Medicare Program; Home Health Prospective Payment System Rate Update for Calendar Year 2012, 40988-41031 [2011-16938]

Agencies

• Centers for Medicare & Medicaid Services
• DEPARTMENT OF HEALTH AND HUMAN SERVICES

[Federal Register Volume 76, Number 133 (Tuesday, July 12, 2011)]
[Proposed Rules]
[Pages 40988-41031]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: 2011-16938]



[[Page 40987]]

Vol. 76

Tuesday,

No. 133

July 12, 2011

Part II





Department of Health and Human Services





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





Centers for Medicare & Medicaid Services





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





42 CFR Parts 409, 424, 440, et al.





Medicare Program; Home Health Prospective Payment System Rate Update 
for Calendar Year 2012; Face-to-Face Requirements for Home Health 
Services; Policy Changes and Clarifications Related to Home Health; 
Proposed Rules

Federal Register / Vol. 76 , No. 133 / Tuesday, July 12, 2011 / 
Proposed Rules

[[Page 40988]]


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

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Centers for Medicare & Medicaid Services

42 CFR Parts 409, 424, and 484

[CMS-1353-P]
RIN 0938-AQ30


Medicare Program; Home Health Prospective Payment System Rate 
Update for Calendar Year 2012

AGENCY: Centers for Medicare & Medicaid Services (CMS), HHS.

ACTION: Proposed rule.

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

SUMMARY: This proposed rule would update the Home Health Prospective 
Payment System (HH PPS) rates, including: The national standardized 60-
day episode rates, the national per-visit rates, the low utilization 
payment amount (LUPA), and outlier payments under the Medicare 
prospective payment system for home health agencies effective January 
1, 2012.

DATES: To be assured consideration, comments must be received at one of 
the addresses provided below, no later than 5 p.m. on September 6, 
2011.

ADDRESSES: In commenting, please refer to file code CMS-1353-P. Because 
of staff and resource limitations, we cannot accept comments by 
facsimile (FAX) transmission.
    You may submit comments in one of four ways (please choose only one 
of the ways listed):
    1. Electronically. You may submit electronic comments on this 
regulation to https://www.regulations.gov. Follow the instructions under 
the ``More Search Options'' tab.
    2. By regular mail. You may mail written comments to the following 
address only: Centers for Medicare & Medicaid Services, Department of 
Health and Human Services, Attention: CMS-1353-P, P.O. Box 8016, 
Baltimore, MD 21244-8016.
    Please allow sufficient time for mailed comments to be received 
before the close of the comment period.
    3. By express or overnight mail. You may send written comments to 
the following address only: Centers for Medicare & Medicaid Services, 
Department of Health and Human Services, Attention: CMS-1353-P, Mail 
Stop C4-26-05, 7500 Security Boulevard, Baltimore, MD 21244-1850.

    4. By hand or courier. If you prefer, you may deliver (by hand or 
courier) your written comments before the close of the comment period 
to either of the following addresses:
    a. For delivery in Washington, DC--Centers for Medicare & Medicaid 
Services, Department of Health and Human Services, Room 445-G, Hubert 
H. Humphrey Building, 200 Independence Avenue, SW., Washington, DC 
20201.

    (Because access to the interior of the Hubert H. Humphrey Building 
is not readily available to persons without Federal government 
identification, commenters are encouraged to leave their comments in 
the CMS drop slots located in the main lobby of the building. A stamp-
in clock is available for persons wishing to retain a proof of filing 
by stamping in and retaining an extra copy of the comments being 
filed.).
    b. For delivery in Baltimore, MD--Centers for Medicare & Medicaid 
Services, Department of Health and Human Services, 7500 Security 
Boulevard, Baltimore, MD 21244-1850.

    If you intend to deliver your comments to the Baltimore address, 
please call (410) 786-7195 in advance to schedule your arrival with one 
of our staff members.
    Comments mailed to the addresses indicated as appropriate for hand 
or courier delivery may be delayed and received after the comment 
period.
    For information on viewing public comments, see the beginning of 
the SUPPLEMENTARY INFORMATION section.

FOR FURTHER INFORMATION CONTACT:

Elizabeth Goldstein, (410) 786-6665, for CAHPS issues.
Mary Pratt, (410) 786-6867, for quality issues.
Randy Throndset, (410)786-0131 (overall HH PPS).


SUPPLEMENTARY INFORMATION:
    Inspection of Public Comments: All comments received before the 
close of the comment period are available for viewing by the public, 
including any personally identifiable or confidential business 
information that is included in a comment. We post all comments 
received before the close of the comment period on the following Web 
site as soon as possible after they have been received: https://www.regulations.gov. Follow the search instructions on that Web site to 
view public comments.
    Comments received timely will also be available for public 
inspection as they are received, generally beginning approximately 3 
weeks after publication of a document, at the headquarters of the 
Centers for Medicare & Medicaid Services, 7500 Security Boulevard, 
Baltimore, Maryland 21244, Monday through Friday of each week from 8:30 
a.m. to 4 p.m. EST. To schedule an appointment to view public comments, 
phone 1-800-743-3951.

Table of Contents

I. Background
    A. Statutory Background
    B. System for Payment of Home Health Services
    C. Updates to the HH PPS
II. Provisions of the Proposed Rule
    A. Case-Mix Measurement
    1. Independent Review of the Models To Assess Nominal Case-Mix 
Growth
    2. Revised Version of Our Models To Assess Nominal Case-Mix 
Growth
    B. Case-Mix Revision to the Case-Mix Weights
    1. Hypertension Diagnosis Coding Under the HH PPS
    2. Proposal for Revision of Case-Mix Weights
    C. Outlier Policy
    1. Background
    2. Regulatory Update
    3. Statutory Update
    4. Loss-Sharing Ratio and Fixed Dollar Loss (FDL) Ratio
    5. Outlier Relationship to the HH Payment Study
    D. CY 2012 Rate Update
    1. Home Health Market Basket Update
    2. Home Health Care Quality Improvement
    a. Background and Quality Reporting Requirements
    b. OASIS Data
    c. Claims Data, Proposed Requirements and Outcome Measure Change
    d. Home Health Care CAHPS Survey (HHCAHPS)
    3. Home Health Wage Index
    4. Proposed CY 2012 Annual Payment Update
    a. National Standardized 60-Day Episode Rate
    b. Proposed Updated CY 2012 National Standardized 60-Day Episode 
Payment Rate
    c. National Per-Visit Rates Used To Pay LUPAs and Compute 
Imputed Costs Used in Outlier Calculations
    d. LUPA Add-on Payment Amount Update
    e. Nonroutine Medical Supply Conversion Factor Update
    5. Rural Add-On
    E. Therapy Corrections and Clarification
    F. Home Health Face-to-Face Encounter
    G. Payment Reform: Home Health Study and Report
    H. International Classification of Diseases 10th Edition (ICD-
10) Coding
    I. Clarification to Benefit Policy Manual Language on ``Confined 
to the Home'' Definition
III. Collection of Information Requirements
IV. Response to Comments
V. Regulatory Impact Analysis
VI. Federalism Analysis
Regulations Text

