, the
Amendments to the HHS-Operated Risk Adjustment Data Validation Under
the Patient Protection and Affordable Care Act's HHS-Operated Risk
Adjustment Program (2020 HHS-RADV Amendments Rule) in the December 1,
2020 Federal Register (85 FR 76979). That final rule revised the
failure rate grouping algorithm, finalized a sliding scale adjustment
in HHS-RADV error rate calculation, and a constraint on risk score
adjustments for low-side failure rate outliers. The final rule also
established a transition from the prospective application of HHS-RADV
adjustments to apply HHS-RADV results to risk scores from the same
benefit year as that being audited.
In the September 2, 2020 Federal Register (85 FR 54820), HHS issued
an interim final rule containing certain policy and regulatory
revisions in response to the COVID-19 public health emergency (PHE),
wherein we set forth risk adjustment reporting requirements for issuers
offering temporary premium credits in the 2020 benefit year (interim
final rule on COVID-19).
In the January 20, 2021 Federal Register (86 FR 6138), HHS issued a
final rule containing certain policy and regulatory revisions related
to the risk adjustment program (hereinafter referred to as ``part 1 of
the 2022 Payment Notice final rule''). In the May 5, 2021 Federal
Register (86 FR 24140), HHS issued another final rule containing policy
and regulatory revisions related to the risk adjustment program,
including approval of the request from Alabama to reduce risk
adjustment transfers by 50 percent in the individual and small group
markets for the 2022 benefit year (hereinafter referred to as ``part 2
of the 2022 Payment Notice final rule''). In addition, part 2 of the
2022 Payment Notice final rule established a revised schedule of
collections for HHS-RADV and updated the provisions regulating second
validation audit (SVA) and initial validation audit (IVA) entities.
2. Program Integrity
In the June 19, 2013 Federal Register (78 FR 37031), we published a
proposed rule that proposed certain program integrity standards related
to Exchanges and the premium stabilization programs (proposed Program
Integrity Rule). The provisions of that proposed rule were finalized in
two rules, the ``first Program Integrity Rule'' published in the August
30, 2013 Federal Register (78 FR 54069) and the ``second Program
Integrity Rule'' published in the October 30, 2013 Federal Register (78
FR 65045).
3. Market Rules
An interim final rule relating to the HIPAA health insurance
reforms was published in the April 8, 1997 Federal Register (62 FR
16894). A proposed rule relating to the 2014 health insurance market
rules was published in the November 26, 2012 Federal Register (77 FR
70584). A final rule implementing the health insurance market rules was
published in the February 27, 2013 Federal Register (78 FR 13406) (2014
Market Rules).
A proposed rule relating to Exchanges and Insurance Market
Standards for 2015 and beyond was published in the March 21, 2014
Federal Register (79 FR 15808) (2015 Market Standards Proposed Rule). A
final rule implementing the Exchange and Insurance Market Standards for
2015 and Beyond was published in the May 27, 2014 Federal Register (79
FR 30240) (2015 Market Standards Rule). The 2018 Payment Notice final
rule in the December 22, 2016 Federal Register (81 FR 94058) provided
additional guidance on guaranteed availability and guaranteed
renewability. In the Market Stabilization final rule that was published
in the April 18, 2017 Federal Register (82 FR 18346), we further
interpreted the guaranteed availability provision. In the 2019 Payment
Notice final rule in the April 17, 2018 Federal Register (83 FR 17058),
we clarified that certain exceptions to the special enrollment periods
only apply with respect to coverage offered outside of the Exchange in
the individual market.
[[Page 591]]
In the Nondiscrimination in Health and Human Education Programs or
Activities final rule on section 1557 of the ACA, published in the June
19, 2020 Federal Register (85 FR 37160), we removed nondiscrimination
protections on the basis of gender identity and sexual orientation from
the guaranteed availability regulation.
In part 2 of the 2022 Payment Notice final rule in the May 5, 2021
Federal Register (86 FR 24140), we made additional amendments to the
guaranteed availability regulation regarding special enrollment periods
and finalized new special enrollment periods related to untimely notice
of triggering events, cessation of employer contributions or government
subsidies to COBRA continuation coverage, and loss of APTC eligibility.
In the final rule Updating Payment Parameters, Section 1332 Waiver
Implementing Regulations, and Improving Health Insurance Markets for
2022 and Beyond published in the September 27, 2021 Federal Register
(86 FR 53412) (part 3 of the 2022 Payment Notice) by HHS and the
Department of the Treasury, HHS finalized additional amendments to the
guaranteed availability regulations regarding special enrollment
periods.
4. Exchanges
We published a request for comment relating to Exchanges in the
August 3, 2010 Federal Register (75 FR 45584). We issued initial
guidance to states on Exchanges on November 18, 2010. We proposed a
rule in the July 15, 2011 Federal Register (76 FR 41865) to implement
components of the Exchanges, and a rule in the August 17, 2011 Federal
Register (76 FR 51201) regarding Exchange functions in the individual
market and Small Business Health Options Program (SHOP), eligibility
determinations, and Exchange standards for employers. A final rule
implementing components of the Exchanges and setting forth standards
for eligibility for Exchanges, as well as network adequacy and ECP
certification standards, was published in the March 27, 2012 Federal
Register (77 FR 18309) (Exchange Establishment Rule).
In the 2014 Payment Notice and in the Amendments to the HHS Notice
of Benefit and Payment Parameters for 2014 interim final rule,
published in the March 11, 2013 Federal Register (78 FR 15541), we set
forth standards related to Exchange user fees. We established an
adjustment to the FFE user fee in the Coverage of Certain Preventive
Services under the Affordable Care Act final rule, published in the
July 2, 2013 Federal Register (78 FR 39869) (Preventive Services Rule).
In the 2016 Payment Notice, we also set forth the ECP certification
standard at Sec. 156.235, with revisions in the 2017 Payment Notice in
the March 8, 2016 Federal Register (81 FR 12203) and the 2018 Payment
Notice in the December 22, 2016 Federal Register (81 FR 94058).
In an interim final rule, published in the May 11, 2016 Federal
Register (81 FR 29146), we made amendments to the parameters of certain
special enrollment periods (2016 Interim Final Rule). We finalized
these in the 2018 Payment Notice final rule, published in the December
22, 2016 Federal Register (81 FR 94058).
In the April 18, 2017 Market Stabilization final rule Federal
Register (82 FR 18346), we amended standards relating to special
enrollment periods and QHP certification. In the 2019 Payment Notice
final rule, published in the April 17, 2018 Federal Register (83 FR
16930), we modified parameters around certain special enrollment
periods. In the April 25, 2019 Federal Register (84 FR 17454), the
final 2020 Payment Notice established a new special enrollment period.
In the February 6, 2020 Federal Register (85 FR 7088), we published
a proposed rule (proposal 2021 Payment Notice). We published the final
rule in the May 14, 2020 Federal Register (85 FR 29164) (2021 Payment
Notice).
In the December 4, 2020 Federal Register (85 FR 78572), we issued a
proposed rule containing certain policy and regulatory revisions
related to user fees (proposed 2022 Payment Notice). In the January 19,
2021 Federal Register (86 FR 6138), HHS issued a rule finalizing
certain of the provisions in the proposed 2022 Payment Notice (part 1
of the 2022 Payment Notice final rule). In the May 5, 2021 Federal
Register (86 FR 24140), HHS published a second final rule addressing
the remainder of the proposed provisions (part 2 of the 2022 Payment
Notice final rule). In the July 1, 2021 Federal Register (86 FR 35156),
HHS and the Department of the Treasury released a proposed rule
proposing to amend certain policies in part 1 of the 2022 Payment
Notice final rule, and finalized the rule in the September 27, 2021
Federal Register (86 FR 53412) (part 3 of the 2022 Payment Notice final
rule).
5. Essential Health Benefits
On December 16, 2011, HHS released a bulletin that outlined an
intended regulatory approach for defining EHB, including a benchmark-
based framework.\14\ A proposed rule relating to EHBs was published in
the November 26, 2012 Federal Register (77 FR 70643). We established
requirements relating to EHBs in the Standards Related to Essential
Health Benefits, Actuarial Value, and Accreditation Final Rule, which
was published in the February 25, 2013 Federal Register (78 FR 12833)
(EHB Rule). In the 2019 Payment Notice, published in the April 17, 2018
Federal Register (83 FR 16930), we added Sec. 156.111 to provide
states with additional options from which to select an EHB-benchmark
plan for PYs 2020 and beyond.
---------------------------------------------------------------------------
\14\ ``Essential Health Benefits Bulletin.'' December 16, 2011.
Available at <a href="https://www.cms.gov/CCIIO/Resources/Files/Downloads/essential_health_benefits_bulletin.pdf">https://www.cms.gov/CCIIO/Resources/Files/Downloads/essential_health_benefits_bulletin.pdf</a>.
---------------------------------------------------------------------------
6. Medical Loss Ratio (MLR)
We published a request for comment on section 2718 of the PHS Act
in the April 14, 2010 Federal Register (75 FR 19297), and published an
interim final rule with a 60-day comment period relating to the MLR
program on December 1, 2010 (75 FR 74863). A final rule with a 30-day
comment period was published in the December 7, 2011 Federal Register
(76 FR 76573). An interim final rule with a 60-day comment period was
published in the December 7, 2011 Federal Register (76 FR 76595). A
final rule was published in the Federal Register on May 16, 2012 (77 FR
28790). The MLR program requirements were amended in final rules
published in the March 11, 2014 Federal Register (79 FR 13743), the May
27, 2014 Federal Register (79 FR 30339), the February 27, 2015 Federal
Register (80 FR 10749), the March 8, 2016 Federal Register (81 FR
12203), the December 22, 2016 Federal Register (81 FR 94183), the April
17, 2018 Federal Register (83 FR 16930), the May 14, 2020 Federal
Register (85 FR 29164), and the May 5, 2021 Federal Register (86 FR
24140), and an interim final rule that was published in the September
2, 2020 Federal Register (85 FR 54820).
7. Quality Improvement Strategy
We promulgated regulations in 45 CFR 155.200(d) to direct Exchanges
to evaluate quality improvement strategies, and 45 CFR 156.200(b) that
direct QHP issuers to implement and report on a quality improvement
strategy or strategies consistent with section 1311(g) standards as a
QHP certification criteria for participation in an Exchange. In the
2016 Payment Notice, published in the February 27, 2015 Federal
Register (80 FR 10749), we finalized
[[Page 592]]
regulations at Sec. 155.1130 to establish standards and the associated
timeframe for QHP issuers to submit the necessary information to
implement QIS standards for QHPs offered through an Exchange.
8. Nondiscrimination
Section 1311(b) and section 1321(b) of the ACA provide that each
state has the opportunity to establish an Exchange. In the July 15,
2011 Federal Register (76 FR 41866), HHS published the ``Patient
Protection and Affordable Care Act; Establishment of Exchanges and
Qualified Health Plans'' proposed rule to implement section 1311(b) and
section 1321(b) of the ACA. In the March 27, 2012 Federal Register (77
FR 18310), HHS published the ``Patient Protection and Affordable Care
Act; Establishment of Exchanges and Qualified Health Plans; Exchange
Standards for Employers'' final rule and interim final rule
(hereinafter referred to as the ``Exchange Standards final rule''),
which included nondiscrimination protections.
Section 1302 of the ACA provides for the establishment of an EHB
package that includes coverage of EHB and actuarial value requirements.
In the November 26, 2012 Federal Register (77 FR 70644), HHS published
the ``Patient Protections and Affordable Care Act; Standards Related to
Essential Health Benefits, Actuarial Value, and Accreditation''
proposed rule to implement section 1302 of the ACA. In the February 25,
2013 Federal Register (78 FR 12834), HHS published the ``Patient
Protections and Affordable Care Act; Standards Related to Essential
Health Benefits, Actuarial Value, and Accreditation'' final rule, which
included nondiscrimination protections.
Sections 2701, 2702, and 2703 of the PHS Act and Section 1312(c) of
the ACA provide protections to individuals and employers in obtaining
health insurance coverage. In the November 26, 2012 Federal Register
(77 FR 70584), HHS published the ``Patient Protection and Affordable
Care Act; Health Insurance Market Rules; Rate Review'' proposed rule to
implement sections 2701, 2702, and 2703 of the PHS Act and section
1312(c) of the ACA. In the February 27, 2013 Federal Register (78 FR
13406), HHS published the ``Patient Protections and Affordable Care
Act; Health Insurance Market Rules; Rate Review'' final rule, which
included nondiscrimination protections.
In the HHS Notice of Benefit and Payment Parameters for 2017
proposed rule, published in the December 2, 2015 Federal Register (80
FR 75488), HHS proposed policies for nondiscrimination protections into
the relevant notice of benefit and payment parameters. In the March 8,
2016 Federal Register (81 FR 12204), HHS published the HHS Notice of
Benefit and Payment Parameters for 2017 final rule, which included
nondiscrimination protections.
In the Nondiscrimination in Health and Human Education Programs or
Activities final rule on section 1557 of the ACA, published in the June
19, 2020 Federal Register (85 FR 37160), HHS removed nondiscrimination
protections on the basis of gender identity and sexual orientation from
various CMS nondiscrimination regulations. In the HHS Notice of
Interpretation and Enforcement of Section 1557 of the Affordable Care
Act and Title IX of the Education Amendments of 1972, published in the
May 25, 2021 Federal Register (86 FR 27984), HHS informed the public
that HHS will interpret and enforce section 1557's and Title IX's
prohibition on discrimination on the basis of sex to include
discrimination based on sexual orientation and gender identity.
B. Stakeholder Consultation and Input
HHS has consulted with stakeholders on policies related to the PHS
Act federal market reform requirements, the operation of Exchanges and
the risk adjustment (including HHS-RADV) program. We have held a number
of meetings with consumers, providers, employers, health plans,
advocacy groups and the actuarial community to gather public input. We
have solicited input from state representatives on numerous topics,
particularly EHBs, state mandates, and risk adjustment. We consulted
with stakeholders through regular meetings with the National
Association of Insurance Commissioners (NAIC), regular contact with
states through the Exchange Blueprint approval and general Exchange
oversight processes, and meetings with Tribal leaders and
representatives, health insurance issuers, trade groups, consumer
advocates, employers, and other interested parties. We considered all
public input we received as we developed the policies in this proposed
rule.
C. Structure of Proposed Rule
The regulations outlined in this proposed rule would be codified in
45 CFR parts 144, 147, 153, 155, 156 and 158.
The proposed changes to 45 CFR part 144 would remove superfluous
language from the definition of large group market.
