CMS-10537.FINAL - Supporting Statment Part A (10-29-14)

SUPPORTING STATEMENT CAHPS^® HOSPICE SURVEY

Introduction

The Centers for Medicare & Medicaid Services (CMS) requests a three-year clearance from the Office of Management and Budget (OMB) under the Paperwork Reduction Act of 1995 to implement the Consumer Assessment of Healthcare Providers and Systems (CAHPS^®) Hospice Survey (also called the Hospice Experience of Care Survey, HECS), and to conduct an assessment of the effects of survey administration mode. Under Contract Number HHSM-500-2014-00350G, the project team will implement and analyze a hospice experience of care survey for primary caregivers (i.e., bereaved family members or close friends) of patients who died while receiving hospice care (“decedents”). Specifically, CMS will (1) implement a survey to collect data on experiences of hospice care, and (2) conduct an experiment to examine effects of survey mode (i.e., mail-only, telephone-only, and mail with telephone follow-up).

A. Justification

A1. Necessity of Information Collection

In 2012, CMS launched the development of the CAHPS^® Hospice Survey, to measure the experiences of patients and their caregivers with hospice care. The survey was developed to:

• Provide a source of information from which selected measures could be publicly reported to beneficiaries and their family members as a decision aid for selection of a hospice program;

• Aid hospices with their internal quality improvement efforts and external benchmarking with other facilities; and

• Provide CMS with information for monitoring the care provided.

CMS developed the Hospice Survey following the same procedures we used for other CAHPS surveys. We started with a public call for measures published in the Federal Register (78 FR 5458). We reviewed the relevant literature regarding surveys of hospice patients and their caregivers. We also convened a technical expert panel made of up scholars, representatives of professional associations, industry representatives, and government experts. We then created three draft survey instruments for patients receiving hospice care in the home setting, the nursing home setting, and the in-patient setting (hospital or hospice stand-alone in-patient unit). The three questionnaires included very similar items, but also had items that applied specifically to each setting. We cognitively tested the questionnaires in stages, focusing on different issues for each stage.

Once the questionnaires were finalized we conducted a field test (OMB number 0938-1208). Data collection occurred from November 12 through December 13, 2013. The survey was administered between two and five months after the death of the hospice patient. Thirty-three hospice programs participated in the field test. (A copy of the development and field test report can be found in Appendix A.)

After the field test we did preliminary analysis of the data, including psychometric analyses and formation of composite measures. We also reviewed the questionnaires in light of the field test results. The results of that review were that the three questionnaires were consolidated into a single instrument, which included some questions for specific settings. We consolidated into a single instrument for ease of administration. Certain items were deleted from the questionnaire because they did not add to the composite measures. We also made minor wording changes in a few items. We then finalized the questionnaire.

CMS announced its intention to implement the CAHPS^® Hospice Survey in the FY 2014

Hospice Wage Index and Payment Rate Update; Hospice Quality Reporting Requirements; and Updates on Payment Reform. (79 FR 50452) According to the final rule, national implementation of the survey will launch on January 1, 2015 with hospices administering the survey for a “dry run” for at least one month in the first quarter of 2015. Starting

April 1, 2015 (second quarter), hospices will be required to participate on a monthly basis in order to receive the full Annual Payment Update (APU).

Hospices will be permitted to implement the survey in one of three survey administration modes: mail-only, telephone-only, and mail with telephone follow-up of non- respondents. In addition to national implementation of the survey, a mode experiment will be conducted to assess the degree to which results from the three modes of survey administration are comparable, and to develop analytic adjustments to compensate for any differences across modes if needed.

