Home/Federal/Federal Register/2022-23722

Rule2022-23722

Medicare Program; Calendar Year (CY) 2023 Home Health Prospective Payment System Rate Update; Home Health Quality Reporting Program Requirements; Home Health Value-Based Purchasing Expanded Model Requirements; and Home Infusion Therapy Services Requirements

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

November 4, 2022

Effective

January 1, 2023

Issuing agencies

Health and Human Services DepartmentCenters for Medicare & Medicaid Services

Abstract

This final rule sets forth routine updates to the Medicare home health payment rates for calendar year (CY) 2023 in accordance with existing statutory and regulatory requirements. This final rule also finalizes a methodology for determining the impact of the difference between assumed versus actual behavior change on estimated aggregate expenditures for home health payments as result of the change in the unit of payment to 30 days and the implementation of the Patient Driven Groupings Model (PDGM) case-mix adjustment methodology and finalizes a corresponding permanent prospective adjustment to the CY 2023 home health payment rate. This rule finalizes the reassignment of certain diagnosis codes under the PDGM case-mix groups, and establishes a permanent mitigation policy to smooth the impact of year-to-year changes in home health payments related to changes in the home health wage index. This rule also finalizes recalibration of the PDGM case-mix weights and updates the low utilization payment adjustment (LUPA) thresholds, functional impairment levels, comorbidity adjustment subgroups for CY 2023, and the fixed-dollar loss ratio (FDL) used for outlier payments. Additionally, this rule discusses comments received on the future collection of data regarding the use of telecommunications technology during a 30-day home health period of care on home health claims. This rule also finalizes changes to the Home Health Quality Reporting Program (HH QRP) requirements; changes to the expanded Home Health Value-Based Purchasing (HHVBP) Model; and updates to the home infusion therapy services payment rates for CY 2023.

Full Text

<html>
<head>
<title>Federal Register, Volume 87 Issue 213 (Friday, November 4, 2022)</title>
</head>
<body><pre>
[Federal Register Volume 87, Number 213 (Friday, November 4, 2022)]
[Rules and Regulations]
[Pages 66790-66887]
From the Federal Register Online via the Government Publishing Office [<a href="http://www.gpo.gov">www.gpo.gov</a>]
[FR Doc No: 2022-23722]

[[Page 66789]]

Vol. 87

Friday,

No. 213

November 4, 2022

Part II

Department of Health and Human Services

-----------------------------------------------------------------------

Centers for Medicare & Medicaid Services

-----------------------------------------------------------------------

42 CFR Part 484

Medicare Program; Calendar Year (CY) 2023 Home Health Prospective
Payment System Rate Update; Home Health Quality Reporting Program
Requirements; Home Health Value-Based Purchasing Expanded Model
Requirements; and Home Infusion Therapy Services Requirements; Final
Rule

Federal Register / Vol. 87 , No. 213 / Friday, November 4, 2022 /
Rules and Regulations

[[Page 66790]]

-----------------------------------------------------------------------

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Centers for Medicare & Medicaid Services

42 CFR Part 484

[CMS-1766-F]
RIN 0938-AU77

AGENCY: Centers for Medicare & Medicaid Services (CMS), Department of
Health and Human Services (HHS).

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: This final rule sets forth routine updates to the Medicare
home health payment rates for calendar year (CY) 2023 in accordance
with existing statutory and regulatory requirements. This final rule
also finalizes a methodology for determining the impact of the
difference between assumed versus actual behavior change on estimated
aggregate expenditures for home health payments as result of the change
in the unit of payment to 30 days and the implementation of the Patient
Driven Groupings Model (PDGM) case-mix adjustment methodology and
finalizes a corresponding permanent prospective adjustment to the CY
2023 home health payment rate. This rule finalizes the reassignment of
certain diagnosis codes under the PDGM case-mix groups, and establishes
a permanent mitigation policy to smooth the impact of year-to-year
changes in home health payments related to changes in the home health
wage index. This rule also finalizes recalibration of the PDGM case-mix
weights and updates the low utilization payment adjustment (LUPA)
thresholds, functional impairment levels, comorbidity adjustment
subgroups for CY 2023, and the fixed-dollar loss ratio (FDL) used for
outlier payments. Additionally, this rule discusses comments received
on the future collection of data regarding the use of
telecommunications technology during a 30-day home health period of
care on home health claims.
This rule also finalizes changes to the Home Health Quality
Reporting Program (HH QRP) requirements; changes to the expanded Home
Health Value-Based Purchasing (HHVBP) Model; and updates to the home
infusion therapy services payment rates for CY 2023.

DATES: These regulations are effective on January 1, 2023.

FOR FURTHER INFORMATION CONTACT:
Brian Slater, (410) 786-5229, for home health and home infusion
therapy payment inquiries.
For general information about home infusion payment, send your
inquiry via email to <a href="/cdn-cgi/l/email-protection#c088afada589aea6b5b3a9afae90afaca9a3b980a3adb3eea8a8b3eea7afb6">[email&#160;protected]</a>.
For general information about the Home Health Prospective Payment
System (HH PPS), send your inquiry via email to
<a href="/cdn-cgi/l/email-protection#80c8efede5c8e5e1ecf4e8d0efece9e3f9c0e3edf3aee8e8f3aee7eff6">[email&#160;protected]</a>.
For information about the Home Health Quality Reporting Program (HH
QRP), send your inquiry via email to <a href="/cdn-cgi/l/email-protection#195151484b49686c7c6a6d7076776a597a746a3771716a377e766f">[email&#160;protected]</a>.
For more information about the expanded Home Health Value-Based
Purchasing Model, please visit the Expanded HHVBP Model web page at
<a href="https://innovation.cms.gov/innovation-models/expanded-home-health-value-based-purchasing-model">https://innovation.cms.gov/innovation-models/expanded-home-health-value-based-purchasing-model</a>.

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Executive Summary
A. Purpose
B. Summary of the Provisions of This Rule
C. Summary of Costs, Transfers, and Benefits
II. Home Health Prospective Payment System
A. Overview of the Home Health Prospective Payment System
B. Provisions for Payment Under the HH PPS
III. Home Health Quality Reporting Program (HH QRP) and Other Home
Health Related Provisions
A. End of the Suspension of OASIS Data Collection on Non-
Medicare/Non-Medicaid HHA Patients and Requirement for HHAs To
Submit All-Payer OASIS Data for Purposes of the HH QRP, Beginning
With the CY 2027 Program Year
B. Technical Changes
C. Codification of the HH QRP Measure Removal Factors
D. Request for Information: Health Equity in the HH QRP
IV. Expanded Home Health Value-Based Purchasing (HHVBP) Model
A. Background
B. Changes to the Baseline Years and New Definitions
C. Request for Comment on a Future Approach to Health Equity in
the Expanded HHVBP Model
V. Home Infusion Therapy Services: Annual Payment Updates for CY
2023
A. Home Infusion Therapy Payment Categories
B. Payment Adjustments for CY 2023 Home Infusion Therapy
Services
C. CY 2023 Payment Amounts for Home Infusion Therapy Services
XI. Collection of Information Requirements and Waiver of Final
Rulemaking
A. Statutory Requirement for Solicitation of Comments
B. Collection of Information Requirements
C. Submission of PRA-Related Comments
D. Waiver of Final Rulemaking
XII. Regulatory Impact Analysis
A. Statement of Need
B. Overall Impact
C. Detailed Economic Analysis
D. Limitations of Our Analysis
E. Regulatory Review Cost Estimation
F. Alternatives Considered
G. Accounting Statement and Tables
H. Regulatory Flexibility Act (RFA)
I. Unfunded Mandates Reform Act (UMRA)
J. Federalism
K. Conclusion
Regulations Text

I. Executive Summary and Advancing Health Information Exchange

A. Executive Summary

1. Purpose and Legal Authority
a. Home Health Prospective Payment System (HH PPS)
As required under section 1895(b) of the Social Security Act (the
Act), this final rule updates the payment rates for HHAs for CY 2023.
In addition, the rule recalibrates the case-mix weights under section
1895(b)(4)(A)(i) and (b)(4)(B) of the Act for 30-day periods of care in
CY 2023; finalizes a methodology to determine the impact of differences
between assumed behavior changes and actual behavior changes on
estimated aggregate Medicare home health expenditures, in accordance
with section 1895(b)(3)(D)(i) of the Act; finalizes a permanent payment
adjustment to the CY 2023 30-day period payment rate; updates the case-
mix weights, LUPA thresholds, functional impairment levels, and
comorbidity subgroups for CY 2023; and updates the CY 2023 fixed-dollar
loss ratio (FDL) for outlier payments (so that outlier payments as a
percentage of estimated total payments are not to exceed 2.5 percent,
as required by section 1895(b)(5)(A) of the Act). This final rule also
discusses the comments received on the collection of data on the use of
telecommunications technology from home health claims.
b. Home Health (HH) Quality Reporting Program (QRP)
This final rule finalizes the end of the suspension of the
collection of Outcome and Assessment Information Set (OASIS) data from
non-Medicare/non-Medicaid patients pursuant to section 704 of the
Medicare Prescription Drug, Improvement, and Modernization Act of 2003
and requires HHAs to report all-payer OASIS data for purposes of the

[[Page 66791]]

HH QRP. In response to concerns raised by commenters on the burden
associated with the proposed new data collection, we are finalizing
that the new OASIS data reporting for the HH QRP will begin with the CY
2027 program year, with two quarters of data required for that program
year. We are finalizing a phase-in period is in place for January 1,
2025 through June 30, 2025 in which failure to submit the data will not
result in a penalty. We are finalizing as proposed regulatory text
change that consolidates the statutory references to data submission.
We are also finalizing as proposed the codification of the measure
removal factors we adopted in the CY 2019 HH PPS final rule. Finally,
this rule summarizes the comments we received in response to our
Request for Information regarding health equity in the HH QRP.
c. Expanded Home Health Value Based Purchasing (HHVBP) Model
In accordance with the statutory authority at section 1115A of the
Act, we are finalizing proposed policy updates, new definitions and
modifications of existing definitions, conforming regulation text
changes for the expanded Home Health Value-Based Purchasing (HHVBP)
expanded Model. We also summarize the comments received on our request
for comment on a potential future approach to health equity in the
expanded HHVBP Model included in the proposed rule.
d. Medicare Coverage of Home Infusion Therapy
This final rule discusses updates to the home infusion therapy
services payment rates for CY 2023 under section 1834(u) of the Act.
2. Summary of the Provisions of This Rule
a. Home Health Prospective Payment System (HH PPS)
In section II.B.2. of this rule, we are finalizing our proposed
behavioral adjustment methodology to reflect the impact of differences
between assumed behavior changes and actual behavior changes on
estimated aggregate payment expenditures under the HH PPS. We are also
finalizing a -3.925 percent permanent payment adjustment for CY 2023
(half of the proposed -7.85 percent adjustment), as we recognize the
potential hardship of implementing the proposed full permanent
adjustment in a single year. In section II.B.3 of this rule, we are
finalizing the proposed reassignment of certain ICD-10-CM codes related
to the PDGM clinical groups and comorbidity subgroups.
In section II.B.4. of this rule, we are finalizing the proposed
recalibration of the PDGM case-mix weights, LUPA thresholds, functional
levels, and comorbidity adjustment subgroups for CY 2023.
In section II.B.5. of this rule, we are finalizing our proposals to
update the home health wage index, the CY 2023 national, standardized
30-day period payment rates, and the CY 2023 national per-visit payment
amounts by the home health payment update percentage. The final home
health payment update percentage for CY 2023 will be 4.0 percent. This
rule also finalizes a permanent 5-percent cap on wage index reductions
in order to smooth the impact of year-to-year changes in home health
payments related to changes in the home health wage index.
Additionally, this rule finalizes the FDL ratio to ensure that
aggregate outlier payments do not exceed 2.5 percent of the total
aggregate payments, as required by section 1895(b)(5)(A) of the Act.
In section II.B.6. of this final rule, we respond to the comment
solicitation on the collection of data on the use of telecommunications
technology from home health claims.
b. HH QRP
In section III.D. of this final rule, we are finalizing our
proposal to end the temporary suspension on our collection of non-
Medicare/non-Medicaid data, in accordance with section 704 of the
Medicare Prescription Drug, Improvement, and Modernization Act of 2003
and, in accordance with section 1895(b)(3)(B)(v) of the Act, to require
HHAs to submit all-payer OASIS data for purposes of the HH QRP. In
response to concerns raised by commenters on the burden associated with
the proposed new data collection, we are finalizing that the new OASIS
data reporting for the HH QRP will begin January 1, 2025 with a phase-
in period for January 1, 2025 through June 30, 2025 in which failure to
submit the data will not result in a penalty. In section III.E. of this
rule, we are finalizing technical changes to Sec. 484.245(b)(1). In
section III.F. of this rule, we are finalizing codification of the
factors we adopted in the CY 2019 HH PPS final rule as the factors we
will consider when determining whether to remove measures from the HH
QRP measure set. Lastly, in section III.G. of this rule, we are
summarizing the comments we received on our Request for Information
regarding health equity in the HH QRP.
c. Expanded Home Health Value Based Purchasing (HHVBP) Model
In section IV. of this final rule, we are finalizing as proposed
changes the HHA baseline year to CY 2022 for all HHAs that were
certified prior to January 1, 2022 starting in the CY 2023 performance
year. We are also making conforming regulation text changes at Sec.
484.350(b) and (c). In addition, we are finalizing proposed amendments
to the Model baseline year from CY 2019 to CY 2022 starting in the CY
2023 performance year to enable CMS to measure competing HHAs
performance on benchmarks and achievement thresholds that are more
current. We are finalizing conforming amendments to definitions in
Sec. 484.345. In section IV.C. of this final rule, we have included a
discussion of comments received in response to the RFI related to a
potential future approach to health equity in the expanded HHVBP Model
that was included in the proposed rule.
d. Medicare Coverage of Home Infusion Therapy
In section V. of this final rule, we discuss updates to the home
infusion therapy services payment rates for CY 2023, under section
1834(u) of the Act.
3. Summary of Costs, Transfers, and Benefits
BILLING CODE 4120-01-P

Table 1--Summary of Costs, Transfers, and Benefits

[[Page 66792]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.000

BILLING CODE 4120-01-C

B. Advancing Health Information Exchange

The Department of Health and Human Services (HHS) has a number of
initiatives designed to encourage and support the adoption of
interoperable health information technology and to promote nationwide
health information exchange to improve health care and patient access
to their digital health information.
To further the goal of data interoperability in post-acute care
settings, CMS and the Office of the National Coordinator for Health
Information Technology (ONC) participate in the Post-Acute Care
Interoperability Workgroup (PACIO) to facilitate collaboration with
industry stakeholders to develop Health Level Seven
International[supreg] (HL7) Fast Healthcare Interoperability
Resources[supreg] (FHIR) standards.\1\ These standards could support
the exchange and reuse of patient assessment data derived from the
Minimum Data Set (MDS), Inpatient Rehabilitation Facility-Patient
Assessment Instrument (IRF-PAI), LTCH Continuity Assessment Record and
Evaluation (CARE) Data Set (LCDS), Outcome and Assessment Information
Set (OASIS), and other sources. The PACIO Project has focused on HL7
FHIR implementation guides for functional status, cognitive status and
new use cases on advance directives, re-assessment timepoints, and
Speech, Language, Swallowing, Cognitive communication and Hearing
(SPLASCH) pathology. We encourage PAC provider and health IT vendor
participation as the efforts advance.
---------------------------------------------------------------------------

\1\ <a href="http://pacioproject.org/">http://pacioproject.org/</a>.
---------------------------------------------------------------------------

The CMS Data Element Library (DEL) continues to be updated and
serves as a resource for PAC assessment data elements and their
associated mappings to health IT standards, such as Logical Observation
Identifiers Names and Codes (LOINC) and Systematized Nomenclature of
Medicine Clinical Terms (SNOMED). The DEL furthers CMS' goal of data
standardization and interoperability. Standards in the DEL (<a href="https://del.cms.gov/DELWeb/pubHome">https://del.cms.gov/DELWeb/pubHome</a>) can be referenced on the CMS website and in
the ONC Interoperability Standards Advisory (ISA). The 2022 ISA is
available at <a href="https://www.healthit.gov/isa">https://www.healthit.gov/isa</a>.
The 21st Century Cures Act (Cures Act) (Pub. L. 114-255, enacted
December 13, 2016) required HHS and ONC to take steps to further
interoperability for providers in settings across the care continuum.
Section 4003(b) of the Cures Act required ONC to take steps to advance
interoperability through the development of a trusted exchange
framework and common agreement aimed at establishing a universal floor
of interoperability across the country. On January 18, 2022, ONC
announced a significant milestone by releasing the Trusted Exchange
Framework \2\ and Common Agreement

[[Page 66793]]

(TEFCA) Version 1.\3\ The Trusted Exchange Framework is a set of non-
binding principles for health information exchange, and the Common
Agreement is a contract that advances those principles. The Common
Agreement and the Qualified Health Information Network Technical
Framework Version 1 \4\ (incorporated by reference into the Common
Agreement) establish the technical infrastructure model and governing
approach for different health information networks and their users to
securely share clinical information with each other--all under commonly
agreed to terms. The technical and policy architecture of how exchange
occurs under the Trusted Exchange Framework and the Common Agreement
follows a network-of-networks structure, which allows for connections
at different levels and is inclusive of many different types of
entities at those different levels, such as health information
networks, healthcare practices, hospitals, public health agencies, and
Individual Access Services (IAS) Providers.\5\ For more information, we
refer readers to <a href="https://www.healthit.gov/topic/interoperability/trusted-exchange-framework-and-common-agreement">https://www.healthit.gov/topic/interoperability/trusted-exchange-framework-and-common-agreement</a>.
---------------------------------------------------------------------------

\2\ The Trusted Exchange Framework (TEF): Principles for Trusted
Exchange (Jan. 2022), <a href="https://www.healthit.gov/sites/default/files/page/2022-01/Trusted_Exchange_Framework_0122.pdf">https://www.healthit.gov/sites/default/files/page/2022-01/Trusted_Exchange_Framework_0122.pdf</a>.
\3\ Common Agreement for Nationwide Health Information
Interoperability Version 1 (Jan. 2022), <a href="https://www.healthit.gov/sites/default/files/page/2022-01/Common_Agreement_for_Nationwide_Health_Information_Interoperability_Version_1.pdf">https://www.healthit.gov/sites/default/files/page/2022-01/Common_Agreement_for_Nationwide_Health_Information_Interoperability_Version_1.pdf</a>.
\4\ Qualified Health Information Network (QHIN) Technical
Framework (QTF) Version 1.0 (Jan. 2022), <a href="https://rce.sequoiaproject.org/wp-content/uploads/2022/01/QTF_0122.pdf">https://rce.sequoiaproject.org/wp-content/uploads/2022/01/QTF_0122.pdf</a>.
\5\ The Common Agreement defines Individual Access Services
(IAS) as ``with respect to the Exchange Purposes definition, the
services provided utilizing the Connectivity Services, to the extent
consistent with Applicable Law, to an Individual with whom the QHIN,
Participant, or Subparticipant has a Direct Relationship to satisfy
that Individual's ability to access, inspect, or obtain a copy of
that Individual's Required Information that is then maintained by or
for any QHIN, Participant, or Subparticipant.'' The Common Agreement
defines ``IAS Provider'' as: ``Each QHIN, Participant, and
Subparticipant that offers Individual Access Services.'' See Common
Agreement for Nationwide Health Information Interoperability Version
1, at 7 (Jan. 2022), <a href="https://www.healthit.gov/sites/default/files/page/2022-01/Common_Agreement_for_Nationwide_Health_Information_Interoperability_Version_1.pdf">https://www.healthit.gov/sites/default/files/page/2022-01/Common_Agreement_for_Nationwide_Health_Information_Interoperability_Version_1.pdf</a>.
---------------------------------------------------------------------------

We invite readers to learn more about these important developments
and how they are likely to affect HHAs.