The proposed changes to 45 CFR part 147 would prohibit issuers from
discriminating against individuals in issuer marketing practices and
benefit designs based on sexual orientation and gender identity. We
also propose to reinterpret the guaranteed availability requirements in
Sec. 147.104 such that issuers could not refuse to effectuate new
coverage based on failure of an individual or employer to pay premiums
owed for prior coverage.
The proposed changes to 45 CFR part 153 would recalibrate the 2023
benefit year risk adjustment models using the 2017, 2018, and 2019
enrollee-level External Data Gathering Environment (EDGE) data. We also
propose to update the adult and child risk adjustment models for 2023
and beyond to better predict plan liability for certain subpopulations.
We propose to update the adult risk adjustment models by removing the
current severity illness factors and replacing the current enrollment
duration factors with enrollment duration factors contingent on the
enrollee having at least one HCC. In addition, we propose to update the
adult and child risk adjustment models by adding a two-stage weighted
approach to model recalibrations and an interacted HCC count model
specification for 2023 and beyond. We propose to continue applying a
market pricing adjustment to the plan liability associated with
Hepatitis C drugs in the risk adjustment models, consistent with the
approach adopted beginning with the 2020 models. We discuss removing
the mapping of hydroxychloroquine sulfate to RXC 09 (Immune
Suppressants and Immunomodulators) in the 2018 and 2019 benefit year
enrollee-level EDGE data used for the annual recalibration of the HHS
risk adjustment models. We also propose for the 2024 benefit year and
beyond to recalibrate the models using the final, fourth quarter (Q4)
RXC mapping document that was applicable for each benefit year of data
that is included in the current year's model recalibration. We propose
using this approach for recalibration of the 2023 adult risk adjustment
models with the exception of the 2017 enrollee-level EDGE data year,
for which we propose to use the most recent RXC mapping document that
was available when we first processed the 2017 enrollee-level EDGE data
(that is, Q2 2018).We also propose to collect and extract five new data
elements including ZIP code, race, ethnicity, ICHRA indicator, and a
subsidy indicator as part of the required risk adjustment data that
issuers must make accessible to HHS in states where HHS is operating
the risk adjustment program. We also propose to extract three new data
elements issuers already
[[Page 593]]
provide to HHS as part of the required risk adjustment data submissions
(plan ID, rating area, and subscriber indicator) and to expand the
permitted uses of the risk adjustment data and reports. Additionally,
we propose an amendment to Sec. 153.730 to address situations when
April 30 does not fall on a business day and to provide that when this
occurs, the deadline for issuers to submit the required risk adjustment
data in states where HHS operates the program would be the next
applicable business day.
The proposals in part 153 also relate to risk adjustment state
flexibility requests. We propose to repeal the ability of states to
request a reduction in risk adjustment transfers calculated by HHS
under the state payment transfer formula starting with the 2024 benefit
year, while proposing to create an exception for any state that has
requested a reduction in prior benefit years. In addition, we solicit
comments on the requests from Alabama to reduce risk adjustment state
transfers for the 2023 benefit year in the individual (including the
catastrophic and non-catastrophic risk pools) and small group markets.
In part 153 we also propose the risk adjustment user fee for the
2023 benefit year and modifications to the error estimation methodology
applied in HHS-RADV. We propose updating the HHS-RADV error estimation
process to extend the application of Super HCCs beyond the sorting step
that assigns HCCs to failure rate groups to also apply throughout the
HHS-RADV error rate calculation processes and to specify that Super
HCCs will be defined separately according to the model (infant, child,
adult) to which an enrollee is subject. We also propose to constrain to
zero any failure rate group outlier negative failure rate, regardless
of whether the outlier issuer has a negative or positive error rate.
Finally, we propose that whenever HHS recoups high-cost risk pool funds
as a result of audits of risk adjustment covered plans, an actionable
discrepancy, or a successful administrative appeal, the recouped high-
cost risk pool funds will be used to reduce high-cost risk pool charges
for that national high-cost risk pool beginning for the next benefit
year for which a high cost risk pool payment has not already been
calculated.
In addition, the proposals regarding part 153 also relate to MLR
reporting requirements and clarify how issuers should report certain
ACA program amounts that could be subject to reconsideration for MLR
reporting purposes. We propose to separately address and reference HHS-
RADV adjustments to make clear that HHS expects issuers to report HHS-
RADV adjustments as part of their MLR reports in the same manner as
they report risk adjustment payment and charge amounts.
The proposed changes to 45 CFR part 155 would allow Exchanges to
implement a verification process for enrollment in or eligibility for
an eligible employer sponsored plan based on the Exchange's assessment
of risk for inappropriate payments of APTC/CSR. In part 155 we also
propose to require all Exchanges to prorate when administering APTC for
enrollees enrolled in a particular policy for less than the full
coverage month, including when the enrollee is enrolled in multiple
policies within a month, each lasting less than the full coverage
month. We also propose new requirements in part 155 related to the QHP
comparative information and standardized disclaimer required to be
displayed on web-broker non-Exchange websites, a prohibition on
displaying QHP advertisements or otherwise providing favored or
preferred placement in the display of QHPs on web-broker non-Exchange
websites based on compensation agents, brokers, or web-brokers receive
from QHP issuers, and a requirement regarding the prominent display of
a clear explanation of the rationale for explicit QHP recommendations
and the methodology for the default display of QHPs on web-broker non-
Exchange websites to better inform and protect consumers using such
websites. We also propose changes to part 155, to clarify the FFE
standards of conduct and what it means for agents, brokers, and web-
brokers to provide the Exchange with correct information under section
1411(b) of the ACA, including ensuring that accurate consumer
information is being entered on Exchange applications. Finally, we
propose changes to part 155 to set forth prohibited agent, broker, and
web-broker business practices commonly observed by HHS and to create
enforceable standards under which HHS may take enforcement action
against agents, brokers, and web-brokers when these prohibited business
practices are discovered.
In 45 CFR part 156, as we do every year in the HHS notice of
benefit and payment parameters, we propose to update the user fee rates
for the 2023 benefit year for all issuers participating on the
Exchanges using the Federal platform. We note that we intend to publish
the 2023 premium adjustment percentage index and related payment
parameters in guidance as finalized in part 2 of the 2022 Payment
Notice. The proposed changes to part 156 also include technical
amendments to Sec. 156.50 to conform the user fee regulations with the
repeal of Exchange Direct Enrollment (DE) option finalized in part 3 of
the 2022 Payment Notice.\15\ We are proposing changes to Sec. 156.430
to clarify that the CSR data submission process is mandatory only for
those issuers that receive CSR payments from HHS for any part of the
benefit year as a result of HHS possessing a valid appropriation to
make CSR payments, and voluntary for other issuers.
---------------------------------------------------------------------------
\15\ 86 FR 53412.
---------------------------------------------------------------------------
In part 156, we also propose an evergreen deadline for EHB-
benchmark plan applications by states, as well as proposing to remove
the ability for states to permit issuers to substitute benefits between
EHB categories, proposing to change de minimis thresholds for the AV of
plans subject to the AV requirements, as well as narrower de minimis
thresholds for individual market silver QHPs and income-based CSR plan
variations; and proposing to remove the annual reporting requirement on
states to report state-required benefits in addition to the EHB to HHS.
In part 156, we also propose to require issuers of QHPs in FFEs and
SBE-FPs to offer through the Exchange standardized QHP options
beginning in PY 2023. We also propose to update the QIS standards in
part 156 to require QHP issuers to address health and health care
disparities as a specific topic area within their QIS beginning with PY
2023.
The proposed changes to part 158 would clarify that only those
provider incentives and bonuses that are tied to clearly defined,
objectively measurable, and well-documented clinical or quality
improvement standards that apply to providers may be included in
incurred claims for MLR reporting and rebate calculation purposes. The
proposed changes to part 158 would also specify that only expenses
directly related to activities that improve health care quality may be
included as QIA expenses for MLR reporting and rebate calculation
purposes. In addition, the proposed changes to part 158 would make a
technical amendment to Sec. 158.170(b) to correct an oversight and
remove the reference to the percentage of premium QIA reporting option
described in Sec. 158.221(b)(8), a provision that was vacated by the
United States District Court for the District of Maryland in City of
Columbus, et al. v.
[[Page 594]]
Cochran,\16\ and thus deleted in part 2 of the 2022 Payment Notice
final rule.
---------------------------------------------------------------------------
\16\ 523 F. Supp. 3d 731 (D. Md. 2021).
---------------------------------------------------------------------------
III. Provisions of the Proposed HHS Notice of Benefit and Payment
Parameters for 2023
A. Part 144--Requirements Relating to Health Insurance Coverage
1. Definitions (Sec. 144.103)
We propose to remove superfluous language from the definition of
large group market. The definition currently provides that ``Large
group market'' means the health insurance market under which
individuals obtain health insurance coverage (directly or through any
arrangement) on behalf of themselves (and their dependents) through a
group health plan maintained by a large employer, unless otherwise
provided under State law. We propose to amend the definition by
deleting the phrase ``unless otherwise provided under State law.'' The
phrase has no meaning or application, and does not appear in the
statutory definition of the term in section 2791(e)(3) of the PHS Act.
That phrase was initially included in the PHS Act regulatory
definitions of large group market, large employer, and small employer
adopted by HHS under HIPAA.\17\ However, in final rules published on
October 30, 2013 (78 FR 65045), we amended the definitions of large
employer and small employer to make them consistent with PHS Act
section 2791(e), as amended by the ACA, and in so doing, removed that
phrase from the definitions. At that time, we inadvertently neglected
to delete the phrase from the regulatory definition of large group
market, and we now propose to do so, in order to align these
definitions and make the regulatory definition for large group market
consistent with the definition under the ACA.
---------------------------------------------------------------------------
\17\ 62 FR 16894 (April 8, 1997) and 69 FR 78720 (Dec. 30,
2004).
---------------------------------------------------------------------------
B. Part 147--Health Insurance Reform Requirements for the Group and
Individual Health Insurance Markets
1. Guaranteed Availability of Coverage (Sec. 147.104)
a. Past-Due Premiums
We propose to re-interpret the guaranteed availability requirement
at section 2702 of the PHS Act and its implementing regulation at Sec.
147.104 to require issuers to accept individuals and employers who
apply for coverage, even where the individual or employer owes past-due
premiums for coverage from the same issuer or another issuer in the
same controlled group. On January 28, 2021, President Biden issued
Executive Order 14009, ``Strengthening Medicaid and the Affordable Care
Act'' (E.O. 14009).\18\ Section 3 of E.O. 14009 directs HHS, and the
heads of all other executive departments and agencies with authorities
and responsibilities related to Medicaid and the ACA, to review all
existing regulations, orders, guidance documents, policies, and any
other similar agency actions to determine whether they are inconsistent
with policy priorities described in Section 1 of E.O. 14009, to include
protecting and strengthening the ACA and making high-quality health
care accessible and affordable for all individuals. Consistent with
E.O. 14009, specifically section 3(iv), this proposal intends to remove
an unnecessary barrier to individuals and families attempting to enroll
into health coverage in the individual market.
---------------------------------------------------------------------------
\18\ E.O. 14009; 86 FR 7793 (Feb. 2, 2021).
---------------------------------------------------------------------------
Specifically, we propose to redesignate Sec. 147.104(i) as Sec.
147.104(j) and add a new Sec. 147.104(i) to specify that a health
insurance issuer that denies coverage to an individual or employer due
to the individual's or employer's failure to pay premium owed under a
prior policy, certificate, or contract of insurance, including by
attributing payment of premium for a new policy, certificate, or
contract of insurance to the prior policy, certificate, or contract of
insurance, violates Sec. 147.104(a). The guaranteed availability
provisions require health insurance issuers offering non-grandfathered
coverage in the individual or group market to accept every individual
and employer in the state that applies for such coverage unless an
exception applies. Individuals and employers typically are required to
pay the first month's premium to effectuate coverage. Under the current
interpretation of the guaranteed availability requirement stated in the
Market Stabilization final rule, to the extent permitted by applicable
state law, an issuer does not violate the guaranteed availability
requirements under Sec. 147.104 where the issuer attributes a premium
payment made for new coverage to any past-due premiums owed for
coverage from the same issuer or another issuer in the same controlled
group within the prior 12-month period before effectuating enrollment
in the new coverage. This policy addressed concerns that individuals
might take unfair advantage of the rules regarding grace periods.\19\
However, in part 3 of the 2022 Payment Notice proposed rule, we stated
our intention to reassess this interpretation to analyze whether this
policy presents unnecessary barriers to accessing health coverage.\20\
---------------------------------------------------------------------------
\19\ QHP issuers are required, under Sec. 156.270, to provide a
grace period of 3 consecutive months for an enrollee, who, when
failing to timely pay premiums, is receiving APTC. If the enrollee
exhausts the grace period without paying all outstanding premiums,
subject to a premium payment threshold implemented under Sec.
155.400(g), then the QHP issuer must terminate the enrollee's
enrollment back to the last day of the first month of the 3-month
grace period. As a result, an individual receiving APTC whose
coverage is terminated after the exhaustion of a grace period would
owe at most 1 month of premiums, net of any APTC paid on their
behalf to the issuer; however, an individual who attempts to enroll
in new coverage while in a grace period, and whose coverage has not
yet been terminated, could owe up to 3 months of premium, net of any
APTC paid on their behalf to the issuer.
\20\ 86 FR 35156, 36071.
---------------------------------------------------------------------------
After reevaluating our interpretation of the guaranteed
availability requirement, we propose reinstating our previous
interpretation of the guaranteed availability rules with respect to
non-payment of premiums.\21\ Under this interpretation, an issuer may
not apply any premium payment made for new coverage in the same or a
different plan or product to any outstanding debt owed from any
previous coverage and then refuse to effectuate the new enrollment
based on failure to pay premiums. Thus, the guaranteed availability
requirement would prohibit issuers from refusing to effectuate new
coverage due to failure to pay outstanding premium debt from the
previous year.
---------------------------------------------------------------------------
\21\ Federally-facilitated Marketplace (FFM) and Federally-
facilitated Small Business Health Options Program Enrollment Manual,
Section 6.3 Terminations for Non-Payment of Premiums, <a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/ENR_FFMSHOP_Manual_080916.pdf">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/ENR_FFMSHOP_Manual_080916.pdf</a> (describing operational requirements
effective as of July 19, 2016, which were superseded by subsequent
publications).