Acronyms

    In addition, because of the many terms to which we refer by 
abbreviation in this proposed rule, we are listing these abbreviations 
and their

[[Page 40989]]

corresponding terms in alphabetical order below:

ACH LOS Acute Care Hospital Length of Stay
ADL Activities of Daily Living
APU Annual Payment Update
BBA Balanced Budget Act of 1997, Public Law 105-33
BBRA Medicare, Medicaid, and SCHIP Balanced Budget Refinement Act of 
1999, Public Law 106-113
CAD Coronary Artery Disease
CAH Critical Access Hospital
CBSA Core-Based Statistical Area
CHF Congestive Heart Failure
CMI Case-Mix Index
CMS Centers for Medicare and Medicaid Services
CoPs Conditions of Participation
COPD Chronic Obstructive Pulmonary Disease
CVD Cardiovascular Disease
DM Diabetes Mellitus
DRA Deficit Reduction Act of 2005, Public Law 109-171, enacted 
February 8, 2006
FDL Fixed Dollar Loss
FI Fiscal Intermediaries
FR Federal Register
FY Fiscal Year
HCC Hierarchical Condition Categories
HCIS Health Care Information System
HHCAHPS Home Health Care Consumer Assessment of Healthcare Providers 
and Systems Survey
HH PPS Home Health Prospective Payment System
HHAs Home Health Agencies
HHRG Home Health Resource Group
HIPPS Health Insurance Prospective Payment System
IH Inpatient Hospitalization
IRF Inpatient Rehabilitation Facility
LTCH Long-Term Care Hospital
LUPA Low Utilization Payment Amount
MEPS Medical Expenditures Panel Survey
MMA Medicare Prescription Drug, Improvement, and Modernization Act 
of 2003, Public Law 108-173, enacted December 8, 2003
MSA Metropolitan Statistical Areas
MSS Medical Social Services
NRS Non-Routine Supplies
OBRA Omnibus Reconciliation Act of 1981, Public Law 97-35, enacted 
August 13, 1981
OCESAA Omnibus Consolidated and Emergency Supplemental 
Appropriations Act, Public Law 105-277, enacted October 21, 1998
OES Occupational Employment Statistics
OIG Office of Inspector General
OT Occupational Therapy
OMB Office of Management and Budget
PAC-PRD Post-Acute Care Payment Reform Demonstration
PEP Partial Episode Payment Adjustment
PT Physical Therapy
QAP Quality Assurance Plan
PRRB Provider Reimbursement Review Board
RAP Request for Anticipated Payment
RF Renal Failure
RFA Regulatory Flexibility Act, Public Law 96-354
RHHIs Regional Home Health Intermediaries
RIA Regulatory Impact Analysis
SLP Speech Language Pathology Therapy
SNF Skilled Nursing Facility
UMRA Unfunded Mandates Reform Act of 1995

I. Background

A. Statutory Background

    The Balanced Budget Act of 1997 (BBA) (Pub. L. 105-33, enacted 
August 5, 1997), significantly changed the way Medicare pays for 
Medicare home health (HH) services. Section 4603 of the BBA mandated 
the development of the home health prospective payment system (HH PPS). 
Until the implementation of a HH PPS on October 1, 2000, home health 
agencies (HHAs) received payment under a retrospective reimbursement 
system.
    Section 4603(a) of the BBA mandated the development of a HH PPS for 
all Medicare-covered HH services provided under a plan of care (POC) 
that were paid on a reasonable cost basis by adding section 1895 of the 
Social Security Act (the Act), entitled ``Prospective Payment For Home 
Health Services''. Section 1895(b)(1) of the Act requires the Secretary 
to establish a HH PPS for all costs of HH services paid under Medicare.
    Section 1895(b)(3)(A) of the Act requires the following: (1) The 
computation of a standard prospective payment amount include all costs 
for HH services covered and paid for on a reasonable cost basis and 
that such amounts be initially based on the most recent audited cost 
report data available to the Secretary; and (2) the standardized 
prospective payment amount be adjusted to account for the effects of 
case-mix and wage levels among HHAs.
    Section 1895(b)(3)(B) of the Act addresses the annual update to the 
standard prospective payment amounts by the HH applicable percentage 
increase. Section 1895(b)(4) of the Act governs the payment 
computation. Sections 1895(b)(4)(A)(i) and (b)(4)(A)(ii) of the Act 
require the standard prospective payment amount to be adjusted for 
case-mix and geographic differences in wage levels. Section 
1895(b)(4)(B) of the Act requires the establishment of an appropriate 
case-mix change adjustment factor for significant variation in costs 
among different units of services.
    Similarly, section 1895(b)(4)(C) of the Act requires the 
establishment of wage adjustment factors that reflect the relative 
level of wages, and wage-related costs applicable to HH services 
furnished in a geographic area compared to the applicable national 
average level. Under section 1895(b)(4)(c) of the Act, the wage-
adjustment factors used by the Secretary may be the factors used under 
section 1886(d)(3)(E) of the Act.
    Section 1895(b)(5) of the Act gives the Secretary the option to 
make additions or adjustments to the payment amount otherwise paid in 
the case of outliers because of unusual variations in the type or 
amount of medically necessary care. Section 3131(b) of the Patient 
Protection and Affordable Care Act of 2010 (the Affordable Care Act) 
(Pub. L. 111-148, enacted March 23, 2010) revised section 1895(b)(5) of 
the Act so that total outlier payments in a given fiscal year (FY) or 
year may not exceed 2.5 percent of total payments projected or 
estimated. The provision also makes permanent a 10 percent agency level 
outlier payment cap.
    In accordance with the statute, as amended by the BBA, we published 
a final rule in the July 3, 2000 Federal Register (65 FR 41128) to 
implement the HH PPS legislation. The July 2000 final rule established 
requirements for the new HH PPS for HH services as required by section 
4603 of the BBA, as subsequently amended by section 5101 of the Omnibus 
Consolidated and Emergency Supplemental Appropriations Act (OCESAA) for 
Fiscal Year 1999, (Pub. L. 105-277, enacted October 21, 1998); and by 
sections 302, 305, and 306 of the Medicare, Medicaid, and SCHIP 
Balanced Budget Refinement Act (BBRA) of 1999, (Pub. L. 106-113, 
enacted November 29, 1999). The requirements include the implementation 
of a HH PPS for HH services, consolidated billing requirements, and a 
number of other related changes. The HH PPS described in that rule 
replaced the retrospective reasonable cost-based system that was used 
by Medicare for the payment of HH services under Part A and Part B. For 
a complete and full description of the HH PPS as required by the BBA, 
see the July 2000 HH PPS final rule (65 FR 41128 through 41214).
    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 HHAs to submit data for purposes of measuring 
health care quality, and links the quality data submission to the 
annual applicable percentage increase. This data submission requirement 
is applicable for CY 2007 and each subsequent year. If an HHA does not 
submit quality data, the HH market basket percentage increase is 
reduced 2 percentage points. In the November 9, 2006 Federal Register 
(71 FR 65884, 65935), we published a final rule to implement the

[[Page 40990]]

pay-for-reporting requirement of the DRA, which was codified at Sec.  
484.225(h) and (i) in accordance with the statute.
    Section 421(a) of the Medicare Prescription Drug, Improvement, and 
Modernization Act of 2003 (MMA) (Pub. L. 108-173, enacted December 8, 
2003) provides an increase of 3 percent of the payment amount otherwise 
made under section 1886(d)(2)(D) of the Act for HH services furnished 
in a rural area with respect to episodes and visits ending on or after 
April 1, 2010, and before January 1, 2016.