A2. Purpose and Use of Information

This survey supports the National Quality Strategy developed by the U.S. Department of Health and Human Services (HHS) that was called for under the Affordable Care Act to create national aims and priorities to guide local, state, and national efforts to improve the quality of health care. This strategy has established six priorities that support a three-part aim focusing on better care, better health, and lower costs through improvement. The six priorities include: making care safer by reducing harm caused by the delivery of care; ensuring that each person and family are engaged as partners in their care; promoting effective communication and coordination of care; promoting the most effective prevention and treatment practices for the leading causes of mortality, starting with cardiovascular disease; working with communities to promote wide use of best practices to enable healthy living; and making quality care more affordable for individuals, families, employers, and governments by developing and spreading new health care delivery models. Because the Hospice Survey focuses on experiences of care, implementation of the survey supports the following national priorities for improving care: engaging patients and families in care and promoting effective communication and coordination. In addition, national implementation and public reporting of Hospice Survey results will provide data on experiences with hospice care that enable consumers to make meaningful comparisons between hospices across the nation.

A3. Technological Collection Techniques

National implementation survey vendors will collect the data from primary caregivers (i.e., bereaved family members or close friends) of patients who died while receiving hospice care. Data will be collected via one of three modes of survey administration: mail-only (a mailed survey followed by a second survey mailed 21 days later), telephone-only (up to 5 telephone attempts), and mixed mode (mailed survey followed by up to 5 telephone attempts beginning 21 days later). The mailed survey will be formatted for data scanning, and data from all returned surveys will be scanned into an electronic data file. Computer Assisted Telephone Interviewing (CATI) will be used for the telephone mode.

A4. Identifying Duplication

The CAHPS^® Hospice Survey includes items addressing communication, care coordination, respect, spiritual and emotional support, symptom management, and overall care ratings. Some survey items were adapted from Nursing Home, Home Health and Clinician and Group CAHPS^® item sets and revised for the hospice care setting. Other items were adapted from the Family Evaluation of Hospice Care survey and revised to reflect CAHPS^® design principles for question wording and response options. Still other items are new, and reflect input from qualitative interviews with hospice caregivers, a technical expert panel, and/or a Call for Topic Areas published in the Federal Register. The survey is designed to gather only the data that CMS needs for assessing experiences with hospice care. To date, no standardized instrument and set of standard protocols has been used to collect such data from all hospices across the country for the purposes of comparability and accountability.

A5. Impact on Small Businesses

Survey respondents are primary caregivers of patients who died while receiving hospice care. The administrative procedures for the Hospice Survey state that small Medicare-certified hospices, defined as hospices that served fewer than 50 survey-eligible decedents/caregivers in the year prior to the survey administration period may be exempted from conducting the survey. This accounts for approximately 18% of Medicare certified hospices, according to CMS claims data for 2013. This means the survey should have no impact on the smallest approximately one-fifth of hospices.

A6. Consequences of Less Frequent Data Collection

The survey will be administered once to each primary caregiver following the death of his or her family member or friend.

A7. Special Circumstances

There are no special circumstances associated with this information collection request.

_{A8. CMS Federal Register Notice}

The 60-day Federal Register notice was published on May 8, 2014. We received comments on the proposed rule. Most of the comments were favorable. The most frequent comment questioned CMS’ proposed exclusion of respondents who were family caregivers of patients who died within 48 hours or less of their admission to hospice care. We responded that we excluded these patients/caregivers because it was not clear that caregivers would have enough time in hospice care to form opinions about it. In addition, our technical expert panel suggested that an abbreviated questionnaire would be more appropriate for these patients. We may potentially consider developing and implementing an abbreviated CAHPS® Hospice Survey in the future.

Other comments were made by single commenters.

A9. Respondent Payments or Gifts

This data collection will not include respondent incentive payments or gifts.

A10. Assurance of Confidentiality

CMS is providing sample cover letters for vendors to use when contacting potential respondents. The full text of the letter can be found in Appendix B.

If a vendor wants to provide the responses of individuals to the hospice, they must add a question to the survey requesting permission to share the individual’s responses with the hospice: The responses to this question are not reported to CMS.

“The hospice that provided care to your family member may want to review your answers so that they can decide how to address any concerns that you have. We will not share your answers to this survey linked to your name unless you give your permission for this information to be shared with the hospice.

“Do you give your permission to provide your answers to this survey linked to your name to the hospice?

“ Yes, I give my permission to share my name and survey responses with the hospice.

: No, I do not give permission to share my name and survey responses with the hospice.”