---------------------------------------------------------------------------
Based on HHS' experience since we codified the currently-effective
interpretation of guaranteed availability, we believe the current
policy, has the unintended consequence of creating barriers to health
coverage that disproportionately affect low-income individuals, and is
therefore inconsistent with the intent of the guaranteed availability
statutory requirements. The current policy heightens the risk of
economic hardships for low-income individuals enrolled in health
insurance coverage with APTC. Individuals stop paying premiums (and
lose coverage due to nonpayment of premiums) for a variety of reasons
throughout the year. For example, commenters to the Market
Stabilization proposed rule stated that individuals who are victims of
crime, or those grappling with domestic violence,
[[Page 595]]
medical emergencies, incarceration, or other urgent circumstances are
often forced to make difficult financial decisions that may lead to
failure to pay their health insurance premiums. Even for some middle-
income families, the high cost of health care for multiple family
members with chronic health conditions may result in non-payment of
premiums.\22\ Requiring such individuals to pay back past-due premium
plus a binder payment prior to enrollment may present an insurmountable
barrier leading to gaps in coverage. For this reason, HHS is of the
view that the current interpretation of the guaranteed availability
requirement creates unnecessary barriers to accessing health coverage.
---------------------------------------------------------------------------
\22\ John Tozzi. (March 2018). ``Why Some Americans Are Risking
It and Skipping Health Insurance.'' Bloomberg News. Retrieved from
<a href="https://www.bloomberg.com/news/features/2018-03-26/why-some-americans-are-risking-it-and-skipping-health-insurance">https://www.bloomberg.com/news/features/2018-03-26/why-some-americans-are-risking-it-and-skipping-health-insurance</a>.
---------------------------------------------------------------------------
HHS is also concerned that the barriers created by the current
interpretation of guaranteed availability disproportionately affect
low-income enrollees for whom APTC is paid. Under federal law governing
grace periods for enrollees for whom APTC is paid, QHP issuers must
provide a 3-month grace period before they are allowed to terminate an
enrollee's coverage for non-payment of premiums and must continue to
provide coverage during the first month of the grace period. As a
result, those enrollees who are unable to satisfy outstanding premium
payments by the end of the 3-month grace period generally may owe at
least one month of past due premium after their coverage is terminated.
In contrast, grace period rules for individuals who are not eligible
for APTC are governed by state law. Many state laws allow for
termination back to the end of the period for which an enrollee paid
premium, in which case an enrollee without APTC whose coverage is
terminated for nonpayment would not owe past-due premium when they
attempt to enroll in coverage during a subsequent open enrollment or
special enrollment period. Enrollees for whom APTC is paid generally
may have household incomes as low as 100 percent of the federal poverty
level (FPL) (which, for the 2021 benefit year, is $12,760 for a single
person household).\23\ Thus, premium payment policies that require
payment of past-due premiums prior to effectuation of new coverage are
likely to disproportionately affect low-income enrollees with APTC, the
individuals who may be least able to pay all outstanding premium debt
among those seeking coverage in the individual market.
---------------------------------------------------------------------------
\23\ See 2021 Poverty Guidelines for the 48 Contiguous States
and the District of Columbia, available at <a href="https://aspe.hhs.gov/topics/poverty-economic-mobility/poverty-guidelines/prior-hhs-poverty-guidelines-federal-register-references/2020-poverty-guidelines">https://aspe.hhs.gov/topics/poverty-economic-mobility/poverty-guidelines/prior-hhs-poverty-guidelines-federal-register-references/2020-poverty-guidelines</a>.
---------------------------------------------------------------------------
Conditioning health insurance enrollment on the payment of past-due
premiums could disincentivize health insurance enrollment altogether,
reducing the rate of enrollment for low-income individuals. The
economic burden associated with being required to pay past-due premiums
prior to enrolling in new coverage may prevent low-income individuals
from enrolling in coverage and affect the demographics of the risk
pool. Various studies have found that low-income families often
struggle to balance out-of-pocket health care costs alongside rent or
mortgage payments, and other necessary living expenses.\24\ Maintaining
the current interpretation of the guaranteed availability rules would
uphold barriers to health insurance coverage for low-income
individuals, who face a greater risk of poorer health outcomes.\25\
Reverting to the previous interpretation of the guaranteed availability
rules would ensure individuals who stand to benefit the most from
health insurance coverage can enroll in coverage, and would promote
more equitable access to health insurance coverage. In addition, the
public health and economic crises caused by the COVID-19 pandemic
exacerbated the hardships facing low-income individuals and families.
The resulting financial and health insecurity caused by the pandemic
underscores the critical role that access to continuous health coverage
will continue to play during the ongoing and often unpredictable
challenges of the pandemic and beyond. Returning to the previous
interpretation of the guaranteed availability rule would remove a
barrier to accessing health coverage that compounds the economic
challenges from the COVID-19 crisis.
---------------------------------------------------------------------------
\24\ Tim Thomas, Ph.D.; Jose Hernandez, Ph.D.; et al. (2019).
The Evictions Study. The University of California Berkeley and the
University of Washington. Retrieved from <a href="https://evictions.study/index.html">https://evictions.study/index.html</a>.
\25\ P.J. Cunningham; T.L. Green; R.T. Braun. (February 2018).
Income Disparities in the Prevalence, Severity, and Costs of Co-
Occurring Chronic and Behavioral Health Conditions. Medical Care.
Retrieved from <a href="https://www.commonwealthfund.org/publications/journal-article/2018/feb/income-disparities-prevalence-severity-and-costs-co-occurring">https://www.commonwealthfund.org/publications/journal-article/2018/feb/income-disparities-prevalence-severity-and-costs-co-occurring</a>.
---------------------------------------------------------------------------
In the Market Stabilization rule, we noted concern that enrollees
with APTC may take advantage of guaranteed availability by declining to
make premium payments for coverage at the end of a benefit year without
losing coverage. Although this remains possible, we are of the view
that the disparate negative impact on low-income populations outweighs
the possible deterrent effect on individuals who may try taking
advantage of the guaranteed availability rules. We seek comment
regarding the frequency of any potential gaming behavior, as well as
information on the primary diagnoses and services that may be involved
in suspected gaming situations so that we may better assess any
contributing causes of such non-payment. For example, non-payment may
not be the result of gaming, but could be indicative of contextual
challenges individuals face in satisfying payment obligations. We are
particularly interested in comments from issuers that have not adopted
a premium payment policy that requires payment of past-due premiums
prior to effectuating enrollment. In addition, we note that issuers are
generally not permitted to forgive past-due premium debt, and can
pursue other mechanisms to collect past-due premiums. We believe this
mitigates the risk that some enrollees may take advantage of the
guaranteed availability rules.
We seek comment on this proposal.
b. Nondiscrimination Based on Sexual Orientation and Gender Identity
We propose to amend 45 CFR 147.104(e) such that its
nondiscrimination protections would explicitly prohibit discrimination
based on sexual orientation and gender identity. HHS previously
codified such nondiscrimination protections at Sec. 147.104(e), but
amendments made in 2020 to Sec. 147.104(e) removed any reference to
sexual orientation and gender identity. If finalized, this proposal
would revert Sec. 147.104(e) to the pre-2020 nondiscrimination
protections.
Section 147.104(e) states that a health insurance issuer and its
officials, employees, agents, and representatives must not employ
marketing practices or benefit designs that would have the effect of
discouraging the enrollment of individuals with significant health
needs in health insurance coverage or discriminate based on race,
color, national origin, present or predicted disability, age, sex,
expected length of life, degree of medical dependency, quality of life,
or other health conditions. Previously, in the 2014 Market Rules, we
finalized Sec. 147.104(e) to also prohibit discrimination based on
sexual orientation and gender
[[Page 596]]
identity.\26\ However, in the 2020 final rule that revised regulations
implementing section 1557 of the ACA, HHS also revised certain CMS
regulations, including those at Sec. 147.104(e), by removing sexual
orientation and gender identity as bases of discrimination subject to
the CMS regulations' nondiscrimination protections.\27\ The 2020
section 1557 final rule is the subject of ongoing litigation.\28\
---------------------------------------------------------------------------
\26\ 78 FR 13406 (February 27, 2013).
\27\ 85 FR 37160 (June 19, 2020); See id. at 37218-21 (the 2020
section 1557 final rule revised the following CMS regulations: 45
CFR 147.104, 155.120, 155.220, 156.200, 156.1230).
\28\ The 2020 section 1557 final rule is the subject of several
lawsuits and court orders. For more information, see <a href="https://www.hhs.gov/civil-rights/for-individuals/section-1557/index.html">https://www.hhs.gov/civil-rights/for-individuals/section-1557/index.html</a>.
---------------------------------------------------------------------------
Pursuant to section 1311(c)(1)(A) of the ACA, the HHS Secretary was
required to establish by regulation criteria for certification that
require QHP issuers to meet marketing requirements and not employ
marketing practices or benefit designs that will have the effect of
discouraging the enrollment of individuals with significant health
needs in QHPs. Under the authority of section 1321(a) of the ACA, which
provides the HHS Secretary broad rulemaking authority with respect to
the establishment and operation of Exchanges and the offering of QHPs
through such Exchanges, in the 2012 Exchange Standards final rule, CMS
codified a regulation implementing this requirement at Sec. 156.225.
Under the general rulemaking authority in section 2792 of the PHS Act,
which provides the HHS Secretary broad rulemaking authority to
promulgate regulations as may be necessary or appropriate to carry out
the provisions of title XXVII of the PHS Act, the 2014 Market Rules
adopted a similar standard in Sec. 147.104(e), applying this
requirement to the group and individual health insurance markets.
Furthermore, in order to ensure consistency against employing
discriminatory marketing practices and benefit designs, HHS finalized
Sec. 147.104(e) to align with other prohibitions on discrimination
that HHS had already codified at that time with respect to EHB in Sec.
156.125, with respect to standards applicable to QHPs under Sec.
156.200(e) that included protections against discrimination on the
basis of sexual orientation and gender identity, and with respect to
marketing standards in Sec. 156.225. The 2014 Market Rules further
clarified that discriminatory marketing practices or benefit designs
represent a failure by issuers to comply with the guaranteed
availability requirements in PHS Act section 2702, as such practices or
designs can have the effect of discouraging or preventing the
enrollment of individuals in health insurance coverage.
In the 2020 section 1557 final rule, HHS revised the section 1557
implementing regulation. Among other things, the rule removed the
definition of ``on the basis of sex,'' which included gender identity,
and instead purported to rely upon the ``plain meaning'' of the word
``sex'' in the underlying Title IX regulation.\29\ However, as HHS
noted in the 2020 section 1557 final rule, CMS possesses statutory
authority independent of section 1557 of the ACA to prohibit
discrimination in the group and individual markets.\30\
---------------------------------------------------------------------------
\29\ 85 FR 37160, 37166 (June 19, 2020). The 2016 and 2020
section 1557 final rules are the subject of several lawsuits and
court orders. For more information, see <a href="https://www.hhs.gov/civil-rights/for-individuals/section-1557/index.html">https://www.hhs.gov/civil-rights/for-individuals/section-1557/index.html</a>, <a href="https://www.hhs.gov/civil-rights/for-individuals/section-1557/index.html">https://www.hhs.gov/civil-rights/for-individuals/section-1557/index.html</a>.
\30\ 85 FR 37160, 37219, 37218-21 (June 19, 2020).
---------------------------------------------------------------------------
Following public posting of the 2020 section 1557 final rule on the
agency's website, the Supreme Court held in Bostock v. Clayton County,
140 S. Ct. 1731 (2020), that discrimination on the basis of sex under
Title VII of the Civil Rights Act of 1964 includes discrimination on
the basis of sexual orientation and gender identity. On January 20,
2021, the President signed Executive Order 13988 stating that it is the
Administration's policy to prevent and combat discrimination on the
basis of gender identity and sexual orientation, and that under
Bostock's reasoning, laws that prohibit sex discrimination also
prohibit discrimination on the basis of gender identity and sexual
orientation, so long as the laws do not contain sufficient indications
to the contrary.\31\ The Executive Order (E.O.) also instructed all
agency heads, including the HHS Secretary, to review all existing
regulations, guidance documents, and other agency actions to determine
whether they are consistent with the aforementioned policy, and to
consider whether to suspend, revise, or rescind any agency actions that
are inconsistent with it. The Department of Justice (DOJ) issued a
memorandum on March 26, 2021 that determined the court's reasoning in
Bostock applies to Title IX and thus that Title IX's prohibition on
discrimination on the basis of sex includes discrimination on the basis
of gender identity and sexual orientation.\32\ Following the E.O. and
DOJ's memorandum, HHS released on May 10, 2021 a Notice that HHS will
interpret and enforce section 1557's and Title IX's prohibition on
discrimination on the basis of sex to include: (1) Discrimination on
the basis of sexual orientation; and (2) discrimination on the basis of
gender identity.\33\
---------------------------------------------------------------------------
\31\ Executive Order 13988 on Preventing and Combating
Discrimination on the Basis of Gender Identity or Sexual
Orientation, 86 FR 7023 (Jan. 20, 2021).
\32\ U.S. Dep't of Justice, Memorandum on Application of Bostock
v. Clayton County to Title IX of the Education Amendments of 1972
(Mar. 26, 2021), <a href="https://www.justice.gov/crt/page/file/1383026/download">https://www.justice.gov/crt/page/file/1383026/download</a>. On June 16, 2021, the Department of Education's Office for
Civil Rights issued a similar Notice explaining that it too will
enforce Title IX's prohibition on discrimination on the basis of sex
to include: (1) Discrimination based on sexual orientation; and (2)
discrimination based on gender identity (available at <a href="https://www2.ed.gov/about/offices/list/ocr/docs/202106-titleix-noi.pdf">https://www2.ed.gov/about/offices/list/ocr/docs/202106-titleix-noi.pdf</a>).
\33\ 86 FR 27984.
---------------------------------------------------------------------------
Likewise, CMS is not relying on authority from section 1557 of the
ACA for the proposal at Sec. 147.104(e) or the parallel proposals to
nondiscrimination regulations at Sec. Sec. 155.120(c), 155.220(j),
156.125(b), 156.200(e), and 156.1230(b). We will further elaborate in
the respective preambles to Sec. Sec. 147.104(e), 155.120(c),
155.220(j), 156.125(b), 156.200(e), and 156.1230(b) the specific ACA
authority CMS is relying on to prohibit discrimination in the group and
individual markets. CMS proposes to exercise the same authority as it
exercised in the 2014 Market Rules to amend Sec. 147.104(e) to again
prohibit a health insurance issuer and its officials, employees,
agents, and representatives from discriminating in its marketing
practices or benefit designs on the basis of sexual orientation and
gender identity. Specifically, CMS proposes to again rely on section
2702 of the PHS Act, as well as section 2792 of the PHS Act, which
provides the HHS Secretary broad rulemaking authority to promulgate
regulations as may be necessary or appropriate to carry out the
provisions of title XXVII of the PHS Act. These are the same
authorities CMS relies upon for implementation of existing
nondiscrimination protections at Sec. 147.104(e). Utilizing these same
authorities to again prohibit discrimination based on sexual
orientation and gender identity would be consistent with the authority
CMS relies upon for those existing protections at Sec. 147.104(e) that
currently prohibit discrimination on the basis of race, color, national
origin, present or predicted disability, age, sex, expected length of
life, degree of medical dependency, quality of life, or other health
conditions.