B. System for Payment of Home Health Services

    Generally, Medicare makes payment under the HH PPS on the basis of 
a national standardized 60-day episode payment rate that is adjusted 
for the applicable case-mix and wage index. The national standardized 
60-day episode rate includes the six HH disciplines (skilled nursing, 
HH aide, physical therapy, speech-language pathology, occupational 
therapy, and medical social services). Payment for non-routine medical 
supplies (NRS), is no longer part of the national standardized 60-day 
episode rate and is computed by multiplying the relative weight for a 
particular NRS severity level by the NRS conversion factor (See section 
II.D.4.e). Payment for durable medical equipment covered under the HH 
benefit is made outside the HH PPS payment system. To adjust for case-
mix, the HH PPS uses a 153-category case-mix classification to assign 
patients to a home health resource group (HHRG). The clinical severity 
level, functional severity level, and service utilization are computed 
from responses to selected data elements in the OASIS assessment 
instrument and are used to place the patient in a particular HHRG. Each 
HHRG has an associated case-mix weight which is used in calculating the 
payment for an episode.
    For episodes with four or fewer visits, Medicare pays based on a 
national per-visit rate, adjusted by the discipline(s) providing the 
services; an episode consisting of four or fewer visits within a 60-day 
period receives what is referred to as a low utilization payment 
adjustment (LUPA). Medicare also adjusts the national standardized 60-
day episode payment rate for certain intervening events that are 
subject to a partial episode payment adjustment (PEP adjustment). For 
certain cases that exceed a specific cost threshold, an outlier 
adjustment may also be available.

C. Updates to the HH PPS

    As required by section 1895(b)(3)(B) of the Act, we have 
historically updated the HH PPS rates annually in the Federal Register. 
The August 29, 2007 final rule with comment period set forth an update 
to the 60-day national episode rates and the national per-visit rates 
under the Medicare prospective payment system for HHAs for CY 2008. The 
CY 2008 rule included an analysis performed on CY 2005 HH claims data, 
which indicated a 12.78 percent increase in the observed case-mix since 
2000. The case-mix represented the variations in conditions of the 
patient population served by the HHAs. Subsequently, a more detailed 
analysis was performed on the 12.78 percent increase in case-mix to 
evaluate if any portion of the increase was associated with a change in 
the actual clinical condition of HH patients. We examined data on 
demographics, family severity, and non-HH Part A Medicare expenditures 
to predict the average case-mix weight for 2005. We identified 8.03 
percent of the total case-mix change as real and decreased the 12.78 
percent of total case-mix change by 8.03 percent to get a final nominal 
case-mix increase measure of 11.75 percent (0.1278 * (1-0.0803) = 
0.1175).
    To account for the changes in case-mix that were not related to an 
underlying change in patient health status, we implemented a reduction 
over 4 years in the national standardized 60-day episode payment rates 
and the NRS conversion factor. That reduction was to be 2.75 percent 
per year for 3 years beginning in CY 2008 and 2.71 percent for the 
fourth year in CY 2011.
    For CY 2011, we published the November 17, 2010 final rule (75 FR 
70372) (hereinafter referred to as the CY 2011 HH PPS final rule) that 
set forth the update to the 60-day national episode rates and the 
national per-visit rates under the Medicare prospective payment system 
for HH services.
    As discussed in the CY 2011 rule, our analysis indicated that there 
was a 19.40 percent increase in overall case-mix from 2000 to 2008 and 
that only 10.07 percent of that overall observed case-mix percentage 
increase was due to real case-mix change. As a result of our analysis, 
we identified a 17.45 percent nominal increase in case-mix. To fully 
account for the 17.45 percent nominal case-mix growth which was 
identified from 2000 to 2008, we proposed 3.79 percent payment 
reductions in both CY 2011 and CY 2012. However, we deferred finalizing 
a payment reduction for CY 2012 until a further study of the case-mix 
data was completed. Independent review of the case-mix model has been 
conducted and the results are discussed in section II.A. of this 
proposed rule.

II. Provisions of the Proposed Rule

A. Case-Mix Measurement

    Every year, since the HH PPS CY 2008 proposed rule, we have stated 
in HH PPS rulemaking that we would continue to monitor case-mix changes 
in the HH PPS and to update our analysis to measure change in case-mix, 
both real changes in case-mix and changes which are unrelated to 
changes in patient acuity (nominal). We have continued to monitor case-
mix changes, and our latest analysis continues to support the need to 
make payment adjustments to account for nominal case-mix growth.
    Before measuring nominal case-mix growth, we examined the total 
case-mix growth every year from 2000 to 2009. Our latest analysis 
indicates that there was a large 1-year increase, 2.6 percent, in the 
average case-mix weight from 2008 to 2009. Specifically, the 2008 
average case-mix was 1.3095 and the 2009 average case-mix was 1.3435. 
It should be noted that the average case-mix for 2008 is slightly 
different than the average case-mix for 2008 that was reported in the 
CY 2011 HH PPS final rule. The difference in case-mix is due to the 
increased availability of data and inclusion of more episodes in the 
2008 sample. As we did last year, we sought to describe how much of the 
1-year change was due to a change in the distribution of episodes 
according to the number of therapy visits and how much was due to a 
change in the average case-mix weight at each level of therapy visits.
    The method we used first holds the average case-mix weight constant 
(at the 2008 values) at each level of therapy visits, and measures the 
effect of the shift to the new distribution of therapy visits. The 
method then holds the distribution of therapy visits constant (at the 
2008 distribution) and measures the effect of the change in average 
case-mix weight at each level of therapy visits. The results were that 
0.0254 or about 75 percent (0.0254/0.0340 = 0.75) of the total change 
in average case-mix weights from 2008 to 2009 was due to the shift in 
the distribution of therapy visits per episode. The remaining 0.0086 or 
about 25 percent (0.0086/0.0340 = 0.25) in overall average case-mix 
weight from 2008 to 2009 was due to an increase in the average case-mix 
weight at each level of therapy visits per episode.