II. Home Health Prospective Payment System

A. Overview of the Home Health Prospective Payment System

1. Statutory Background
Section 1895(b)(1) of the Act requires the Secretary to establish a
Home Health Prospective Payment System (HH PPS) for all costs of home
health services paid under Medicare. Section 1895(b)(2) of the Act
requires that, in defining a prospective payment amount, the Secretary
will consider an appropriate unit of service and the number, type, and
duration of visits provided within that unit, potential changes in the
mix of services provided within that unit and their cost, and a general
system design that provides for continued access to quality services.
In accordance with the statute, as amended by the Balanced Budget Act
of 1997 (BBA) (Pub. L. 105-33, enacted August 5, 1997), we published a
final rule in the July 3, 2000 Federal Register (65 FR 41128) to
implement the HH PPS legislation.
Section 5201(c) of the Deficit Reduction Act of 2005 (DRA) (Pub. L.
109-171, enacted February 8, 2006) added new section 1895(b)(3)(B)(v)
to the Act, requiring home health agencies (HHAs) to submit data for
purposes of measuring health care quality, and linking the quality data
submission to the annual applicable payment percentage increase. This
data submission requirement is applicable for CY 2007 and each
subsequent year. If an HHA does not submit quality data, the home
health market basket percentage increase is reduced by 2 percentage
points. In the November 9, 2006 Federal Register (71 FR 65935), we
published a final rule to implement the pay-for-reporting requirement
of the DRA, which was codified at Sec. 484.225(h) and (i) in
accordance with the statute. The pay-for-reporting requirement was
implemented on January 1, 2007.
Section 51001(a)(1)(B) of the Bipartisan Budget Act of 2018 (BBA of
2018) (Pub. L. 115-123) amended section 1895(b) of the Act to require a
change to the home health unit of payment to 30-day periods beginning
January 1, 2020. Section 51001(a)(2)(A) of the BBA of 2018 added a new
subclause (iv) under section 1895(b)(3)(A) of the Act, requiring the
Secretary to calculate a standard prospective payment amount (or
amounts) for 30-day units of service furnished that end during the 12-
month period beginning January 1, 2020, in a budget neutral manner,
such that estimated aggregate expenditures under the HH PPS during CY
2020 are equal to the estimated aggregate expenditures that otherwise
would have been made under the HH PPS during CY 2020 in the absence of
the change to a 30-day unit of service. Section 1895(b)(3)(A)(iv) of
the Act requires that the calculation of the standard prospective
payment amount (or amounts) for CY 2020 be made before the application
of the annual update to the standard prospective payment amount as
required by section 1895(b)(3)(B) of the Act.
Additionally, section 1895(b)(3)(A)(iv) of the Act requires that in
calculating the standard prospective payment amount (or amounts), the
Secretary must make assumptions about behavior changes that could occur
as a result of the implementation of the 30-day unit of service under
section 1895(b)(2)(B) of the Act and case-mix adjustment factors
established under section 1895(b)(4)(B) of the Act. Section
1895(b)(3)(A)(iv) of the Act further requires the Secretary to provide
a description of the behavior assumptions made in notice and comment
rulemaking. CMS finalized these behavior assumptions in the CY 2019 HH
PPS final rule with comment period (83 FR 56461).
Section 51001(a)(2)(B) of the BBA of 2018 also added a new
subparagraph (D) to section 1895(b)(3) of the Act. Section
1895(b)(3)(D)(i) of the Act requires the Secretary to annually
determine the impact of differences between assumed behavior changes,
as described in section 1895(b)(3)(A)(iv) of the Act, and actual
behavior changes on estimated aggregate expenditures under the HH PPS
with respect to years beginning with 2020 and ending with 2026. Section
1895(b)(3)(D)(ii) of the Act requires the Secretary, at a time and in a
manner determined appropriate, through notice and comment rulemaking,
to provide for one or more permanent increases or decreases to the
standard prospective payment amount (or amounts) for applicable years,
on a prospective basis, to offset for such increases or decreases in
estimated aggregate expenditures, as determined under section
1895(b)(3)(D)(i) of the Act. Additionally, 1895(b)(3)(D)(iii) of the
Act requires the Secretary, at a time and in a manner determined
appropriate, through notice and comment rulemaking, to provide for one
or more temporary increases or decreases to the payment amount for a
unit of home health services for applicable years, on a prospective
basis, to offset for such increases or decreases in estimated aggregate
expenditures, as determined under section 1895(b)(3)(D)(i) of the Act.
Such a temporary increase or decrease shall apply only with respect to
the year for which such temporary increase or decrease is made, and the
Secretary shall not take into account such a temporary increase or
decrease in computing the payment amount for a unit of home health
services for a subsequent year. Finally, section

[[Page 66794]]

51001(a)(3) of the BBA of 2018 amends section 1895(b)(4)(B) of the Act
by adding a new clause (ii) to require the Secretary to eliminate the
use of therapy thresholds in the case-mix system for CY 2020 and
subsequent years.
2. Current System for Payment of Home Health Services
For home health periods of care beginning on or after January 1,
2020, Medicare makes payment under the HH PPS on the basis of a
national, standardized 30-day period payment rate that is adjusted for
case-mix and area wage differences in accordance with section
51001(a)(1)(B) of the BBA of 2018. The national, standardized 30-day
period payment rate includes payment for the six home health
disciplines (skilled nursing, home health aide, physical therapy,
speech-language pathology, occupational therapy, and medical social
services). Payment for non-routine supplies (NRS) is also part of the
national, standardized 30-day period rate. Durable medical equipment
(DME) provided as a home health service, as defined in section 1861(m)
of the Act, is paid the fee schedule amount or is paid through the
competitive bidding program and such payment is not included in the
national, standardized 30-day period payment amount. Additionally, the
30-day period payment rate does not include payment for certain
injectable osteoporosis drugs and negative pressure wound therapy
(NPWT) using a disposable device, but such drug and services must be
billed separately by the HHA and paid under Part B, while a patient is
under a home health plan of care, as the law requires consolidated
billing of osteoporosis drugs and NPWT using a disposable device.
To better align payment with patient care needs and to better
ensure that clinically complex and ill beneficiaries have adequate
access to home health care, in the CY 2019 HH PPS final rule with
comment period (83 FR 56406), we finalized case-mix methodology
refinements through the Patient-Driven Groupings Model (PDGM) for home
health periods of care beginning on or after January 1, 2020. The PDGM
did not change eligibility or coverage criteria for Medicare home
health services, and as long as the individual meets the criteria for
home health services as described at 42 CFR 409.42, the individual can
receive Medicare home health services, including therapy services. For
more information about the role of therapy services under the PDGM, we
refer readers to the Medicare Learning Network (MLN) Matters article
SE2000 available at <a href="https://www.cms.gov/regulations-and-guidanceguidancetransmittals2020-transmittals/se20005">https://www.cms.gov/regulations-and-guidanceguidancetransmittals2020-transmittals/se20005</a>. To adjust for
case-mix for 30-day periods of care beginning on and after January 1,
2020, the HH PPS uses a 432-category case-mix classification system to
assign patients to a home health resource group (HHRG) using patient
characteristics and other clinical information from Medicare claims and
the Outcome and Assessment Information Set (OASIS) assessment
instrument. These 432 HHRGs represent the different payment groups
based on five main case-mix categories under the PDGM, as shown in
Figure 1. Each HHRG has an associated case-mix weight that is used in
calculating the payment for a 30-day period of care. For periods of
care with visits less than the low-utilization payment adjustment
(LUPA) threshold for the HHRG, Medicare pays national per-visit rates
based on the discipline(s) providing the services. Medicare also
adjusts the national standardized 30-day period payment rate for
certain intervening events that are subject to a partial payment
adjustment (PEP). For certain cases that exceed a specific cost
threshold, an outlier adjustment may also be available.
Under this case-mix methodology, case-mix weights are generated for
each of the different PDGM payment groups by regressing resource use
for each of the five categories (admission source, timing, clinical
grouping, functional impairment level, and comorbidity adjustment)
using a fixed effects model. A detailed description of each of the
case-mix variables under the PDGM have been described previously, and
we refer readers to the CY 2021 HH PPS final rule (85 FR 70303 through
70305).
BILLING CODE 4120-01-P

[[Page 66795]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.001

BILLING CODE 4120-01-C

B. Provisions for CY 2023 Payment Under the HH PPS

1. Monitoring the Effects of the Implementation of PDGM
In the CY 2023 HH PPS proposed rule (87 FR 37605), CMS provided
data analysis on Medicare home health benefit utilization, including
overall total 30-day periods of care and average periods of care per
HHA user; distribution of the type of visits in a 30-day period of care
for all Medicare fee-for-service (FFS) claims; the percentage of
periods that receive the LUPA; estimated costs for 30-day periods of
care; the distribution, by percentage, of 30-day periods of care, using
the five clinical variables (clinical group, comorbidity adjustment,
admission source, timing, and functional impairment level); the OASIS
``GG'' functional items by response type; and the proportion of 30-day
periods of care with and without any therapy visits, nursing visits,
and/or aide/social worker visits.
We will continue to monitor and analyze home health trends and
vulnerabilities within the home health payment system.
2. PDGM Behavioral Assumptions and Adjustments Under the HH PPS
a. Background
As discussed in section II.A.1. of this rule, the Secretary was
statutorily required to change the unit of payment under the HH PPS
from a 60-day episode of care to a 30-day period of care, starting with
payments for services made on and after January 1, 2020. In determining
the CY 2020 standard prospective 30-day payment amount, CMS was also
required to make assumptions about behavior changes that could occur as
a result of the implementation of the 30-day unit of payment and
changes in case-mix adjustment factors, including the elimination of
therapy thresholds as a factor in determining case-mix adjustments. In
the CY 2019 HH PPS final rule with comment period (83 FR 56455), we
finalized the following three behavior assumptions:
<bullet<ls-thn-eq> Clinical Group Coding: The clinical group is
determined by the principal diagnosis code for the patient as

[[Page 66796]]

reported by the HHA on the home health claim. This behavior assumption
assumes that HHAs will change their documentation and coding practices
and put the highest paying diagnosis code as the principal diagnosis
code in order to have a 30-day period be placed into a higher-paying
clinical group.
<bullet<ls-thn-eq> Comorbidity Coding: The PDGM further adjusts
payments based on patients' secondary diagnoses as reported by the HHA
on the home health claim. The OASIS only allows HHAs to designate 1
principal diagnosis and 5 secondary diagnoses while the home health
claim allows HHAs to designate 1 principal diagnosis and up to 24
secondary diagnoses. This behavior assumption assumes that by
considering additional ICD-10-CM diagnosis codes listed on the home
health claim (beyond the 6 allowed on the OASIS), more 30-day periods
of care will receive a comorbidity adjustment.
<bullet<ls-thn-eq> LUPA Threshold: This behavior assumption assumes
that for one-third of LUPAs that are 1 to 2 visits away from the LUPA
threshold HHAs will provide 1 to 2 extra visits to receive a full 30-
day payment.
As described in the CY 2020 HH PPS final rule with comment period
(84 FR 60512), in order to calculate the CY 2020 30-day base payment
rates both with and without behavior assumptions, we first calculated
the total, aggregate amount of expenditures that would occur under the
pre-PDGM case-mix adjustment methodology (60-day episodes under 153
case-mix groups). We then calculated what the 30-day payment amount
would need to be set at in order for CMS to pay the estimated aggregate
expenditures in CY 2020 with the application of a 30-day unit of
payment under the PDGM.
We initially determined a -8.389 percent behavior change adjustment
to the base payment rate would be needed in order to ensure that the
payment rate in CY 2020 would be budget neutral, as required by law.
However, based on the comments received and reconsideration as to the
frequency of the assumed behaviors during the first year of the
transition to a new unit of payment and case-mix adjustment
methodology, we believed it was reasonable to apply the three behavior
change assumptions to only half of the 30-day periods in our analytic
file (randomly selected). Therefore, we finalized in the CY 2020 HH PPS
final rule with comment period (84 FR 60519), a -4.36 percent behavior
change assumption adjustment (``assumed behaviors'') in order to
calculate the 30-day payment rate in a budget-neutral manner for CY
2020. After applying the wage index budget neutrality factor and the
home health payment update, the CY 2020 30-day payment rate was set at
$1,864.03.
Our data analysis in section II.B.1. of the CY 2023 HH PPS proposed
rule compares the CY 2018 and CY 2019 simulated 30-day periods of care
with behavior assumptions applied and actual CY 2020 and CY 2021 30-day
periods of care. Specifically, Tables B4, B6, and B7 (87 FR 37607
through 37609) indicate that the three assumed behavior changes did
occur as a result of the implementation of the PDGM. Additionally, this
monitoring shows that other behaviors, such as changes in the provision
of therapy, also occurred. Overall, the CYs 2020 and 2021 actual 30-day
periods are similar to the simulated CYs 2018 and 2019 30-day periods
with the behavior assumptions applied, which is supporting evidence
that HHAs did make behavior changes. We reminded readers that, by law,
we are required to ensure that estimated aggregate expenditures under
the HH PPS are equal to our determination of estimated aggregate
expenditures that otherwise would have been made under the HH PPS in
the absence of the change to a 30-day unit of payment and changes in
case-mix adjustment factors. Regardless of the magnitude and frequency
of individual behavior change (for example, LUPAs, therapy, etc.), the
occurrence of any behavior change is captured by the methodology to
determine the impact on aggregate expenditures.
We also reminded readers that in the CY 2020 HH PPS final rule with
comment period (84 FR 60513), we stated that we interpret actual
behavior changes to encompass both the assumed behavior changes that
were previously identified by CMS, as well as other behavior changes
not identified at the time the budget-neutral 30-day payment rate for
CY 2020 was established. Subsequently, as noted previously, our
analysis resulted in the identification of other behavior changes that
occurred after the implementation of the PDGM. Although not originally
one of the three finalized behavior assumptions, a decline in therapy
utilization is indicative of an additional behavior change. For
example, Table B10 and Figure B3 in section II.B.1. of the CY 2023 HH
PPS proposed rule (87 FR 37612 through 37613) indicates the number of
therapy visits declined in CYs 2020 and 2021. However, the data, as
depicted in Figure B3, also indicates a slight decline in therapy
visits began in CY 2019 after the finalization of the removal of
therapy thresholds and the PDGM, but prior to implementation. This
suggests HHAs were already beginning to decrease their therapy
provision in anticipation of the new payment system.
Each Health Insurance Prospective Payment System (HIPPS) code is
assigned a case-mix weight which determines the base payment of non-
LUPA claims prior to any other adjustments (for example, outlier
payment adjustments). Prior to the PDGM, the first position of the
HIPPS code was a numeric value that represented the interaction of
episode timing and number of therapy visits (grouping step). The
second, third, and fourth positions of the pre-PDGM HIPPS code
reflected clinical severity, functional severity, and service
utilization respectively. Therefore, to evaluate how the decrease in
therapy visits related to payments, we compared the average case-mix
weights of CY 2018 actual 60-day episodes and updated CY 2021 simulated
60-day episodes. Prior to the PDGM, the average case-mix weight for CY
2018 actual 60-day episodes was 1.0176 and the average case-mix weight
for CY 2021 simulated 60-day episodes was 0.9682. Using the updated CY
2021 simulated 60-day episodes, we set therapy levels at the pre-PDGM
(that is, CY 2018) levels and kept the clinical and functional levels
at the PDGM levels (that is, CY 2021). This resulted in an average
case-mix weight of 1.0389, slightly higher than the actual CY 2018 60-
day episodes. Next, we kept therapy levels at the PDGM (that is, CY
2021) levels and set the clinical and functional levels at the pre-PDGM
levels (that is, CY 2018) and found the average case-mix weight was
0.9383, much lower than the CY 2018 actual 60-day episodes. By
controlling for therapy levels, we were able to determine the change in
60-day episode case-mix weights was largely driven by therapy
utilization. The decrease in therapy visits led to a decrease in case-
mix weight, and therefore, a decrease in aggregate expenditures under
the pre-PDGM HH PPS.
b. Method To Annually Determine the Impact of Differences Between
Assumed Behavior Changes and Actual Behavior Changes on Estimated
Aggregate Expenditures
To evaluate if the national, standardized 30-day payment rate and
resulting estimated aggregate expenditures maintained budget neutrality
after the implementation of the PDGM, we used actual 30-day period
claims data to simulate 60-day episodes and estimate what aggregate
expenditures would have been under the 153-group case-mix system and
60-day unit of payment. Using the

[[Page 66797]]

estimated aggregate expenditures under the 153-group case-mix system
(simulated 60-day episodes from 30-day periods) we are able to
calculate permanent and temporary adjustments as discussed in section
II.B.2.c of this final rule. We used the following steps:
The first step in repricing PDGM claims was to calculate estimated
aggregate expenditures under the pre-PDGM, 153-group case-mix system
and 60-day unit of payment, by determining which PDGM 30-day periods of
care could be grouped together to form simulated 60-day episodes of
care. To facilitate grouping, we made some exclusions and assumptions
as described later in this section prior to pricing out the simulated
60-day episodes of care. We note in the early months of CY 2020, there
were 60-day episodes which started in 2019 and ended in 2020 and
therefore, some of these exclusions and assumptions may be specific to
the first year of the PDGM. We identify, through footnotes, if an
exclusion or assumption is specific to CY 2020 only. The following
describes the steps in determining the annual estimated aggregate
expenditures including the exclusions and assumptions made when
simulating 60-day episodes from actual 30-day periods.
(1) Exclusions
<bullet> Claims where the claim occurrence code 50 date (OASIS
assessment date) occurred on or after October 31 of that year. This
exclusion was applied to ensure the simulated 60-day episodes contained
both 30-day periods from the same year and would not overlap into the
following year (for example, 2021, 2022, 2023). This is done because
any 30-day periods with an OASIS assessment date in November or
December might be part of a simulated 60-day episode that would
continue into the following year and where payment would have been made
based on the ``through'' date. For CYs 2021 through 2026, we also
excluded claims with an OASIS assessment date before January 1 of that
year.\6\ Again, this is to ensure a simulated 60-day episode (simulated
from two 30-day periods) does not overlap years.
---------------------------------------------------------------------------

\6\ There are no 30-day PDGM claims which started in CY 2019 and
ended in CY 2020, and therefore this exclusion would not apply to
the CY 2020 dataset.
---------------------------------------------------------------------------

<bullet> Beneficiaries and all of their claims if they have
overlapping claims from the same provider (as identified by CMS
Certification Number (CCN)). All of a beneficiary's claims are dropped
so as not to create problems with assigning episode timing if only a
subset of claims is dropped
<bullet> Beneficiaries and all of their claims if three or more
claims from the same provider are linked to the same occurrence code 50
date. This is done because if three or more claims link to the same
OASIS it would not be clear which claims should be joined to simulate a
60-day episode.
(2) Assumptions
<bullet> If two 30-day periods of care from the same provider
reference the same OASIS assessment date (using occurrence code 50),
then we assume those two 30-day periods of care would have been billed
as a 60-day episode of care under the 153-group system.
<bullet> If two 30 day-periods of care reference different OASIS
assessment dates and each of those assessment dates is referenced by a
single 30-day period of care, and those two 30-day periods of care
occur together close in time (that is, the ``from'' date of the later
30-day period of care is between 0 to 14 days after the ``through''
date of the earlier 30-day period of care), then we assume those two
30-day periods of care also would have been billed as a 60-day episode
of care under the 153-group system.
<bullet> For all other 30-day periods of care, we assume that they
would not be combined with another 30-day period of care and would have
been billed as a single 30-day period.
(3) Calculating Estimated Aggregate Expenditures--Pricing Simulated 60-
Day Episode Claims
After applying the exclusions and assumptions described previously,
we have the simulated 60-day episode dataset for each year.
Starting with CY 2020 claims, we assign each simulated 60-day
episode of care as a normal episode, PEP, LUPA, or outlier based on the
payment parameters established in the CY 2020 HH PPS final rule with
comment period (84 FR 60478) for 60-day episodes of care. Next, using
the October 2019 3M Home Health Grouper (v8219) \7\ we assign a HIPPS
code to each simulated 60-day episode of care using the 153-group
methodology. Finally, we price the CY 2020 simulated 60-day episodes of
care using the payment parameters described in the CY 2020 HH PPS final
rule with comment period (84 FR 60537) for 60-day episodes of care. For
CYs 2021 through 2026, we would adjust the simulated 60-day base
payment rate to align with current payments for the analysis year (that
is, wage index budget neutrality factor, home health payment update).
For example, to calculate the CY 2021 simulated 60-day episode base
payment rate, we started with the final CY 2020 60-day base payment
rate ($3,220.79) multiplied by the final CY 2021 wage index budget
neutrality factor (0.9999) and the CY 2021 home health payment update
(1.020) to get an adjusted 60-day base payment rate ($3,284.88) for CY
2021. We used the adjusted 60-day base payment rate ($3,284.88) to
price the CY 2021 simulated 60-day claims under the pre-PDGM HH PPS
(60-day episodes under 153 case-mix groups).
---------------------------------------------------------------------------

\7\ <a href="https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/HomeHealthPPS/CaseMixGrouperSoftware">https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/HomeHealthPPS/CaseMixGrouperSoftware</a>.
---------------------------------------------------------------------------

Once each simulated 60-day claim is priced under the pre-PDGM HH
PPS, we calculate the estimated aggregate expenditures for all
simulated 60-day episodes. That is, using actual behavior (using the
most current year of PDGM claims) we determine what the aggregate
expenditures would have been under the prior 153 group case-mix system.
Next, to control for utilization, we calculate the PDGM aggregate
expenditures using those specific 30-day periods that were used to
create the simulated 60-day episodes. That is, both the actual PDGM
aggregate expenditures and the simulated pre-PDGM aggregate
expenditures are based on the same number of claims. We received 770
comments on the methodology and implementation of a permanent
prospective behavior change adjustment on the CY 2023 home health
payment rate.
Comment: A few commenters stated that CMS' proposal would violate
three separate statutory requirements. The commenters stated that: (1)
the proposal uses therapy thresholds to determine payment despite the
statute's mandate to eliminate this practice; (2) ignores the statutory
provision by failing to correct its assumptions about how home health
agencies would change behaviors in response to the new payment system;
and (3) violates the statute's budget-neutrality requirement by
reducing overall aggregate expenditures.
Response: The BBA of 2018 tasked CMS with ensuring that Medicare
spending under the new 30-day payment system is the same as the
estimated spending under the old 60-day home health payment system.
Section 1895(b)(3)(A)(iv) of the Act directed the Secretary to
calculate a standard prospective payment amount for CY 2020,
incorporating assumptions about behavior changes, that could occur as a
result of the implementation of a 30-day unit of payment and changes in
case-mix adjustment factors. In other

