CMS-10298 Response to Comments

RESPONSES TO 60-DAY PUBLIC COMMENT PERIOD
The Centers for Medicare and Medicaid Services established a research project titled
“Developing Outpatient Therapy Payment Alternatives” (DOTPA) designed to identify,
collect, and analyze therapy-related information tied to beneficiary need and the
effectiveness of outpatient therapy services. The immediate goals are to develop and
assess the feasibility of a comprehensive and uniform therapy-related data collection
instrument and to determine the subset of the measures that CMS can routinely and
reliably collect in support of payment alternatives. The ultimate goal is to develop
payment method alternatives to the current financial cap on outpatient therapy services.
Data collection instruments will be administered to Medicare beneficiaries upon
admission and discharge from outpatient therapy providers.
The data collection instruments (CARE-C for patients in community settings and CAREF for patients in institutional settings) are comprised of a set of common assessment
items administered to all patients and a set of supplemental items only administered for
specific conditions or at particular times (i.e., discharge only).
Analysis of and Response to Public Comments
We received 11 comments from physical therapists, occupational therapists, hospitals,
professional associations, and health care associations. All comments were reviewed,
analyzed and grouped into categories based on subject matter. Summaries of the public
comments and our responses to those comments follow.
General Comments
CMS received several comments related to concern over burden to beneficiaries and
providers related to this data collection. CMS also received comments suggesting
general changes and other comments suggesting changes to and additions of specific
assessment items. Additionally, several commenters requested clarification of terms
and underscored the need to provide sufficient staff training. The following summarizes
comments received and responses to those comments.
1. Study Purpose
Comment 1.1: One commenter felt that CMS needs to provide more clarification on the
purpose of the study and research questions and that one purpose should be collecting
data associated with determining appropriateness/need for outpatient therapy services.
Response: The purpose of this one-time data collection is to collect data
associated with patient utilization of therapy services not available in claims
submitted to Medicare that are feasible for use in Medicare payment for therapy
services. The data collection instruments are not intended to be the final
instruments used for payment purposes. Rather, this one-time data collection is
intended to provide information on what data are feasible to collect as well as
what measures are valid and statistically reliable. Establishing criteria for
appropriateness or coverage of services is not within the scope of this project.

Comment 1.2: One commenter felt that CMS should implement instead the study
design and data collection plan based on a report by Ciolek and Hwang (2004),
conducted under contract to CMS.
Response: CMS refers interested parties to the Statement of Work for the
Request for Task Order Proposal (RTOP CMS-07-033) for the scope of work
specified by CMS for this project.
2. Respondent Burden
Comment 2.1: Several commenters felt that the proposed data collection instruments
were too long (had too many items or too many response categories).
Response: The purpose of this one-time data collection is to identify patient
assessment items for Medicare outpatient therapy services that are: (1) best
predictive of utilization of Medicare outpatient therapy services and/or reliable for
use in measures of therapy outcomes; (2) minimally gamable; and (3) as
minimally burdensome as possible, conditional on usefulness for case mix or
outcomes measurement. CMS understands that provider and beneficiary burden
is an important consideration, especially for an assessment instrument that would
be used in a payment system. The data collection results, including data
collection burden, will guide the development of any instrument used for payment
purposes. CMS expects that the data collection instruments in this one-time data
collection will contain items that will not be used in any assessment instrument
used for Medicare outpatient therapy payment.
Comment 2.2: Several commenters felt that the proposed data collection instruments
would take too much time for patients and clinicians to complete and would reduce
treatment time.
Response: It is not the intention of CMS for this data collection to interfere with
patient care. The CARE Tool, on which the CARE-F instrument in this data
collection was based, has been successfully used by approximately 150
providers, including skilled nursing facilities, in the Medicare Post Acute Care
Payment Reform Demonstration (PAC:PRD). This data collection is testing a
modified (and shortened) version of the CARE Tool. The CARE-C instrument,
intended to be used with community-based providers, includes items based on or
derived from those found in assessment instruments used in a number of
outpatient therapy settings (e.g., the AM-PAC and the NOMS). This instrument
also features a core-and-supplemental approach in which many of the items are
completed only for patients with certain conditions, functional status, or
impairments. However, understanding the items’ burden on providers and
beneficiaries is an important component of data analysis. CMS anticipates a total
data collection burden (reporting plus record keeping) of 20 minutes per
assessment for providers using the CARE-C instrument and 35 minutes per
assessment for providers using the CARE-F instrument.