[[Page 40991]]

    The decomposition suggests that agencies in 2009 were still 
responding to the 2008 refinements in terms of both coding practices 
and the definition of therapy treatment plans for patients. This 
analysis by itself, however, does not isolate real case-mix change 
within total case-mix change. We discuss our latest analysis of real 
and nominal case-mix change in the remainder of this section.
    Section 1895(b)(3)(B)(iv) of the Act gives CMS the authority to 
implement payment reductions for nominal case-mix growth, changes in 
case-mix that are not related to actual changes in patient 
characteristics over time. Nominal case-mix growth was assessed and 
reported in CY 2008 and CY 2011 rulemaking, and payment reductions to 
the base rate were implemented to account for the nominal case-mix 
growth observed.
    In CY 2008 rulemaking, to assess nominal case-mix growth, we first 
estimated real case-mix growth, changes in case-mix which are related 
to changes in patient characteristics, using a regression-based, 
predictive model of individual case-mix weights. The predictive model 
contained measures of patients' demographic characteristics, clinical 
status, inpatient history, and Part A Medicare costs in the time period 
leading up to their home health episodes. The regression coefficients 
for the predictive model were developed using 2000 as a base year and 
were applied to episodes from 2005, allowing estimation of the change 
in real case-mix. We then determined the nominal case-mix growth from 
2000 to 2005 using the regression model-predicted real case-mix change 
and the total case-mix change for the time period of interest.
    In 2000, the average case-mix was 1.0960 and in 2005, the average 
case-mix was 1.2361. As such, the total measure of case-mix change from 
2000 to 2005 was 12.78 percent ((1.2361 - 1.0960)/1.0960 = 0.1278). 
Using the regression-based predictive model, we identified 8.03 percent 
of the total case-mix change as real case-mix change from 2000 to 2005, 
and we adjusted the 12.78 percent of total change in case-mix, 
downward, by 8.03 percent to get a final nominal case-mix change 
measure of 11.75 percent (0.1278 * (1 - 0.0803) = 0.1175). To account 
for the 11.75 percent increase in nominal case-mix, we implemented a 
payment reduction of 2.75 percent each year for 3 years, beginning in 
2008, and we planned to implement a payment reduction of 2.71 in CY 
2011.
    Since the HH PPS CY 2008 proposed rule, we have continued to 
monitor case-mix changes in the HH PPS, and in CY 2011 rulemaking we 
updated our analysis to measure change in real and nominal case-mix. In 
CY 2011 rulemaking, we developed two regression-based models to assess 
nominal case-mix growth from 2000 to 2008. One model was developed 
using 2000 as a base year and the 80 grouper case-mix system. The 
regression coefficients in the model were applied to 2007 data to 
determine the change in real case-mix from 2000 to 2007. The second 
model was developed using 2008 as a base year and the 153 grouper case-
mix system. The regression coefficients in the model were applied to 
2007 data to determine the change in real case-mix from 2007 to 2008. 
The data from both of the models were then used to calculate the 
overall real and nominal case-mix change from 2000 to 2008. Our 
analysis indicated that there was a 19.40 percent increase in overall 
case-mix from 2000 to 2008 and 10.07 percent of that overall observed 
case-mix change was identified as real case-mix change. Consequently, 
as a result of our analysis, we identified a 17.45 percent nominal 
increase in case-mix (0.1940 * (1 - 0.1007) = 0.1745) from 2000 to 
2008. In other words, there was a growth in case-mix of 17.45 percent 
that was unrelated to differences in patient characteristics and 
reflects changes in coding procedures and documentation rather than the 
treatment of more resource-intensive patients. This 17.45 percent 
increase was larger than expected. Previously, there was about 1 
percent annual case-mix growth from 2000 to 2007. Between 2007 and 
2008, we observed a 4 percent overall case-mix growth. As a result of 
our analysis, in CY 2011, we proposed an increase to the planned 2.71 
percent payment reduction in 2011 to a 3.79 percent payment reduction 
and we proposed another 3.79 percent payment reduction in 2012 to fully 
account for the 17.45 percent nominal case-mix growth which was 
identified from 2000 to 2008.
    We received many comments on our CY 2011 HH PPS proposed rule that 
criticized our methodology for assessing real case-mix change. The 
criticisms from commenters centered on the idea that we underestimated 
the percentage of case-mix growth that was real. Multiple commenters 
stated that our model for assessing real case-mix change relies too 
heavily on hospital discharge data. Commenters stated that we should 
include more variables which capture the severity of patients entering 
home health from the community since more than half of Medicare home 
health patients are admitted to home health from a setting other than a 
hospital. Also, commenters suggested that the acute care hospital APR-
DRG and other prior use variables in our models may not be relevant for 
patients with more than one home health episode. Another criticism was 
that our model should consider that there are shorter hospital stays, 
and therefore, the patients who are discharged from the hospital into 
home health may have a higher level of severity of illness than the 
model recognizes. Moreover, commenters stated that all of the HHAs were 
being penalized for the actions of a few HHAs and that the nominal 
case-mix change reductions should be limited to certain types of 
agencies (such as by region or for-profit/non-profit status or by case-
mix index [CMI]). Furthermore, one commenter stated that a recent study 
by Dr. Partha Deb of Hunter College used data from a nationally 
representative survey (the Medical Expenditures Panel Survey--MEPS) and 
found that the health status of Medicare beneficiaries worsened, 
suggesting a possible increase in real case-mix in the Medicare 
population from 2000 through 2007 (the study by Partha Deb can be found 
at https://www.aha.org/aha/content/2010/pdf/100715-CMItrends.pdf). 
Commenters inferred that the change in real case-mix was larger than 
the change we measured for the home health population, and therefore, 
commenters doubted whether our model accounted for the entire real 
case-mix change in the home health population. The study by Dr. Deb 
constructed a case-mix measure from medical expenditures and diagnosis-
related data and compared results for 2000 and 2007.
    In the CY 2011 HH PPS final rule, we implemented the proposed 
payment reduction of 3.79 percent to the national standardized episode 
rate in CY 2011. However, due to the extensive comments we received, we 
deferred finalizing a payment reduction for CY 2012 until further study 
of the case-mix data and methodology was completed.
1. Independent Review of the Models To Assess Nominal Case-Mix Growth
    To assess the validity of the criticisms we received about our 
models to measure real and nominal case-mix change, we procured an 
independent review of our methodology by a team at Harvard University 
led by Dr. David Grabowski. The review included an examination of the 
predictive regression models and data used in CY 2011 rulemaking, and 
further analysis consisting of extensions of the model to allow a 
closer look at nominal case-mix

[[Page 40992]]

growth by categorizing the growth according to provider types and 
subgroups of patients. The extensions showed a similar rate of nominal 
case-mix growth from 2000 to 2008 (Table 1A) for the various categories 
and subgroups. Below, we discuss these results in terms of the 
criticisms we received.