[[Page 66798]]

words, using the data available at the time of rulemaking, we were
required to estimate a national, standardized payment rate so that
estimated aggregate expenditures with assumed behavior changes
(clinical group coding, comorbidity coding, and LUPA thresholds) for CY
2020 would be the same under the PDGM as they would have been under the
prior payment system (153 group). In the CY 2020 HH PPS final rule with
comment period (84 FR 60513), we estimated that this would mean a -
8.389 percent payment adjustment to the base payment rate in order to
avoid overestimating payments under the 30-day system. In response to
commenter concerns that the pervasiveness of expected behavioral
changes among HHAs was overestimated, we stated that given the scale of
the payment system changes, we agree that it might take HHAs more time
before they fully changed their behaviors in ways expected by CMS.
Therefore, we finalized a policy that applied the three behavioral
assumptions only to half (randomly selected) of the simulated 30-day
periods of care. This reduction in the application of the assumptions
resulted in a -4.36 percent behavior assumption adjustment. Therefore,
we met the initial requirement of section 1895(b)(3)(A)(iv) by setting
the CY 2020 national, standardized 30-day payment rate ($1,864.03) in a
budget-neutral manner, based on available data (simulated 30-day
periods) at the time of rulemaking.
Following the implementation of the new payment system, the BBA of
2018 tasks CMS with determining the impact of the difference between
our assumed behavior changes and actual behavior changes on estimated
aggregate expenditures beginning with CY 2020 through CY 2026, as set
out in section 1895(b)(3)(D)(i) of the Act.
As the Act requires CMS to look at actual behavior, the methodology
uses actual claims data for 30-day periods under the 432-group case-mix
model (PDGM claims) to simulate 60-day episodes under the 153-group
case-mix model (representing pre-PDGM HH PPS claims) in order to
estimate what the aggregate expenditures would have been in the absence
of the PDGM. In other words, CMS used the same claims (actual PDGM 30-
day periods and simulated 60-day episodes from the 30-day periods) to
compare estimated aggregate expenditures under both systems in order to
determine the estimated aggregate impact of behavior change. This
allows us to control for actual utilization, not predicted utilization,
to determine the impact of differences between what we estimate
aggregate expenditures would have been in the absence of the PDGM using
actual data and what the expenditures actually were under the PDGM.
As stated previously, CMS is not required to correct each of its
original assumptions regarding home health agency behavior changes or
itemize each behavior change for which its methodology accounts, as
commenters asserted. For example, while paragraph (3)(D)(i) clarifies
that the ``assumed behavior changes'' CMS must use in its calculations
are those ``described in paragraph (3)(A)(iv),'' it contains no such
qualification for the ``actual behavior changes'' to which CMS compares
the assumed behavior. CMS accordingly ensured that the payment rate
accurately accounts for all ``actual behavior changes'', in the
aggregate, that occurred in a given year.
Neither this provision, nor section 1895(b)(3)(A)(iv) of the Act,
requires CMS to ensure that it actually spends the amount of the
original estimated aggregate expenditures (that is, $16.2 billion)
based on simulated 30-day periods for CY 2020. Rather, section
1895(b)(3)(D)(i) of the Act requires that CMS compare the estimated
aggregate expenditures resulting from the 30-day payment rate with
estimated assumed behavior changes (resulting in a $1,864.03
standardized rate) to the new estimated aggregate expenditures derived
from actual data--incorporating actual behavior changes--that would
have occurred under the prior 60-day system. In other words, we are not
required to compare our original estimated aggregate expenditures
(estimated at $16.2 billion) to actual expenditures (that is, $15.1
billion), and make up the difference. Rather, under the statute, we re-
estimate aggregate expenditures under the pre-PDGM based on actual
behavior changes, as derived from actual claims. This is because, the
original estimated aggregate expenditures ($16.2 billion) were based on
predicted utilization, not actual utilization.
With regard to therapy, CMS received comments in the CY 2022 HH PPS
final rule (86 FR 62247) and in response to the CY 2023 HH PPS proposed
rule that the decrease in therapy utilization, including termination of
therapy staff, is related to the removal of the therapy payment
incentive. In their comment letter, a leading industry association
detailed how HHAs have responded to changes in the benefit structure
and have altered their operations, affecting the level of care received
by patients. For instance, prior to the PDGM, the industry notes that
HHAs were incentivized to provide the highest volume of therapy visits
possible, and a low volume of other services. The industry association
goes on to note that under the PDGM, the elimination of the therapy
volume adjustment as a case mix measure will likely lead to a reduction
in therapy services to patients. In an article published in February
2020,\8\ the National Association for Home Care and Hospice (NAHC) was
quoted as saying ``categorically, across the board, we're going to
reduce our therapy services'' as a result of the PDGM. More recently in
an article in April 2022,\9\ it was estimated that nearly half of HHAs
had planned to decrease therapy utilization after the implementation of
the PDGM. In that article, NAHC was quoted as saying ``There was a
precipitous drop in therapy visits in January and February of 2020
before the pandemic hit.'' In addition, their consulting firm stated,
``Importantly, note that the reduction in therapy visits began before
COVID-19 PHE started in March 2020--indicating that HHA providers were
already experiencing significant declines in therapy visits as a result
of PDGM, even before the onset of the pandemic. Thus, the PDGM effect
on therapy is not a COVID effect, but rather a PDGM effect.'' These
comments from interested parties confirm that the decrease in therapy
is a concerted provider behavior change in response to a financial
incentive rather than the COVID-19 PHE. Anecdotal evidence and the data
presented in the CY 2023 HH PPS proposed rule (87 FR 37612 through
37613) supports the conclusion there has been a significant change
(decline) in therapy visits due to the implementation of the PDGM.
---------------------------------------------------------------------------

\8\ Why Home Health Care Is Suddenly Harder to Come by For
Medicare Patients. <a href="https://khn.org/news/why-home-health-care-is-suddenly-harder-to-come-by-for-medicare-patients/">https://khn.org/news/why-home-health-care-is-suddenly-harder-to-come-by-for-medicare-patients/</a>.
\9\ Home Health Agencies Should Brace for PDGM Battle Later This
Year. <a href="https://homehealthcarenews.com/2022/04/home-health-agencies-should-brace-for-pdgm-battle-later-this-year/">https://homehealthcarenews.com/2022/04/home-health-agencies-should-brace-for-pdgm-battle-later-this-year/</a>.
---------------------------------------------------------------------------

If we were to artificially inflate aggregate expenditures in CYs
2020 and 2021 by including payments for therapy visits that may have
occurred under the old thresholds, but that were in fact not provided
under the new system (as shown by actual data), we would be setting
payment based on how providers would have presumably behaved under the
old system rather than actual behaviors under the new system, which we
believe is not the best reading of the law. It would be inappropriate
to manipulate the data so that old behaviors (in this case, inflated
therapy visits to reach payment thresholds)

[[Page 66799]]

would change the resulting payment adjustment for assumed versus actual
behavior changes under the PDGM. It would be inappropriate for CMS to
continue to pay for therapy as if HHAs were still inflating therapy
provision based on the former therapy thresholds, when the number of
therapy visits after the implementation of the PDGM has actually
declined. Despite the commenters' argument that CMS cannot use the
reduction in therapy to determine payment because the BBA of 2018
mandated the elimination of therapy thresholds, the law did not mandate
a reduction in the provision of therapy or even decrease the payment
rates for therapy disciplines. It simply removed a payment incentive
structured around the quantity of therapy visits, which had resulted in
provider behavior to maximize payment, exactly the type of actual
behavior change that CMS is tasked to consider when setting the base
payment rate.
We disagree with commenters who read sections 1895(b)(3)(A)(iv) and
1895(b)(3)(D) of the Act to require payments based on earlier, higher
therapy utilization rates instead of permitting us to re-run the
calculations we used to predict aggregated expenditures with actual
2020 data. Subparagraph (A)(iv) required CMS, in determining budget
neutrality for 2020, to estimate a payment amount so that the
``estimated aggregate amount of expenditures'' under the new 30-day
case-mix system--after including ``assumptions about behavior changes
that could occur'' because of the changed methodology--was ``equal to
the estimated aggregate amount of expenditures that otherwise would
have been made'' if the new 30-day case-mix system ``had not been
enacted.'' And subparagraph (D) requires CMS, for years 2020-2026, to
adjust payments based on how differences between the ``assumed''
behavior changes that CMS originally predicted and the ``actual''
behavior changes CMS now observes impact original ``estimated aggregate
expenditures.'' CMS followed subparagraph (A)(iv) by estimating
aggregate expenditures for CY 2020 using simulated 30-day case-mix
system claims (as this was the only data available at the time of CY
2020 rulemaking) to calculate a 30-day base payment rate as if the 30-
day case-mix system ``had not been enacted''. CMS followed subparagraph
(D) by determining the impact of assumed behavior changes to actual
behavior changes by comparing the 30-day base payment rate and
aggregate expenditures (based on assumed behaviors) to what the 30-day
base payment rate and aggregate expenditures should have been (based on
actual behaviors).
Some commenters read the requirement in subparagraph (A)(iv) to
calculate estimated aggregate expenditures as if one of Congress'
payment reforms ``had not been enacted'' to require payments based on
pre-2020 therapy utilization rates--pointing also to subparagraph
(A)(iv)'s title of ``budget neutrality for 2020.'' But that reading
ignores the requirement in subparagraph (D) to adjust estimated
aggregate expenditures based on ``actual behavior changes,'' as well as
its instruction in subparagraph (A)(iv) to incorporate into CMS's
estimated aggregate expenditures ``assumptions about behavior changes
that could occur as a result of'' implementing these payment reforms.
These provisions authorize CMS to account for how behavior changes,
like therapy utilization, would have affected payments under the old
60-day system and do not require CMS to pay for therapy that never
actually occurred. This ensures that HHAs were still paid the same
amount they would have been under the old system for services they
actually did provide--thus achieving budget neutrality.
We also disagree with the commenter who suggests that subparagraph
(D) prohibits CMS from recalculating estimated aggregate expenditures
and instead requires CMS to compare the aggregate expenditures CMS
estimated in 2019 to actual expenditures CMS observed in 2020.
Subparagraph (D) requires CMS to evaluate how using actual behavior
changes rather than assumed behavior changes affects predicted
expenditures.
Comment: Multiple commenters stated that CMS' proposed rule
violates notice and comment rulemaking because ``an agency must provide
the public with the relevant data and technical studies on which it
relies to form decisions''. Commenters indicated that CMS did not
disclose to the public both the data model and the post-manipulation
data and they were therefore unable to replicate and test the CMS'
findings and conclusions. Specifically, commenters requested the
baseline payments at the claim level used by CMS to calculate the CY
2023 impacts, any additional adjustments to the CY 2021 data to roll it
forward to CY 2022, home health agency level impacts, the dataset CMS
used to determine budget neutrality and the adjustment factors for CYs
2020 and 2021, a spreadsheet analogue to the SNF parity-adjustment, and
the input data supporting its calculations. In addition, a few
commenters stated that the methodology was not clear and did not
provide the specific claims to use in analysis. Some commenters stated
that agency-level impacts should have been provided and that they could
not fully analyze the methodology without such agency-level impacts.
Response: We disagree with commenters that we violated notice and
comment rulemaking by not providing the public with relevant data and
technical studies. We also remind commenters that this methodology, the
corresponding data files and step-by-step instructions also were
detailed in the CY 2022 HH PPS proposed rule (86 FR 35889) and CMS
solicited comments on this methodology in that proposed rule.
Interested parties did not state that the data and instructions
provided at that time were insufficient to provide comments on the
methodology. Moreover, in the CY 2023 HH PPS proposed rule, we made
available sufficient data and methodological descriptions for
interested commenters to replicate our calculations to provide comments
on this rule. These are further described below.
First, in the CY 2023 HH PPS proposed rule (87 FR 37616 through
37620), CMS provided a detailed methodology and described the results
of applying that methodology, citing the year and the source of the
home health claims data obtained from the Chronic Conditions Warehouse
(CCW) and the Home Health Claims--OASIS limited data set (LDS) file.
The CY 2022 HH PPS proposed rule (86 FR 35889 through 35892) also
included a comment solicitation on this same detailed methodology,
citing the LDS file, a publicly-available claims database. The OASIS
LDS includes the same data as the CCW, except de-identified for public
release. CMS repeatedly states that at the HH PPS LDS web page \10\
such raw data are available, and agency records reflect that multiple
commenters in fact received the CY 2021 Home Health Claims--OASIS LDS
data at issue in this rule. That file provides the variables and their
descriptions for the CY 2023 HH PPS proposed rule as well as
diagnostics that provide basic statistics for each variable CMS
considered.
---------------------------------------------------------------------------

\10\ <a href="https://www.cms.gov/Research-Statistics-Data-and-Systems/Files-for-Order/LimitedDataSets/Home_Health_PPS_LDS">https://www.cms.gov/Research-Statistics-Data-and-Systems/Files-for-Order/LimitedDataSets/Home_Health_PPS_LDS</a>.
---------------------------------------------------------------------------

Second, CMS detailed each methodological step it took in the rules,
including the exclusions and assumptions that CMS used to calculate
estimated aggregate expenditures. As such, commenters had access to
both

[[Page 66800]]

the dataset (including baseline payments at the claim level, and the
exact number of claims and the payment rates used in calculating the CY
2020 and CY 2021 proposed permanent and temporary adjustments) they
requested, as well as how CMS used that data to calculate the
adjustments. Interested parties were thus able to replicate CMS'
calculations with the information that CMS made available to them.
Commenters' requests for additional information go beyond the
critical factual material needed to comment on CMS' proposals. CMS did
not adjust the data to ``roll'' the CY 2021 data to CY 2022, and so
information about CY 2022 data is irrelevant to CMS's calculations. Nor
did CMS need to generate an analog to the SNF parity adjustment
spreadsheet, which was not part of the critical factual materials the
agency considered when making the calculations in the rule. Similarly,
commenters did not need home health agency level impacts data, because
impacts estimate how the national payment rate may affect HHAs overall,
which was not a metric CMS used to calculate the adjustments. Finally,
CMS did not need to release the simulated 60-day episodes because CMS
provided the detailed instructions on how commenters could simulate
those claims themselves based on the data CMS provided. We are aware
that some courts have read a procedural requirement into the
Administrative Procedure Act (Pub. L. 89-554) mandating that agencies
provide for public comment the critical factual materials on which they
rely.\11\ By releasing sufficient raw data files and methodological
descriptions that allowed commenters to replicate CMS's process, CMS
has more than satisfied any legal requirements to disclose factual
materials.
---------------------------------------------------------------------------

\11\ See, for example, Am. Radio Relay League, Inc. v. F.C.C.,
524 F.3d 227, 236 (DC Cir. 2008); but cf. id. at 246 (Kavanaugh, J.,
concurring in the judgment in relevant part) (noting critical
factual material doctrine ``stands on a shaky legal foundation'').
---------------------------------------------------------------------------

Comment: Multiple commenters expressed concerns that the COVID-19
PHE may have impacted CY 2020 and 2021 data. Commenters stated the
COVID-19 PHE required a shift in priorities, thereby changing
utilization patterns.
Response: The proposed methodology controls for changes in
utilization as a result of exogenous factors such as the COVID-19 PHE
by using the same claims dataset, that is the same basket of services,
under both payment systems. This ensures any difference in aggregate
expenditures is not related to the COVID-19 PHE or other exogenous
factors. It may be helpful to review the comments received from MedPAC
on the proposed rule.\12\ MedPAC stated in its comments that the
methodology presented in the proposed rule was reasonable because
applying the case-mix system in effect prior to 2020 reflects how
Medicare would have paid in the absence of the BBA 2018 changes. MedPAC
explained that any effect of the COVID-19 PHE is included in both
estimated aggregate expenditures (that is, 60-day episodes and 30-day
periods). Therefore, they noted that methodology presented ensures that
any differences between the two calculated spending amounts would not
be attributable to the COVID-19 PHE.
---------------------------------------------------------------------------

\12\ <a href="https://www.medpac.gov/wp-content/uploads/2022/08/08152022_HomeHealth_MedPAC_COMMENT_SEC.pdf">https://www.medpac.gov/wp-content/uploads/2022/08/08152022_HomeHealth_MedPAC_COMMENT_SEC.pdf</a>.
---------------------------------------------------------------------------

In addition, while the initial onset of the COVID-19 PHE in the
early months of CY 2020 may have had an impact on home health
utilization, the healthcare system has since begun to return to normal
and stabilize. For example, studies have shown that elective surgeries
and other medical treatments have resumed to pre-pandemic capacity.\13\
As shown in the CY 2023 HH PPS proposed rule (87 FR 37605 through
37614), many aspects of home health utilization (volume, visits,
clinical groups, comorbidity adjustment, admission source, timing, and
functional impairment level) are similar throughout CYs 2020 and 2021.
Furthermore, in the CY 2023 HH PPS proposed rule, we solicited data
from interested parties showing how COVID-19 affected these aspects of
home health utilization and we did not receive any empirical
information on this issue specifically. Therefore, we find the CYs 2020
and 2021 data are sufficient and complete, for the purpose of this
methodology, and we believe the data are not significantly impacted as
a result of the COVID-19 PHE.
---------------------------------------------------------------------------

\13\ Aviva S. Mattingly, BA; Liam Rose, Ph.D.; Hyrum S.
Eddington, BS; Amber W. Trickey, Ph.D.; Mark R. Cullen, MD; Arden M.
Morris, MD, MPH; Sherry M. Wren, MD. Trends in US Surgical
Procedures and Health Care System Response to Policies Curtailing
Elective Surgical Operations During the COVID-19. December 8, 2021.
JAMA Network Open. 2021;4(12):e2138038. doi:10.1001/
jamanetworkopen.2021.38038.
---------------------------------------------------------------------------

Comment: A commenter stated CMS' data shows that after
implementation of the PDGM, HHAs continued to provide therapy, but the
pattern of therapy provision changed. For example, they noted the most
significant decline was for episodes with 13 or more therapy visits. In
addition, several commenters stated there has been a decline in therapy
visits since the implementation of the PDGM. However, several
commenters stated that even if therapy visits were reduced in CYs 2020
and 2021, but outcomes (for example, hospitalizations, meeting goals of
the plan of care) did not worsen, then payment reductions should not be
made.
Response: We appreciate the commenters' recommendation. However,
CMS does not have the authority to tie this payment adjustment to
outcomes or other quality measures, or to modify this adjustment on an
agency level.
Comment: A commenter suggested using Hierarchical Condition
Categories (HCC) scores within the behavioral assumptions.
Response: We appreciate the commenter's recommendation; however, we
note that the HCC scores are dependent on beneficiaries having a claims
history (which may be limited for those newly enrolled in Medicare),
and therefore, do not think they would be appropriate to use in this
methodology as it may limit our ability to capture beneficiary
characteristics needed for case-mix adjustment.
Comment: A few commenters questioned why CMS did not include
therapy utilization as one of the original three behavior change
assumptions when setting the CY 2020 payment rate.
Response: We have noted in past rules that we use the functional
impairment level case-mix adjustment, developed as part of the PDGM
case-mix, to provide the necessary payment adjustments to ensure that
functional care needs necessitating therapy, are met based on actual
patient characteristics (84 FR 60497). The functional impairment case-
mix factor was not meant to be a direct proxy for the therapy
thresholds; however, we expected that functional impairment along with
other case-mix factors (for example, admission source), would
appropriately compensate HHAs for therapy.
Likewise, we expected the functional impairment adjustment, along
with other case-mix factors (for example, admission source), to not
only alleviate concerns that removal of the therapy thresholds would
dissuade providers from delivering needed therapy, but to assure
providers that patients can and should still receive the necessary type
and amount of therapy based on patient characteristics. In this
respect, while we did note that we were aware of how payment may affect
practice patterns and that visits vary in response to financial
incentives, we also stated that the therapy thresholds promoted the
provision of care based on increased payment associated with each of
these