In addition the Quality Assurance Guidelines manual, which specifies survey protocols in detail, also lists requirements for the mailings (p 31)

“Mailings must include a personalized cover letter, a questionnaire, and a business reply envelope. The cover letters may be sent in both English and Spanish and may be two-sided, English on one side and Spanish on the other. Cover letters sent to respondents must be personalized with the decedent’s name, the caregiver’s name and the hospice’s name. The letter must also provide a toll-free number for respondents to call if they have questions. The cover of the questionnaire must include a label indicating the name of the hospice, and if applicable, may include the specific hospice inpatient unit, acute care hospital or nursing home facility in which their family member or friend resided.

“For CAHPS Hospice Survey administration, the OMB Paperwork Reduction Act language must appear in the mailing, either on the cover letter or on the front or back of the questionnaire. In addition, the OMB control number must appear on the front page of the questionnaire.”

The Quality Assurance Guidelines manual also includes similar language in the telephone script. (The manual is available on the survey web site, www.hospicecahpssurvey.org.)

A system of records notice is not relevant for the national implementation of the Hospice Survey because all data received by CMS is de-identified. The sample files used by vendors will include patient and caregiver information. However, these sample files are not submitted to CMS and the vendors are required to destroy sample files immediately after survey procedures using them are completed. Similarly, for the mode experiment, CMS will only receive de-identified data from our contractor.

A11. Sensitive Questions

As this survey requests information from bereaved family members or close friends of a deceased patient, the potential for distress is possible; however, CMS’ experience with a field test of the CAHPS^® Hospice Survey, and the extensive experience of the Department of Veterans Affairs with the Bereaved Family Survey and the National Hospice and Palliative Care Organization with the Family Evaluation of Hospice Care survey over the past ten years, indicates that this is a rare event in retrospective surveys with bereaved family members.

During the field test of the survey, the telephone interviewers reported that no respondents expressed distress. The field test included the same procedures proposed for national implementation of the survey. Field test interviewers were trained on these procedures to recognize and handle distressed respondents.

In addition CMS has taken steps to reduce the likelihood of distress among respondents. First, we have added a minimum two-month lag between the month in which the death occurs and the month in which vendors may start contacting caregivers. Here is an example: if an eligible patient died in January, the hospice would include them in their January sample. However, the vendor would not be able to contact the respondent until April 1. This is a lag of at least two months, depending upon when in January the patient died. We also require that the cover letter accompanying mailed surveys include a toll-free number which respondents may use for questions or concerns; should a respondent experience distress significant enough for him or her to request additional support, we recommend that survey vendors’ telephone staff put the respondent in contact with the appropriate local resource (generally a bereavement counselor or social worker on the hospice team that provided care to their family member or friend). Support for bereavement is among the services covered under the Medicare Hospice Benefit. In addition we require vendors to have an established distressed respondent protocol and that they train interviewers to use it.

A12. Burden of Information Collection

Estimated annualized burden hours and costs to respondents for the national implementation of the CAHPS^® Hospice Survey are shown in Tables 1 and 2. Per CMS’ final rule in FY 2014 Hospice Wage Index and Payment Rate Update; Hospice Quality Reporting Requirements; and Updates on Payment Reform, these estimates assume that only those hospices with 50 decedents or more in the prior year will sponsor administration of the CAHPS^® Hospice Survey. Specifically, we assume that the 2,326 hospices with 50 to 699 decedents in the prior year will administer the survey to all cases with an average of 151 cases per hospice; and that the 274 hospices with 700 or more decedents in the prior year will administer the survey to a sample of an average of 700 cases. Thus, we estimate that the CAHPS^® Hospice Survey will be administered to a maximum of 543,026 individuals in 2015 for the purposes of national implementation. As not all sampled cases will complete the survey, this estimate reflects the maximum

burden possible.