           Table 1A--Models for Assessing Real Case-Mix Change
------------------------------------------------------------------------
                                                       Nominal case-mix
                       Model                           percent increase
                                                      from 2000 to 2008
------------------------------------------------------------------------
(ALL) Total Nominal growth using Full Data Set                     17.45
 (Replication).....................................
(ALL) Full Data Set using MEDIAN ACH LOS                           17.38
 (Replication).....................................
(ALL) Full Data Set using Q3 ACH LOS (Replication).                17.47
(1a) Pre-HHA: With IH in prior 14 days.............                21.16
(1b) Pre-HHA: With IH in prior 15-120 days.........                16.81
(2a) Pre-HHA: Without IH in prior 14 days..........                15.85
(2b) Pre-HHA: Without IH in prior 15-120 days......                18.19
(3a) Pre-HHA: With IRF/SNF/LTCH in prior 14 days...                13.90
(3b) Pre-HHA: With IRF/SNF/LTCH in prior 15-120                    14.11
 days..............................................
(4a) Pre-HHA: Without IRF/SNF/LTCH in prior 14 days                18.51
(4b) Pre-HHA: Without IRF/SNF/LTCH in prior 15-120                 18.33
 days..............................................
(5a) Pre-HHA: With IH/IRF/SNF/LTCH in prior 14 days                18.97
(5b) Pre-HHA: With IH/IRF/SNF/LTCH in prior 15-120                 16.74
 days..............................................
(6a) Pre-HHA: Without IH/IRF/SNF/LTCH in prior 14                  16.95
 days..............................................
(6b) Pre-HHA: Without IH/IRF/SNF/LTCH in prior 15-                 18.29
 120 days..........................................
(7a) AGENCY-LEVEL: Owner: Non-Profit...............                14.49
(7b) AGENCY-LEVEL: Owner: For-Profit...............                18.63
(7c) AGENCY-LEVEL: Owner: Government...............                15.22
(8a) AGENCY-LEVEL: Facility-Based HHA..............                14.17
(8b) AGENCY-LEVEL: Free-Standing HHA...............                17.86
(9a) AGENCY-LEVEL: West Region.....................                17.51
(9b) AGENCY-LEVEL: Midwest Region..................                16.76
(9c) AGENCY-LEVEL: South Region....................                18.01
(9d) AGENCY-LEVEL: Northeast Region................                14.81
(10a) AGENCY-LEVEL: Large Agency...................                17.21
(10b) AGENCY-LEVEL: Small Agency...................                17.53
(11a) AGENCY-LEVEL: Urban HHA......................                17.75
(11b) AGENCY-LEVEL: Rural HHA......................                15.36
(12a) AGENCY-LEVEL: Treats predominantly post-acute                16.67
 patients..........................................
(12b) AGENCY-LEVEL: Treats predominantly community                 18.87
 patients..........................................
(13) First Episode Only............................                19.06
------------------------------------------------------------------------
HHA = home health agency; IH = Inpatient hospitalization; IRF =
  inpatient rehabilitation facility; SNF = skilled nursing facility;
  LTCH = long-term care hospital, ACH LOS = acute care hospital length
  of stay.

    To address the concern about our current models' robustness when 
there is no prior inpatient or post-acute care setting (when patients 
are admitted from the community), the Harvard team re-ran our models 
for separate subgroups; in most cases, subgroups were defined by the 
prior hospital and post-acute care use measures present on the data 
file. Specifically, they defined prior inpatient/post-acute care use in 
six different ways (shown in lines 1a through 6b of Table 1A): Any 
hospital use over the past 14 days (yes/no); any post-acute use over 
the prior 14 days (yes/no); any hospital use over the past 15-120 days 
(yes/no); any post-acute care use over the past 15-120 days (yes/no); 
any hospital or post-acute care use in the preceding 14 days (yes/no); 
and any hospital or post-acute care use in the preceding 15-120 days 
(yes/no). As another test, the team separated agencies according to 
whether they treated predominantly post-acute patients or not. To 
calculate this measure, the Harvard team split agencies above/below the 
median based on their percentage of home health episodes in 2007 with 
an inpatient hospital stay in the preceding 14 days.
    Across all models, there was evidence of significant and similar 
nominal case-mix growth, suggesting that high rates of nominal case-mix 
growth exist regardless of whether there was a preceding inpatient or 
post-acute stay. Agencies classified as serving predominantly community 
patients had a slightly higher nominal case-mix percentage increase 
compared to agencies classified as serving predominately post-acute 
patients (as shown in lines 12a and 12b in Table 1A). (For a full 
description of the Harvard team's analysis and results, please see the 
L&M final report located at https://www.cms.gov/center/hha.asp).
    Also, to evaluate the validity of the comment that the acute care 
hospital APR-DRG and other prior use variables in our model may not be 
relevant for patients with more than one home health episode, the 
Harvard team re-ran our current predictive models using only the first 
home health episode for each patient (shown in line 13 of Table 1A). 
Once again, results based on this first episode were similar to the 
overall results of our current model, suggesting that the model is 
relatively stable across home health episodes. The results show that 
the inclusion of the later episodes does not dramatically alter the 
primary finding of significant nominal case-mix growth.
    To evaluate the comment that our models should take into account 
the fact that there are shorter hospital stays and therefore, the 
patients who are discharged from the hospital into home health may have 
a higher level of severity of illness than the model recognizes, our 
predictions were calculated assuming there was a different average 
length of stay than the actual average length of stay found for the LOS 
predictor variables in the 2007 and 2008 follow-up years. Harvard 
developed predictions of real and nominal case-mix growth using the

[[Page 40993]]