[[Page 66801]]

thresholds as opposed to actual patient needs (83 FR 56485). It was our
belief, when setting the original behavior change assumptions, that the
functional impairment adjustment would effectively offset reductions in
therapy visits that could result from the elimination of the therapy
thresholds, especially those patients requiring multiple therapy
disciplines or patients with significant functional impairment. As a
result, we did not initially contend that removal of the therapy
thresholds would significantly alter provider behavior, as we were
still compensating therapy through the functional impairment case-mix
adjustment. Our expectation was that therapy utilization would reflect
actual patient acuity.
Comment: Commenters stated they support the structure of the PDGM,
but the budget neutrality adjustment methodology is inconsistent with
other methodologies applied to other health care providers and would
result in a loss of access to care.
Response: We thank interested parties for their comments. However,
the commenters did not clarify what they meant by ``inconsistent with
other methodologies applied to other health care providers''. We
believe that the proposed methodology satisfies the budget neutrality
requirements at section 1895(b)(3)(A)(iv) of the Act, as well as the
requirements at section 1895(b)(3)(D)(i) of the Act, to determine the
impact of differences between assumed behavior changes and actual
behavior changes on estimated aggregate expenditures for home health
periods of care. Furthermore, MedPAC stated in their March, 2022 report
\14\ that the Commission found positive access, quality, and financial
indicators for the sector. As such, we do not believe that this
methodology and its resulting payment adjustment would result in a loss
of access to care.
---------------------------------------------------------------------------

\14\ <a href="https://www.medpac.gov/wp-content/uploads/2022/03/Mar22_MedPAC_ReportToCongress_v2_SEC.pdf">https://www.medpac.gov/wp-content/uploads/2022/03/Mar22_MedPAC_ReportToCongress_v2_SEC.pdf</a>.
---------------------------------------------------------------------------

Comment: Several commenters recommended CMS hold a Technical Expert
Panel (TEP) to determine a methodology for calculating the budget
neutrality adjustment.
Response: We thank commenters for their suggestion. However, CMS
solicited comments on the CY 2022 HH PPS proposed rule (86 FR 35892)
for alternative methodologies, and interested parties were able to
submit comments on the CY 2023 HH PPS proposed rule. We received 75
comments on the CY 2022 proposed rule and 770 comments on the CY 2023
proposed rule. We also note that a TEP is not required by statute, and
there is insufficient time to obtain such input.
Comment: Many commenters stated the proposed methodology was
``technically flawed'' because the methodology does not compare
behaviors assumed by CMS in establishing the CY 2020 rate to actual
behaviors observed on aggregate expenditures. A commenter stated the
methodology was based on faulty data and that the methodology uses an
outdated logic, therefore the behavioral adjustment is based on ``poor
logic''.
Response: As stated previously, CMS is not required to correct or
quantify each original assumption regarding home health agency behavior
change, but rather, ensure that the payment rate is accurately
accounting for all behaviors that actually occurred in a given year. As
required by law, CMS determined the base payment rate for CY 2020
incorporating assumptions about behavior changes that could occur as a
result of the PDGM. It is unclear why the commenter believes the data
were faulty or how the methodology was outdated. The proposed
methodology for adjusting for behavioral changes compares the payment
rate and aggregate expenditures based on assumed behaviors to the what
the payment rate and estimated aggregate expenditures would have been
using actual behaviors. Therefore, CMS' proposed methodology is
comparing assumed behaviors to actual behaviors on estimated aggregate
expenditures, as required by law. Further, as stated in the CY 2023 HH
PPS proposed rule (87 FR 37616), we continue to assert that the best
reading of the law requires us to retrospectively determine if the 30-
day payment amount in CY 2020 resulted in the same estimated aggregate
expenditures that would have been made if the change in the unit of
payment and the PDGM case-mix adjustment methodology had not been
implemented. It does not require that our rates be retrospectively
adjusted to mirror estimated aggregate spending.
Comment: Several commenters recommended including changes that
affect other aspects of Medicare home health spending such as Medicare
enrollment; modification/improvement of enforcement of coverage
standards (for example, maintenance therapy, home infusion therapy);
behavior changes in other PAC services that affect home health
utilization; technological advances; and other factors that may
contribute to Medicare spending changes not specifically related to the
implementation of the PDGM. Some commenters suggesting adjusting for
nominal versus real case-mix change. A commenter recommended replacing
the proposed methodology, which they stated focused on a change in
average case-mix weight, to a methodology which focuses on behavior
changes.
Response: We thank the commenters for their suggestions. While we
recognize other factors affect the utilization of home health services,
we believe the statute is best read to instruct us to consider only
changes related to provider behavior in response to the 30-day unit of
payment and case-mix changes. As stated in the CY 2023 HH PPS proposed
rule (87 FR 37616), while changes in nominal case-mix may be
supplemental to our findings, the law requires CMS to determine the
impact of differences between assumed versus actual behavioral changes
on estimated aggregate expenditures, which are not factored into our
calculations of case-mix adjustment authority. Section
1895(b)(3)(B)(iv) of the Act states that CMS has the authority to
adjust for case-mix changes that are a result of changes in the coding
or classification of different units of services that do not reflect
real changes in case mix. Therefore, at this time we believe analyses
of nominal case-mix change are provided under a separate authority than
the statutory requirement to evaluate what aggregate expenditures would
have been in absence of the PDGM and the elimination of therapy
thresholds.
We disagree the methodology focuses on the change in average case-
mix weight. Instead, the methodology compares assumed behavior to
actual behavior and determines the impact of those differences on
estimated aggregate expenditures, as required by law. Our discussion of
case-mix in section II.B.2. of this final rule is only used as
supporting evidence in the decrease of therapy utilization.
Comment: A commenter stated the proposed methodology fails to
account for the reduction in average per-episode therapy services under
the PDGM, which would have substantially reduced payments under the
prior case-mix system. The commenter stated that this resulted in a
behavioral offset in CY 2020 that was too high and would carry over
into subsequent years.
Response: We recognize commenters are concerned that the
methodology does not control for therapy. However, as stated
previously, we believe it would be inappropriate to manipulate the data
to assume that behaviors (that is, therapy provision) remain the same
between both payment systems, when calculating the behavior change
adjustment. The commenter is correct that the same methodology will be
used

[[Page 66802]]

in subsequent years, meaning we will not control for therapy in
subsequent years either; however, we remind commenters that the law
requires we annually determine the impact of the assumed versus actual
behavior changes on estimated aggregate expenditures for CY 2020
through CY 2026 and adjust the payment rate to offset for such
increases or decreases in a time and manner determined appropriate.
Keeping behaviors constant when they changed in between payment systems
is inconsistent with this instruction.
It is unclear what the commenter suggested by a ``carry over''
effect. To clarify, the methodology analyzes each year of data
independently and captures any behavior changes which occurred in that
year, including any changes in therapy provision. As such, if any
behaviors continue into subsequent years, these will be captured in the
methodology. We also remind readers the permanent adjustment is based
on the percent change between the actual 30-day base payment rate and
the repriced 30-day base payment rate for the same year of data (for
example, CY 2021).
Comment: Multiple commenters recommended modifying the proposed
methodology to account for changes in therapy utilization and the onset
of the COVID-19 PHE. Specifically, many commenters stated that the
therapy provision under the prior 153-group payment system would be
higher than seen under the PDGM and that CMS should control for the
change in therapy utilization. Many commenters recommended that CMS
adopt the methodology presented by a consulting firm hired by several
interested parties. The consulting firm recommended applying the
Patient Driven Payment Model (PDPM) parity adjustment methodology used
in the CY 2023 Skilled Nursing Facility (SNF) PPS proposed \15\ and
final rule (87 FR 47502) \16\ to CY 2020 PDGM data. The consulting firm
stated ``based on this approach, we found that CY 2020 PDGM payments
were approximately 2.5 percent below budget neutrality (with COVID-19
cases included) and 2.4 percent below budget neutrality with COVID-19
cases excluded.''
---------------------------------------------------------------------------

\15\ <a href="https://www.federalregister.gov/documents/2022/04/15/2022-07906/medicare-program-prospective-payment-system-and-consolidated-billing-for-skilled-nursing-facilities">https://www.federalregister.gov/documents/2022/04/15/2022-07906/medicare-program-prospective-payment-system-and-consolidated-billing-for-skilled-nursing-facilities</a>.
\16\ <a href="https://www.govinfo.gov/content/pkg/FR-2022-08-03/pdf/2022-16457.pdf">https://www.govinfo.gov/content/pkg/FR-2022-08-03/pdf/2022-16457.pdf</a>.
---------------------------------------------------------------------------

Response: We appreciate the commenters' recommendation to modify
the proposed methodology to control for therapy utilization in
alignment with the SNF parity adjustment methodology. However, the SNF
PPS and HH PPS are fundamentally different; SNFs are paid a per-diem
payment with different case-mix variables, and HHAs are paid under a
bundled payment system. In addition, unlike the requirements of the SNF
PPS parity adjustment, CMS is required, by law, to account for behavior
changes related to the implementation of the PDGM, which CMS did by
comparing actual PDGM claims to what the same utilization (for example,
visits, OASIS responses, etc.) would look like under a 60-day unit of
payment.
Section 1895(b)(4)(B)(ii) of the Act statutorily required the
removal of therapy thresholds in establishing payment, but CMS stated
multiple times (83 FR 56481, 84 FR 60497, 86 FR 62247, and 87 FR 37615)
that therapy must be provided in accordance with the plan of care and
that the PDGM is not limiting or prohibiting the provision of therapy
services. As the data, as well as commenters, indicate that HHAs are
decreasing therapy utilization in response to the removal of a payment
incentive, and not the COVID-19 PHE, we disagree with commenters who
suggest adjusting attributing decreased therapy to the COVID-19 PHE.
Given CMS has not directed HHAs to modify the amount of services
provided, but rather continue providing services in accordance with the
plan of care, then any changes (operational or otherwise) by HHAs are
actual behavior changes due to the implementation of the PDGM. As
stated earlier, this type of response to a new payment system is what
CMS is required by law to evaluate and account for with subsequent
payment rate adjustments. If CMS were to implement the method presented
by the consulting firm, we would need to artificially inflate the
number of therapy visits in CYs 2020 and 2021. As noted above, doing so
is inconsistent with how we read the statute. Instead, the methodology
presented by the consulting firm would be comparing the payment rate
and aggregate expenditures based on the previous assumed behavior
assumptions to a payment rate and aggregate expenditures based on new
assumed behavior assumptions. In other words, any method which controls
for therapy provision (or other behaviors) would result in CMS
comparing assumed versus assumed behavior, which would be inconsistent
with what the statute requires.
Comment: Several commenters stated the proposed methodology does
not compare the behaviors assumed by CMS in establishing the initial
payment rate, but rather creates an artificial target amount to reduce
payments as an attempt to rebase the 30-day payment amount. As such,
many commenters also recommended the alternative methodology presented
by the consulting firm. This methodology recommended comparing the
average CY 2020 30-day episode payments to the estimated average CY
2020 payments with behavioral assumptions used by CMS to set CY 2020
payment rates (based on data from CY 2018 60-day episodes converted to
30-day episodes).
Response: We appreciate the commenters' recommendation; however,
the law requires us to determine the difference between assumed versus
actual behaviors on estimated aggregate expenditures. Therefore, we
continue to believe that the best reading of the law requires us to
retrospectively determine if the 30-day payment amount in CY 2020 and
CY 2021 resulted in the same estimated aggregate expenditures if the
change in the unit of payment and the PDGM case-mix adjustment had not
been implemented and the visits and OASIS responses did not change. As
stated previously, the proposed methodology compares the payment rate
and aggregate expenditures based on assumed behaviors to what the
payment rate and estimated aggregate expenditures would have been using
actual behaviors, which we believe is what the law requires.
Comment: Several commenters stated the PDGM claims cannot be
reasonably regrouped under an alternative payment system.
Response: We disagree with this comment, as both payment systems
(153-group and PDGM) group claims into case-mix groups based on
information available on the claim, the OASIS, and other accessible
administrative data. While the PDGM removed the payment incentive for
excess therapy, it is not only reasonable, but required by law, to
compare the same claims under two different case-mix systems.
Additionally, the proposed methodology is consistent with the original
methodology used in establishing the PDGM. As stated in the CY 2020 HH
PPS final rule with comment period (84 FR 60512), we divided actual 60-
day episodes from the 153-group payment system into two 30-day periods
in order to calculate the 30-day payment amounts. Specifically, we
simulated 9,336,898 30-day periods from 5,471,454 60-day episodes and
using estimated aggregate expenditures we calculated what we thought
the CY 2020 payment rate would need to be, based on assumed behavior
changes. We are replicating this method in reverse to

[[Page 66803]]

evaluate what the CY 2020 base payment rate should have been based on
actual behavior changes and actual utilization.
Comment: Several commenters indicated that CMS did not provide
enough information, specifically the OASIS assessments, to replicate
the methodology. In addition, a commenter stated certain OASIS items
used to group the 60-day episodes are optional in CYs 2020 and 2021,
which may impact the adjustment calculations.
Response: CMS provided a detailed explanation of the methodology in
the CY 2023 HH PPS proposed rule (87 FR 37616) and data that can be
used to carry out the methodology is made available via the Home Health
Claims--OASIS LDS. The LDS file contains all necessary information,
including OASIS, and the proposed rule described the necessary steps
and the methodology used to allow interested parties the ability to
replicate the 60-day simulated episodes. Those replicated 60-day
simulated episodes and the actual 30-day periods would have resulted in
the ability to calculate estimated aggregate expenditures, a repriced
base payment rate, and the permanent and temporary adjustments. If a
particular OASIS item did not have a response, then that item would not
contribute to the functional or clinical score under the 153-group
payment system. If there were certain OASIS items missing on claims,
those items may not have affected the overall functional or clinical
score and corresponding level. Additionally, based on the analysis
shown in the CY 2023 HH PPS proposed rule (87 FR 37615), the data
showed the difference in case-mix weights was largely driven by therapy
utilization and not functional or clinical score. Therefore, if a small
subset of claims had missing OASIS items, it would not significantly
change the overall aggregate expenditures and resulting adjustments.
Comment: A commenter noted approximately 40 percent of diagnosis
codes, which were previously allowed under the 153 case-mix group
system, are no longer accepted as a principal diagnosis under the PDGM.
This commenter stated that this systematic change may have impacted a
provider's coding behavior and could have potentially led to the
simulated 60-day episodes being inaccurately assigned a ``clinical
domain.''
Response: We thank this commenter for their review of the diagnosis
codes. While we acknowledge 41 percent (29,948) of all the diagnosis
codes are not assigned a clinical group under the PDGM,\17\ we disagree
that those unassigned codes would have created any significant
difference in assigning the clinical level in the 153-group case-mix
system. For example, out of all the diagnosis codes available in the
final grouper for the 153-group case mix system, only 22 percent
(15,936) of the diagnosis codes could potentially contribute to the
clinical score. Of those codes which could have contributed to the
clinical score, only 6.99 percent (1,114) of the diagnosis codes are
not accepted as a principal diagnosis under the PDGM. In addition,
there are only three clinical dimensions (Diabetes, Skin 1, and Neuro
1) under the 153-group system which produced a different score when the
diagnosis was counted as a principal diagnosis instead of a secondary
diagnosis. The other clinical dimensions awarded the same points with
either a primary or other diagnosis listed on the OASIS. Therefore,
while approximately 7 percent of the diagnosis codes that contributed
to the clinical score under the 153 case-mix group system are no longer
accepted as principal under the PDGM, many of these codes could still
be used as a secondary diagnosis code and counted towards the clinical
score. Additionally, there were thresholds for the clinical level, and
even if the diagnosis code was accepted as principal, it would not
automatically increase the clinical score to the point where it would
have triggered a new clinical level. In the CY 2023 HH PPS proposed
rule (87 FR 37615), we described an analysis that shows the decline in
the average case-mix weight for simulated 60-day episodes were largely
driven by reductions in therapy utilization instead of the clinical
score (which may be impacted by diagnoses). That means, even if all the
diagnosis codes were accepted under the PDGM, we find it would be
unlikely for the case-mix weight to have increased enough to counteract
the reduction in therapy.
---------------------------------------------------------------------------

\17\ Using V03.2.22 of the home health grouper.
---------------------------------------------------------------------------

Comment: A few commenters detailed their interpretation of our
proposed methodology for CY 2020 describing a calculation that uses the
number of 30-day periods (7,618,061) multiplied by the 30-day base
payment rate ($1.936.38) subtracted from actual expenditures ($14.2
million) multiplied by the number of 30-day periods. They stated that
this calculation resulted in a different payment adjustment and
expressed concern that CMS inaccurately calculated the adjustment or
did not provide sufficient detail to allow commenters to accurately
replicate the methodology.
Response: The calculations presented by commenters make several
incorrect assumptions and do not accurately replicate the detailed
methodology described in the CY 2023 HH PPS proposed rule. As stated in
the CY 2023 HH PPS proposed rule (87 FR 37617), after all exclusions
and assumptions were applied, we designated each 60-day episode of care
as a normal episode, PEP, LUPA, or outlier based on the payment
parameters established in the CY 2020 HH PPS final rule with comment
period (84 FR 60478) for 60-day episodes of care. Next, using the
October 2019 3M Home Health Grouper (v8219), we assigned a HIPPS code
to each simulated 60-day episode of care using the 153-group
methodology. Finally, we priced the CY 2020 simulated 60-day episodes
of care using the payment parameters described in the CY 2020 HH PPS
final rule with comment period (84 FR 60537) for 60-day episodes of
care.\18\ The CY 2023 HH PPS proposed rule states that each claim is
paid based on the type of claim (that is, normal, PEP, LUPA, outlier)
and assigned a HIPPS code, which would result in a specific case-mix
weight for each claim. Next, each claim (determined by claim type,
HIPPS) was priced based on the parameters previously described in the
CY 2020 rule for 60-day episodes. CMS did not simply multiply each
claim by the base payment rate, as the commenters suggested, as this
would miscalculate aggregate expenditures. As stated earlier, the
available Home Health Claims--OASIS LDS dataset included all
information for interested parties to determine the claim type and the
associated HIPPS code to accurately estimate aggregate expenditures.
---------------------------------------------------------------------------

\18\ Note, we also performed similar calculations using CY2021
data. When doing this calculation for CY2021 data, we updated the
C2020 payment rates by the payment parameters used to establish the
CY2021 PDGM payment.
---------------------------------------------------------------------------

In addition, the commenters referenced two unrelated numbers. As
stated in the CY 2023 HH PPS proposed rule (87 FR 37618), the 7,618,061
claims were the actual 30-day periods after all exclusions and
assumptions were applied to create the 4,463,549 simulated 60-day
episodes. We then determined what the payment rate should have been to
equal the aggregate expenditures that we calculated from the simulated
CY 2020 60-day episodes. We stated to determine the difference in
aggregate expenditures, we calculated the ``aggregate expenditures for
all CY 2020 PDGM 30-day claims'' using both payment rates (87 FR
37618). In other