People who identify as part of the lesbian, gay, bisexual,
transgender, and
[[Page 597]]
queer (LGBTQI+) community face pervasive health and health care
disparities, and are at higher risk for many concomitant conditions,
including substance use and \34\ mental health disorders, sexually
transmitted infections,\35\ HIV,\36\ cancer, cardiovascular disease,
and obesity.\37\ Overall, LGBTQI+ people report being in poorer health
than non-LGBTQI+ individuals. LGBTQI+ people of all genders are more
likely to become disabled at a younger age than heterosexual
individuals.\38\ In addition to disparities in health outcomes, LGBTQI+
people face barriers to obtaining appropriate health care and
transgender people who can access insurance may nonetheless be denied
coverage for needed services. For example, nearly half of transgender
respondents in one survey said their health insurance company denied
them gender affirming surgery,\39\ and a similar proportion reported
that they were denied coverage for hormone therapy.\40\ Beyond health
coverage issues, LGBTQI+ people may struggle to access care because of
cost barriers. LGBTQI+ people are also more likely than others to
report postponing or forgoing health care due to costs, and costs were
an even greater obstacle for younger LGBTQI+ people and those who are
transgender--especially transgender people of color.\41\
---------------------------------------------------------------------------
\34\ Hilary Daniel et al, Annals of Internal Med. Position
Papers, Lesbian, Gay, Bisexual, and Transgender Health Disparities:
Executive Summary of a Policy Position Paper From the American
College of Physicians (July 21, 2105), <a href="https://www.acpjournals.org/doi/full/10.7326/M14-2482?journalCode=aim">https://www.acpjournals.org/doi/full/10.7326/M14-2482?journalCode=aim</a>.
\35\ Hilary Daniel et al, Annals of Internal Med. Position
Papers, Lesbian, Gay, Bisexual, and Transgender Health Disparities:
Executive Summary of a Policy Position Paper From the American
College of Physicians (July 21, 2105), <a href="https://www.acpjournals.org/doi/full/10.7326/M14-2482?journalCode=aim">https://www.acpjournals.org/doi/full/10.7326/M14-2482?journalCode=aim</a>.
\36\ U.S. Dep't of Health & Human Servs., Ctrs. for Disease
Control and Prevention, HIV Surveillance Report, 2019; Vol. 32 (May
2021), <a href="https://www.cdc.gov/hiv/pdf/library/reports/surveillance/cdc-hiv-surveillance-report-2018-updated-vol-32.pdf">https://www.cdc.gov/hiv/pdf/library/reports/surveillance/cdc-hiv-surveillance-report-2018-updated-vol-32.pdf</a>.
\37\ See, for example, Lesbian, Gay, Bisexual, and Transgender
Health, Healthy People 2020, https://www.healthypeople.gov/2020/
topics-objectives/topic/lesbian-gay-bisexual-and-transgender-
health#:~:text=Research%20suggests%20that%20LGBT%20individuals,%2C2%2
C%203%20and%20suicide; Hafeez, Hudaisa et al. ``Healthcare
Disparities Among Lesbian, Gay, Bisexual, and Transgender Youth: A
Literature Review.'' Cureus vol. 9,4 e1184. 20 Apr. 2017,
doi:10.7759/cureus.1184 (<a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5478215/">https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5478215/</a>); Fredriksen-Goldsen KI, Kim H-J, Barkan SE, Muraco A
and Hoy-Ellis CP (2013) Health disparities among lesbian, gay, and
bisexual older adults: Results from a population-based study.
American Journal of Public Health 103, 1802-1809; Billy A. Caceres
et al. ``A Systematic Review of Cardiovascular Disease in Sexual
Minorities'', American Journal of Public Health 107, no. 4 (April 1,
2017): pp. e13-e21.
\38\ Hilary Daniel et al, Annals of Internal Med. Position
Papers, Lesbian, Gay, Bisexual, and Transgender Health Disparities:
Executive Summary of a Policy Position Paper From the American
College of Physicians (July 21, 2105), <a href="https://www.acpjournals.org/doi/full/10.7326/M14-2482?journalCode=aim">https://www.acpjournals.org/doi/full/10.7326/M14-2482?journalCode=aim</a>.
\39\ For purposes of this preamble, the term ``gender affirming
care'' means gender affirming care for transgender individuals. This
may also be referred to as ``transition related care.''
\40\ Sharita Gruberg et al, Center for American Progress, The
State of the LGBTQ Community in 2020 (Oct. 6, 2020), <a href="https://www.americanprogress.org/issues/lgbtq-rights/reports/2020/10/06/491052/state-lgbtq-community-2020/">https://www.americanprogress.org/issues/lgbtq-rights/reports/2020/10/06/491052/state-lgbtq-community-2020/</a>.
\41\ Sharita Gruberg et al, Center for American Progress, The
State of the LGBTQ Community in 2020 (Oct. 6, 2020), <a href="https://www.americanprogress.org/issues/lgbtq-rights/reports/2020/10/06/491052/state-lgbtq-community-2020/">https://www.americanprogress.org/issues/lgbtq-rights/reports/2020/10/06/491052/state-lgbtq-community-2020/</a>.
---------------------------------------------------------------------------
We believe that prohibiting discrimination based on sexual
orientation or gender identity can lead to improved health outcomes for
this community \42\ and that the removal of such protections in the
2020 section 1557 final rule frustrated not only guaranteed
availability requirements, but also the broader aim of improving health
equity. Without protection from discrimination, individuals may
continue to face barriers to accessing medically necessary health care.
For example, without protection from discrimination, transgender
individuals may face barriers or be denied medically necessary gender-
affirming care. We believe amending the nondiscrimination protections
as proposed at Sec. 147.104(e) to again explicitly prohibit
discrimination based on sexual orientation and gender identity is
warranted in light of the existing trends in health care discrimination
and to better address barriers to health equity for LGBTQI+
individuals.\43\ As proposed, such revisions to Sec. 147.104(e) would
also support the original objective of ensuring consistency against
employing discriminatory marketing practices and benefit designs, as we
are proposing parallel changes to nondiscrimination regulations at
Sec. Sec. 147.104(e), 155.120(c), 155.220(j), 156.125(b), 156.200(e),
and 156.1230(b).
---------------------------------------------------------------------------
\42\ Ward, BW, Dahlhamer, JM, Galinsky, AM, and Joestl, SS.
Sexual Orientation & Health Among U.S. Adults: National Health
Interview Survey, CDC National Health Statistics Report 77, 2014.
\43\ Nguyen, T.T., Vable, A.M., Glymour, M.M. et al. Trends for
Reported Discrimination in Health Care in a National Sample of Older
Adults with Chronic Conditions. J GEN INTERN MED 33, 291-297 (2018).
<a href="https://doi.org/10.1007/s11606-017-4209-5">https://doi.org/10.1007/s11606-017-4209-5</a>.
---------------------------------------------------------------------------
If any of the provisions at Sec. Sec. 147.104(e), 155.120(c),
155.220(j), 156.125(b), 156.200(e), and 156.1230(b) are held to be
invalid or unenforceable by its terms, or as applied to any person or
circumstance, it shall be severable from this part and shall not affect
the remainder thereof or the application of the provision to other
persons not similarly situated or to other dissimilar circumstances. In
enforcing the nondiscrimination provisions in the corresponding CMS
regulations, HHS will comply with laws protecting the exercise of
conscience and religion, including the Religious Freedom Restoration
Act (42 U.S.C. 2000bb through 2000bb-4) and all other applicable legal
requirements.
We seek comment on this proposal.
C. Part 153--Standards Related to Reinsurance, Risk Corridors, and Risk
Adjustment
In subparts A, D, G, and H of part 153, we established standards
for the administration of the risk adjustment program. The risk
adjustment program is a permanent program created by section 1343 of
the ACA that transfers funds from lower-than-average risk, risk
adjustment covered plans to higher-than-average risk, risk adjustment
covered plans in the individual, small group markets, or merged
markets, inside and outside the Exchanges. In accordance with Sec.
153.310(a), a state that is approved or conditionally approved by the
Secretary to operate an Exchange may establish a risk adjustment
program, or have HHS do so on its behalf.\44\ HHS did not receive any
requests from states to operate risk adjustment for the 2023 benefit
year. Therefore, HHS will operate risk adjustment in every state and
the District of Columbia for the 2023 benefit year.
---------------------------------------------------------------------------
\44\ Also see 42 U.S.C. 18041(c)(1).
---------------------------------------------------------------------------
1. Sequestration
In accordance with the OMB Report to Congress on the Joint
Committee Reductions for Fiscal Year 2022, the permanent risk
adjustment program is subject to the fiscal year 2022
sequestration.\45\ The federal government's 2022 fiscal year begins
October 1, 2021. Therefore, the risk adjustment program will be
sequestered at a rate of 5.7 percent for payments made from fiscal year
2022 resources (that is, funds collected during the 2022 fiscal year).
---------------------------------------------------------------------------
\45\ <a href="https://www.whitehouse.gov/wp-content/uploads/2021/05/BBEDCA_251A_Sequestration_Report_FY2022.pdf">https://www.whitehouse.gov/wp-content/uploads/2021/05/BBEDCA_251A_Sequestration_Report_FY2022.pdf</a>.
---------------------------------------------------------------------------
HHS, in coordination with OMB, has determined that, under section
256(k)(6) of the Balanced Budget and Emergency Deficit Control Act of
1985 (Pub. L. 99-177, enacted December 12, 1985), as
[[Page 598]]
amended, and the underlying authority for the risk adjustment program,
the funds that are sequestered in fiscal year 2022 from the risk
adjustment program will become available for payment to issuers in
fiscal year 2023 without further Congressional action. If Congress does
not enact deficit reduction provisions that replace the Joint Committee
reductions, the program would be sequestered in future fiscal years,
and any sequestered funding would become available in the fiscal year
following that in which it was sequestered.
Additionally, we note that the Coronavirus Aid, Relief, and
Economic Security (CARES) Act amended section 251A(6) of the Balanced
Budget and Emergency Deficit Control Act of 1985 and extended
sequestration for the risk adjustment program through fiscal year 2030
at a rate of 5.7 percent per fiscal year.\46\
---------------------------------------------------------------------------
\46\ <a href="https://www.congress.gov/116/bills/s3548/BILLS-116s3548is.pdf">https://www.congress.gov/116/bills/s3548/BILLS-116s3548is.pdf</a>.
---------------------------------------------------------------------------
2. HHS Risk Adjustment (Sec. 153.320)
The HHS risk adjustment models predict plan liability for an
average enrollee based on that person's age, sex, and diagnoses (also
referred to as hierarchical condition categories (HCCs)), producing a
risk score. The HHS risk adjustment methodology utilizes separate
models for adults, children, and infants to account for clinical and
cost differences in each age group. In the adult and child models, the
relative risk assigned to an individual's age, sex, and diagnoses are
added together to produce an individual risk score. Additionally, to
calculate enrollee risk scores in the adult models, we added enrollment
duration factors beginning with the 2017 benefit year, and prescription
drug categories (RXCs) beginning with the 2018 benefit year.\47\ Infant
risk scores are determined by inclusion in one of 25 mutually exclusive
groups, based on the infant's maturity and the severity of diagnoses.
If applicable, the risk score for adults, children, or infants is
multiplied by a CSR factor. The enrollment-weighted average risk score
of all enrollees in a particular risk adjustment covered plan (also
referred to as the plan liability risk score) within a geographic
rating area is one of the inputs into the risk adjustment state payment
transfer formula, which determines the state transfer payment or charge
that an issuer will receive or be required to pay for that plan for the
applicable state market risk pool. Thus, the HHS risk adjustment models
predict average group costs to account for risk across plans, in
keeping with the Actuarial Standards Board's Actuarial Standards of
Practice for risk classification.
---------------------------------------------------------------------------
\47\ For the 2018 benefit year, there were 12 RXCs, but starting
with the 2019 benefit year, the two severity-only RXCs were removed
from the adult risk adjustment models. See, for example, 83 FR
16941.
---------------------------------------------------------------------------
a. Data for Risk Adjustment Model Recalibration for 2023 Benefit Year
and Beyond
We are proposing to recalibrate the 2023 benefit year risk
adjustment models with the 2017, 2018, and 2019 enrollee-level EDGE
data. Consistent with the approach outlined in the 2020 Payment Notice
to no longer rely upon MarketScan[supreg] data for recalibrating the
risk adjustment models, we will recalibrate the risk adjustment models
for the 2023 benefit year using only enrollee-level EDGE data, and we
will continue to use blended, or averaged, coefficients from the 3
years of separately solved models for the 2023 benefit year model
recalibration.\48\ Additionally, as outlined in the 2022 Payment
Notice, we will use the 3 most recent consecutive years of enrollee-
level EDGE data that are available at the time we incorporate the data
in the draft recalibrated coefficients published in the proposed rule
for the applicable benefit year,\49\ and will not update the
coefficients between the proposed and final rules if an additional year
of enrollee-level EDGE data becomes available for incorporation.\50\ We
believe this promotes stability, better meets the goal of the risk
adjustment program, and allows issuers more time to incorporate this
information when pricing their plans for the upcoming benefit year.
---------------------------------------------------------------------------
\48\ 84 FR 17463 through 17466.
\49\ While we do receive the next year of enrollee-level EDGE
data prior to the proposed rule, that data must go through several
quality and analysis checks before it is useable for risk adjustment
model recalibration.
\50\ 86 FR 24140 at 24152.
---------------------------------------------------------------------------
As such, we propose to determine coefficients for the 2023 benefit
year based on a blend of separately solved coefficients from the 2017,
2018, and 2019 benefit years' enrollee-level EDGE data.\51\ The draft
coefficients listed in Tables 1 through 6 reflect the use of 2017,
2018, and 2019 benefit year enrollee-level EDGE data, as well as other
risk adjustment model updates proposed in this proposed rule (including
changes to the model specifications, the pricing adjustment to
Hepatitis C drugs, and the removal of the mapping of hydroxychloroquine
sulfate to an RXC). However, we note that the coefficients could change
if we identify an error or if some or all of the proposed model changes
are not finalized or are modified in response to comments. In addition,
consistent with Sec. 153.320(b)(1)(i), if we are unable to finalize
the final coefficients in time for publication in the final rule, we
would publish the final coefficients for the 2023 benefit year in
guidance soon after the publication of the final rule. We seek comment
on the proposal to determine 2023 benefit year coefficients based on a
blend of separately solved coefficients from the 2017, 2018, and 2019
enrollee-level EDGE data.