median acute care hospitalization length of stay, instead of the mean 
length of stay which is used in our current model. The median is lower 
than the mean acute care hospitalization length of stay. Harvard also 
developed predictions of real and nominal case-mix growth using the 
third quartile acute care hospitalization length of stay, which is 
longer than the mean. The results were very similar to the overall 
nominal case-mix percentage increase and therefore, the analysis 
suggests that our methodology is not particularly limited in capturing 
length of stay effects, because acute care hospitalization length of 
stay does not play a big role in determining average patient severity.
    To evaluate the suggestion that we should limit nominal case-mix 
change reductions to certain types of agencies (such as by region or 
for-profit/non-profit status or by CMI), the Harvard team re-ran our 
model based on ownership type (non-profit, government, for-profit), 
agency type (facility-based, freestanding), region of the country 
(Northeast, South, Midwest, West), urban vs. rural status, and agency 
size (large vs. small; based on the number of initial episodes), shown 
in lines 7a through 11b in Table 1A. As noted earlier, the team also 
examined case-mix growth by whether the agency had a particular focus 
on post-acute vs. community patients. Across all these different 
categories (ownership, agency type, region, urban vs. rural status, 
agency size, agency focus), nominal case-mix growth was present. As 
expected, nominal case-mix growth was larger for some sub-groups. For 
example, nominal case-mix growth was higher for for-profit agencies 
(18.63 percent) than non-profit (14.49 percent) and government agencies 
(15.22 percent); however, these latter ownership types still exhibited 
high rates of nominal case-mix growth. As such, the Harvard team 
asserted that similar high rates of nominal case-mix growth exist for 
all types of HHAs.
    To address the comment that a study which used MEPS data showed a 
higher rate of real case-mix growth in the entire Medicare population 
than our model estimated for Medicare home health patients, a more 
detailed analysis of the MEPS data was performed. The trends in health 
status of four different populations from 2000 to 2008 were analyzed. 
The data for the analysis were obtained from the MEPS 2000 and 2008 
Full Year Consolidated Data files. The four populations that were 
analyzed were: (1) The full MEPS sample; (2) all Medicare 
beneficiaries, defined as all respondents ever having Medicare in a 
given year; (3) all home health patients, defined as having at least 
one home health provider day in a given year; and (4) all home health 
Medicare beneficiaries, defined as all respondents with any Medicare 
home health charges. Two measures of self-reported health status and 
one measure derived from patient information that screened for 
activities of daily living (ADL) limitations were used to determine the 
trends in health status. These types of measures have been shown to be 
highly correlated with actual health (Ware and Sherbourne, 1992; 
McHorney, Ware, and Raczek, 1993). The three measures which were 
analyzed for each of the populations were: (1) Whether the respondent 
indicated perceived health status of ``poor'' or ``fair'' as opposed to 
those indicating health status as ``good'', ``very good'', or 
``excellent''; (2) whether the respondent indicated if pain limited 
normal work (including work in the home) in the past 4 weeks 
``extremely'' or ``quite a bit'' as opposed to those indicating pain 
limited work ``moderately'', ``a little bit'', or ``not at all''; and 
(3) whether respondents had a positive screen for needing assistance 
with ADL. In all cases, responses such as ``refused'', ``don't know'', 
or ``not ascertained'' were omitted from the analysis. The Medicare 
analysis samples consisted of 3,371 and 4,144 beneficiaries in 2000 and 
2008, respectively. The Medicare home health subsamples consisted of 
174 and 289 beneficiaries in 2000 and 2008, respectively. The survey 
responses were then weighted using pre-constructed MEPS survey weights 
to estimate nationally representative changes in the three health 
status variables.
    All three measures indicated a slight increase in the overall 
health status of the Medicare home health population. Two of these 
results were not statistically significant, but the percent of home 
health Medicare beneficiaries experiencing ``extreme'' or ``quite a 
bit'' of work-limiting pain decreased substantially, from 56.6 percent 
in 2000 to 45.4 percent in 2008 (p = 0.039). Unlike Dr. Deb's original 
study, the new MEPS analysis focuses specifically on Medicare home 
health users (as opposed to the entire Medicare population), and it is 
not reliant on expenditure data. A limitation of the Debs case-mix 
measure, which relies on expenditure data, is that it could reflect 
large increases in expenditures, such as drug expenditures, but any 
relationship to actual increases in impairments and other reasons for 
using home health resources is unclear. A possible limitation of the 
new MEPS analysis is that the sample of Medicare home health 
respondents is relatively small, notwithstanding that the result of one 
of the three measures was statistically significant. Also, the ADL 
screening item may not capture a change in the frequency of very severe 
ADL limitations since the measure may be insensitive to changes at high 
levels of disability. However, the Harvard team asserted that the 
methods of the new MEPS analysis are more appropriate for assessing 
whether there are increases in the severity of illness burden that 
would specifically indicate a need for more resources in the Medicare 
home health population. Based on the two kinds of evidence, and a 
recognition of the limitations of both, we conclude that the MEPS data 
provide no evidence of an increase in patient severity from 2000 to 
2008.
    Based on the findings from the extensions of the current model that 
were tested, including the finding that the two nominal case-mix 
percentage increases for the post-acute and community patients are 
similar (Table 1A), and the results of the MEPS analysis which do not 
provide evidence to suggest that the Medicare home health population 
has experienced a decrease in their health status over time, the 
Harvard team concluded that the current model adequately measures real 
case-mix growth for home health patients, including patients admitted 
to home health from the community.
    When reviewing the model, the Harvard team found that overall, our 
models are robust. However, one area of potential refinement to our 
models that the Harvard team suggested was to incorporate variables 
derived from Hierarchical Condition Categories (HCC) data, which is 
used by CMS to risk-adjust payments to managed care organizations in 
the Medicare program. Currently, the HCC model includes 70 HCCs, each 
of which is defined based on the presence of particular ICD-9-CM codes 
identified from Medicare claims data (inpatient and outpatient hospital 
claims and Part B Physician Claims). Some of the HCCs reflect 
hierarchies among related conditions, but, for unrelated diseases, each 
HCC is separately defined. The HCC model also includes demographic 
items related to gender, age, Medicaid enrollment, and whether Medicare 
eligibility was originally based on disabled status. We have augmented 
our modeling data with HCC information, as described in the next 
section.
2. Revised Version of Our Models To Assess Nominal Case-Mix Growth
    In the past, we have considered using HCC data to assess real and 
nominal

[[Page 40994]]

case-mix change; however, we have yet to implement a change to our 
models which would incorporate the HCC data. Based on Dr. Grabowski and 
his team's recommendation and our previous consideration to incorporate 
HCC data in our models to assess real case-mix change, we explored the 
effects of adding the managed care data to our models. To incorporate 
HCC data into our models, we augmented our analytic files used to 
measure real case-mix change. We obtained HCC data on all home health 
users for 2004-2009. There were several different types of HCC 
variables that could be added to our models to assess real case-mix. 
Some of the variables we considered are the HCC risk score, binary 
variables for each of the HCCs, demographic variables, and disease 
indicators.
    In the HCC model used for managed care risk adjustment, each HCC 
has an associated regression coefficient. Regression coefficients for 
each beneficiary's HCCs, along with the regression coefficients for 
their demographic and enrollment characteristics, are summed to 
calculate predicted expenditures. A risk score for each record can then 
be calculated based on expected expenditures for the patient divided by 
the mean expenditures for all patients. The HCC data include several 
risk score measures, including the HCC community risk score, the 
institutional risk score, and the risk score for new Medicare 
enrollees. Because home health patients live in the community, the 
community risk score seemed more appropriate than the institutional 
risk score. An alternative to using the HCC risk score was to include 
binary variables for each of the 70 HCCs, which may better capture a 
patient's severity. Along with the HCC risk score and the individual 
HCCs, we considered other elements of the HCC data such as the 
demographic variables, whether disability was the original reason for 
Medicare entitlement, and an indicator for whether the individual is a 
Medicaid beneficiary. Furthermore, we examined interactions involving a 
number of disease conditions that are included with the HCC data, such 
as congestive heart failure (CHF), diabetes mellitus (DM), chronic 
obstructive pulmonary disease (COPD), cardiovascular disease (CVD), 
renal failure (RF), and coronary artery disease (CAD).
    To test the usefulness of these different HCC variables, we 
developed several models to examine real case-mix and which contained 
different types of HCC data. We examined models in which we added the 
HCC community score to our CY 2005 data so that the HCC score was 
included with the APR-DRG variables in an equation explaining 2005 
case-mix weights. We also examined models which incorporated individual 
HCCs, instead of the HCC risk score. Furthermore, we examined models in 
which either the HCC risk score or individual HCCs were added to our 
model along with demographic and disease indicator variables. Moreover, 
we examined models which did not include APR-DRGs, but rather the HCC 
risk score or individual HCCs replaced the APR-DRGs in the model. When 
we replaced the APR-DRGs in the models with the HCC risk score, there 
was a low R-squared value, lower than any of the other models we 
examined. When we replaced the APR-DRG variables in our models with the 
individual HCC indicators, we observed a negative change in real case-
mix. This negative change in real case-mix would indicate that the 
health status of the Medicare home health population has improved over 
time and that all of the change in case-mix from 2000-2009 would be 
nominal case-mix change. As a result of the findings from the various 
models, we decided to augment our current model with the HCC variables 
rather than replace our APR-DRG variables with HCC variables.
    It should be noted that in addition to examining which HCC 
variables we should include in our models, we also examined which year 
of HCC data we should use in our models. There is a 1 year look-back 
period with HCC data in that the HCC data are based on the previous 
calendar year's claims history for an individual. Therefore, when 
developing our models, we assessed whether we should use HCC data from 
the previous year or HCC data in the same year as when the home health 
episode occurred (the home health episode is the unit of observation in 
our models). Our concern was that if we used HCC data in the same year 
as the episode, the HCC data may partially reflect diseases and 
conditions identified after a home health episode. However, we decided 
to use HCC data in the same year as the episode since we thought it 
best reflected the health status of the patients in that year.
    For this year's analysis, we used a similar approach to our 
previous methods. The basic method is to estimate a prediction model 
and use coefficients from that model along with predictor variables 
from a different year to predict the average case-mix for that year. It 
should be noted that we chose to enhance our models with HCC data 
starting in 2005 due to the availability of HCC data in our analytic 
files. Therefore, we analyzed real case-mix change for three different 
periods, from 2000 to 2005, from 2005 to 2007, and from 2007 to 2009. 
The real case-mix change in the period from 2005 to 2007 and the period 
from 2007 to 2009 were assessed using enhanced models, which included 
HCC data. The real case-mix change from 2000 to 2005 was assessed using 
the same variables used in the model described in last year's 
regulation (75 FR 43238), a variable list consisting of measures of 
patients' demographic characteristics, clinical status, inpatient 
history, and Part A Medicare costs in the time period leading up to 
their home health episodes. The regression coefficients from the model 
without HCC variables were applied to episodes from 2005, allowing us 
to estimate how much of the change in observed case-mix was 
attributable to changes in patient characteristics between the IPS 
period and 2005.
    We added HCC variables for the 2005 to 2007 period, estimating the 
model using data from 2005. The enhanced model includes HCC community 
scores, HCC demographic variables, and disease indicator variables for 
2005 and later. We chose this version of the HCC-enhanced case-mix 
change model largely based on its ability to predict higher real case-
mix change relative to the other HCC enhanced models. We applied the 
regression coefficients to means from 2007, allowing estimation of real 
case-mix change between 2005 and 2007.
    For the 2007 to 2009 period, we used the 153 HHRG case-mix weights 
and data from 2009 to estimate the same set of models as we did for 
2005. Using the backwards prediction method that we used in CY 2011 
rulemaking, the coefficients from this model were developed using 2009 
data and were applied to episodes from 2007. This procedure allows us 
to estimate how much of the 2007 through 2009 change (based on the 
HHRG153 case-mix for both periods) was associated with changes in 
patient characteristics between 2007 and 2009.
    From 2000 to 2009, we identified a total change in case-mix of 
0.2476 (1.3435-1.0959 = 0.2476), which results in a case-mix growth of 
22.59 percent ((1.3435-1.0959)/1.0959 = 0.2259). We then estimated the 
real and nominal change in case-mix for each of the three periods. The 
change in real case-mix from 2000 to 2005 was 0.0207 case-mix units. 
The change in real case-mix from 2005 to 2007 was 0.0061 case-mix 
units. The change in real case-mix from 2007 to 2009 was 0.0122 case-
mix units. After adding together the estimated real case-mix change in 
case-mix units for the three periods, the total