[[Page 66804]]

words, the $14.2 billion referenced by the commenter was determined
using the $1,742.52 PDGM payment rate for all 8,423,688 30-day periods,
rather than pricing the 7,618,061 claims at their adjusted (for
example, wage index, case-mix) rate.
Comment: A few commenters stated it was unclear how episode timing
and LUPA thresholds were assigned to the simulated 60-day episodes.
Response: As described in the CY 2023 HH PPS proposed rule, we used
the October 2019 3M Home Health Grouper (v8219) to group 60-day
episodes (87 FR 37617). Episode timing, early and late, were based on
the number of 60-day episodes that occur within a sequence of 60-day
episodes. Additionally, under the 153-group system, any 60-day episode
with 4 or fewer visits was classified as a LUPA (84 FR 60519).
Comment: A commenter recommended recalibrating the regression
coefficients for the 153-group payment model using the simulated 60-day
episodes from the CY 2020 and 2021 data to create an equivalent
approach to compare PDGM to the hypothetical pre-PDGM. The commenter
stated that this would be consistent with CMS's policy to annually
recalibrate and control for changes in home health resource use and
changes in utilization patterns.
Response: Any change in the average case-mix weight is counteracted
through a corresponding change in the payment rate so that aggregate
expenditures are budget neutral regardless of whether recalibration is
applied. Recalibration ensures that payment incentives for future
utilization are aligned with the design of the payment system (for
example, recalibration ensures roughly a third of periods and episodes
are in a particular functional level). While we currently do not
believe there would be any benefit in recalibrating the case-mix
weights for the simulated 60-day episodes, we may consider it in future
rulemaking.
Comment: A few commenters were concerned the exclusions of certain
categories of claim used in the proposed methodology may have biased
the results.
Response: As stated in the CY 2023 HH PPS proposed rule, exclusions
were made to the CY 2020 and 2021 claims data in order to simulate 60-
day episodes of care (87 FR 37617). These exclusions included
overlapping claims, three or more claims linked to the same OASIS, and
whether it was unclear if there would have been a prior or subsequent
30-day period that would have been a part of a simulated 60-day
episode. All of these exclusions were thoroughly discussed in previous
rulemaking cycles. Without these exclusions, we would not be confident
we were appropriately grouping 30-day periods into simulated 60-day
episodes. It is also important to note, for CY 2020 we excluded 9.5
percent of 30-day periods and for CY 2021 we excluded 16.3 percent of
30-day periods. That is, we kept the majority of 30-day periods in each
year (over 90 percent for CY 2020 and over 83 percent for CY 2021). The
excluded 30-day periods would need to show large differences compared
to the episodes that were not excluded in order to significantly change
the estimated aggregate expenditures from the 60-day episodes to
produce significant revisions to our calculations. As we showed in the
monitoring section of the CY 2023 HH PPS proposed rule, utilization
patterns look largely the same in both CYs 2020 and 2021 (87 FR 37605).
Additionally, the permanent adjustment is based on the percent change
between the payment rates (which utilizes the same claims) and the
temporary adjustment is based on the aggregate expenditures of all
claims (that is, no exclusions) using the two payment rates (that is,
the actual payment rate and the budget neutral payment rate with the
permanent adjustment applied). Therefore, we do not expect the small
portion of excluded claims significantly biased our results.
Comment: A commenter stated that in their own analysis of CMS data
they excluded 30-day claims with a primary diagnosis of COVID-19
because they were unable to assign it a HIPPS code.
Response: We appreciate the diligence of the commenter, and are
grateful that they were able to make full analytical use of the
publicly available data. However, simulated 60-day episodes with a
primary diagnosis of COVID-19 would still be assigned a HIPPS under the
V8219 Home Health Grouper from 3M and would not have been excluded from
the repricing analysis unless there was another unrelated issue with
the claim that prevented grouping.
Final Decision: After consideration of all the comments received
and thorough review of section 1895(b) of the Act, we are finalizing
the proposed methodology to evaluate the impact of the differences of
assumed versus actual behavior changes on estimated aggregate
expenditures.
c. Calculating Permanent and Temporary Payment Adjustments
To offset for such increases or decreases in estimated aggregate
expenditures as a result of the impact of differences between assumed
behavior changes and actual behavior changes, in any given year, we
calculate a permanent prospective adjustment by determining what the
30-day base payment amount should have been in order to achieve the
same estimated aggregate expenditures as obtained from the simulated
60-day episodes. This would be our recalculated base payment rate. The
percent change between the actual 30-day base payment rate and the
recalculated 30-day base payment rate would be the permanent
prospective adjustment.
To calculate a temporary retrospective adjustment for each year we
would determine the dollar amount difference between the estimated
aggregate expenditures from all 30-day periods using the recalculated
30-day base payment rate, and the aggregate expenditures for all 30-day
periods using the actual 30-day base payment rate for the same year. In
determining the temporary retrospective dollar amount, we use the full
dataset of actual 30-day periods using both the actual and recalculated
base payment rates to ensure utilization and distribution of claims are
the same. In accordance with section 1895(b)(3)(D)(iii) of the Act, the
temporary adjustment is to be applied on a prospective basis and shall
apply only with respect to the year for which such temporary increase
or decrease is made. Therefore, after we determine the dollar amount to
be reconciled in any given year, we calculate a temporary adjustment
factor to be applied to the base payment rate. The temporary adjustment
factor is based on an estimated number of 30-day periods in the next
year using historical data trends, and as applicable, we control for a
permanent adjustment factor, case-mix weight recalibration neutrality
factor, wage index budget neutrality factor, and the home health
payment update. The temporary adjustment factor is applied last.
d. CY 2020 Results
Using the methodology described previously, we simulated 60-day
episodes using actual CY 2020 30-day periods to determine what the CY
2020 permanent and temporary payment adjustments should be to offset
for such increases or decreases in estimated aggregate expenditures.
For CY 2020, we began with 8,423,688 30-day periods and dropped 603,157
30-day periods that had a claim occurrence code 50 date after October
31, 2020. We also eliminated 79,328 30-day periods that didn't appear
to group with another 30-day period to form a 60-day episode if the 30-
day period had a ``from date'' before January 15, 2020 or a ``through

[[Page 66805]]

date'' after November 30, 2020. This was done to ensure a 30-day period
would not have been part of a 60-day episode that would have overlapped
into CY 2021. Applying the additional exclusions and assumptions as
described previously, an additional 14,062 30-day periods were excluded
from this analysis. Additionally, we excluded 66,469 simulated 60-day
episodes of care where no OASIS information was available in the CCW
VRDC or could not be grouped to a HIPPS due to a missing primary
diagnosis or other reason. Our simulated 60-day episodes of care
produced a distribution of two 30-day periods of care (70.6 percent)
and single 30-day periods of care (29.4 percent). This distribution is
similar to what we found when we simulated 30-day periods of care for
implementation of the PDGM. After all exclusions and assumptions were
applied, the final dataset included 7,618,061 actual 30-day periods of
care and 4,463,549 simulated 60-day episodes of care for CY 2020.
Using the final dataset for CY 2020 (7,618,061 actual 30-day
periods which made up the 4,463,549 simulated 60-day episodes) we
determined the estimated aggregate expenditures using the pre-PDGM HH
PPS data were lower than the estimated aggregate expenditures using the
PDGM HH PPS data (see Table 2). This indicates that actual aggregate
expenditures under the PDGM were higher than if the 153-group payment
system was still in place in CY 2020. As described previously, we
recalculated what the CY 2020 30-day base payment rate should have been
to equal aggregate expenditures that we calculated using the simulated
CY 2020 60-day episodes. The percent change between the two payment
rates would be the permanent adjustment. To calculate the temporary
adjustment for CY 2020, we calculated the difference in aggregate
expenditures for all CY 2020 PDGM 30-day claims using the actual and
recalculated payment rates. This difference between these two aggregate
expenditures, based on actual and recalculated payment rates, is the
retrospective dollar amount needed to offset any increase or decrease
in the estimated aggregate expenditures. Our results are shown in Table
2.

Table 2--CY 2020 Proposed Permanent and Temporary Adjustments
[GRAPHIC] [TIFF OMITTED] TR04NO22.002

As shown in Table 2, a permanent prospective adjustment of -6.52
percent to the CY 2023 30-day payment rate would be required to offset
for such increases in estimated aggregate expenditures in future years.
Additionally, we determined that our initial estimate of base payment
rates required to achieve budget neutrality resulted in excess payments
to HHAs of approximately $873 million in CY 2020. This would require a
temporary adjustment to offset for such increase in estimated aggregate
expenditures for CY 2020.
e. CY 2021 Results
We will continue the practice of using the most recent complete
home health claims data at the time of rulemaking. The CY 2021 analysis
presented in the CY 2023 HH PPS proposed rule was considered
``preliminary'' and as more data became available from the latter half
of CY 2021, we updated our results. Using the methodology described
previously, we simulated 60-day episodes using actual CY 2021 30-day
periods to determine what the permanent and temporary payment
adjustments should be to offset for such increases or decreases in
estimated aggregate expenditures as a result of the impact of
differences between assumed behavior changes and actual behavior
changes. For CY 2021, we began with 9,269,971 30-day periods of care
and dropped 570,882 30-day periods of care that had claim occurrence
code 50 date after October 31, 2021. We also excluded 968,434 30-day
periods of care that had claim occurrence code 50 date before January
1, 2021 to ensure the 30-day period would not be part of a simulated
60-day episode that began in CY 2020. Applying the additional
exclusions and assumptions as described previously, an additional 5,868
30-day periods were excluded.
Additionally, we excluded 14,302 simulated 60-day episodes of care
where no OASIS information was available in the CCW VRDC or could not
be grouped to a HIPPS due to a missing primary diagnosis or other
reason. Our simulated 60-day episodes of care produced a distribution
of two 30-day periods of care (70.0 percent) and single 30-day periods
of care (30.0 percent) that was similar to what we found when we
simulated two 30-day periods of care for implementation of the PDGM.
After all exclusions and assumptions were applied, the final dataset
included 7,703,261 actual 30-day periods of care and 4,529,498
simulated 60-day episodes of care for CY 2021.
Using the final dataset for CY 2021 (7,703,261 actual 30-day
periods which made up the 4,529,498 simulated 60-day episodes) we
determined the estimated aggregate expenditures under the pre-PDGM HH
PPS was lower than the actual estimated aggregate expenditures under
the PDGM HH PPS. This indicates that aggregate expenditures under the
PDGM were higher than if the 153-group payment system was still in
place in CY 2021. As described previously, we recalculated what the CY
2021 30-day base payment rate should have been to equal aggregate
expenditures that we calculated using the simulated CY 2021 60-day
episodes. We note, the actual CY 2021 base payment rate of $1,901.12
does not account for any adjustments previously made for CY 2020 and
therefore, to evaluate changes for only CY 2021 we need to control for
the -6.52 percent prospective adjustment that we determined for CY
2020. Therefore, using the recalculated CY 2020 base

[[Page 66806]]

payment rate of $1,742.52, multiplied by the CY 2021 wage index budget
neutrality factor (0.9999) and the CY 2021 home health payment update
(1.020), the CY 2021 base payment rate for assumed behavior would have
been $1,777.19. The percent change between the two payment rates would
be the permanent adjustment (assuming the -6.52 percent adjustment was
already taken). Next, we calculated the difference in aggregate
expenditures for all CY 2021 PDGM 30-day claims using the actual
($1,901.12) and recalculated ($1,751.90) payment rates. This difference
is the retrospective dollar amount needed to offset payment. Our
results are shown in Table 3.

Table 3--CY 2021 Proposed Permanent and Temporary Adjustments
[GRAPHIC] [TIFF OMITTED] TR04NO22.003

As shown in Table 3, an additional permanent prospective adjustment
of -1.42 percent (assuming the -6.52 percent adjustment was already
taken) would be required to offset for such increases in estimated
aggregate expenditures in future years. Additionally, we determined
that our initial estimate of the base payment rates required to achieve
budget neutrality resulted in excess expenditures of approximately $1.2
billion in CY 2021. This would require a temporary adjustment factor to
offset for such increases in estimated aggregate expenditures for CY
2021.
f. CY 2023 Permanent and Temporary Adjustments
The percent change between the actual CY 2021 base payment rate of
$1,901.12 and the CY 2021 recalculated base payment rate of $1,751.90
is the total permanent adjustment for CYs 2020 and 2021, because no
previous adjustments were applied to the CY 2020 rate to reset the CY
2021 rate. The summation of the dollar amount for CYs 2020 and 2021 is
the amount that represents the temporary payment adjustment to offset
for increased aggregate expenditures in both CYs 2020 and 2021. Our
results are shown in Table 4 and 5.

Table 4--Total Permanent Adjustment for CYs 2020 and 2021
[GRAPHIC] [TIFF OMITTED] TR04NO22.004

Table 5--Total Temporary Adjustment for CYs 2020 and 2021
[GRAPHIC] [TIFF OMITTED] TR04NO22.005

To offset the increase in estimated aggregate expenditures for CYs
2020 and 2021 based on the impact of the differences between assumed
and actual behavior changes, CMS would need to apply a -7.85 percent
permanent adjustment to the CY 2023 base payment rate as well as
implement a temporary adjustment of approximately $2.1 billion to
reconcile retrospective overpayments in CYs 2020 and 2021. We recognize
that applying the full permanent and temporary adjustment immediately
would result in a significant negative adjustment in a single year.
However, if the PDGM base 30-day payment rate remains higher than it
should be, then there would likely be a compounding effect, potentially
creating the need for an even larger reduction to adjust for behavioral

[[Page 66807]]

changes in future years. Therefore, we proposed to apply only the
permanent adjustment to the CY 2023 base payment rate. We believed this
could mitigate the need for a larger permanent adjustment and could
reduce the amount of any additional temporary adjustments in future
years. We solicited comments on the application of only the permanent
payment adjustment to the CY 2023 30-day payment rate. Additionally, we
solicited comments on how best to collect the temporary payment
adjustment of approximately $2.0 billion for CYs 2020 and 2021.
Comment: MedPAC supported the proposed payment reduction and stated
it is consistent with their recommendation of a five percent reduction
to the base payment rate in the March 2022 report to Congress.\19\
MedPAC commented CMS should decrease home health payments to better
align payments with actual incurred costs, as they found that Medicare
margins for freestanding agencies averaged more than 20 percent from
2001 to 2020.
---------------------------------------------------------------------------

\19\ <a href="https://www.medpac.gov/wp-content/uploads/2022/03/Mar22_MedPAC_ReportToCongress_v2_SEC.pdf">https://www.medpac.gov/wp-content/uploads/2022/03/Mar22_MedPAC_ReportToCongress_v2_SEC.pdf</a>.
---------------------------------------------------------------------------

Response: We appreciate the supportive comment by MedPAC.
Comment: Several commenters expressed concern that the proposed
permanent behavior assumption adjustment would negatively impact home
health providers' business operations. These commenters stated that the
negative adjustment does not consider operational and financial
challenges providers are currently experiencing related to inflation,
staffing shortages, rising costs of gasoline, and medical supplies,
including personal protective equipment (PPE). Commenters also stated
that staffing shortages could be the reason for the decline in visits.
They stated that a negative 7.69 percent behavior assumption adjustment
will cause many agencies to operate with negative margins. Commenters
also expressed concerns that the proposed behavior assumption
adjustment penalizes HHAs and would put access to home health in
jeopardy and impact the quality of care given to home health
beneficiaries. Other commenters stated that CMS should utilize the
existing program integrity measures to identify and target specific
agencies that have excess profit margins rather than impose an across
the board reduction for all agencies, and that CMS should use its
enforcement authority to target HHAs that are cutting utilization or
engaged in other payment-driven behaviors to the detriment of patients.
Another commenter stated that CMS should look for ways to reward ``good
provider behavior.''
Response: We recognize concerns around staffing and appreciate the
commenters' recommendation. However, the statutorily required permanent
and temporary adjustments due to behavior changes is neither to
``reward'' nor ``penalize'' providers. The proposed methodology
controls for overall utilization by using a single year of utilization
data priced under two payment systems to estimate aggregate
expenditures. As such, any effects of staffing issues would be present
in the data under both systems. The payment adjustment is solely to
offset for any increase or decrease in estimated aggregate expenditures
between the two payment systems.
We also recognize the impact inflation and the COVID-19 PHE has had
on healthcare providers, however, we note that in its March 2022 Report
to the Congress,\20\ MedPAC states that Medicare margins increased
under the PDGM, from 15.4 percent in 2019 to 20.2 percent in 2020.
Additionally, they projected margins for home health agencies in 2022
will be roughly 17.0 percent. Furthermore, MedPAC stated in their
report that the Commission found positive access, quality, and
financial indicators for the sector, with average margins of 20.2
percent for freestanding HHAs in 2020, even though the cost per 30-day
period increased by 3.1 percent in this year. We believe that these
margins, despite economic challenges, demonstrate that the payment
rate, along with the market basket update, are more than adequate to
support business operations. Finally, while we appreciate the
commenters' suggestion regarding targeted claim review for specific
home health agencies, we do not believe targeted program integrity
efforts would mitigate behavioral changes resulting from a case-mix
system. We previously addressed this suggestion in the CY 2016 HH PPS
and CY 2019 HH PPS final rules (80 FR 68421 and 83 FR 56455,
respectively). As we previously noted, this strategy is not viable,
given the widespread nature of coding changes and improvements, small
sample sizes of agencies with significant nominal case-mix across
different classes of agencies, and difficulty in precisely
distinguishing the agencies that engage in abusive coding from all
others. Additionally, we reiterate that we are required to make
temporary and permanent payment adjustments to the national,
standardized 30-day period payment rate based on the impact of
differences between assumed versus actual behavior change, in
accordance with sections 1895(b)(3)(D)(ii) and (iii) to offset for such
increases or decreases in estimated aggregate expenditures. These
adjustments are not intended to account for coding abuses, but rather
behavior changes CMS observes across the system. As such, we do not
believe that reducing the 30-day payment rate only for agencies with
high margins is the best way to implement the by statute.
---------------------------------------------------------------------------

\20\ <a href="https://www.medpac.gov/wp-content/uploads/2022/03/Mar22_MedPAC_ReportToCongress_v2_SEC.pdf">https://www.medpac.gov/wp-content/uploads/2022/03/Mar22_MedPAC_ReportToCongress_v2_SEC.pdf</a>.
---------------------------------------------------------------------------

Comment: A few commenters also stated that reduced payment from the
permanent behavior assumption adjustment would exacerbate the already
reduced payment that home health agencies receive from Medicare
Advantage and Medicaid. A commenter stated that CMS fails to consider
that the margins associated with a traditional Medicare beneficiary
subsidize the care of managed Medicare Advantage and Medicaid patients.
Response: While industry representatives contend that Medicare
payments should subsidize payments from other payers (Medicare
Advantage and Medicaid), we disagree. Medicare has never set payments
in order to cross-subsidize other payers. Section 1861(v)(1)(A) of the
Act states ``under the methods of determining costs, the necessary
costs of efficiently delivering covered services to individuals covered
by the insurance programs established by this title will not be borne
by individuals not so covered, and the costs with respect to
individuals not so covered will not be borne by such insurance
programs.'' There is no statutory authority to take the payment rates
of other payers into account when setting Medicare fee-for-service
payment rates.
Comment: Many commenters recommended a phased-in approach over
several years for the permanent and temporary adjustments.
Specifically, a commenter indicated that a phase-in should reduce
payments by no more than 2 percent annually until the adjustment is
achieved. Another commenter recommended the temporary adjustment
starting no earlier than 2026. A few commenters recommended postponing
any adjustments until more data are made available.
Response: We thank the commenters for their recommendations. We
recognize the desire to reduce the payment adjustment; however, note
that any delay in the permanent adjustment

[[Page 66808]]

through a phase-in approach may require larger temporary and permanent
adjustments in the future. While we didn't propose a temporary
adjustment in CY 2023, we will consider the best approach, including a
phase-in, when we do propose the temporary adjustment in future rule-
making.
Final Decision: We stand by the methodology as described previously
and maintain our authority to finalize the adjustment as proposed. But
we recognize the potential hardship of implementing the full -7.85
percent permanent adjustment in a single year. As we have the
discretion to implement any adjustment in a time and manner determined
appropriate, we are finalizing only a -3.925 percent (half of the -7.85
percent) permanent adjustment for CY 2023. However, we note the
permanent adjustment to account for actual behavior changes in CYs 2020
and 2021 should be -7.85 percent. Therefore, applying a -3.925 percent
permanent adjustment to the CY 2023 30-day payment rate would not
adjust the rate fully to account for differences in behavior changes on
estimated aggregate expenditures during those years, as well as in CYs
2022 and 2023. We would have to account for that difference, and any
other potential adjustments needed to the base payment rate, to account
for behavior change based on data analysis in future rulemaking.
While we did not propose to adjust the CY 2023 payment rate using
our temporary adjustment authority for CYs 2020 and 2021, we did
solicit comments on how best to implement the temporary adjustment.
Comment: MedPAC recommended CMS adjust temporary payment rates over
several years, such as adjusting the aggregate rate by $502.5 million
per year for CYs 2023 through 2026. MedPAC strongly recommended
beginning these reductions immediately to avoid potential larger
reductions in future years.
Response: We thank MedPAC for their recommendation. However, while
CMS proposed the methodology for calculating both the permanent and
temporary adjustments, in the CY 2023 HH PPS proposed rule we did not
propose collecting the $2.0 billion temporary adjustment for CYs 2020
and 2021 beginning in CY 2023. We did solicit comments on how best to
collect the temporary payment adjustment and will take these comments
into consideration when we propose any temporary adjustments in future
rulemaking.
Comment: Many commenters recommended a phase-in over several years
for the temporary adjustment and another year delay before recovering
any overpayments. Another commenter stated the recoupment should not be
applied equally to all HHAs, but rather CMS should target recoupment
based on agency level analyses to determine those HHAs who had high
margins, egregious behavior changes, and ``cherry pick'' patients.
Response: We appreciate the commenters recommendation. We note that
this is not a recoupment in the legal sense, but, as the statute
specifies at section 1895(b)(3)(D)(iii) of the Act, a temporary
adjustment to account for retrospective behavior. While there may be
different business models between HHAs, those practices are outside the
scope of this policy. Specifically, we believe the best way to
interpret the statute is to apply any adjustments (permanent and
temporary) to the national, standardized 30-day period payment rate on
a prospective basis.
Final Decision: We thank commenters for their suggestions about how
to implement the temporary payment adjustments and will consider them
in future rulemaking.
3. Reassignment of Specific ICD-10-CM Codes Under the PDGM
a. Background
The 2009 final rule, ``HIPAA Administrative Simplification:
Modifications to Medical Data Code Set Standards To Adopt ICD-10-CM and
ICD-10-PCS'' \21\ (74 FR 3328, January 16, 2009), set October 1, 2013,
as the compliance date for all covered entities under the Health
Insurance Portability and Accountability Act of 1996 (HIPAA) to use the
International Classification of Diseases, 10th Revision, Clinical
Modification (ICD-10-CM) and the International Classification of
Diseases, 10th Revision, Procedure Coding System (ICD-10-PCS) medical
data code sets. The ICD-10-CM diagnosis codes are granular and
specific, and provide HHAs a better opportunity to report codes that
best reflect the patient's conditions that support the need for home
health services. However, as stated in the CY 2019 HH PPS final rule
with comment period (83 FR 56473), because the ICD-10-CM is
comprehensive, it also contains many codes that may not support the
need for home health services. For example, diagnosis codes that
indicate death as the outcome are Medicare covered codes, but are not
relevant to home health. In addition, diagnosis and procedure coding
guidelines may specify the sequence of ICD-10-CM coding conventions.
For example, the underlying condition must be listed first (for
example, Parkinson's disease must be listed prior to Dementia if both
codes were listed on a claim). Therefore, not all the ICD-10-CM
diagnosis codes are appropriate as principal diagnosis codes for
grouping home health periods into clinical groups or to be placed into
a comorbidity subgroup when listed as a secondary diagnosis. As such,
each ICD-10-CM diagnosis code is assigned, including those diagnosis
codes designated as ``not assigned'' (NA), to a clinical group and
comorbidity subgroup within the HH PPS grouper software (HHGS). We
reminded commenters the ICD-10-CM diagnosis code list is updated each
fiscal year with an effective date of October 1st and therefore, the HH
PPS is generally subject to a minimum of two HHGS releases, one in
October and one in January of each year, to ensure that claims are
submitted with the most current code set available. Likewise, there may
be new ICD-10-CM diagnosis codes created (for example, codes for
emergency use) or a new or revised edit in the Medicare Code Editor
(MCE) so an update to the HHGS may occur on the first of each quarter
(January, April, July, October).
---------------------------------------------------------------------------