---------------------------------------------------------------------------
\51\ As discussed later in this proposed rule, we propose to
remove the mapping of hydroxychloroquine to RXC 09 (Immune
Suppressants and Immunomodulators) and the related RXC 09
interactions.
---------------------------------------------------------------------------
We also solicit comments on the future use of the 2020 enrollee-
level EDGE data due to the COVID-19 PHE. Under current policy, 2020
enrollee-level EDGE data would be used in recalibration of the HHS risk
adjustment models for the 2024 benefit year and that data would
continue to be used for the 2025 and 2026 benefit year models.\52\
Although HHS has not analyzed the 2020 enrollee-level EDGE data yet, we
solicit comment on the future use of the 2020 enrollee-level EDGE data
for the annual recalibration of the HHS risk adjustment models.
---------------------------------------------------------------------------
\52\ Consistent with the approach finalized in the 2022 Payment
Notice, use of the 3 most recent consecutive years of enrollee-level
EDGE data would result in the use of 2018, 2019, and 2020 enrollee-
level EDGE data for the recalibration of the 2024 benefit year
models; the use of 2019, 2020, and 2021 enrollee-level EDGE data for
recalibration of the 2025 benefit year models; and the use of 2020,
2021, and 2022 enrollee-level EDGE data for recalibration of the
2026 benefit year models.
---------------------------------------------------------------------------
b. Risk Adjustment Model Updates
Beginning with the 2023 benefit year, we are proposing three
modeling updates to the risk adjustment models. Consistent with the
potential model updates discussed in the 2021 RA Technical Paper, we
propose the following model updates, which are the same as those
proposed but not finalized in the 2022 Payment Notice: \53\ (1) Adding
a two-stage weighted model specification to the adult and child models;
(2) removing the severity illness factors in the adult models and
[[Page 599]]
replacing them with new severity and transplant indicators interacted
with HCC count factors in the adult and child models; and (3) replacing
the current enrollment duration factors in the adult models with HCC-
contingent enrollment duration factors in the adult models.
---------------------------------------------------------------------------
\53\ See 85 FR 78572 at 78583-78586. In the 2022 Payment Notice
Final Rule, in response to comments, we did not finalize the
proposed updates and announced that we would publish a technical
paper on the proposed model changes; see 86 FR 24140 at 24151-24162.
See also the 2021 HHS-Operated Risk Adjustment Technical Paper on
Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a> and the HHS-Operated Risk
Adjustment Technical Paper on Possible Model Changes: Summary
Results for Transfer Simulations, available at <a href="https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs">https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs</a>.
---------------------------------------------------------------------------
As described in prior rulemakings and in the 2021 RA Technical
Paper, the current HHS-HCC models, which are linear models,
underpredict plan liability for enrollees without HCCs and the lowest
expected expenditures, underpredict plan liability for enrollees with
the highest HCC counts and the highest expected expenditures, and
underpredict plan liability for partial-year enrollees with HCCs.\54\
The proposals in this proposed rule are intended to improve the risk
adjustment adult and child models' prediction for these subpopulations.
We released the 2021 RA Technical Paper in response to stakeholder
requests for more information on the impacts of these proposals before
they were adopted and released simulated transfer estimates reflecting
the combination of these proposed changes in December 2021.\55\ We
continue to believe the combination of these proposed model changes
will improve the current models' predictive accuracy for the lowest-
risk enrollees, certain partial-year adult enrollees, and the very
highest-risk enrollees, while limiting trade-offs in other areas of
model performance and complexity. As such, we are re-proposing these
combined model specification changes in this rule, and the following
sections describe these proposed model specification changes in detail.
---------------------------------------------------------------------------
\54\ See, for example, 85 FR 29164 at 29188-29190; 85 FR 78572
at 78583-78586; and 86 FR 24140 at 24151-24162. See also the 2021
HHS-Operated Risk Adjustment Technical Paper on Possible Model
Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\55\ See the 2021 HHS-Operated Risk Adjustment Technical Paper
on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a> and the HHS-Operated Risk
Adjustment Technical Paper on Possible Model Changes: Summary
Results for Transfer Simulations, available at <a href="https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs">https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs</a>.
Issuers that participated in the simulation also received issuer-
specific data, including risk score and transfer estimates for the
simulated results.
---------------------------------------------------------------------------
i. Two-Stage Weighted Model Specification
We propose to use a two-stage weighted model specification to
recalibrate the adult and child risk adjustment models starting with
the 2023 benefit year to improve the underprediction of plan liability
for the lowest-risk enrollees (that is, enrollees in low risk deciles
and enrollees without HCCs).\56\ Since approximately 80 percent of
enrollees in the individual and small group (or merged) markets do not
have HCCs, this underprediction, while small in magnitude, represents a
large number of enrollees.\57\
---------------------------------------------------------------------------
\56\ When we refer to the enrollees without HCCs, we are
referring to enrollees without payment HCCs.
\57\ See Chapter 2 of the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>, and the HHS-
Operated Risk Adjustment Technical Paper on Possible Model Changes:
Summary Results for Transfer Simulations, available at <a href="https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs">https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs</a>.
---------------------------------------------------------------------------
To improve prediction for the lowest-risk enrollees, we explored
calibrating the adult and child models in two stages to reweight the
healthier enrollees more heavily. In the first-stage estimation, the
model coefficients would be estimated using the current model
specifications; and in the second stage, we would re-estimate the model
weighting enrollees in the recalibration sample by the capped
reciprocal of the predicted values of relative expenditures from the
first step estimation with the same model specification. More
specifically, the first stage of this proposed weighted estimation
method for the adult models involves a linear regression (weighted by
the person-specific eligibility fraction of the number of months
enrolled divided by 12) of simulated plan liability \58\ on age-sex
factors, payment HCC factors, severity illness factors,\59\ the
enrollment duration factors,\60\ and RXCs. For the child models, the
first stage of the proposed weighted estimation method involves a
linear regression of simulated plan liability on age-sex factors and
payment HCC factors.\61\ The methodology for conducting the proposed
first stage regression would be essentially identical to the current
adult and child risk adjustment recalibrations. The second stage of the
proposed two-stage weighted model specification involves using
recalibration sample enrollees' inverse (also referred to as
reciprocal) capped predictions from the first stage as weights for a
second linear regression. As such, this step has the material effect of
weighting healthier enrollees more heavily so that the statistical
model predicts their expenditures more accurately. It also
systematically reduces the influence of very expensive enrollees on the
final model factors.
---------------------------------------------------------------------------
\58\ We simulate plan liability expenditures for each metal
level for each enrollee in the recalibration dataset (that is, we
apply different standardized benefit design parameters to the same
sample for each metal level). See <a href="https://www.cms.gov/mmrr/Downloads/MMRR2014_004_03_a03.pdf">https://www.cms.gov/mmrr/Downloads/MMRR2014_004_03_a03.pdf</a>.
\59\ We are also proposing to remove the current severity
illness indicators in the adult models and add new severity and
transplant indicators interacted with HCC count factors in the adult
and child models, as described elsewhere in this proposed rule.
\60\ We are also proposing to modify the enrollment duration
factors in the adult models, as described elsewhere in this proposed
rule.
\61\ See supra note 58.
---------------------------------------------------------------------------
To help provide stability to the proposed two-stage weighted model
specification, we imposed lower and upper bound caps on the first-stage
predictions at the 2.5th and 97.5th percentiles in the adult models,
and the 2.5th and 99.5th percentiles in the child models. This capped
weighted approach avoids excessively large or small weights for any
observations for the second stage estimation, and therefore mitigates
the potential to underpredict at the high end for expensive enrollees,
as well as any possible low-end overprediction of healthier enrollees.
We tested various caps for the weights based on the distribution of
costs and found these lower and upper bound caps achieved better
prediction on average.\62\
---------------------------------------------------------------------------
\62\ See Section 2.2 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>. Also see 85
FR at 78667 and 86 FR at 24283.
---------------------------------------------------------------------------
Additionally, in our consideration of the two-stage weighted model
specification, we tested various methods of determining weights for the
second stage, including reciprocals of the square root of predictions,
log of predictions, and residuals from the first stage estimation, but
the reciprocal of the capped predictions from the first stage resulted
in better predictive ratios for low-cost enrollees compared to any of
these alternative weighting functions.\63\
---------------------------------------------------------------------------
\63\ Ibid.
---------------------------------------------------------------------------
Our conceptual reasoning for pursuing the two-stage weighted model
specification is to retain the simple linear, additive structure of the
current models while forcing the model to better predict lowest-risk
enrollees, who our analyses identified as underpredicted in the current
adult and child models. Based on analyses using 2018 enrollee-level
EDGE data, the two-stage weighted approach significantly improves the
predictive ratios (PRs) of the lower deciles and the PRs for enrollees
without HCCs compared to the current models.\64\ Similar results were
also seen when using 2016 and 2017 enrollee-
[[Page 600]]
level EDGE data.\65\ In addition, the two-stage weighted approach
eliminated the overprediction observed in risk decile 8.\66\ We also
found that the two-stage weighted approach did not meaningfully change
factor coefficients for most HCCs, providing stability to the risk
adjustment model factors.
---------------------------------------------------------------------------
\64\ See Figure 2.2 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\65\ The PRs calculated in the 2021 RA Technical Paper are
calculated using the same samples on which the models were
calibrated. However, as is common practice in evaluating model fit,
we also tested splitting the sample for calibration and validation
purposes and the results were unchanged. Further, for purposes of
the analysis in the 2021 RA Technical Paper, we calculated PRs for
at least three data years and the results always appear the same. We
therefore generally only reported results in the 2021 RA Technical
Paper from the 2018 data year, which was the most recently available
dataset at the time that we ran these analyses in preparation for
announcing the proposed model changes in the proposed 2022 Payment
Notice.
\66\ See Figure 2.2 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
---------------------------------------------------------------------------
At the same time, we also considered whether the two-stage weighted
approach worsens the fit of the models along other dimensions,
identifying three areas that had minor, negative impacts on the model
fit. First, the two-stage weighted approach predicts plan liability by
age-sex factor less accurately than the current models, especially for
younger and older women. Overall, we considered this to be an
acceptable trade-off, because across all age and sex factors, most PRs
were within a tolerable threshold of +/-5 percent (for example, 0.95 to
1.05), and the two-stage weighted approach has the major benefit of
more accurately predicting the age-sex factors for the enrollees
without HCCs, which is a much larger population than enrollees with
HCCs. Second, the two-stage weighted approach is somewhat less accurate
at predicting certain HCCs, with the two-stage weighted approach
worsening adult model silver plan PRs by at least 5 percentage points
for 14 (out of 91) ungrouped HCCs and 3 (out of 18) grouped HCCs. For
the vast majority of HCCs, the impact is very small and most affected
HCCs or HCC groups have small sample sizes.\67\ Again, we considered
this reduced accuracy to be an acceptable trade-off because most of the
PRs for the two-stage weighted approach were within a tolerable
threshold of +/-5 percent (for example, 0.95 to 1.05), most enrollees
do not have HCCs, and the two-stage weighted approach predicts plan
liability better for those no HCC enrollees. Third, the two-stage
weighted approach had lower R-squared values compared to the current
models. However, the decrease in R-squared is at most 0.1 percentage
points for all metal levels, which is a minor reduction in fit across
models.\68\ Similar to the worsening of the age-sex cell and the HCC
PRs, we were not concerned about the lower R-squared as the reduction
in fit was minor at all metal levels, the values remained within the
range of R-squared statistics of other concurrent models predicting
expenditures for commercial insurance enrollees,\69\ and the proposed
two-stage weighted model specification better predicts plan liability
for enrollees with no HCCs, which is the majority of enrollees. After
considering the impact of the approach on model performance, we
determined that the proposed two-stage weighted model specification
does not have material unintended consequences in model performance and
achieves the aim of improving the predictive accuracy of the current
adult and child models for enrollees in the lowest risk deciles and for
enrollees without HCCs. For these reasons, we believe that the two-
stage weighted approach can improve prediction for lowest-risk
enrollees with limited trade-offs in other parts of the models'
performance. Therefore, we are proposing to add the two-stage weighted
model specification to the adult and child models beginning with the
2023 benefit year in combination with the proposed interacted HCC
counts model specification and the updated adult model enrollment
duration factors described later in this proposed rule.
---------------------------------------------------------------------------
\67\ For example, only one HCC or HCC group whose PR was
identified in our analysis as worsening by at least 5 percentage
points was present in greater than 1 percent of the adult silver
plan enrollees in the 2018 enrollee-level EDGE dataset (HCC 142
Specified Heart Arrhythmias). Our analysis found that all other HCCs
had recalibration dataset frequencies of less than 0.5 percent of
enrollees. See Chapter 2.3 and Table 2.1 in the 2021 HHS-Operated
Risk Adjustment Technical Paper on Possible Model Changes, available
at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\68\ See Figure 2.6 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\69\ See Winkelman, R., & Mehmud, S. (2007). A Comparative
Analysis of Claims-Based Tools for Health Risk Assessment.
Schaumberg, IL: Society of Actuaries.
---------------------------------------------------------------------------
In the 2021 RA Technical Paper, we explained that we believe that
by addressing the underprediction of costs associated with lowest-risk
enrollees in the adult and child models, we could further encourage the
retention and offering of plans that enroll a higher proportion of this
subpopulation of enrollees. We believe issuers offering these types of
plans are at greater risk of exiting the market if transfers calculated
under the state payment transfer formula undercompensate for the true
plan liability of the lowest-risk enrollees. We received stakeholder
comments in this regard, noting that the underprediction of the lowest-
risk enrollees could disincentivize issuers from attracting healthy
enrollees to their plans, thereby undermining the goals of developing a
healthy and stable market and encouraging competition on the basis of
high quality rather than risk selection. However, other stakeholders
have questioned if we should focus model changes on improving
prediction for the lowest-risk enrollees when the risk adjustment
program is intended to reduce incentives for issuers to avoid enrolling
individuals with higher risk.
We also received comments concerned that the two-stage weighted
model would be redundant of other elements in the state payment
transfer formula, which stated that the administrative cost adjustment
to statewide average premium \70\ already addresses some of the
underprediction of the lowest-risk enrollees in the risk adjustment
models. We clarify that the proposed two-stage weighted model
specification and existing administrative cost adjustment to statewide
average premium are not redundant and address separate considerations.