[[Page 40995]]

estimated change in real case-mix from 2000 to 2009 was 0.0390 (0.0207 
+ 0.0061 + 0.0122 = 0.0390). Therefore, we estimate that 15.76 percent 
of the total percentage change in the national average case-mix weight 
since the IPS baseline through 2009 is due to change in real case-mix 
(0.0390/0.2476 = ~0.1576). It should be noted that due to rounding, 
there is a 0.01 percentage point difference between the calculated and 
actual value. When taking into account the total measure of case-mix 
change (22.59 percent) and the 15.76 percent of total case-mix change 
estimated as real from 2000 to 2009, we obtained a final nominal case-
mix change measure of 19.03 percent from 2000 to 2009 (0.2259 * (1-
0.1576) = 0.1903). Please see Table 1B for additional information about 
the calculations used to make the real and nominal case-mix change 
estimates from 2000 to 2009.
    Our estimates of real and nominal case-mix change are consistent 
with past results. Most of the case-mix change has been due to improved 
coding, coding practice changes, and other behavioral responses to the 
prospective payment system, such as increased use of high therapy 
treatment plans.

 Table 1B--Summary of Real and Nominal Case-Mix Change Estimates: 2000-
                                  2009
------------------------------------------------------------------------
                           Measure                               Model
------------------------------------------------------------------------
Actual case-mix: 2000........................................     1.0959
Actual case-mix: 2009........................................     1.3435
Total change in case-mix.....................................     0.2476
Total percentage change......................................     22.59%
Estimated real change in case-mix............................     0.0390
Percent of total change estimated as real....................     15.76%
Percent of total change estimated as nominal (creep).........     84.24%
Real case-mix percent increase...............................      3.56%
Nominal case-mix percent increase............................     19.03%
------------------------------------------------------------------------

    As we described earlier in this proposed rule, our CY 2008 HH PPS 
final rule finalized a reduction over 4 years in the national 
standardized 60-day episode payment rates to account for a large 
increase in case-mix from 2000 to 2005 which we determined was not 
related to treatment of more intense patients. We implemented a 2.75 
percent reduction each year for 2008, 2009, and 2010 and planned to 
reduce payments by 2.71 percent in 2011. In CY 2011 rulemaking, we 
updated our analysis of nominal case-mix growth through 2008 and 
determined that there was 17.45 percent nominal case-mix growth from 
2000 to 2008. Therefore, we proposed and finalized an increase in the 
planned 2.71 percent reduction to 3.79 percent for CY 2011. Also, in 
the CY 2011 proposed rule, we stated that if we were to identify 
further increases in nominal case-mix as more current data becomes 
available, it would be our intent to account fully for those increases 
when they are identified, rather than continuing to phase in the 
reductions over more than 1 year. For the CY 2012 proposed rule, after 
updating our models to incorporate HCC data, we have determined that 
there was a 19.03 percent nominal case-mix change from 2000 to 2009. To 
account for the remainder of the 19.03 percent residual increase in 
nominal case-mix beyond that which has been accounted for in previous 
payment reductions, we estimate that the percentage reduction to the 
national standardized 60-day episode rates for nominal case-mix change 
for CY 2012 will be 5.06 percent. Therefore, for CY 2012, we propose to 
implement a 5.06 percent payment reduction to the national standardized 
60-day episode rates to fully account for growth in nominal case-mix 
from the inception of HH PPS through 2009.