Table 1 shows the estimated annualized burden for the respondents' time to participate in the national implementation data collection. The survey contains 47 items and is estimated to require an average administration time of 10.4 minutes in English (at a pace of 4.5 items per minute) and 12.5 minutes in Spanish (assuming 20 percent more words in the Spanish translation), for an average response time of 10.505 minutes or 0.175 hours (assuming that 5% of survey respondents complete the survey in Spanish). These burden and pace estimates are based on CMS’ experience with surveys of similar length that were fielded with Medicare beneficiaries. As indicated below, the annual total burden hours for survey participants are estimated to be 95,029.55 for the national implementation of the survey.

_{Table 1. Estimated Annualized Burden Hours for Respondents: National}

_{Implementation of the CAHPS}^® _{Hospice Survey}

Survey Version		Number of Respondents	Number of Responses per Respondent	Hours per Response	Total Burden Hours
CAHPS® Hospice Survey		543,026	1	.175	95,029.55
	Total	543,026	1	.175	95,029.55

Table 2 shows the cost burden to respondents associated with their time to complete a survey as part of national implementation. The annual total cost burden is estimated to be

$2,163,823.

Table 2. Estimated Annualized Cost Burden for Respondents: National

_{Implementation}

Form Name	Number of Respondents	Total Burden Hours	Average Hourly Wage Rate*	Total Cost Burden
CAHPS® Hospice Survey	543,026	95,029.55	$22.77	$2,163,823
Total	543,026	95,029.55	$22.77	$2,163,823

*Based upon mean hourly wages, “National Compensation Survey: All United States December 2009 – January 2011,” U.S. Department of Labor, Bureau of Labor Statistics.

Estimated annualized burden hours and costs for the CAHPS^® Hospice Survey mode experiment are shown in Tables 3 and 4. For the mode experiment, we estimate that 60 hospices will sample 300 cases each. Thus, we estimate that the survey will be administered to a maximum of 18,000 respondents in 2015 for the purposes of the mode experiment. The annual total cost burden to respondents for participation in the mode experiment is estimated to be $71,725.50.

Table 3. Estimated Annualized Burden Hours for Respondents: Mode Experiment for the CAHPS^® Hospice Survey

Survey Version		Number of Respondents	Number of Responses per Respondent	Hours per Response	Total Burden Hours
CAHPS® Hospice Survey		18,000	1	.175	3,150
	Total	18,000	1	.175	3,150

_{Table 4. Estimated Annualized Cost Burden for Respondents: Mode Experiment}

Form Name		Number of Respondents	Total Burden Hours	Average Hourly Wage Rate*	Total Cost Burden
CAHPS® Hospice Survey		18,000	3,150	$22.77	$71,725.50
	Total	18,000	3,150	$22.77	$71,725.50

*Based upon mean hourly wages, “National Compensation Survey: All United States December 2009 –

January 2011,” U.S. Department of Labor, Bureau of Labor Statistics.

In addition to respondent burden, both hospices required to participate in national implementation of the survey, and those exempted from participation, will face a burden.

Per CMS’ final rule in FY 2014 Hospice Wage Index and Payment Rate Update; Hospice Quality Reporting Requirements; and Updates on Payment Reform, hospices with 50 decedents or more in a year will be required to administer the CAHPS^® Hospice Survey. The burden to these hospices will consist of their time and effort to prepare and submit decedent data files to their approved CAHPS^® Hospice Survey vendor. The data files will contain data on patients who died in the hospice’s care in the prior month, and will be used by the survey vendor to select the sample and field the survey. Hospices will use existing databases to generate these files, and should be able to do so with minimal effort. To keep the burden to hospices as low as possible, the list of required data elements for the file will be as parsimonious as possible.

Table 5 shows the estimated annualized cost burden for the provision of the monthly data file by hospices required to participate in national implementation. We estimate that preparation and submission of the monthly file will take 16.0 hours annually for each hospice. CMS arrived at this estimate after discussing our sampling protocol and variables with our technical expert panel. The panel advised us that the majority of hospices would have the electronic capability to pull a sample list and associated variables and that the biggest concern would be establishing the initial query. Based on the information we received, CMS allowed four hours for setting up the query and one hour per month for running the query.