As detailed in the 2018 Payment Notice, the purpose of the
administrative cost adjustment to statewide average premium is to
exclude fixed administrative costs that are not dependent on enrollee
risk, such as taxes.\71\ In contrast, and as previously described
elsewhere,\72\ the purpose of the proposed two-stage weighed model
specification is to improve the current adult and child models'
prediction for the lowest risk enrollees.
---------------------------------------------------------------------------
\70\ 81 FR at 94099-94100.
\71\ See 81 FR at 61488-61489. Also see 81 FR at 94099-94100.
\72\ See Section 2.2 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>. Also see 85
FR at 78667 and 86 FR at 24283.
---------------------------------------------------------------------------
We seek comment on the two-stage weighted model specification
proposal, specifically regarding whether we should implement the
proposed two-stage weighted model specification alone, independent of
the other proposed model specification changes outlined in this rule,
beginning with the 2023 benefit year; whether we should implement the
proposed two-stage weighted model specification in conjunction with
these other proposals; or whether we should not implement the two-stage
weighted model specification at all. Additionally, given the
stakeholder comments we received
[[Page 601]]
questioning the need for this type of model update, we also generally
solicit comments on whether we should seek to improve the current
models' prediction for the lowest-risk enrollees.
ii. Interacted HCC Counts Model Specification
In addition to the two-stage weighted approach, we are proposing to
add an interacted HCC counts model specification to the adult and child
risk adjustment models starting with the 2023 benefit year to address
the current models' underprediction of plan liability for the very
highest-risk enrollees (that is, those in the top risk percentile and
those enrollees with the most HCCs). While this highest-risk
subpopulation represents a small number of enrollees, it represents a
large portion of expenditures. As described in the 2021 RA Technical
Paper, enrollees in risk decile 10 represent roughly 74.29 percent of
actual plan liability, compared to only 1.36 percent for enrollees in
risk decile 1.\73\ We found that for enrollees with a high HCC count,
there is an increasing, non-linear effect that leads to higher costs
than are currently predicted by adding up the incremental effects of
each HCC.
---------------------------------------------------------------------------
\73\ See Table 4.1 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
---------------------------------------------------------------------------
Therefore, to address the underprediction of the highest-cost
enrollees, we explored the addition of severity and transplant factors
interacted with HCC counts in the adult and child models, wherein a
factor flagging the presence of at least one severe or transplant
payment HCC is interacted with counts of the enrollee's payment
HCCs.\74\ The purpose of adding severity and transplant factors
interacted with HCC count factors to the adult and child models is to
address the underprediction of the highest risk enrollees (as the
proposed two-stage-weighted model specification addresses the
underprediction of the healthiest enrollees) by accounting for the fact
that costs of certain HCCs rise significantly when they occur with
multiple other HCCs. Specifically, the goals of this approach were to:
---------------------------------------------------------------------------
\74\ For HCCs in a coefficient estimation group, the group is
counted at most once. These groups of HCCs in the HHS risk
adjustment adult and child models are detailed in the HHS-Developed
Risk Adjustment Model Algorithm ``Do It Yourself (DIY)'' Software
``Additional Adult Variables'' and ``Additional Child Variables''
table logic (Tables 6 and 7 in the 2021 Benefit Year DIY Software).
The August 3, 2021 version of the DIY software is available at
<a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance</a>.
---------------------------------------------------------------------------
1. Address the non-linearity in costs between enrollees without
HCCs or with very low costs and enrollees with multiple HCCs or with
high costs;
2. Empirically incorporate the cost impact of multiple complex
diseases; and
3. Reduce incentives for coding proliferation to mitigate the
gaming concerns with HCC counts models.
In developing this interacted HCC counts approach, we identified
common HCCs for enrollees with extremely high costs, as well as HCCs
that were being underpredicted in the current risk adjustment adult and
child models. We found that many of the HCCs that were flagged as being
underpredicted were the current severe illness HCCs, the transplant
HCCs, and other HCCs related to the severity of disease. Therefore, we
considered dropping the current severity illness factors in the adult
models and replacing them with severity and transplant factors
interacted with HCC count factors in the adult models, as well as
adding the severity and transplant factors interacted with HCC count
factors to the child models.
We propose the inclusion of the factors in Tables 1 and 2 as the
interacted severity and transplant factors in the adult and child
models starting with the 2023 benefit year. We separated out transplant
HCCs and severity HCCs into their own separate set of interacted
factors, as expressed in Tables 1 and 2, because we found that this
approach improved prediction for high-cost enrollees better than an
approach that combined severity and transplant HCCs into a single set
of factors. Furthermore, under the current risk adjustment models,
adult severity illness interaction factors are collapsed into a single
binary variable indicating the presence of any severity illness
interaction. In contrast, the proposed severity factors would not be
collapsed and would instead be separated out by the HCC count with
which the severity or transplant illness indicator was interacted.
We defined the new proposed interaction factors such that an
enrollee would receive one or more of these factors if they had any
HCCs in the severity or transplant indicator groups in Table 3 and
according to how many HCCs were recorded in the enrollee's data in
total. As such, the proposed severity and transplant interaction
factors would express the presence of one or more of the selected
severity or transplant HCCs in Table 3. That is, an enrollee must have
at least one HCC in the ``severity'' or ``transplant'' indicator groups
in Table 3 to receive the interacted HCC count factor toward their risk
score, but would not receive any additional flags for having more than
one of the ``severity'' or ``transplant'' HCCs in an indicator group
beyond the total HCC count.
The proposed severity-HCC-count-interaction factors were calculated
as 10 separate factors for the adult models, and seven separate factors
for the child models. In the adult models, the first nine factors
specified the presence of (1) an HCC in the severity list in Table 3
and (2) exactly one payment HCC in the enrollee's data, exactly two,
exactly three, and so on, up to exactly nine payment HCCs. The tenth
factor specified the presence of (1) an HCC in the severity list in
Table 3 and (2) ten or more payment HCCs in the enrollee's data. For
the child models, the first five factors represented the presence of
(1) an HCC in the severity list in Table 3 and (2) exactly one payment
HCC in the enrollee's data, exactly two, exactly three, and so on, but
the sixth factor represents the presence of (1) an HCC in the severity
list in Table 3 and (2) six to seven payment HCCs, and the seventh
factor represents the presence of (1) an HCC in the severity list in
Table 3 and (2) eight or more payment HCCs in the enrollee's data.
The proposed transplant-HCC-count-interaction factors were
calculated similarly. However, the transplant factors were calculated
using a different range of HCC counts. In the adult models, five
separate transplant interaction factors were created, representing the
presence of (1) an HCC in the transplant list in Table 3 and (2)
payment HCC counts of exactly four, exactly five, exactly six, exactly
seven, and eight or more payment HCCs in the enrollee's data. For the
child models, we created only one transplant interaction factor
indicating the presence of (1) an HCC in the transplant list in Table 3
and (2) a total of four or more payment HCCs in the enrollee's data. As
detailed later in this section, this treatment of transplant-HCC-count-
interaction factors stabilized the child model estimates by increasing
the sample size used to estimate the factor coefficients.
To illustrate how the proposed severity- (or transplant-) HCC-
count-interaction factors would be assigned to an enrollee, consider an
adult enrollee with four payment HCCs, one of which is HCC 34 ``Liver
Transplant Status/Complications''. Because HCC 34 appears in both the
severity and transplant indicator groups in Table 3, this enrollee
would receive the following factor coefficients toward their risk score
in the adult models: (1) The four factor coefficients for their
individual HCCs (the three non-transplant HCC factors and the HCC 34
[[Page 602]]
transplant HCC factor), (2) the factor coefficient for the severity-
HCC-count-interaction indicating four payment HCCs, and (3) the factor
coefficient for the transplant-HCC-count-interaction indicating four
payment HCCs.\75\ The child model would operate similarly. For a child
enrollee with a transplant HCC in the transplant factor group and three
other payment HCCs, the following would be used to calculate the
enrollee's risk score: (1) The factor coefficients for all four HCCs
(that is, the three non-transplant HCCs and the transplant HCC), (2)
the factor coefficient for the severity-HCC-count-interaction
indicating four payment HCCs, and (3) the factor coefficient for the
transplant-HCC-count-interaction indicating four or more payment HCCs.
---------------------------------------------------------------------------
\75\ This is in addition to other factors that the adult
enrollee has that are used to calculate their risk score (such as
the applicable demographic factors, RXCs (if any), and the
applicable enrollment duration factors).
---------------------------------------------------------------------------
To implement the severity- and transplant-HCC-count-interaction
factors in the regression model and estimate the value of their factor
coefficients, we are proposing to remove the current severity illness
factors in the adult models, and add severity- and transplant-HCC-
count-interaction factors for the adult and child models beginning with
the 2023 benefit year. Although the severity (or transplant) HCC-count-
interaction factor coefficients may be estimated as having negative
values, the combination of these interaction factor coefficients with
the factor coefficient of the HCC that triggered the severity factor
will always be positive. For example, the proposed adult silver metal
level model factor coefficient for Viral or Unspecified Meningitis (HCC
04), which is proposed as a severe illness HCC, is 6.914, when combined
with the proposed severity-HCC-count-interaction factor coefficient for
one HCC of -4.603 (indicating that the enrollee only has HCC 04 present
in their data), would increase the enrollee's risk score by 2.311.
Moreover, an increase in the count of HCCs would lead to a monotonic
increase in the enrollee risk score, because the severity-HCC-count-
interaction factor coefficients are less negative (and sometimes
positive) with a larger number of payment HCCs.
One potential concern with this proposed model specification change
is that the severity- and transplant-HCC-count-interaction factor
coefficients might be based on small sample sizes. In recognition of
this issue, we considered sample sizes of the various interacted HCC
count factors when developing this proposal and the proposed factor
coefficients. We explored alternative methods of interacting HCC counts
with severity and transplant HCCs, including interacting the HCC counts
with individual selected severity and transplant HCCs, but found that
interacting the HCC counts with a factor indicating the presence of at
least one of the selected HCCs in each group produced PR improvements
and sufficient sample sizes for reasonably stable factor coefficient
estimates. To that end, we analyzed 2016, 2017, and 2018 enrollee-level
EDGE data and chose the model specifications that grouped the HCC
counts interacted with individual severity and transplant HCCs into two
sets of aggregated factors to maximize sample size, reduce concerns of
overfitting the model, and reduce the number of factors being added to
the models. More specifically, in the adult models, we found that
starting with 4+ HCCs for the transplant interacted factors improved
predictions of enrollees at the very high end in terms of risk and cost
and ending at 8+ HCCs for the transplant interacted factors, instead of
10+ HCCs, addressed the small sample sizes of enrollees with a
transplant and 9 or more HCCs. For the child models, we found having
one transplant interacted factor for 4+ HCCs provided more stable
estimates given the smaller sample sizes for children than those for
adults. With the proposed structure for transplant and severity
interacted factors in place, the resulting sample sizes for both
proposed sets of factors in the child and adult models in the proposed
2022 Payment Notice and in this rule are consistent with the sample
sizes used for individual HCCs in the adult and child risk adjustment
models.
We also considered potential gaming concerns in developing the
proposed interacted HCC counts factors. We believe that the proposal to
restrict the incremental risk score adjustment to enrollees with at
least one severe illness HCC, which accounts for less than 2 percent of
the adult enrollee-level EDGE data population across the 2016, 2017,
and 2018 benefit years, helps mitigate the concern that issuers may
attempt to inflate HCC counts to influence their transfers under the
state payment transfer formula. In other words, the scope for
potentially inflating HCC coding frequency under this proposal would be
limited to a small fraction of total enrollees, in contrast to an
approach that would interact HCC counts for any payment HCC, where a
payment HCC is present in approximately 20 percent of the adult
enrollee population across the same three benefit years of enrollee-
level EDGE data.\76\ We also note that enrollees with interacted HCCs
are likely to have more HCCs and higher risk scores and therefore are
more likely to be sampled and have their risk scores reviewed in the
HHS-operated risk adjustment data validation (HHS-RADV) process due to
our use of stratified sampling and application of the Neyman
allocation.\77\
---------------------------------------------------------------------------
\76\ This analysis was based on 2016, 2017, and 2018 enrollee-
level EDGE data. See Chapter 4.2 in the 2021 HHS-Operated Risk
Adjustment Technical Paper on Possible Model Changes, available at
<a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\77\ For a discussion of our use of stratified sampling and
application of the Neyman allocation, see 79 FR at 13756-13758; and
84 FR at 17494-17495.
---------------------------------------------------------------------------
Our analysis of the proposed interacted HCC counts factors combined
with the proposed HCC-contingent enrollment duration factors in the
adult models (discussed in the following section) significantly
improves predictions across most deciles and HCC counts for the very
highest-risk enrollees, as well as the lowest-risk enrollees without
HCCs. Specifically, as described in the 2021 RA Technical Paper, the
proposed interacted HCC counts approach improves the PRs for enrollees
across most HCC counts, with significant improvements for enrollees
with high numbers of HCCs (greater than 6).\78\ The proposed interacted
HCC counts approach also demonstrated improved R-squared statistics
across all metal levels in the adult and child models using 2016, 2017,
and 2018 enrollee-level EDGE data.\79\
---------------------------------------------------------------------------
\78\ See Figure 4.3 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\79\ See Figure 4.4 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
---------------------------------------------------------------------------
Some commenters on the 2021 RA Technical Paper were concerned about
potential data bias because of the exclusion of enrollees with
capitated claims from the analytic sample used to test the model
specification changes. As previously stated in the 2016 RA White
Paper,\80\ we have historically excluded enrollees with capitated
claims from the recalibration sample due to concerns that methods for
computing and reporting derived amounts from capitated claims would not
result in
[[Page 603]]
reliable data for recalibration or analysis.\81\
---------------------------------------------------------------------------
\80\ See the March 2016 Risk Adjustment Methodology White Paper
(March 24, 2016), available at <a href="https://www.cms.gov/CCIIO/Resources/Forms-Reports-and-Other-Resources/Downloads/RA-March-31-White-Paper-032416.pdf">https://www.cms.gov/CCIIO/Resources/Forms-Reports-and-Other-Resources/Downloads/RA-March-31-White-Paper-032416.pdf</a>.
\81\ See Chapter 1.4 in the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
---------------------------------------------------------------------------
Beyond the predictive improvements, an additional benefit of the
proposed interacted HCC count model specification is that it would not
overhaul the existing risk adjustment factors and would instead build
upon the current models. Additionally, the factors would remain fairly
stable, could be used in combination with other refinements and model
updates, and could be easily modified, adjusted, expanded, or
constrained in the future to include additional HCCs or to remove HCCs.
For all of these reasons, we are proposing to add the proposed
interacted HCC counts model specification as outlined above to the
adult and child risk adjustment models beginning with the 2023 benefit
year.