B. Case-Mix Revision to the Case-Mix Weights

1. Hypertension Diagnosis Coding Under the HH PPS
    In CY 2011 rulemaking, we proposed to remove ICD-9-CM code 401.1, 
Benign Essential Hypertension, and ICD-9-CM code 401.9, Unspecified 
Essential Hypertension, from the HH PPS case-mix model's hypertension 
group. Beginning with the HH PPS refinements in 2008, hypertension was 
included in the HH PPS system because data suggested it was associated 
with elevated resource use. As a result, the diagnoses Unspecified 
Essential Hypertension and Benign Essential Hypertension were 
associated with additional points from the four-equation model and 
subsequently, potentially higher case-mix weights in the HH PPS case-
mix system. When examining the trends in reporting of hypertension 
codes from 2000 to 2008, our analysis showed a large increase in the 
reporting of codes 401.1 and 401.9 in 2008. However, when looking at 
2008 claims data, the average number of visits for claims with code 
401.9 was slightly lower than the average for claims not reporting 
these hypertension codes. In last year's proposed rule, we proposed to 
remove codes 401.1 and 401.9 from our case-mix model based on 
preliminary analysis of the trends in coding and resource use of 
patients with these codes. We suspected that the 2008 refinements, 
which newly awarded points for the diagnosis codes 401.1 and 401.9, led 
to an increase in reporting of these codes and that this reporting was 
a key driver of the high 2008 growth in nominal case-mix. In response 
to this proposed policy change, we received numerous comments, many of 
which stated that additional analysis was needed to substantiate the 
rationale for removing hypertension codes 401.1 and 401.9. In the CY 
2011 HH PPS final rule, we withdrew our proposal to eliminate 401.1 and 
401.9 from our model and described our plans to do a more comprehensive 
analysis of the resource use of patients with these two hypertension 
codes. We have since completed a more thorough analysis. Based on the 
results of our latest analyses, we propose to remove codes 401.1 and 
401.9 from the HH PPS case-mix system.
    We performed several analyses of the resource use and prevalence of 
patients with Benign Essential Hypertension and Unspecified Essential 
Hypertension (codes 401.1 and 401.9) to assess the appropriateness of 
these codes in our case-mix model. We looked at the HH PPS episode data 
using two samples to more accurately assess the trends in hypertension 
prevalence over time. In one sample, we excluded episodes from 
providers in areas exhibiting suspect billing practices. For the other 
sample, we excluded outlier episodes. In all of the analyses that 
follow, we report the results from the sample that excludes outliers 
because results from the alternate analysis were highly similar. Also, 
the sample that excludes outliers is more appropriate than one that 
includes outliers because our case-mix research has been conducted on 
samples without outliers.
    One of our analyses looked at the prevalence of various 
hypertension codes over time. We compared the change in prevalence of 
401.1 and 401.9 diagnoses to the prevalence of other diagnoses in the 
hypertension group--401.0 (malignant essential hypertension), 402 
(hypertensive heart disease), 403 (hypertensive chronic kidney 
disease), 404 (hypertensive heart and chronic kidney disease), and 405 
(secondary hypertension)--from 2005 to 2009 (Table 2). Our analysis 
shows that the prevalence of episodes with a 401.9 diagnosis continued 
to increase in 2009, from 50.58 percent of episodes in 2008 to 55.52 
percent in 2009, and more than doubled between 2005 and 2009. The 
prevalence of episodes with a 401.1 diagnosis decreased from 2008 to 
2009 but the prevalence remained slightly higher than the prevalence in 
2005.

[[Page 40996]]



                                 Table 2--Prevalence of Hypertension--2005-2009
                                                  [In percent]
----------------------------------------------------------------------------------------------------------------
                           Diagnosis                              2005      2006      2007      2008      2009
----------------------------------------------------------------------------------------------------------------
Any hypertension..............................................     33.32     40.22     46.26     60.37     65.65
401.0 Malignant essential hypertension........................      0.56      0.54      0.53      0.56      0.47
401.1 Benign essential hypertension...........................      2.89      3.36      3.44      3.79      2.95
401.9 Essential hypertension, unspecified.....................     27.23     33.22     38.74     50.58     55.52
402 Hypertensive heart disease................................      2.19      2.38      2.49      2.99      2.76
403 Hypertensive renal disease................................      0.31      0.56      0.92      2.24      3.66
404 Hypertensive heart and renal disease......................      0.14      0.17      0.20      0.31      0.39
405 Secondary hypertension....................................      0.04      0.04      0.03      0.03      0.04
----------------------------------------------------------------------------------------------------------------
Outlier episodes are excluded.
Source: Abt Associates analysis of 20 percent sample of Home Health Datalink file for 2005-2009.

    We also examined the prevalence of hypertension coding by various 
agency characteristics, such as agency type, region, and provider size, 
in 2005 versus 2009 (Tables 3 and 4). We compared the 2005 data (Table 
3) to more current data (Table 4) because the 2005 data were used to 
simulate the 2008 refinements for the CY 2008 HH PPS final rule 
implementing the 153-group case-mix system (72 FR 49762 through 49945). 
Based on our analysis, except for government-owned agencies and 
agencies in a few regions, agencies (regardless of type) had a similar 
prevalence of episodes with a 401.9 diagnosis across the board in 2009 
(Table 4). Also, agencies had a relatively similar prevalence of 
episodes with a 401.1 diagnosis across the board in 2009, except for 
West South Central, which had a high prevalence of 6.68 percent (Table 
4)--about 9 times the region's prevalence in 2005. In addition, small 
facilities with less than 19 home health episodes in a year in the 20 
percent sample of the Home Health Datalink file had a high prevalence 
of diagnosis 401.1; 8.30 percent of their episodes had a 401.1 
diagnosis. All categories of agencies appear to have a significant 
increase in the reporting of a 401.9 diagnosis when comparing 2005 HH 
PPS claims and OASIS data to 2009 HH PPS claims and OASIS data. The 
reporting of a 401.9 diagnosis in 2009 was typically 1.8 to 2.1 times 
the reporting of a 401.9 diagnosis in 2005, with the exception of the 
East North and the West North Central regions which had an increase of 
around 1.7 and 1.5 fold respectively. Also, it should be noted that the 
Mid-Atlantic region had around a 2.4 fold increase in the reporting of 
a 401.9 diagnosis between 2005 and 2009 and the West South Central 
region had almost a threefold increase in the reporting of a 401.9 
diagnosis between 2005 and 2009. Furthermore, many categories had an 
increase in the reporting of a 401.1 diagnosis when comparing 2005 data 
to 2009.

                   Table 3--Prevalence of Hypertension by Various Agency Characteristics--2005
                                                  [In percent]
----------------------------------------------------------------------------------------------------------------
                                     Any      401.0     401.1     401.9      402       403       404       405
----------------------------------------------------------------------------------------------------------------
All Agencies....................     33.59      0.56      2.96     27.34      2.26      0.32      0.15      0.04
----------------------------------------------------------------------------------------------------------------
                                                Type of Facility
----------------------------------------------------------------------------------------------------------------
Free-Standing/Other Vol/NP......     27.50      0.21      0.63     25.49      0.83      0.30      0.06      0.01
Free-Standing/Other Prop........     39.35      0.86      4.86     29.63      3.48      0.30      0.19      0.06
Free-Standing/Other Govt........     29.01      0.41      1.35     25.36      1.51      0.22      0.17      0.04
Hospital-Based Vol/NP...........     25.11      0.17      0.68     23.33      0.51      0.35      0.09      0.01
Hospital-Based Prop.............     29.79      0.30      0.68     27.50      0.83      0.37      0.16      0.01
Agency-Based Govt...............     30.94      0.80      3.04     24.46      1.92      0.53      0.23      0.02
----------------------------------------------------------------------------------------------------------------
                                                Facility Location
------------------------------------------------------------------------------------------------------