In addition, a number of commenters remarked that there were additional items
that they would like on the data collection instruments to measure other aspects
of patients’ function or impairments. Design of this instrument is the result of
careful attention to creating a balance between comprehensiveness, so that no
potentially relevant measure is excluded, and response burden, so that the
instrument will be completed by sufficient numbers of beneficiaries to produce
valid and reliable information.
Comment 2.3: Some commenters questioned whether CMS would compensate
providers for time spent assisting patients with completing the data collection or
otherwise would provide a stipend for participation.
Response: At this time, for budgetary reasons, CMS cannot provide any
stipends for data collection.
3. Data Collection Instruments’ Ability to Identify Intended Factors
Comment 3.1: A number of commenters remarked that it is very important for factors
associated with outpatient therapy utilization and outcomes be collected in a reliable
and valid manner.
Response: CMS agrees that the reliability and validity of items is a very
important consideration in developing an assessment instrument used for
payment. To the extent possible, these data collection instruments have been
based on existing research on and approaches to patient assessment. A multidisciplinary team of contract staff, consultants, technical expert panel members,
and nominees of therapy associations contributed to the development of the data
collection instruments. In addition, CMS received extensive input from
associations, and there are several items in the instruments based on joint work
with the associations.
CMS is using a modified version of the CARE Tool for patients in nursing facility
settings. CMS also consulted extensively with developers of existing instruments
to identify or develop items for use in the CARE-C instrument for this data
collection. The core of the CARE-C instrument is based on the CARE-F and the
AM-PAC short form. The AM-PAC is recognized as one of the more
comprehensive of existing therapy-related proprietary instruments. The AM-PAC
short form item pools have been donated to the public domain. Furthermore,
testing these items’ reliability and validity is an important part of developing
alternative assessment item sets for use in Medicare outpatient therapy payment.
Comment 3.2: Some commenters felt that other, existing, assessment instruments
should be used in the data collection. Examples cited by commenters include the Focus
on Therapeutic Outcomes (FOTO) Patient Inquiry tool and the American SpeechLanguage-Hearing Association (ASHA) NOMS instrument.
Response: For this data collection, CMS’ goal is to develop a cross-setting,
cross-discipline, and cross-condition set of items for assessing patients’ medical,

functional, and cognitive condition that could be used for payment. This approach
is consistent with the CARE initiative to develop a comprehensive set of
assessment items to be used across treatment settings. These items are divided
into 4 data collection instruments (admission and discharge versions of
assessment instruments for patients in community versus institutional settings).
This was done to: (1) reduce the number of items on each instrument; (2) provide
items most relevant in each setting to avoid ceiling and floor effects; and (3)
create a single data collection instrument for each setting that covers the full
range of conditions among patients receiving outpatient therapy. It is CMS’
understanding that existing instruments, including the FOTO and NOMS
instruments, have strengths in assessing particular patient populations but not
the full range of Medicare beneficiaries receiving outpatient therapy. An
assessment instrument that addresses the range of medical, functional, and
cognitive conditions should improve comparability across patients to better
ensure consistency of case mix measurement, especially for patients receiving
care from multiple disciplines. In addition, CMS intends to avoid using items not
in the public domain.
Comment 3.3: A number of commenters felt that completion of certain sections/items
should be discipline-specific.
Response: CMS understands that Medicare beneficiaries receive outpatient
therapy services in a variety of single-discipline and multidisciplinary
environments. Which specific discipline assesses and treats a particular patient
would be based on disciplines’ scopes of practice; local treatment patterns; and
patients’ specific medical, functional, and cognitive issues. CMS intends for
individual providers to complete the assessments consistent with their current
workflow practices. Consistent with the CARE initiative, CMS intends for case
mix adjustors and outcomes items based on items that multiple disciplines can
complete. Whether certain items are completed by one discipline or another
should be based on disciplines’ scope of practice and clinical judgment rather
than CMS requirements.
Comment 3.5: Two commenters were concerned that patients’ self-assessments of
conditions and outcomes may disagree, at least in some cases, with clinicians’
assessments of those patients.
Response: There has been extensive research understanding sources of
differences between patient-reported function and clinician-reported function, and
this research indicates broad agreement between patient self-reported function
and clinician-reported function of the patient. However, since the purpose of this
data collection is to investigate assessment items useful for case mix and/or
outcomes measurement for payment, CMS intends to investigate differences
between patient self-assessment, non-clinician proxy-reported assessment, and
clinician-reported assessment in this data collection, review the literature to
understand reasons for differences, and consider implications for their use in a
payment system.