Assuming that 2,600 hospices will be eligible to conduct the CAHPS^® Hospice Survey, the burden would be 41,600 hours (2,600 hospices * 16 hours). CMS believes that this labor can be conducted by a Medical Records Reviewer. The U.S. Bureau of Labor Statistics estimates that the hourly wage of a Medical Records Reviewer is $24.94. Therefore, the annual cost of the wage labor would be $399.04 for each hospice (16 hours * $24.94) for a total of $1,037,504.00 (41,600 hours * $24.94 per hour) for all hospices participating in national implementation.

Table 5. Estimated Annualized Cost Burden to Hospices Eligible to Participate in

_{National Implementation}

Form Name		Number of Hospices	Total Burden Hours per Hospice	Average Hourly Wage Rate*	Total Cost Burden
Data File for CAHPS® Hospice Survey		2,600	16.0	$24.94	$1,037,504.00
	Total	2,600	16.0	$24.94	$1,037,504.00

*Based upon mean hourly wages for medical & clinical laboratory technologists, “National Compensation

Survey: All United States December 2009 – January 2011,” U.S. Department of Labor, Bureau of Labor

Statistics.

Table 6 shows the estimated annualized cost burden for the provision of the monthly data file by hospices participating in the mode experiment. We estimate that those hospices participating in the mode experiment will incur an additional burden of 4 hours of Medical Record Reviewer time in order to provide a data file to RAND Corporation, which will be acting as the survey vendor for the mode experiment. Assuming that 60

hospices will be participating in the mode experiment, the burden would be 240 hours (60 hospices * 4 hours). The annual cost of the wage labor would be $99.76 for each of the

60 hospices participating in the mode experiment (4 hours * $24.94) for a total of

$5,985.60 (240 hours * $24.94 per hour) for all hospices participating in the mode experiment.

Table 6. Estimated Annualized Cost Burden to Hospices Eligible to Participate in the Mode Experiment

Form Name		Number of Hospices	Total Burden Hours per Hospice	Average Hourly Wage Rate*	Total Cost Burden
Data File for CAHPS® Hospice Survey		60	4.0	$24.94	$5,985.60
	Total	60	4.0	$24.94	$5,985.60

*Based upon mean hourly wages for medical & clinical laboratory technologists, “National Compensation

Survey: All United States December 2009 – January 2011,” U.S. Department of Labor, Bureau of Labor

Statistics.

Table 7 shows the estimated annualized cost burden for hospices applying for exemption from participation in national implementation. Hospices with fewer than 50 decedents in the prior year will need to file an exemption form on the CAHPS^® Hospice Survey web site. We estimate that preparation of the annual decedent count and completion of

the form will take 20 minutes. Assuming that there are 1,101 hospices with fewer than

50 decedents in the prior year, the burden would be 367 hours (1,101 hospices * 0.333 hours). CMS believes that this labor can be conducted by a Medical Records Reviewer. The annual cost of the wage labor would be $8.23 for each hospice (0.33 hours * $24.94) for a total of $9,152.98 (367 hours * $24.94 per hour) for all hospices completing the exemption form.

Table 7. Estimated Annualized Cost Burden to Hospices Applying for Exemption from Participation in National Implementation

Form Name		Number of Hospices	Total Burden Hours per Hospice	Average Hourly Wage Rate*	Total Cost Burden
Exemption from CAHPS® Hospice Survey		1,101	0.33	$24.94	$9,061.45
	Total	1,101	0.33	$24.94	$9,061.45

*Based upon mean hourly wages for medical & clinical laboratory technologists, “National Compensation

Survey: All United States December 2009 – January 2011,” U.S. Department of Labor, Bureau of Labor

Statistics.

A13. Capital Costs

Survey participants will incur no capital costs as a result of participation. Hospices participating in national implementation will bear an annual cost to secure the services of an approved CAHPS^® Hospice Survey vendor to conduct the survey on their behalf. Table 8 summarizes the estimated annual cost burden to hospices of paying the survey vendor. We estimate that the average annual survey vendor cost will be $3,300 for each of the 2,600 hospices participating in national implementation of the CAHPS^® Hospice Survey, for a total of $8,580,000 (2,600 hospices * $3,300). The CMS cost estimate is based on information from hospice experts familiar with the costs hospices have experienced on other surveys. We also used information we have developed regarding charges for other CAHPS surveys. Hospices currently have a total of 32 approved vendors from which to choose. Typically the cost of a survey varies among vendors, as well as with the mode of administration (mail tends to be least expensive), with the size of the sample (larger samples cost more), and with the types of additional services provided by the vendor. Our estimate of $3,300 is an attempt to average the costs across the full range of hospice sizes, survey modes, and services.