\21\ <a href="https://www.federalregister.gov/documents/2009/01/16/E9-743/hipaa-administrative-simplification-modifications-to-medical-data-code-set-standards-to-adopt">https://www.federalregister.gov/documents/2009/01/16/E9-743/hipaa-administrative-simplification-modifications-to-medical-data-code-set-standards-to-adopt</a>.
---------------------------------------------------------------------------

b. Methodology for ICD-10-CM Diagnosis Code Assignments
Although it is not our intent to review all ICD-10-CM diagnosis
codes each year, we recognize that occasionally some ICD-10-CM
diagnosis codes may require changes to their assigned clinical group
and/or comorbidity subgroup. For example, there may be an update to the
MCE unacceptable principal diagnosis list, or we receive public
comments from interested parties requesting specific changes. Any
addition or removal of a specific diagnosis code to the ICD-10-CM code
set (for example, three new diagnosis codes, Z28.310, Z28.311 and
Z28.39, for reporting COVID-19 vaccination status were effective April
1, 2022) or minor tweaks to a descriptor of an existing ICD-10-CM
diagnosis code generally would not require rulemaking and may occur at
any time. However, if an ICD-10-CM diagnosis code is to be reassigned
from one clinical group and/or a comorbidity subgroup to another, which
may affect payment, then we believe it is appropriate to propose these
changes through notice and comment rulemaking.
We rely on the expert opinion of our clinical reviewers (for
example, nurse

[[Page 66809]]

consultants and medical officers) and current ICD-10-CM coding
guidelines to determine if the ICD-10-CM diagnosis codes under review
for reassignment are significantly similar or different to the existing
clinical group and/or comorbidity subgroup assignment. As we stated in
the CY 2018 HH PPS proposed rule (82 FR 35313), the intent of the
clinical groups is to reflect the reported principal diagnosis,
clinical relevance, and coding guidelines and conventions. Therefore,
for the purposes of assignment of ICD-10-CM diagnosis codes into the
PDGM clinical groups we would not conduct additional statistical
analysis as such decisions are clinically based and the clinical groups
are part of the overall case-mix weights.
As we noted in the CY 2019 HH PPS final rule with comment period
(83 FR 56486), the home health-specific comorbidity list is based on
the principles of patient assessment by body systems and their
associated diseases, conditions, and injuries to develop larger
categories of conditions that identified clinically relevant
relationships associated with increased resource use meaning the
diagnoses have at least as high as the median resource use and are
reported in more than 0.1 percent of 30-day periods of care. If
specific ICD-10-CM diagnosis codes are to be reassigned to a different
comorbidity subgroup (including NA), we will first evaluate the
clinical characteristics (as discussed previously for clinical groups)
and if the ICD-10-CM diagnosis code does not meet the clinical
criteria, then no reassignment will occur. However, if an ICD-10-CM
diagnosis code does meet the clinical criteria for a comorbidity
subgroup reassignment, then we will evaluate the resource consumption
associated with the ICD-10-CM diagnosis codes, the current assigned
comorbidity subgroup, and the proposed (reassigned) comorbidity
subgroup. This analysis is to ensure that any reassignment of an ICD-
10-CM diagnosis code (if reported as secondary) in any given year would
not significantly alter the overall resource use of a specific
comorbidity subgroup. For resource consumption, we use non-LUPA 30-day
periods to evaluate the total number of 30-day periods for the
comorbidity subgroup(s) and the ICD-10-CM diagnosis code, the average
number of visits per 30-day periods for the comorbidity subgroup(s) and
the ICD-10-CM diagnosis code, and the average resource use for the
comorbidity subgroup(s) and the ICD-10-CM diagnosis code. The average
resource use measures the costs associated with visits performed during
a home health period, and was previously described in the CY 2019 HH
PPS final rule with comment period (83 FR 56450).
c. ICD-10-CM Diagnosis Code Reassignments to a PDGM Clinical Group or
Comorbidity Subgroup
The following section proposed reassignment of 320 diagnosis codes
to a different clinical group when listed as a principal diagnosis,
reassignment of 37 diagnosis codes to a different comorbidity subgroup
when listed as a secondary diagnosis, and the establishment of a new
comorbidity subgroup for certain neurological conditions and disorders.
Due to the amount of diagnosis codes proposed for reassignment this
year, we posted the ``CY 2023 Proposed Reassignment of ICD-10-CM
Diagnosis Codes for HH PDGM Clinical Groups and Comorbidity Subgroups''
supplemental file on the Home Health Prospective Payment System
Regulations and Notices web page.\22\
---------------------------------------------------------------------------

\22\ Home Health Prospective Payment System Regulations and
Notices web page. <a href="https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/HomeHealthPPS/Home-Health-Prospective-Payment-System-Regulations-and-Notices">https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/HomeHealthPPS/Home-Health-Prospective-Payment-System-Regulations-and-Notices</a>.
---------------------------------------------------------------------------

Comment: Several commenters supported the general refinement of
coding assignments, including all the proposed coding changes. A
commenter stated that the changes will help to more accurately reflect
patients' needs and why they need home health services, rather than
using ``pain'' as a diagnosis.
Response: We thank these commenters for their support and agree
that the changes will provide more specific information related to the
needs of the patient under a home health plan of care.
Comment: Several commenters expressed concern that reassignment of
clinical groups for principal diagnosis codes would result in an access
to care issue. For example, commenters were concerned that a
reassignment of principal diagnosis codes from a clinical group to no
clinical group, would change the case-mix weight and reimbursement, and
that the HHA may refuse the patient, thus restricting access to care.
There was also concern that if the clinical group changed (for example,
MS-Rehab to Wounds), the HHA would restrict the type of services
provided, such as physical therapy, also restricting access to care.
Response: It is unclear why commenters believe any reassignments
would restrict access to care, and note that the CoPs at Sec. 484.60
state that the individualized plan of care must specify the care and
services necessary to meet the patient-specific needs as identified in
the comprehensive assessment, including identification of the
responsible discipline(s), and the measurable outcomes that the HHA
anticipates will occur as a result of implementing and coordinating the
plan of care. Services must be furnished in accordance with accepted
standards of practice. The purpose of any reassignment is to ensure
that diagnoses are assigned to the appropriate clinical group or
comorbidity subgroup and to align as closely as possible to ICD-10-CM
coding conventions and MCE edits. These edits may have payment effects
but should not result in any change in clinical practice or
availability of services, unless the agency is failing to act in
accordance with the plan of care.
Comment: A few commenters requested that CMS modify the clinical
groups to accept and include diagnosis codes which may drive a home
health need. Specifically, commenters requested allowing R29.6
(repeated falls), R54 (age-related physical debility), R26.89 (other
abnormalities of gait and mobility), R42.82 (altered mental status,
unspecified), and M62.81(muscle weakness (generalized)) to be accepted
as a principal diagnosis and placed into a clinical group for payment.
Response: We thank the commenters for their coding recommendations.
However, we did not propose to assign any of the R-codes to a clinical
group and therefore, such suggestions are out of scope for this rule.
We remind commenters that R-codes are codes describing symptoms, signs,
and abnormal clinical and laboratory findings, not elsewhere
classified) and are generally not allowed as a principal diagnosis
(except for a few) in accordance with ICD-10-CM coding guidelines. Any
changes to the acceptable principal diagnosis list for home health,
including the addition of new ICD-10 codes, would have to go through
notice and comment rulemaking.
(1) Clinical Group Reassignment of Certain Unspecified Diagnosis Codes
We reminded readers that in the CY 2019 HH PPS final rule with
comment period (83 FR 56473) we stated that whenever possible, the most
specific code that describes a medical disease, condition, or injury
should be used. Generally, ``unspecified'' codes are used when there is
lack of information about location or severity of medical conditions in
the medical record. However, we would expect a provider to

[[Page 66810]]

use a precise code whenever more specific codes are available.
Furthermore, if additional information regarding the diagnosis is
needed, we would expect the HHA to follow-up with the referring
provider in order to ensure the care plan is sufficient in meeting the
needs of the patient. For example, T14.90 ``Injury, unspecified'' does
not provide sufficient information (for example, the type and extent of
the injury) that would be necessary in care planning for home health
services. The ICD-10-CM code set also includes laterality. We believe a
home health clinician should not report an ``unspecified'' code if that
clinician can identify the side or site of a condition. For example, a
home health clinician should be able to state whether a fracture of the
arm is on the right or left arm. In the FY 2022 Inpatient Prospective
Payment System/Long-Term Care Hospital Prospective Payment System
(IPPS/LTCH PPS) final rule (86 FR 44940 through 44943), CMS finalized
the implementation of a new MCE to expand the list of unacceptable
principal diagnoses for ``unspecified'' ICD-10-CM diagnosis codes when
there are other diagnosis codes available in that diagnosis code
subcategory that further specify the anatomic site. As such, we
reviewed all the ICD-10-CM diagnosis codes where ``unspecified'' is
used and not just the ones listed on the new MCE edit. We identified
159 ICD-10-CM diagnosis codes that are currently accepted as a
principal diagnosis that have more specific codes available for such
medical conditions that would more accurately identify the primary
reason for home health services. For example, S59.109A (Unspecified
physeal fracture of upper end of radius, unspecified arm, initial
encounter for closed fracture) does not specify which arm has the
fracture; whereas, S59.101A (Unspecified physeal fracture of upper end
of radius, right arm, initial encounter for closed fracture) does
indicate the fracture is on the right arm and therefore more accurately
identifies the primary reason for home health services. Therefore, in
accordance with our expectation that the most precise code be used, we
stated that we believe these 159 ICD-10 CM diagnosis codes are not
acceptable as principal diagnoses and we proposed to reassign them to
``no clinical group'' (NA). We refer readers to Table 1.A of the CY
2023 Proposed Reassignment of ICD-10-CM Diagnosis Codes supplemental
file \23\ for the list of the 159 unspecified diagnosis codes.
---------------------------------------------------------------------------

\23\ Home Health Prospective Payment System Regulations and
Notices web page. <a href="https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/HomeHealthPPS/Home-Health-Prospective-Payment-System-Regulations-and-Notices">https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/HomeHealthPPS/Home-Health-Prospective-Payment-System-Regulations-and-Notices</a>.
---------------------------------------------------------------------------

We also determined that B78.9 strongyloidiasis, unspecified was
assigned to clinical group C (Wounds), and should be reassigned to
clinical group K (MMTA--Infectious Disease, Neoplasms, and Blood-
Forming Diseases) because it would be consistent with the assignment of
the other strongyloidiasis codes. We also identified that N83.201
unspecified ovarian cyst, right side was assigned to clinical group A
(MMTA--Other) and should be reassigned to clinical group J (MMTA--
Gastrointestinal Tract and Genitourinary System) because it would be
consistent with the assignment of other ovarian cyst codes. We proposed
to reassign these two ICD-10-CM diagnosis codes' clinical groups as
shown in Table 6.

Table 6--Reassignment of Clinical Group for ``Unspecified'' ICD-10-CM
Diagnosis Codes
[GRAPHIC] [TIFF OMITTED] TR04NO22.006

Comment: Several commenters were concerned about the proposal to
reassign the 159 ICD-10-CM codes to no clinical group (NA) when listed
as a principal diagnosis. Commenters stated that only 45 of the 159
ICD-10-CM codes were listed on the MCE 20 list of unacceptable
principal diagnoses and that the home health Grouper would be
inconsistent with the other MCE edits. While commenters agreed the most
specific documentation should be reflected in medical records to assign
the most specific code available, they noted that there are certain
circumstances in which an unspecified code should be accepted as a
principal diagnosis according to the MCE manual and ICD-10-CM Official
Guidelines for Coding and Reporting.\24\ In addition, commenters stated
that obtaining additional information may be burdensome to certain
HHAs.
---------------------------------------------------------------------------

\24\ <a href="https://www.cms.gov/files/document/fy-2022-icd-10-cm-coding-guidelines-updated-02012022.pdf">https://www.cms.gov/files/document/fy-2022-icd-10-cm-coding-guidelines-updated-02012022.pdf</a>.
---------------------------------------------------------------------------

Response: We thank interested parties for their comments. As we
noted in the CY 2023 HH PPS proposed rule and previously in this final
rule, we did not limit our review of unspecified codes only to those on
the MCE edit list. Instead, the release of the MCE 20 edit prompted our
review of all unspecified codes currently assigned to a clinical group
when listed as a principal diagnosis.
We also recognize the desire for a consistent unspecified edit for
all health care entities; however, this is not feasible given the vast
differences across Medicare benefits and their associated payment
systems. As such, CMS has created different groupers to institute edits
to a specific program. For example, home health uses the Home Health
Resource Group (HHRG), while inpatient rehabilitation facilities use
Case Mix Group (CMG), both of which are different from the inpatient
and outpatient grouper software.
We acknowledge the ICD-10-CM Official Guidelines for Coding and
Reporting Section I.B.18 states ``If a definitive diagnosis has not
been established by the end of the encounter, it is appropriate to
report codes for sign(s) and/or symptom(s) in lieu of a definitive
diagnosis. When sufficient clinical information is not known or
available about a particular health condition to assign a more specific
code, it is acceptable to report the appropriate ``unspecified'' code
(for example, a diagnosis of pneumonia has been determined, but not the
specific type). Unspecified codes should be reported when they are the
codes that most accurately reflect what is known about the patient's
condition at the time of that particular encounter.'' However, as
previously stated in the CY 2019 HH PPS final rule with comment period
(83

[[Page 66811]]

FR 56473) and the CY 2023 HH PPS proposed rule, ``unspecified'' codes
are used when the record lacks information about location or severity
of medical conditions if additional information regarding the diagnosis
is needed, we would expect the HHA to follow-up with the referring
provider in order to ensure the care plan is sufficient in meeting the
needs of the patient. Of the proposed 159 ICD-10-CM diagnosis codes, 85
percent (136 codes) lacked information about location (that is,
laterality) while the remaining 15 percent (23 codes) lacked
information about severity. We understand commenters concerns that many
home health visits may be subsequent to the initial injury or disease
and the medical record may lack information. However, we still believe
this supports the need for more specific codes in order for the
provider to appropriately provide services in alignment with the plan
of care.
In addition, per the FY 2022 IPPS/LTCH final rule (86 FR 44943),
if, upon review, additional information to identify the laterality from
the available medical record documentation by any other clinical
provider is unable to be obtained, or there is documentation in the
record indicating that the physician is clinically unable to determine
the laterality because of the nature of the disease/condition, then the
provider must enter that information into the remarks section. If there
is no language entered into the remarks section as to the availability
of additional information to specify laterality and the provider
submits the claim for processing, the claim would then be returned to
the provider. While Medicare systems may allow an edit to be bypassable
(for example, the NOA timelines extension), it does not currently allow
an unacceptable home health principal diagnosis to be bypassable. We
may consider adding certain additional edits as bypassable in future
rulemaking.
In response to the 15 codes where more specific codes identify
severity, rather than laterality, we further evaluated if a more
specific code would be appropriate in determining the plan of care and
home health services required. We determined that 11 of the codes not
only had more specific codes, but there are similar unspecified codes
in the same subchapter which we do not accept as a principal diagnosis.
For example, for pregnancy-related codes, we expect the trimester to be
specified. However, based on comments and further review we determined
the four codes listed in Table 7 below should remain with their current
assigned clinical group when listed as a principal diagnosis as we
believe the information in these codes is sufficient to establish a
home health plan of care to address such conditions.

Table 7--Unspecified Diagnosis Codes Remaining in Clinical Groups
[GRAPHIC] [TIFF OMITTED] TR04NO22.007

Final Decision: After consideration of the public comments
received, we are modifying our proposal of the 159 ICD-10 CM
``unspecified'' diagnosis codes to be reassigned to N/A by excluding
the four codes listed in Table 7. Instead we are finalizing the
reassignment of the remaining 155 ICD-10 CM diagnosis codes from their
current assigned clinical group to NA when the codes are listed as a
principal diagnosis. We remind readers that if a claim cannot be
assigned a clinical group, the claim will be returned to the provider
for further information. We are also finalizing the reassignment of
B78.9 (strongyloidiasis, unspecified) from clinical group C (Wounds) to
clinical group K (MMTA--Infectious Disease, Neoplasms, and Blood-
Forming Diseases) and the reassignment of N83.201 (unspecified ovarian
cyst, right side) from clinical group A (MMTA-Other) to clinical group
J (MMTA--Gastrointestinal Tract and Genitourinary System) when listed
as the principal diagnoses. We urge interested parties to review the
final HH Clinical Group and Comorbidity Adjustment Diagnosis list
released with this final rule, as well as the 3M Grouper January 2023
HH PPS Grouper Software HH PDGM v04.0.23, when determining if an ICD-10
CM diagnosis code is accepted as a principal diagnosis and assigned a
clinical group.
(2) Clinical Group Reassignment of Gout-Related Codes
We identified that certain groups of gout-related ICD-10-CM
diagnosis codes, such as idiopathic gout and drug-induced gout, were
assigned to clinical group E (musculoskeletal rehabilitation) when
listed as a principal diagnosis. However, other groups of gout related
ICD-10-CM diagnosis codes, such as gout due to renal impairment, were
assigned to ``no clinical group'' (NA). Therefore, we reviewed all
gout-related codes and determined there are 144 gout related codes with
an anatomical site specified, not currently assigned to a clinical
group that should be moved to clinical group E (musculoskeletal
rehabilitation) for consistency with the aforementioned gout codes. In
the ICD-10-CM code set, gout codes and osteoarthritis codes are found
in chapter 13 Diseases of the Musculoskeletal System and Connective
Tissue (M00-M99). Gout and osteoarthritis affect similar joints such as
the fingers, toes, and knees and they can initially be treated with
medications. However, generally, as a part of a treatment program, once
the initial inflammation

[[Page 66812]]

is reduced, physical therapy can be started to stretch and strengthen
the affected joint to restore flexibility and joint function. Because
those cases may require therapy, we believe gout codes are more
appropriately placed into MS rehab along with other codes affecting the
musculoskeletal system. We refer readers to Table 1.B of the CY 2023
Proposed Reassignment of ICD-10-CM Diagnosis Codes supplemental file
for the list of the 144 gout related codes. We did not receive comments
on this proposal and therefore are finalizing the reassignment of these
144 gout-related ICD-10-CM diagnosis codes to clinical group E
(musculoskeletal rehabilitation) without modification.
(3) Clinical Group Reassignment of Crushing Injury-Related Codes
We identified 12 ICD-10-CM diagnosis codes related to crushing
injury of the face, skull, and head that warrant reassignment. These
codes are listed in Table 8.

Table 8--ICD-10-CM Diagnosis Codes Related to Crushing Injury of Face,
Skull, and Head
[GRAPHIC] [TIFF OMITTED] TR04NO22.008

Our clinical advisors reviewed the 12 ICD-10-CM diagnosis codes
related to crushing injury of the face, skull, and head and determined
that reassignment of these codes to clinical group B (Neurological
Rehabilitation) is clinically appropriate because they are consistent
with other diagnosis codes in clinical group B that describe injuries
requiring neurological rehabilitation. We did not receive comments on
this proposal and therefore are finalizing the reassignment of the ICD-
10-CM diagnosis codes listed in Table 8 from clinical group A (MMTA-
Other) to clinical group B (Neurological Rehabilitation) without
modification.
(4) Clinical Group Reassignment of Lymphedema-Related Codes
We received questions from interested parties regarding three
lymphedema codes with conflicting clinical group assignments when
listed as a principal diagnosis. These codes are listed in Table 9.