Comment 3.6: Some commenters were concerned that the data collection instrument
does not include items to measure patients’ ability to participate in life situations,
including home and the community.
Response: The data collection instrument for community-based patients
features a patient-reported function section that uses items from the AM-PAC
item pool (the item pool has been placed in the public domain). The American
Occupational Therapy Association (AOTA) has endorsed the AM-PAC for use by
its members for measuring daily activities outcomes because it most closely
conforms to occupational therapy scope of practice.
In addition, CMS is adding a small set of items based on the Participation
Measure for Post-Acute Care (PM-PAC), which attempts to measure participation
in a manner consistent with the International Classification of Function (ICF)
definition of participation.
Comment 3.7: Some commenters suggested that additional objective measures of
patient impairment and function (e.g., gait analysis, functional movement analysis, etc.)
be included in the data collection instrument, possibly as supplemental sections specific
to physical therapy.
Response: CMS weighed the value of items versus burden in developing these
data collection instruments based on input from a technical expert panel and
other experts. These experts felt that the items in the data collection instruments
address these impairment issues. However, an important component of the
analyses of the data collected will be to understand if additional items need to be
included in an instrument used for payment purposes.
4. Ability of Providers to Administer the Data Collection Instrument
Comment 4.1: Several commenters were concerned that completing the discharge
assessment for the CARE-C instrument will be difficult for providers. These commenters
remarked that patients “often” do not return for their last scheduled outpatient therapy
visit and that completing a discharge assessment would often not be possible.
Response: CMS understands that providers in community-based settings may
have difficulty administering an assessment upon discharge since, in these
settings, patients may not attend their final visit. In fact, one purpose of this data
collection is to understand the feasibility of collecting a discharge assessment in
community-based settings and use of those data in a payment system. However,
to make collecting these data more likely, CMS will ask that a discharge
assessment be completed on the last or second-to-last planned visit.
Comment 4.2: Some commenters felt that collection of data on admission will be
incomplete or misleading because the therapist may not have a complete understanding
of the patient’s medical history and therapy needs/plan of care by the first visit.

Response: CMS understands that a therapist may not have a complete
understanding of a patient’s condition at the first visit. CMS will therefore ask
that the admission assessment be completed within the first two visits.
5. Data Collection Methods
Comment 5.1: Two commenters were concerned that data collection should include the
full range of conditions for which patients receive outpatient therapy, not on a subset of
diagnoses/conditions.
Response: CMS agrees that it is important to include as much of the full range
of medical, functional, and cognitive issues found among the Medicare outpatient
therapy population as possible. The sampling plan for this data collection
specifically stratifies by setting (nursing facility, hospital outpatient, CORF/ORF,
and private practices). Furthermore, in provider recruitment, CMS will attempt to
identify the type of patient population the provider treats (e.g., orthopedic, poststroke, etc.) in order to include as much of the range of patients as possible.
Comment 5.2: Two commenters remarked on the sample design plan. One commenter
felt that provider types and disciplines should be sampled proportionally to their
Medicare volume. One commenter felt that equal numbers of cases from the three
therapy disciplines (physical therapy, occupational therapy, and speech/language
pathology) should be included in the sample.
Response: CMS is very sensitive to design issues; the data collection design
attempts to balance having sufficient representation of patients treated among all
outpatient therapy settings while not deviating excessively from proportional
representation. It is important that as much of the variation in patient clinical
condition be represented in the sample in sufficient number, especially those
patients with less common conditions. Sampling proportionate to Medicare
volume may result in some small subpopulations being insufficiently represented
(e.g., those with uncommon language disorders), and sampling disciplines in
strictly equal number may result in inefficient statistical estimates and reduced
ability to make inferences about the Medicare outpatient therapy population as a
whole.
Comment 5.4: Two commenters remarked that it is very important for there to be
standardized training procedures for providers participating in the data collection.
Response: CMS agrees with this comment and is developing training manuals
and procedures for data collection. In addition, CMS will hold several regional
training sessions around the country at the beginning of data collection.
Comment 5.5: Two commenters suggested that CMS develop electronic versions of
the data collection instruments to incorporate skip patterns and reduce the burden of
paper-based data collection.