The CAHPS Hospice Survey vendor would be responsible for conducting the CAHPS survey, but might or might not be responsible for additional survey efforts. Hospices may conduct other surveys as long as the CAHPS survey is received first and as long as the CAHPS sample is drawn first.

Table 8. Estimated Annualized Cost Burden to Hospices for Contracting a Survey

_{Vendor for National Implementation of the CAHPS Hospice Survey}

Form Name		Number of Hospices	Average Cost to Contract a Survey Vendor	Total Cost Burden
CAHPS® Hospice Survey		2,600	$3,300	$8,580,000
	Total	2,600	$3,300	$8,580,000

Please note that there is no additional vendor cost for hospices participating in the mode experiment, as these costs are borne by CMS via a contract to RAND Corporation, as indicated in Section A14.

A14. Cost to the Federal Government

The annual cost to the Federal Government for sampling, data collection, analysis and reporting of scores is $1,305,405. The annual cost to the federal government is a result of a contract awarded to RAND Corp for the implementation of the survey. The contract includes such activities as operation of a data warehouse for collecting survey results from vendors, technical assistance to vendors and hospices, training for vendors and hospices, web site maintenance of the survey web site, vendor oversight, and data checks. In addition, RAND will provide analysis as requested by CMS. The cost reported here is the contract amount for the base year of the contract. There is additionally one option year.

A15. Program Changes or Adjustments to Annual Burden

This is a new information collection request.

A16. Tabulation and Publication of Results

We anticipate that the analysis plan will include (1) psychometric evaluation of the survey items; (2) development of sampling and nonresponse weights (3) development and evaluation of case-mix adjustment; (4) development of adjusted hospice-level results, (5) development of national, regional, and subgroup estimates, and (6) mode experiment analysis. All aspects of these analyses will be described in a final project report to CMS.

(1) Psychometric Evaluation. Analyses will include evaluation of item missing data, item distribution (including ceiling and floor effects), and assessment of hospice-level reliability of items. We will compute these statistics overall, and separately by care setting, mode of administration, and language, computing mean scores for composites and global rating items.

(2) Weighting. Analyses will include the calculation of (a) Sampling weights to reflect

the probability that each primary caregiver is selected for the survey and (b) nonresponse weights to reflect the probability that a sampled caregiver responds to the survey. We will also calculate and consider for use poststratification weights to make the characteristics of the respondent sample more similar to the overall population. The sample frame will include administrative data on all hospice patients. The frame will be supplied by the hospice to the vendor.

(3) Case-mix adjustment and nonresponse. In consultation with CMS, we will consider mixed effect regression models of performance measures for hospice care. These models would include fixed effects for case-mix adjustors, such as age, education, and primary diagnosis.

(4) Adjusted Hospice-Level Estimates. We will produce case-mix adjusted estimates of experience at the hospice-level.

(5) National, Regional, and Subgroup Estimates. RAND will use adjustments as appropriate, to produce national and regional estimates of experience with hospice care.

(6) Mode Experiment Analysis. We will compare response rates and measure scores with and without mode adjustment by survey mode on the basis of cases being randomized by mode within hospice.

Publication of Results: CMS intends to publicly report hospice-level data from this survey, and to provide preview reports directly to hospices.

A17. Display of OMB Expiration Date

The expiration date for OMB approval of this information collection will be displayed on the survey.

A18. Exceptions to the Certification Statement

There are no exceptions to the certification statement identified in item 19 of OMB Form

83-I associated with this data collection effort.

Attachments:

Appendix A: Hospice Experience of Care Survey Development and Field Test Report

Appendix B: Sample Cover Letter

12