Table 9--ICD-10-CM Diagnosis Code Related to Lymphedema
[GRAPHIC] [TIFF OMITTED] TR04NO22.009

Our clinical advisors reviewed the three ICD-10-CM diagnosis codes
related to lymphedema and determined that assessing and treating
lymphedema is similar to the assessment and staging of wounds. It
requires the assessment of pulses, evaluation of the color and amount
of drainage, and measurement. In addition, some lymphedema can require
compression bandaging, similar to wound care. Because of these
similarities, we determined the reassignment of the three ICD-10-CM
diagnosis codes related to lymphedema to clinical group C (Wounds) is
clinically appropriate. Therefore, we proposed to reassign the ICD-10-
CM diagnosis codes listed in Table 9 from clinical group E
(Musculoskeletal Rehabilitation) and clinical group A (MMTA-Other) to
clinical group C (Wounds).
Comment: Several commenters questioned whether the reassignment of
lymphedema to clinical group C (wounds) would impact the type of
practitioner who would be able to treat the wound or limit patient
access to resources such as complete decongestive therapy including
manual lymph drain
Response: We thank the commenters for their concern. The
reassignment of lymphedema, or any other code, would not impact the
type of practitioner providing services, as long as the allowed
practitioner can perform the care under their scope of practice. In
addition, per the CoPs, HHAs should continue to provide services in
accordance with the plan of care.

[[Page 66813]]

Comment: A commenter questioned if CMS considers lymphedema a wound
type and if we believe lymphedema is correlated to venous disease/
wounds.
Response: Although CMS does not consider lymphedema to be a wound
type, we believe clinically that the home health services needed to
treat and manage lymphedema are equivalent to the time and services
needed for managing an open wound regardless of the precipitating
condition that resulted in lymphedema. Treatment for lymphedema focuses
on reducing swelling and minimizing complications. As such, treatment
could involve exercises, manual lymphatic drainage, compression
bandages or garments, sequential pneumatic compression, and even wound
care for any skin breakdown. Because the home health treatments can be
similar in terms of care and intensity of care, we believe lymphedema
and wounds are appropriate to be grouped together for clinical
groupings.
Final Decision: After consideration of the public comments we
received, we are finalizing the reassignment of the ICD-10-CM diagnosis
codes listed in Table B19 from clinical group E (Musculoskeletal
Rehabilitation) and clinical group A (MMTA-Other) to clinical group C
(Wounds).
(5) Behavioral Health Comorbidity Subgroups
Our clinical advisors reviewed the ICD-10-CM diagnosis code F60.5
(obsessive-compulsive personality disorder) which is currently assigned
to the comorbidity subgroup behavioral 6 (Schizotypal, Persistent Mood,
and Adult Personality Disorders). However, they noted that behavioral 5
(Phobias, Other Anxiety and Obsessive-Compulsive Disorders) contains
other obsessive-compulsive disorders (for example, F42.8 and F42.9) and
clinically F60.5 should be reassigned to the comorbidity subgroup
behavioral 5. In addition, we evaluated resource consumption related to
the comorbidity subgroup behavioral 5, the comorbidity subgroup
behavioral 6, and F60.5 and found no significant variations negating a
reassignment, meaning the reassignment is still in alignment with the
actual costs of providing care. We did not receive comments on this
proposal, and therefore are finalizing the reassignment of diagnosis
code F60.5 to behavioral 5 when listed as a secondary diagnosis.
(6) Circulatory Comorbidity Subgroups
We reviewed Q82.0 (hereditary lymphedema) for clinical group
reassignment, as described in section II.B.3.4. of this rule. During
this review, we discovered Q82.0 is not currently assigned to a
comorbidity subgroup when listed as a secondary diagnosis. The
comorbidity subgroup circulatory 10 includes ICD-10-CM diagnosis codes
related to varicose veins and lymphedema. Therefore, our clinical
advisors determined that Q82.0 should be assigned to the comorbidity
subgroup circulatory 10 similar to other lymphedema diagnosis codes. In
addition, we evaluated resource consumption related to the comorbidity
subgroup circulatory 10 and Q82.0 and found no significant variations
negating a reassignment. Therefore, we proposed to assign diagnosis
code Q82.0 to circulatory 10 (varicose veins and lymphedema) when
listed as a secondary diagnosis.
Final Decision: We received a comment in support of this
assignment; therefore, we are finalizing the assignment of Q82.0
(hereditary lymphedema) from ``NA'' to circulatory 10 (varicose veins
and lymphedema) when listed as a secondary diagnosis.
(7) Neoplasm Comorbidity Subgroups
(i) Malignant Neoplasm of Upper Respiratory
In response to interested parties' questions regarding upper
respiratory malignant neoplasms, we reviewed 14 ICD-10-CM diagnosis
codes related to malignant neoplasms of the upper respiratory tract
currently assigned to the comorbidity subgroup neoplasm 6 (malignant
neoplasms of trachea, bronchus, lung, and mediastinum). These 14 codes
are listed in Table 10.

Table 10--ICD-10-CM Diagnosis Code Related to Malignant Neoplasms of
Upper Respiratory Tract
[GRAPHIC] [TIFF OMITTED] TR04NO22.010

Our clinical advisors reviewed the codes listed in Table 10 and
determined that C32.3, C32.8, and C32.9 are currently assigned to the
most clinically appropriate neoplasm comorbidity subgroup (neoplasm 6),
and therefore no further analysis was conducted for these three ICD-10
CM diagnosis codes. However, upon review of all the neoplasm
comorbidity subgroups, they determined that the remaining 11 codes
listed in Table 10 should be reassigned

[[Page 66814]]

to neoplasm 1 (malignant neoplasms of lip, oral cavity, and pharynx,
including head and neck cancers) in alignment with clinically similar
diagnosis codes already assigned (for example, C11.0 malignant neoplasm
of superior wall of nasopharynx). In addition, we evaluated resource
consumption related to the comorbidity subgroup, neoplasm 1, as well as
diagnosis codes, C30.0, C30.1, C31.0, C31.1, C31.2, C31.3, C31.8,
C31.9, C32.0, C32.1, or C32.2 and found no significant variations
negating a reassignment.
We did not receive comments on this proposal and therefore are
finalizing the reassignment of diagnosis codes C30.0, C30.1, C31.0,
C31.1, C31.2, C31.3, C31.8, C31.9, C32.0, C32.1, or C32.2 from neoplasm
6 to neoplasm 1 when listed as a secondary diagnosis.
(ii) Malignant Neoplasm of Unspecified Adrenal Gland
While reviewing unspecified codes for a change in clinical group,
we noticed that ICD-10-CM diagnosis codes C74.00 (malignant neoplasm of
cortex of unspecified adrenal gland) and C74.90 (malignant neoplasm of
unspecified part of unspecified adrenal gland) were coded as ``N/A''
instead of placed in a comorbidity subgroup. The comorbidity subgroup
neoplasm 15 currently includes ICD-10-CM diagnosis codes related to
malignant neoplasm of adrenal gland, endocrine glands and related
structures; specifically, C74.10 (malignant neoplasm of medulla of
unspecified adrenal gland). At this time, we believe that C74.00 and
C74.90 should be reassigned to neoplasm 15 based on clinical
similarities of other codes currently assigned. In addition, we
evaluated resource consumption related to the comorbidity subgroup
neoplasm 15, as well as diagnosis codes C74.00, and C74.90 and found no
significant variations negating a reassignment. We did not receive
comments on this proposal and therefore are finalizing the reassignment
of diagnosis codes C74.00 and C74.90 from ``NA'' to neoplasm 15
(malignant neoplasm of adrenal gland, endocrine glands and related
structures) when listed as secondary diagnoses.
(8) New Neurological Comorbidity Subgroup
In response to a comment received, we discussed in the CY 2022
final rule (86 FR 62263, 62264) our review of ICD-10-CM diagnosis codes
related to specified neuropathy or unspecified polyneuropathy. These
include specific ICD-10-CM G-codes. We stated that the codes were
assigned to the most clinically appropriate subgroup at the time.
However, upon further clinical review we believe a new neurological
comorbidity subgroup to include ICD-10-CM diagnosis codes related to
nondiabetic neuropathy is warranted. We identified 18 ICD-10-CM
diagnosis codes for potential reassignment to a proposed new
comorbidity subgroup, neurological 12. We refer readers to Table 1.C of
the CY 2023 Proposed Reassignment of ICD-10-CM Diagnosis Codes
supplemental file for a list of the G-codes related to specified
neuropathy or unspecified polyneuropathy. Of the 18 codes, 11 diagnosis
codes were not currently assigned a comorbidity group and seven
diagnosis codes were assigned to neurological 11 comorbidity subgroup.
Using claims data from the CY 2021 HH PPS analytical file, we
identified that the 18 diagnosis G-codes related to specified
neuropathy or unspecified polyneuropathy would have sufficient claims
(>400,000) for a new comorbidity subgroup. The removal of the seven
codes from the neurological 11 comorbidity subgroup, would still allow
for sufficient claims (>250,000) and include the remaining 146
diagnosis codes currently listed in the neurological 11 comorbidity
subgroup. We evaluated resource consumption related to the comorbidity
subgroup neurological 11, the 18 diagnosis G-codes, and the proposed
comorbidity subgroup neurological 12 and found no significant
variations negating a reassignment. A new neurological comorbidity
subgroup allows more clinically similar codes, nondiabetic neuropathy,
to be grouped together. Therefore, we proposed to reassign the 18
diagnosis codes listed in Table 1.C of the CY 2023 Proposed
Reassignment of ICD-10 CM Diagnosis Codes supplemental file, to the new
comorbidity subgroup neurological 12 (nondiabetic neuropathy) when
listed as secondary diagnoses. In conjunction with the proposed new
comorbidity subgroup, we proposed to change the description of the
current comorbidity subgroup, neurological 11, from ``Diabetic
Retinopathy and Macular Edema'' to ``Disease of the Macula and
Blindness/Low Vision''.
Comment: A few commenters supported the creation of the
neurological subgroup for nondiabetic neuropathy.
Response: We thank the commenters for their support.
Final Decision: After consideration of the public comments we
received, we are finalizing a new neurological comorbidity subgroup,
neurological 12 (nondiabetic neuropathy), and reassigning the 18
diagnosis codes listed in Table 1.C of the CY 2023 Proposed
Reassignment of ICD-10 CM Diagnosis Codes supplemental file to the
neurological 12 (nondiabetic neuropathy). We did not receive comments
on the proposal to change the description of the comorbidity subgroup,
neurological 11, and are therefore finalizing neurological 11, from
``Diabetic Retinopathy and Macular Edema'' to ``Disease of the Macula
and Blindness/Low Vision''.
(9) Respiratory Comorbidity Subgroups
(i) J18.2 Hypostatic Pneumonia, Unspecified Organism
Our clinical advisors reviewed the ICD-10-CM diagnosis code J18.2
(hypostatic pneumonia, unspecified organism) which is currently
assigned to the comorbidity subgroup respiratory 4 (bronchitis,
emphysema, and interstitial lung disease). However, respiratory 2
(whooping cough and pneumonia) contains other pneumonia with
unspecified organism (for example, J18.1 and J18.8). Clinically, J18.2
is similar to the other pneumonias in respiratory 2 and therefore,
should be reassigned from comorbidity subgroup respiratory 4 to
comorbidity subgroup respiratory 2. In addition, we evaluated resource
consumption related to the comorbidity subgroups respiratory 2 and
respiratory 4, and J18.2 and found no significant variations negating a
reassignment.
We did not receive comments on this proposal and therefore are
finalizing the reassignment of diagnosis code J18.2 (hypostatic
pneumonia, unspecified organism) to respiratory 2 when listed as a
secondary diagnosis.
(ii) J98.2 Interstitial Emphysema and J98.3 Compensatory Emphysema
Our clinical advisors reviewed the ICD-10-CM diagnosis codes J98.2
(interstitial emphysema) and J98.3 (compensatory emphysema), which are
currently assigned to the comorbidity subgroup respiratory 9
(respiratory failure and atelectasis). However, respiratory 4
(bronchitis, emphysema, and interstitial lung disease) contains other
emphysema codes (for example, J43.0 through J43.9) and therefore
clinically we believe it is appropriate to reassign J98.2 and J98.3 to
the comorbidity subgroup respiratory 9. In addition, we evaluated
resource consumption related to the comorbidity subgroups respiratory 4
and respiratory 9, as well as diagnosis codes J98.2, and J98.3 and
found no significant variations negating a reassignment. We did not
receive comments on this proposal and therefore are finalizing the
reassignment

[[Page 66815]]

of diagnosis codes J98.2 and J98.3 to respiratory 4 when listed as a
secondary diagnosis.
(iii) U09.9 Post COVID-19 Condition, Unspecified
Our clinical advisors reviewed the ICD-10-CM diagnosis code U09.9
(post COVID-19 condition, unspecified), which is currently assigned to
the comorbidity subgroup, respiratory 2 (whooping cough and pneumonia).
However, respiratory 10 (2019 novel Coronavirus) contains other COVID-
19 codes (for example, U07.1). Therefore, we believe clinically that
U09.9 should be reassigned to the comorbidity subgroup, respiratory 10.
In addition, we evaluated resource consumption related to the
comorbidity subgroups respiratory 2 and respiratory 10, and diagnosis
codes U09.9 and found no significant variations negating a
reassignment. We did not receive comments on this proposal and
therefore are finalizing the reassignment of diagnosis code U09.9 to
respiratory 10 when listed as a secondary diagnosis.
4. CY 2023 PDGM LUPA Thresholds and PDGM Case-Mix Weights
a. CY 2023 PDGM LUPA Thresholds
Under the HH PPS, LUPAs are paid when a certain visit threshold for
a payment group during a 30-day period of care is not met. In the CY
2019 HH PPS final rule with comment period (83 FR 56492), we finalized
setting the LUPA thresholds at the 10th percentile of visits or 2
visits, whichever is higher, for each payment group. This means the
LUPA threshold for each 30-day period of care varies depending on the
PDGM payment group to which it is assigned. If the LUPA threshold for
the payment group is met under the PDGM, the 30-day period of care will
be paid the full 30-day period case-mix adjusted payment amount
(subject to any PEP or outlier adjustments). If a 30-day period of care
does not meet the PDGM LUPA visit threshold, then payment will be made
using the CY 2023 per-visit payment amounts as described in section
II.B.5.c. of this final rule. For example, if the LUPA visit threshold
is four, and a 30-day period of care has four or more visits, it is
paid the full 30-day period payment amount; if the period of care has
three or less visits, payment is made using the per-visit payment
amounts.
In the CY 2019 HH PPS final rule with comment period (83 FR 56492),
we finalized our policy that the LUPA thresholds for each PDGM payment
group would be reevaluated every year based on the most current
utilization data available at the time of rulemaking. However, as CY
2020 was the first year of the new case-mix adjustment methodology, we
stated in the CY 2021 HH PPS final rule (85 FR 70305 through 70306)
that we would maintain the LUPA thresholds that were finalized and
shown in Table 17 of the CY 2020 HH PPS final rule with comment period
(84 FR 60522) for CY 2021 payment purposes. We stated that at that
time; we did not have sufficient CY 2020 data to reevaluate the LUPA
thresholds for CY 2021.
In the CY 2022 HH PPS final rule (86 FR 62249), we finalized the
proposal to recalibrate the PDGM case-mix weights, functional
impairment levels, and comorbidity subgroups while maintaining the LUPA
thresholds for CY 2022. We stated that because there are several
factors that contribute to how the case-mix weight is set for a
particular case-mix group (such as the number of visits, length of
visits, types of disciplines providing visits, and non-routine
supplies) and the case-mix weight is derived by comparing the average
resource use for the case-mix group relative to the average resource
use across all groups, we believe the COVID-19 PHE would have impacted
utilization within all case-mix groups similarly. Therefore, the impact
of any reduction in resource use caused by the COVID-19 PHE on the
calculation of the case-mix weight would be minimized since the impact
would be accounted for both in the numerator and denominator of the
formula used to calculate the case-mix weight. However, in contrast,
the LUPA thresholds are based on the number of overall visits in a
particular case-mix group (the threshold is the 10th percentile of
visits or 2 visits, whichever is greater) instead of a relative value
(like what is used to generate the case-mix weight) that would control
for the impacts of the PHE. We noted that visit patterns and some of
the decrease in overall visits in CY 2020 may not be representative of
visit patterns in CY 2022. Therefore, to mitigate any potential future
and significant short-term variability in the LUPA thresholds due to
the COVID-19 PHE, we finalized the proposal to maintain the LUPA
thresholds finalized and displayed in Table 17 in the CY 2020 HH PPS
final rule with comment period (84 FR 60522) for CY 2022 payment
purposes.
For CY 2023, we proposed to update the LUPA thresholds using CY
2021 Medicare home health claims (as of March 21, 2022) linked to OASIS
assessment data. After reviewing the CY 2021 home health claims
utilization data we determined that visit patterns have stabilized. Our
data analysis indicates that visits in 2021 were similar to visits in
2020. We believe that CY 2021 data will be more indicative of visit
patterns in CY 2023 rather than continuing to use the LUPA thresholds
derived from the CY 2018 data pre-PDGM. Therefore, we proposed to
update the LUPA thresholds for CY 2023 using data from CY 2021.
The final LUPA thresholds for the CY 2023 PDGM payment groups with
the corresponding Health Insurance Prospective Payment System (HIPPS)
codes and the case-mix weights are listed in Table B26. We solicited
public comments on the proposed updates to the LUPA thresholds for CY
2023. The public comments on our proposal to recalibrate the LUPA
thresholds for CY 2023 payment purposes and our responses are
summarized in this section of the rule.
Comment: A commenter expressed concern regarding the proposal to
recalibrate the LUPA thresholds using CY 2021 utilization data. This
commenter stated that while the observed changes in the recalibrated
thresholds may not seem large, they could serve as evidence that visits
during 2020 and 2021 may well be reduced (when compared to pre-PDGM
levels) due to pandemic influence.
Response: We acknowledge the commenter's statement and concerns
regarding the potential impact of the COVID-19 PHE on home health
utilization in CYs 2020 and 2021. However, we continue to believe that
it is important to base the LUPA thresholds on actual PDGM utilization
data and shift away from the use of data prior to the implementation of
the PDGM. Using the most recent data ensures that payment aligns with
the most recent cost of providing home health care services.
Comment: A commenter recommended that CMS reduce the LUPA threshold
in CY 2023 for all case-mix groups to two visits and reassess the
impact using CY 2023 data before making any further adjustments.
Response: We thank the commenter for this recommendation; however,
this recommendation is out of scope for the CY 2023 HH PPS proposed
rule. In the CY 2019 HH PPS final rule with comment period (83 FR
56492), we finalized setting the LUPA thresholds at the 10th percentile
of visits or 2 visits, whichever is higher, for each payment group. Any
changes to the LUPA threshold policy beyond the proposal to recalibrate
the thresholds using the CY 2021 utilization data would need to go
through notice and comment rulemaking.