We seek comment on this proposal, specifically regarding whether we
should implement the proposed interacted HCC counts model specification
alone, independent of the other proposed model specification changes
outlined in this rule, beginning with the 2023 benefit year; whether we
should implement the proposed interacted HCC counts model specification
in conjunction with these other proposals; or whether we should not
implement the proposed interacted HCC counts model specification at
all. We also seek comment on the variations on the HCC counts model
specification discussed in this section, including whether we should
interact severity or transplant factors with individual HCCs, or should
interact HCC counts with individual selected severity and transplant
HCCs, rather than interacting HCC counts with only an indicator of the
presence of severity or transplant HCCs, as proposed. Finally, we seek
comment on the proposed list of severity and transplant HCCs in Table 3
that would be used to calculate the proposed interacted HCC count
factor coefficients and whether other HCCs should be to added to the
proposed list that trigger the interacted HCC count factor coefficients
or whether any of the HCCs on the proposed list should be removed.
iii. Changes to the Adult Model Enrollment Duration Factors \82\
---------------------------------------------------------------------------
\82\ As explained in the 2021 Payment Notice proposed rule, we
found that partial year enrollees in the child models did not have
the same risk differences as partial year enrollees in the adult
models and they tended to have similar risk to full year enrollees
in the child models. See 85 FR 7103-7104. In the infant models, we
found that partial year infants had higher expenditures on average
compared to their full year counterparts; however, the incorporation
of enrollment duration factors created interaction issues with the
current severity and maturity factors and did not have a meaningful
impact on the general predictive accuracy of the infant models.
Ibid. We therefore propose to continue to apply enrollment duration
factors to the adult models only.
---------------------------------------------------------------------------
In addition to the proposed two-stage weighted model specification
and the interacted HCC counts model specification, we are also
proposing to change the enrollment duration factors in the adult risk
adjustment models to improve the prediction for partial-year adult
enrollees with and without HCCs. Although the value for the factors
change from year to year as part of the annual recalibration of the
adult models, we have not made changes to the structure of the
enrollment duration factors since they were first adopted for the 2017
benefit year. To develop the current enrollment duration factors for
the adult models, we reviewed the annualized predicted expenditures,
actual expenditures, and PRs by enrollment duration groups (for each: 1
month, 2 months, and so on up to 12 months) for our risk adjustment
concurrent modeling sample, which was made up of adults in the 2014
MarketScan[supreg] data.\83\ This analysis found that actuarial risk
for adult enrollees with short enrollment periods tended to be
underpredicted in our methodology, and actuarial risk for adult
enrollees with full enrollment periods (12 months) tended to be
overpredicted. We therefore proposed and finalized in the 2018 Payment
Notice that, beginning for the 2017 benefit year, the adult models
would include enrollment duration factors that apply to all adults with
partial-year enrollment.\84\ The value for the enrollment duration
factors have generally decreased since they were first introduced in
the adult models for the 2017 benefit year, reflecting a reduced impact
of enrollment duration on risk scores of partial year enrollees. After
a slight increase between 2017 and 2018, the factors have decreased
significantly from 2018 to 2021, and in some cases (the 10- and 11-
month factors) the factors are now 0.000, relative to a 12-month
enrollment baseline.\85\
---------------------------------------------------------------------------
\83\ See pages 35-39 of the March 2016 Risk Adjustment
Methodology White Paper (March 24, 2016), available at <a href="https://www.cms.gov/CCIIO/Resources/Forms-Reports-and-Other-Resources/Downloads/RA-March-31-White-Paper-032416.pdf">https://www.cms.gov/CCIIO/Resources/Forms-Reports-and-Other-Resources/Downloads/RA-March-31-White-Paper-032416.pdf</a>.
\84\ 81 FR 94058 at 94071-94074.
\85\ In unconstrained models, these factors are negative;
therefore, we constrained them to zero because we do not believe
negative enrollment duration factors are appropriate, as this would
create inappropriate incentives. See Figure 3.1 in the 2021 HHS-
Operated Risk Adjustment Technical Paper on Possible Model Changes,
available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
---------------------------------------------------------------------------
As described in prior rulemakings and the 2021 RA Technical Paper,
we have been considering potential adjustments to the enrollment
duration factors and our more recent analysis of enrollee-level EDGE
data found that the current adult model enrollment duration factors
underpredicted plan liability for partial-year adult enrollees with
HCCs and overpredicted plan liability for partial-year adult enrollees
without HCCs.\86\ \87\ More specifically, our analysis of 2017 and 2018
enrollee-level EDGE data found that the current enrollment duration
factors are driven by enrollees with HCCs.\88\ That is, partial-year
enrollees with HCCs had higher per member, per month (PMPM)
expenditures on average as compared to full-year enrollees with HCCs,
and partial-year enrollees without HCCs were not significantly
different in PMPM expenditures compared to full-year enrollees without
HCCs.\89\
---------------------------------------------------------------------------
\86\ See 85 FR 29164 at 29188-29190.; 86 FR 24140 at 24151-
24162.; and the 2021 HHS-Operated Risk Adjustment Technical Paper on
Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\87\ When we refer to the enrollees with and without HCCs, we
are referring to enrollees without payment HCCs.
\88\ See, for example, Chapters 1.4 and 3.2 of the 2021 HHS-
Operated Risk Adjustment Technical Paper on Possible Model Changes,
available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>. Also see 85 FR at 7103-7104 and 85 FR at 78585-78586.
\89\ See Chapter 1.4 of the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
---------------------------------------------------------------------------
Therefore, beginning with the 2023 benefit year, we are proposing
to eliminate the current monthly enrollment duration factors of up to
11 months for all enrollees in the adult models, and replace them with
new monthly enrollment duration factors of up to 6 months that would
apply only to adult enrollees with HCCs. If finalized as proposed, this
would mean there would be no enrollment duration factors for adult
enrollees without HCCs starting with the 2023 benefit year nor would
there be enrollment duration factors for adult enrollees with HCCs and
more than 6 months of enrollment.
While we considered other enrollment duration factor structures, we
are proposing to limit the enrollment duration factors to 6 months
because we found that the monthly average cost variation by number of
months enrolled is meaningfully reduced after 6 months for adult
enrollees with HCCs, and enrollment duration factors beyond 6 months
did not meaningfully improve
[[Page 604]]
prediction for the adult models. As part of our analysis of enrollment
duration factor options, we also considered adoption of enrollment
duration factors by market, but we did not find a meaningful
distinction in relative costs between markets on average once we
implemented the proposed enrollment duration factors of up to 6 months
for adult enrollees with HCCs.\90\ We also considered HCC-type
contingent enrollment duration factors. Specifically, we found that the
distribution of enrollment duration and PMPM allowed charges by
enrollment duration is similar for adults with any acute HCCs versus
adults with only chronic HCCs.\91\ We therefore determined that, on
balance, it would add unnecessary complexity to introduce enrollment
duration factors by market type or that are contingent on types of HCCs
with little benefit. Therefore, we are not proposing enrollment
duration factors for the adult models by market type or that are
contingent on types of HCCs at this time.
---------------------------------------------------------------------------
\90\ See Chapter 3.3.2 of the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\91\ See Chapter 3.3.3 of the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
---------------------------------------------------------------------------
We also considered previous comments we received that expressed
concerns that certain issuers--particularly small group market issuers,
small issuers, or Medicaid issuers--may have partial-year enrollees
with HCCs that are not coded. These commenters expressed concerns that
these issuers may have difficulty obtaining diagnoses for these
enrollees, creating cases where the issuer may pay claims, and incur
costs, for services associated with a condition for the partial-year
enrollee, but the issuer's limited time with the partial-year enrollee
may not be adequate to capture the diagnosis code associated with the
HCC.\92\ \93\ In response to the 2021 RA Technical Paper, we got
further comment from stakeholders who questioned whether the HCC-
contingent enrollment duration factors would have negative impacts on
small group market issuers that offer non-calendar year coverage and
take on new business later in the year. As we noted in the 2021 RA
Technical Paper, our analysis did not find evidence that issuers are
unable to capture cost-meaningful HCCs for partial-year enrollees in
the individual or small group (including merged) market.\94\
---------------------------------------------------------------------------
\92\ See Chapter 3.4 of the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\93\ This issue differs from situations where issuers may not
have a complete diagnostic profile for a partial-year enrollee
because the services received were not related to the diagnoses that
were not captured. For example, if an enrollee received services due
to a condition while enrolled with a different issuer, then the
current issuer may not have all diagnosis codes for a partial-year
enrollee. However, such cases do not have cost implications for the
current issuer since the partial-year enrollee received no services
associated with that diagnosis.
\94\ See Chapter 3.4 of the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
---------------------------------------------------------------------------
We solicit comments on the proposed changes to the enrollment
duration factors for the adult models. We also solicit comments
regarding whether we should implement the proposed changes to
enrollment duration factors alone, independent of the other proposed
model specification changes outlined in this rule, beginning with the
2023 benefit year; whether we should implement the proposed changes to
enrollment duration factors in conjunction with these other proposals;
or whether we should not implement the proposed changes to enrollment
duration factors at all and maintain the current structure for these
factors.
iv. Combined Impact of the Proposed Model Changes
In sum, we are proposing to modify the HHS risk adjustment model
specifications for the adult and child models beginning with the 2023
benefit year by combining a two-stage weighted approach with the
removal of the current adult model severe illness interaction factors
and the addition of new severe illness and transplant interacted HCC
count factors to the adult and child models. We are also proposing to
replace the current enrollment duration factors in the adult models.
For the two-stage weighted approach, we propose calibrating the adult
and child models in two stages. The first stage of the weighted
estimation method would involve a linear regression of simulated plan
liability on age-sex factors and payment HCC factors for the adult and
child models, with the addition of RXCs and the new proposed enrollment
duration factors for the adult models. The second stage would use the
reciprocal of prediction from the first step to weight a second stage
linear regression. To stabilize the weights from the first stage
predictions, we propose lower and upper bound caps on the predictions
used as weights at the 2.5th and 97.5th percentiles in the adult models
and the 2.5th and 99.5th percentiles in the child models. This two-
stage weighted approach would be combined with the new severity and
transplant indicators from the interacted HCC count factors. For the
severity indicator group, we propose to add separate count factors for
one to 10+ payment HCCs (1, 2, . . . , 10+) for the adult models and
one to 5, 6 or 7, and 8+ payment HCCs (1, 2, . . . 5, 6 or 7, 8+) for
the child models. The proposed HCCs that would flag the severity
indicator are listed in Table 3. For the transplant HCCs, we propose to
incorporate factors for 4 to 8+ payment HCCs (4, 5, 6, 7, 8+) for the
adult models and one factor for 4+ payment HCCs for the child models.
The proposed HCCs that would flag the transplant indicator are listed
in Table 3. The severity- (and transplant-) HCC-count-interaction
factors would be included in both stages of the regressions. We propose
to incorporate the two-stage weighted approach and the interacted HCC
count specification updates beginning with the 2023 benefit year HHS
risk adjustment adult and child models. We also propose to remove the
current severity illness factors in the adult models beginning with the
2023 benefit year. Lastly, we propose to remove the current 11
enrollment duration factors for all enrollees in the adult models and
replace them with new monthly enrollment duration factors of up to 6
months that only apply to enrollees with HCCs. We propose to
incorporate the new HCC-contingent enrollment duration factors
beginning with the 2023 benefit year adult models.
We tested combining these model specifications into an approach
that incorporated the two-stage weighted approach, the severity and
transplant factors interacted with HCC count factors, and the HCC-
contingent enrollment duration factors. We found that, together, these
changes are expected to improve model performance in comparison to the
current models. Our analysis found this combined approach generally
improved prediction for enrollees at both the low and high ends of
expected expenditures and had higher R-squared statistics across metal
levels than the current models, indicating a better individual-level
fit.\95\ Our analysis also found general improvement in PRs for the
models with the combined proposed model specification changes across
each decile of predicted plan liability, by age-sex factor for adult
enrollees with and without HCCs, and by enrollment
[[Page 605]]
length.\96\ We also found that the mean absolute error did not
materially differ between the current adult and child models and the
proposed adult and child models with the combined proposed model
specification changes incorporated.\97\ These observations support our
belief that the best way to comprehensively improve the predictive
accuracy of the current models across the risk spectrum is to implement
all three proposed model specification changes together. To further
assist issuers and other stakeholders with analyzing the impact of the
combination of these proposed model specification changes, HHS also
conducted a transfer simulation and provided summary-level and issuer-
specific risk score and transfer estimates.\98\ \99\
---------------------------------------------------------------------------
\95\ See Chapter 5.1 of the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\96\ Ibid.
\97\ Ibid.
\98\ See the 2021 HHS-Operated Risk Adjustment Technical Paper
on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>. See also the HHS-Operated Risk
Adjustment Technical Paper on Possible Model Changes: Summary
Results for Transfer Simulations, available at <a href="https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs">https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs</a>.
Issuers that participated in the simulation also received detailed
issuer-specific data, including risk score and transfer estimates
for the simulated results.
\99\ If an issuer wishes to use the simulation results to assist
in assessing the impact of these model specification changes on
future benefit year transfer amounts, it should do so with caution
and in combination with other significant data.
---------------------------------------------------------------------------
As detailed in the 2021 RA Technical Paper, this transfer
simulation applied the proposed model specification changes to 2020
benefit year EDGE data to illustrate and estimate what 2020 benefit
year risk adjustment transfers would have been if the combined model
specification changes were applied.\100\ The transfer simulation
provided issuers with detailed, plan-level simulated results.\101\ The
coefficients values presented in Tables 1 and 2 incorporate the
combination of these proposed model specification changes and Table 3
provides the list of the proposed severity and transplant HCCs that
would apply for the proposed interacted HCC counts factors. We seek
comment on the combination of these proposed model changes and the
adoption of these changes beginning with the 2023 benefit year.
---------------------------------------------------------------------------
\100\ See Chapter 5.2 of the 2021 HHS-Operated Risk Adjustment
Technical Paper on Possible Model Changes, available at <a href="https://www.cms.gov/files/document/2021-ra-technical-paper.pdf">https://www.cms.gov/files/document/2021-ra-technical-paper.pdf</a>.
\101\ See the HHS-Operated Risk Adjustment Technical Paper on
Possible Model Changes: Summary Results for Transfer Simulations,
available at <a href="https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs">https://www.cms.gov/CCIIO/Programs-and-Initiatives/Premium-Stabilization-Programs</a>.