Response: CMS understands that electronic data collection may be more
efficient for some providers, especially those already collecting patient
assessments (including patient self-assessment) electronically. In addition, skip
patterns in the assessment would be easier with electronic data collection since
these would skip automatically. However, many potential sample providers may
not have the ability to conduct electronic data collection, especially with patient
self-report. In CMS’ experience, even where electronic data collection is
preferred for patient assessment and self-report, many of the respondents collect
the data via paper and then enter the results at a centralized computer. Very few
providers have the space or resources to enter data electronically at the
immediate site of care. As a result, this data collection is paper-based. However,
providers are free to make an electronic version of the form and transmit the data
to CMS if there is a data format and transmission process acceptable to CMS
and its contractor that meets all CMS requirements for transmission of private
health information.
6. Data Analysis Methods
Comment 6.1: A few commenters had suggestions for the analysis of the data in this
one-time data collection. These include creating discipline- or setting-specific case mix
models as well as developing a plan to analyze missing data (item and section nonresponse).
Response: CMS agrees that the analysis design is important with many
important issues to consider. However, the purpose of this PRA comment period
is to address the need for and reasonableness of the proposed one-time data
collection and to assess respondent burden concerns. Data analysis will be
addressed at a later date through additional stakeholder involvement. CMS has
anticipated this need and has planned a future “CMS Open Door Forum” to
address data analysis.
7. Stakeholder Involvement
Comment 7.1: One commenter felt that the speech/language pathology community has
not been sufficiently represented in the study design and data collection methodology
development. Another commenter felt that the speech/language pathology community
had disproportionate influence on the collection methodology development, particularly
for swallowing, cognition, and communication function items.
Response: CMS believes that the speech/language pathology community was
given extensive involvement in the data collection instrument development
process through inclusion of ASHA outcomes measurement experts in the
technical expert panel and in an advisory group convened to modify the data
collection instruments after the expert panel met. These experts from ASHA were
involved extensively in the development of items related to conditions treated by
speech/language pathologists.

8. Specific Comments on the Data Collection Instrument Items
Comment 8.1: One commenter suggested that an item be included to identify where
the patient is receiving therapy.
Response: RTI will distribute paper data collection instruments to all
participating providers. Each instrument will have a code indicating the provider
to which the data collection instrument was sent. RTI will maintain a database of
participating providers indicating, among other data, the provider’s type of setting
and address of the particular provider location(s) participating.
Comment 8.2: One commenter suggested that CMS include the function items from the
CARE-F instrument on the CARE-C instrument to better measure function and changes
in function for lower-functioning community-based patients.
Response: CMS is changing the data collection plan for patients in day
rehabilitation programs, intended for lower-functioning community-based patients
to use the CARE-F instrument in those settings rather than the CARE-C
instrument. Clinicians in these programs will observe the patient for longer
periods of time during each visit and may be able to respond to those items
better than clinicians in clinic settings. In addition, CMS is concerned that adding
function items to the instrument will be too burdensome for providers.
Comment 8.3: A few commenters suggested changing the patient-reported function
items in the CARE-C instruments. Two providers were concerned that the mobility and
daily activities items did not have sufficient range to measure function in the highestand lowest functioning patients. One commenter suggested that the items be more
specifically be identified with specific ADLs.
Response: CMS is replacing some specific items in these sections to address
the concern about the range of function in which the items, taken together, are
sensitive. However, CMS feels that the three sets of function items cover the full
range of ADLs.
Comment 8.4: Two commenters suggested changes to the Primary Reason for
Therapy and Medical Condition items on both the CARE-C and CARE-F instruments to
increase the specificity of several response categories.
Response: CMS is refining several of the response categories based on the
comments received.