[[Page 66816]]

Final Decision: We are finalizing the proposal to update the LUPA
thresholds for CY 2023. The LUPA thresholds for CY 2023 are located in
table 16 and will also be available on the HHA Center web page.
b. CY 2023 Functional Impairment Levels
Under the PDGM, the functional impairment level is determined by
responses to certain OASIS items associated with activities of daily
living and risk of hospitalization; that is, responses to OASIS items
M1800-M1860 and M1033. A home health period of care receives points
based on each of the responses associated with these functional OASIS
items, which are then converted into a table of points corresponding to
increased resource use. The sum of all of these points results in a
functional score which is used to group home health periods into a
functional level with similar resource use. That is, the higher the
points, the higher the response is associated with increased resource
use. The sum of all of these points results in a functional impairment
score which is used to group home health periods into one of three
functional impairment levels with similar resource use. The three
functional impairment levels of low, medium, and high were designed so
that approximately one-third of home health periods from each of the
clinical groups fall within each level. This means home health periods
in the low impairment level have responses for the functional OASIS
items that are associated with the lowest resource use, on average.
Home health periods in the high impairment level have responses for the
functional OASIS items that are associated with the highest resource
use on average.
For CY 2023, we proposed to use CY 2021 claims data to update the
functional points and functional impairment levels by clinical group.
The CY 2018 HH PPS proposed rule (82 FR 35320) and the technical report
from December 2016, posted on the Home Health PPS Archive web page
located at: <a href="https://www.cms.gov/medicare/home-health-pps/home-health-pps-archive">https://www.cms.gov/medicare/home-health-pps/home-health-pps-archive</a>, provide a more detailed explanation as to the construction
of these functional impairment levels using the OASIS items. We
proposed to use this same methodology previously finalized to update
the functional impairment levels for CY 2023. The updated OASIS
functional points table and the table of functional impairment levels
by clinical group for CY 2023 are listed in Tables 11 and 12,
respectively. We solicited public comments on the updates to functional
points and the functional impairment levels by clinical group.
BILLING CODE 4120-01-P

Table 11--Final Oasis Points Table for CY 2023

[[Page 66817]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.011

[[Page 66818]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.012

Table 12--Final Thresholds for Functional Levels by Clinical Group, for
CY 2023

[[Page 66819]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.013

[GRAPHIC] [TIFF OMITTED] TR04NO22.014

[[Page 66820]]

BILLING CODE 4120-01-C
Comment: Some commenters were concerned that changes caused by
recalibration were reducing resources to home health agencies.
Commenters argued that since the CY 2022 rates were recalibrated, it
should not be done again prior to the availability of the CY 2022 data.
Commenters were particularly concerned that changes to the functional
impairment points and thresholds did not account for the higher acuity
patients they have treated in recent years.
Response: It is important to note that recalibration is calculated
so that changes to case-mix and related items (for example, functional
points) are budget neutral. The adjustments made to functional points,
functional threshold levels, comorbidities, LUPA thresholds, and case-
mix weights are made so that after the application of the case-mix
budget neutrality factor, recalibration does not have any impact on
aggregate payments when using data from CY 2021. Recalibration ensures
there is variation in payment between the 432 case-mix groups so that
those groups with lower resource use get paid less than those with
higher resource use. If we did not adjust the functional points,
functional threshold levels, comorbidities, LUPA thresholds, and case-
mix weights to reflect resource utilization, then payments would be
less accurate. Specifically, if we did not account for changes in
functional points, we could potentially pay the same for the low
functional impairment patients and the high functional impairments
patients (who have more resources associated with their visits). If
that occurred, and since payment would be adjusted in a budget neutral
way, this could mean we would be overpaying for low functional
impairment and underpaying for high functional impairment.
Functional points, functional threshold levels, comorbidities, LUPA
thresholds and case-mix weights can be impacted even if there are no
changes in coding patterns but there are changes in resource use. In
the CY 2019 HH PPS final rule with comment period (83 FR 56486), we
stated that after implementation of the PDGM in CY 2020, we would
continue to analyze the impact of all of the PDGM case mix variables to
determine if any additional refinements need to made. We continue to
believe that updating the functional impairment levels using current
data ensures that all variables used as part of the overall case-mix
adjustment appropriately align home health payment with the actual cost
of providing home health care services. Performing a yearly
recalibration allows us to be as accurate and up-to-date as possible
when measuring relationship between resource use and functional points,
functional threshold levels, comorbidities, LUPA thresholds and case-
mix weights. The most recent year of data that we have is CY 2021. We
feel that relationships seen in the CY 2021 data are going to be more
similar to the relationships that we will eventually in see in CY 2023
data versus if we continued to use the relationships we see in the CY
2020 data. Commenters should note that although functional points did
decrease for many items, the functional thresholds also decreased
(meaning fewer points are needed to qualify for the higher functional
impairment levels).
Comment: Some commenters were concerned that CMS grouped patients
into one of three functional impairment levels even if it meant
potentially reducing resources to patients who previously would have
been classified as medium or high functional impairment.
Response: We remind commenters that the recalibration is
implemented in a budget neutral manner. We set the functional levels so
roughly a third of periods within each clinical group are assigned to
low, medium, and high. This is done to ensure that the case-mix system
pays appropriately for differences in functional impairment level. If
all 30-day periods ended up in one functional impairment level then
we'd be paying the same for the low functional impairment patients and
the high functional impairment patients (who have more resources
associated with their visits). We believe that the functional
impairment level adjustment adequately captures the level of functional
impairment based on patient characteristics reported on the OASIS. The
PDGM not only uses the same five OASIS items used under the previous HH
PPS to determine the functional case-mix adjustment (M1810, M1820,
M1830, M1830, M1850, and M1860), but also adds two additional OASIS
items (M1800 and M1033) to determine the level of functional
impairment. The structure of categorizing functional impairment into
low, medium, and high levels has been part of the home health payment
structure since the implementation of the HH PPS. The previous HH PPS
grouped home health episodes using functional scores based on
functional OASIS items with similar average resource use within the
same functional level, with approximately a third of episodes
classified as low functional score, a third of episodes classified as
medium functional score, and a third of episodes classified as high
functional score. Likewise, the PDGM groups home health periods of care
using functional impairment scores based on functional OASIS items with
similar resource use and has three levels of functional impairment
severity: low, medium, and high. However, the PDGM differs from the
current HH PPS functional variable in that the three functional
impairment level thresholds in the PDGM vary between the clinical
groups. The PDGM functional impairment level structure accounts for the
patient characteristics within that clinical group associated with
increased resource costs affected by functional impairment. This is to
further ensure that payment is more accurately aligned with actual
patient characteristics and resource needs.
Comment: A commenter indicated that Table B21 in the CY 2023 HH PPS
proposed rule (87 FR 37627) showed that a lower functional impairment
response was associated with more points than a higher functional
impairment response (M1860 responses 2 and 3).
Response: For recalibration, we use the data as they are submitted.
Home health agencies should consider the appropriateness of their OASIS
responses in relation to the level of resources that should be required
for certain functional impairments. CMS would expect to find, on
average, that patients who are more functionally impaired would have
higher resource use. However, as noted by the commenter, this
correlation does not always occur when looking at individual OASIS
items and responses.
Final Decision: We are finalizing to update the functional points
and functional impairment levels for CY 2023 as proposed, using CY 2021
claims data. Table 11 includes the final functional points based on the
most available data.
c. CY 2023 Comorbidity Subgroups
Thirty-day periods of care receive a comorbidity adjustment
category based on the presence of certain secondary diagnoses reported
on home health claims. These diagnoses are based on a home-health
specific list of clinically and statistically significant secondary
diagnosis subgroups with similar resource use, meaning the diagnoses
have at least as high as the median resource use and are reported in
more than 0.1 percent of 30-day periods of care. Home health 30-day
periods of care can receive a comorbidity adjustment under the
following circumstances:

[[Page 66821]]

<bullet> Low comorbidity adjustment: There is a reported secondary
diagnosis on the home health-specific comorbidity subgroup list that is
associated with higher resource use.
<bullet> High comorbidity adjustment: There are two or more
secondary diagnoses on the home health-specific comorbidity subgroup
interaction list that are associated with higher resource use when both
are reported together compared to when they are reported separately.
That is, the two diagnoses may interact with one another, resulting in
higher resource use.
<bullet> No comorbidity adjustment: A 30-day period of care
receives no comorbidity adjustment if no secondary diagnoses exist or
do not meet the criteria for a low or high comorbidity adjustment.
In the CY 2019 HH PPS final rule with comment period (83 FR 56406),
we stated that we would continue to examine the relationship of
reported comorbidities on resource utilization and make the appropriate
payment refinements to help ensure that payment is in alignment with
the actual costs of providing care. For CY 2023, we proposed to use the
same methodology used to establish the comorbidity subgroups to update
the comorbidity subgroups using CY 2021 home health data.
For CY 2023, we proposed to update the comorbidity subgroups to
include 23 low comorbidity adjustment subgroups and 94 high comorbidity
adjustment interaction subgroups. The final update to the comorbidity
adjustment subgroups includes 22 low comorbidity adjustment subgroups
as identified in table 13 and 91 high comorbidity adjustment
interaction subgroups as identified in table 14. The final 22 low
comorbidity adjustment subgroups and 91 high comorbidity adjustment
interactions reflect the final coding changes detailed in section
II.B.3.c. of this final rule. The final CY 2023 low comorbidity
adjustment subgroups and the high comorbidity adjustment interaction
subgroups including those diagnoses within each of these comorbidity
adjustments will also be posted on the HHA Center web page at <a href="https://www.cms.gov/Center/Provider-Type/Home-Health-Agency-HHA-Center">https://www.cms.gov/Center/Provider-Type/Home-Health-Agency-HHA-Center</a>.
We invited comments on the proposed updates to the low comorbidity
adjustment subgroups and the high comorbidity adjustment interactions
for CY 2023.
BILLING CODE 4120-01-P

Table 13--Low Comorbidity Adjustment Subgroups for CY 2023
[GRAPHIC] [TIFF OMITTED] TR04NO22.015

[[Page 66822]]

Table 14--High Comorbidity Adjustment Interactions for CY 2023
[GRAPHIC] [TIFF OMITTED] TR04NO22.016

[[Page 66823]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.017

[[Page 66824]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.018

[[Page 66825]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.019

[[Page 66826]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.020

[[Page 66827]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.021

[[Page 66828]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.022

BILLING CODE 4120-01-C
Comment: A commenter expressed support for the proposed updates to
the low and high comorbidity subgroups. This commenter stated that the
changes achieve the stated goal of ensuring that payment is in
alignment with the actual costs of providing care and that the high
comorbidity adjustment interaction subgroups acknowledge the impact of
multiple diagnoses on care delivery complexity and cost.
Response: We thank the commenter for their support.
Final Decision: We are finalizing the proposal to use the same
methodology used to establish the comorbidity subgroups to update the
comorbidity subgroups using CY 2021 home health data. For CY 2023, the
final update to the comorbidity adjustment subgroups includes 22 low
comorbidity adjustment subgroups as identified in Table 13 and 91 high
comorbidity adjustment interaction subgroups as identified in Table 14.
The final 22 low comorbidity adjustment subgroups and 91 high
comorbidity adjustment interactions reflect the final coding changes
detailed in section II.B.3.c. of this final rule.
d. CY 2023 PDGM Case-Mix Weights
As finalized in the CY 2019 HH PPS final rule with comment period
(83 FR 56502), the PDGM places patients into meaningful payment
categories based on patient and other characteristics, such as timing,
admission source, clinical grouping using the reported principal
diagnosis, functional impairment level, and comorbid conditions. The
PDGM case-mix methodology results in 432 unique case-mix groups called
HHRGs. We also finalized a policy in the CY 2019 HH PPS final rule with
comment period (83 FR 56515) to recalibrate annually the PDGM case-mix
weights using a fixed effects model, as outlined in that rule, with the
most recent and complete utilization data available at the time of
annual rulemaking. Annual recalibration of the PDGM case-mix weights
ensures that the case-mix weights reflect, as accurately as possible,
current home health resource use and changes in utilization patterns.
To generate the proposed recalibrated CY 2023 case-mix weights, we used
CY 2021 home health claims data with linked OASIS data (as of March 21,
2021). These data are the most current and complete data available at
this time. We believe that recalibrating the case-mix weights using
data from CY 2021 would be reflective of PDGM utilization and patient
resource use for CY 2023. The proposed recalibrated case-mix weights
were updated based on more complete CY 2021 claims data for this final
rule.
The claims data provide visit-level data and data on whether non-
routine supplies (NRS) were provided during the period and the total
charges of NRS. We determine the case-mix weight for each of the 432
different PDGM payment groups by regressing resource use on a series of
indicator variables for each of the categories using a fixed effects
model as described in the following steps:
Step 1: Estimate a regression model to assign a functional
impairment level to each 30-day period. The regression model estimates
the relationship between a 30-day period's resource use and the
functional status and risk of hospitalization items included in the
PDGM, which are obtained from certain OASIS items. We refer readers to
Table B21 for further information on the OASIS items used for the
functional impairment level under the PDGM. We measure resource use
with the cost-per-minute + NRS approach that uses

[[Page 66829]]

information from 2020 home health cost reports. We use 2020 home health
cost report data because it is the most complete cost report data
available at the time of rulemaking. Other variables in the regression
model include the 30-day period's admission source, clinical group, and
30-day period timing. We also include home health agency level fixed
effects in the regression model. After estimating the regression model
using 30-day periods, we divide the coefficients that correspond to the
functional status and risk of hospitalization items by 10 and round to
the nearest whole number. Those rounded numbers are used to compute a
functional score for each 30-day period by summing together the rounded
numbers for the functional status and risk of hospitalization items
that are applicable to each 30-day period. Next, each 30-day period is
assigned to a functional impairment level (low, medium, or high)
depending on the 30-day period's total functional score. Each clinical
group has a separate set of functional thresholds used to assign 30-day
periods into a low, medium or high functional impairment level. We set
those thresholds so that we assign roughly a third of 30-day periods
within each clinical group to each functional impairment level (low,
medium, or high).
Step 2: A second regression model estimates the relationship
between a 30-day period's resource use and indicator variables for the
presence of any of the comorbidities and comorbidity interactions that
were originally examined for inclusion in the PDGM. Like the first
regression model, this model also includes home health agency level
fixed effects and includes control variables for each 30-day period's
admission source, clinical group, timing, and functional impairment
level. After we estimate the model, we assign comorbidities to the low
comorbidity adjustment if any comorbidities have a coefficient that is
statistically significant (p-value of 0.05 or less) and which have a
coefficient that is larger than the 50th percentile of positive and
statistically significant comorbidity coefficients. If two
comorbidities in the model and their interaction term have coefficients
that sum together to exceed $150 and the interaction term is
statistically significant (p-value of 0.05 or less), we assign the two
comorbidities together to the high comorbidity adjustment.
Step 3: After Step 2, each 30-day period is assigned to a clinical
group, admission source category, episode timing category, functional
impairment level, and comorbidity adjustment category. For each
combination of those variables (which represent the 432 different
payment groups that comprise the PDGM), we then calculate the 10th
percentile of visits across all 30-day periods within a particular
payment group. If a 30-day period's number of visits is less than the
10th percentile for their payment group, the 30-day period is
classified as a Low Utilization Payment Adjustment (LUPA). If a payment
group has a 10th percentile of visits that is less than two, we set the
LUPA threshold for that payment group to be equal to two. That means if
a 30- day period has one visit, it is classified as a LUPA and if it
has two or more visits, it is not classified as a LUPA.
Step 4: Take all non-LUPA 30-day periods and regress resource use
on the 30-day period's clinical group, admission source category,
episode timing category, functional impairment level, and comorbidity
adjustment category. The regression includes fixed effects at the level
of the home health agency. After we estimate the model, the model
coefficients are used to predict each 30-day period's resource use. To
create the case-mix weight for each 30- day period, the predicted
resource use is divided by the overall resource use of the 30-day
periods used to estimate the regression.
The case-mix weight is then used to adjust the base payment rate to
determine each 30-day period's payment. Table 15 shows the coefficients
of the payment regression used to generate the weights, and the
coefficients divided by average resource use.
BILLING CODE 4120-01-P

Table 15--Coefficient of Payment Regression and Coefficient Divided by
Average Resource Use

[[Page 66830]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.023

[[Page 66831]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.024

The case-mix weights proposed for CY 2023 are listed in Table 16
and will also be posted on the HHA Center web page \25\ upon display of
this final rule.
---------------------------------------------------------------------------

\25\ HHA Center web page: <a href="https://www.cms.gov/Center/Provider-Type/Home-Health-Agency-HHA-Center">https://www.cms.gov/Center/Provider-Type/Home-Health-Agency-HHA-Center</a>.
---------------------------------------------------------------------------

Table 16--Final Case-Mix Weights and LUPA Thresholds for Each HHRG
Payment Group

[[Page 66832]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.025

[[Page 66833]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.026

[[Page 66834]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.027

[[Page 66835]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.028

[[Page 66836]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.029

[[Page 66837]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.030

[[Page 66838]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.031

[[Page 66839]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.032

[[Page 66840]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.033

[[Page 66841]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.034

[[Page 66842]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.035

[[Page 66843]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.036

[[Page 66844]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.037

[[Page 66845]]

[GRAPHIC] [TIFF OMITTED] TR04NO22.038

BILLING CODE 4120-01-C
Changes to the PDGM case-mix weights are implemented in a budget
neutral manner by multiplying the CY 2023 national standardized 30-day
period payment rate by a case-mix budget neutrality factor. Typically,
the case-mix weight budget neutrality factor is also calculated using
the most recent, complete home health claims data available. However,
in the CY 2022 HH PPS proposed rule (86 FR 35908), due to the COVID-19
PHE, we discussed using the previous calendar year's home health claims
data (CY 2019) to determine if there were significant differences
between utilizing CY 2019 and CY 2020 claims data. We noted that CY
2020 was the first year of actual PDGM utilization data, therefore, if
we were to use CY 2019 data due to the COVID-19 PHE we would need to
simulate 30-day periods from 60-day episodes under the old system. We
determined that using CY 2020 utilization data was more appropriate
than using CY 2019 utilization data, as it is actual PDGM utilization
data. For CY 2023, we will continue the practice of using the most
recent complete home health claims data at the time of rulemaking,
which is CY 2021 data. The case-mix budget neutrality factor is
calculated as the ratio of 30-day base payment rates such that total
payments when the CY 2023 PDGM case-mix

[[Page 66846]]

weights (developed using CY 2021 home health claims data) are applied
to CY 2021 utilization (claims) data are equal to total payments when
CY 2022 PDGM case-mix weights (developed using CY 2020 home health
claims data) are applied to CY 2021 utilization data. This produces a
case-mix budget neutrality factor for CY 2023 of 0.9904.
We invited comments on the CY 2023 proposed case-mix weights and
proposed case-mix weight budget neutrality factor and these are
summarized below.
Comment: A few commenters expressed support for the proposal to
recalibrate the PDGM case-mix weights for CY 2023 using CY 2021
utilization data.
Response: We thank the commenters for their support.
Comment: Several commenters were opposed to the proposal to
recalibrate the PDGM case-mix weights for CY 2023. A commenter
expressed concerns about the influence of the COVID-19 surges and its
overall effects on the types of patients being served. This commenter
recommended not updating the case-mix weights at this time and resuming
this practice once the pandemic is over.
Response: CMS appreciates the comments received regarding CY 2021
utilization trends and the impact of the COVID-19 PHE on the provision
of home health services. We recognize that commenters have concerns
regarding how the COVID-19 PHE affected the type of home health
patients served as well as care practices. However, as stated in the CY
2023 HH PPS proposed rule (87 FR 37626), we believe that visit patterns
have stabilized as our data analysis indicates that visits in 2021 were
similar to visits in 2020. As such, we believe that CY 2021 data will
be indicative of visit patterns in CY 2023. In the CY 2019 HH PPS final
rule, we finalized our proposal to annually recalibrate the PDGM case-
mix weights (83 FR 56515) to reflect the most recent utilization data
available at the time of rulemaking. We continue to believe that the
annual recalibration of the HH PPS case-mix weights ensures that the
case-mix weights reflect, as accurately as possible, current home
health resource use, changes in utilization patterns, and reflects the
types of patients currently receiving home health services. We believe
that prolonging recalibration could lead to more significant variation
in the case-mix weights than what is observed using CY 2021 utilization
data. Therefore, we believe that utilizing CY 2021 data to recalibrate
the CY 2023 case-mix weights is appropriate.
Comment: A commenter recommended that any recalibration should be
done in a non-budget-neutral manner given the higher-acuity patients,
increasing expenses, increased demand for care, and increased shortage
of labor.
Response: We thank the commenter for this recommendation; however,
consistent with our established policy, we apply a case-mix budget
neutrality factor to the CY 2023 national, standardized 30-day period
payment rate to ensure that there are no changes in aggregate payments
due to the recalibration.
Final Decision: We are finalizing the recalibration of the HH PPS
case-mix weights as proposed for CY 2023. We are also finalizing the
proposal to implement the changes to the PDGM case-mix weights in a
budget neutral manner by applying a case-mix budget neutrality factor
to the CY 2023 national, standardized 30-day period payment rate. As
stated previously, the final case-mix budget neutrality factor for CY
2023 will be 0.9904.
5. CY 2023 Home Health Payment Rate Updates
a. CY 2023 Home Health Market Basket Update for HHAs
Section 1895(b)(3)(B) of the Act requires that the standard
prospective payment amounts for home health be increased by a factor
equal to the applicable home health market basket update for those HHAs
that submit quality data as required by the Secretary. In the CY 2019
HH PPS final rule with comment period (83 FR 56425), we finalized a
rebasing of the home health market basket to reflect 2016 cost report
data. A detailed description of how we rebased the home health market
basket is available in the CY 2019 HH PPS final rule with comment
period (83 FR 56425 through 56436).
Section 1895(b)(3)(B) of the Act requires that in CY 2015 and in
subsequent calendar years, except CY 2018 (under section 411(c) of the
Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) (Pub. L.
114-10, enacted April 16, 2015)), and CY 2020 (under section 53110 of
the Bipartisan Budget Act of 2018 (BBA) (Pub. L. 115-123, enacted
February 9, 2018)), the market basket percentage under the HHA
prospective payment system, as described in section 1895(b)(3)(B) of
the Act, be annually adjusted by changes in economy-wide productivity.
Section 1886(b)(3)(B)(xi)(II) of the Act defines the productivity
adjustment to be equal to the 10-year moving average of changes in
annual economy-wide private nonfarm business multifactor productivity
(MFP) (as projected by the Secretary for the 10-year period ending with
the applicable fiscal year, calendar year, cost reporting period, or
other annual period). The United States Department of Labor's Bureau of
Labor Statistics (BLS) publishes the official measures of productivity
for the United States economy. We note that previously the productivity
measure referenced in section 1886(b)(3)(B)(xi)(II) was published by
BLS as private nonfarm business multifactor productivity. Beginning
with the November 18, 2021 release of productivity data, BLS replaced
the term ``multifactor productivity'' with ``

[…truncated; see source link]

View on FederalRegister.gov →Plain-text source

Indexed from Federal Register on November 4, 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.