---------------------------------------------------------------------------
We seek comment on finalizing each of these proposed model
specification changes as a whole, in part, or in combination or for
example, whether we should finalize the proposed interaction HCC counts
model specification and the proposed changes to the adult model
enrollment duration factors without the proposed two stage weighted
model specification. Finally, we seek comment on finalizing the 2023
models without the proposed model specification changes, but with
updates to the data years used for recalibration, (that is, to use
2017, 2018, and 2019 enrollee-level EDGE data, as detailed elsewhere in
this proposed rule); or, alternatively, using the updated final 2022
risk adjustment model coefficients \102\ for the 2023 benefit year risk
adjustment models, trended forward to project 2023 costs or not trended
forward to project 2023 costs.
---------------------------------------------------------------------------
\102\ See ``Final 2021 Benefit Year Final HHS Risk Adjustment
Model Coefficients.'' May 12, 2020. Available at <a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/Final-2021-Benefit-Year-Final-HHS-Risk-Adjustment-Model-Coefficients.pdf">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/Final-2021-Benefit-Year-Final-HHS-Risk-Adjustment-Model-Coefficients.pdf</a>.
---------------------------------------------------------------------------
c. Pricing Adjustment for the Hepatitis C Drugs
For the 2023 benefit year, we propose to continue applying a market
pricing adjustment to the plan liability associated with Hepatitis C
drugs in the risk adjustment models.\103\ Since the 2020 benefit year
risk adjustment models, we have been making a market pricing adjustment
to the plan liability associated with Hepatitis C drugs to reflect
future market pricing prior to solving for coefficients for the
models.\104\ This market pricing adjustment has been necessary to
account for the significant pricing changes associated with the
introduction of new and generic Hepatitis C drugs between the data
years used for recalibrating the models and the applicable
recalibration benefit year. We also continue to be cognizant that
issuers might seek to influence provider prescribing patterns if a drug
claim can trigger a large increase in an enrollee's risk score that is
higher than the actual plan liability of the drug claim, and therefore,
make the transfer results more favorable for the issuer. We have
committed to reassessing this pricing adjustment with additional years
of enrollee-level EDGE data, as data become available. As part of the
2023 benefit year model recalibration, we reassessed the Hepatitis C
RXC using available enrollee-level EDGE data (including 2019 benefit
year data) to consider whether the adjustment was still needed and if
it is still needed, whether it should be modified. We found that the
data for the Hepatitis C RXC that would be used for the 2023 benefit
year recalibration (that is, the 2017, 2018, and 2019 enrollee-level
EDGE data) still do not account for the significant pricing changes due
to the introduction of new Hepatitis C drugs and, therefore, do not
precisely reflect the average cost of Hepatitis C treatments applicable
to the benefit year in question.
---------------------------------------------------------------------------
\103\ See, for example, 84 FR 17463 through 17466.
\104\ The Hepatitis C drugs market pricing adjustment to plan
liability is applied for all enrollees taking Hepatitis C drugs in
the data used for recalibration.
---------------------------------------------------------------------------
Specifically, we are proposing to recalibrate the 2023 benefit year
risk adjustment models with the 2017, 2018, and 2019 enrollee-level
EDGE data. Generic Hepatitis C drugs did not become available on the
market until 2019.\105\ Due to the lag between the data years used to
recalibrate the risk adjustment models and the applicable benefit year
of risk adjustment, we do not believe that the data used for
recalibrating the models precisely reflect the average cost of
Hepatitis C treatments expected in the 2023 benefit year. Therefore, we
continue to believe a market pricing adjustment for the 2023 benefit
year is necessary to account for the significant pricing changes
associated with the introduction of new and generic Hepatitis C drugs
between the data years used for recalibrating the models and the
applicable recalibration benefit year. We intend to continue to assess
this pricing adjustment in future benefit year recalibrations using
additional years of enrollee-level EDGE data. We seek comment on our
proposal to continue applying a market pricing adjustment to the plan
liability associated with Hepatitis C drugs for the 2023 benefit year.
---------------------------------------------------------------------------
\105\ See <a href="https://www.gilead.com/news-and-press/company-statements/authorized-generics-for-hcv">https://www.gilead.com/news-and-press/company-statements/authorized-generics-for-hcv</a>. See also <a href="https://news.abbvie.com/news/abbvie-receives-us-fda-approval-mavyret-glecaprevirpibrentasvir-for-treatment-chronic-hepatitis-c-in-all-major-genotypes-gt-1-6-in-as-short-as-8-weeks.htm">https://news.abbvie.com/news/abbvie-receives-us-fda-approval-mavyret-glecaprevirpibrentasvir-for-treatment-chronic-hepatitis-c-in-all-major-genotypes-gt-1-6-in-as-short-as-8-weeks.htm</a>.
---------------------------------------------------------------------------
d. Risk Adjustment RXC Mapping for Recalibration
i. Inclusion and Exclusion Criteria for Drugs in RXC Mapping and
Recalibration
This section provides an overview of the inclusion and exclusion
criteria HHS uses to identify drugs for mapping to RXCs in the adult
risk adjustment models, reviews what version of the RXC mapping
document HHS uses when processing the enrollee-level EDGE data for a
benefit year for recalibration of the adult risk adjustment models, and
outlines the criteria that warrant consideration for changes to the
incorporation (or
[[Page 606]]
exclusion) of particular drugs from the RXC mappings in future benefit
year recalibrations. We also propose a change to the approach for
identifying the version of the RXC mapping document HHS would use to
process a given benefit year's enrollee-level EDGE data for
recalibration of the adult risk adjustment models.
In accordance with Sec. 153.320, HHS develops and publishes the
risk adjustment methodology applicable in states where HHS operates the
program, including the draft factors to be employed in the models for
the benefit year. This includes the annual recalibration of the adult
risk adjustment models' RXC coefficients using data from the applicable
prior benefit years trended forwarded to reflect the applicable benefit
year of risk adjustment. Drugs that appear on claims data, either
through National Drug Codes (NDCs) or Healthcare Common Procedural
Coding System (HCPCS), are cross walked to RxNorm Concept Unique
Identifiers (RXCUIs).\106\ RXCUI mappings are always matched to the
NDCs and HCPCS applicable to the particular EDGE data year as the NDC
and HCPCS reflect the drugs that were available in the market during
the benefit year.\107\ Currently, we use the most recent RXC mappings
(RXCUIs that map to RXCs) that are available when we first process the
enrollee-level EDGE data for a benefit year for recalibration of the
adult risk adjustment models. For example, for the 2022 benefit year,
we recalibrated the adult risk adjustment models using 2016, 2017, and
2018 enrollee-level EDGE data and applied the second quarter (Q2) 2018
RXC mapping document for both 2016 and 2017,\108\ and applied the Q2
2019 mapping document for 2018 for recalibration of the adult risk
adjustment models RXC factors.\109\
---------------------------------------------------------------------------
\106\ See, for example, 81 FR at 94074-94080.
\107\ See, for example, Creation of the 2018 Benefit Year HHS-
Operated Risk Adjustment Models Draft Prescription Drug (RXCUIs) to
HHS Drug Classes (RXCs) Crosswalk Memorandum at <a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/Draft-RxC-Crosswalk-Memo-9-18-17.pdf">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/Draft-RxC-Crosswalk-Memo-9-18-17.pdf</a>.
\108\ RXCs were not added to the risk adjustment models until
2018 benefit year; therefore, we used 2018 RXC mappings for both
2016 and 2017 enrollee-level EDGE data as there were no 2016 and
2017 RXC mapping documents. Note that, even though 2018 RXC mappings
were applied to these earlier years, they were cross walked to the
NDCs and HCPCS that describe the applicable drugs during those
earlier years.
\109\ Although the recalibration proposals are typically
released towards the end of the calendar year, we generally receive
the prior benefit year enrollee-level EDGE data in the summer or
fall, at which point we apply the most recently available mapping
document as we begin to prepare the data to recalibrate the models
for the applicable benefit year. This is why, for example, we used
the 2019 Q2 mapping document when processing the 2018 enrollee-level
EDGE data for recalibration of the 2022 benefit year adult models.
---------------------------------------------------------------------------
As noted in the 2022 Payment Notice, we also continuously assess
the availability of drugs in the market and the associated mapping of
those drugs to RXCs in the adult risk adjustment models.\110\ More
specifically, during a benefit year, HHS conducts quarterly reviews of
RXCUIs that map to RXCs in the adult risk adjustment models for that
benefit year. During our annual review of enrollee-level EDGE data for
recalibration purposes, and to a certain extent during quarterly
reviews of RXCUIs that map to RXCs in the adult risk adjustment models,
HHS evaluates the inclusion and exclusion of RXCUIs based on criteria
such as: (1) Whether costs for an individual drug are comparable to the
costs of other drugs in the same class, (2) whether a drug is a good
predictor of the presence of the diseases that map to the HCCs that an
RXC indicates (which can be evaluated through clinical expert review in
the absence of data), (3) whether clinical expert reviews of the
pharmacological properties and prescribing patterns are consistent with
treatment of a particular condition, and (4) stakeholder feedback.\111\
As a result of this on-going assessment, we may make quarterly updates
to the RXC Crosswalk, which identifies the list of NDCs and HCPCS
indicating the presence of an RXC in the current benefit year DIY and
EDGE reference data, to ensure drugs are mapped to RXCs, where
appropriate. This can include the addition or removal of drugs based on
market availability and the other criteria identified above. As such,
the risk adjustment mapping of RXCUIs to RXCs, along with the list of
NDCs and HCPCS that crosswalk to each RXCUI, may be updated throughout
a particular benefit year of risk adjustment. HHS provides information
to issuers on these updates through the DIY software, which is
published on the CCIIO website,\112\ as well as through the EDGE global
reference updates, which are published on the Distributed Data
Collection program page on the Registration for Technical Assistance
Portal (REGTAP).\113\
---------------------------------------------------------------------------
\110\ See 86 FR at 26164.
\111\ See, for example, the Creation of the 2018 Benefit Year
HHS-Operated Risk Adjustment Adult Models Draft Prescription Drug
(RXCUIs) to HHS Drug Classes (RXCs) Crosswalk (September 17, 2017),
available at <a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/Draft-RxC-Crosswalk-Memo-9-18-17.pdf">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/Downloads/Draft-RxC-Crosswalk-Memo-9-18-17.pdf</a>.
\112\ The August 3, 2021 version of the DIY software is
available at <a href="https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance">https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance</a>.
\113\ Available at <a href="https://www.regtap.info/reg_library.php?libfilter_topic=3">https://www.regtap.info/reg_library.php?libfilter_topic=3</a>.
---------------------------------------------------------------------------
This ongoing updating process occurs on a different timeline than
the annual model recalibration activities for a given benefit year.
In this rule, we propose to change the approach for identifying the
version of the RXC mapping document HHS would use to process a given
benefit year's enrollee-level EDGE data for the annual recalibration of
the adult risk adjustment models. More specifically, we propose to
recalibrate the adult risk adjustment models using the final, fourth
quarter (Q4) RXC mapping document that was applicable for each benefit
year of data that is included in the applicable benefit year's model
recalibration, while continuing to engage in annual and quarterly
review processes using the inclusion and exclusion criteria described
above. For example, if we recalibrate the 2024 benefit year adult risk
adjustment models using 2018, 2019, and 2020 benefit years of enrollee-
level EDGE data, we would use the Q4 RXC mapping document for each of
those benefit years (that is, Q4 2018, Q4 2019, and Q4 2020,
respectively) for recalibration purposes. We would also use the
criteria described above to evaluate the inclusion and exclusion of
RXCUIs and may make other updates to the 2024 benefit year RXC
Crosswalk to ensure drugs are mapped to RXCs, where appropriate.
We propose to begin to use this approach for recalibration of the
2023 adult risk adjustment models with the exception of the 2017
enrollee-level EDGE data year, for which we propose to use the most
recent RXC mapping document that was available when we first processed
the 2017 enrollee-level EDGE data (that is, Q2 2018). We propose to use
the applicable benefit year's Q4 RXC mapping documents for both the
2018 and 2019 benefit years of enrollee-level EDGE data for the
recalibration of the adult risk adjustment models for the 2023 benefit
year. Under this proposal, we would hold those mappings constant when
using the 2018 and 2019 enrollee level EDGE data years in future
benefit year model recalibrations--meaning that we would use the
applicable benefit year's Q4 RXC mapping documents when the 2018 or
2019 benefit year of enrollee-level EDGE data is used for future
benefit year model recalibrations.\114\
[[Page 607]]
The purpose of maintaining a specific version of the same RXC mapping
document for future recalibrations under this proposal is to limit the
volatility of some coefficients from year-to-year and to ensure that we
are capturing the utilization and costs observed for the underlying
drugs in use in that year for the condition. Because the final DIY
software update contains the Q4 list, this approach would also have the
added benefit of providing issuers the opportunity to see the mappings/
crosswalk that will be applied to that data year in the final DIY
software release before it is used for recalibration.
---------------------------------------------------------------------------
\114\ Consistent with the approach finalized in the 2022 Payment
Notice, the 2018 and 2019 enrollee-level EDGE data would be used for
the recalibration of the 2024 benefit year models and the 2019
enrollee-level EDGE data would be used for the recalibration of the
2025 benefit year models. See, supra, note 47.
---------------------------------------------------------------------------
For purposes of the 2023 benefit year recalibration, we are
proposing an exception for the 2017 benefit year enrollee-level EDGE
data and would instead use the most recent RXC mapping document that
was available when we first processed the benefit year's enrollee-level
EDGE data for recalibration purposes (that is, Q2 2018). We are
proposing this approach for the 2017 benefit year enrollee-level EDGE
data because we did not include RXCs in the adult risk adjustment
models until 2018 \115\ and therefore, we do not have a Q4 RXC mapping
for the 2017 benefit year. Thus, we propose to use the Q2 2018 RXC
mapping document for the 2017 benefit year enrollee-level EDGE data
year for 2023 model recalibration, consistent with the mapping used for
processing the 2017 data for recalibration of the 2021 and 2022 adult
models. We seek comment on this proposal to change the approach for
identifying the version of the RXC mapping document that would be used
to process a given benefit year's data for the annual recalibration of
the adult models, as well as the proposed applicability beginning with
the 2023 benefit year model recalibration and the proposed exception
for the mapping document for the 2017 benefit year enrollee-level EDGE
data.
---------------------------------------------------------------------------
\115\ See 81 FR at 94075.
---------------------------------------------------------------------------
Alternatively, we seek comment on whether we should take a
different approach to recalibration of the RXC mappings for the adult
risk adjustment models. Under this alternative, we would use t
[…truncated; see source link]Indexed from Federal Register on January 5, 2022.
This is legal information, not legal advice. Laws vary by jurisdiction and change frequently. Always verify current law with official sources and consult a licensed attorney in your jurisdiction for advice on your specific situation.