Supporting Statement Part B

Supporting Statement B Field Test of Survey of End-of-Life Care – 0935-0124

TABLE OF ContentS

SUPPORTING STATEMENT: 1

Field Test of Survey of End-of-Life Care 1

B. Collection of Information Employing Statistical Methods 1

B1. Respondent Universe, Sampling and Respondent Selection 1

B2. Data Collection Procedures 3

B3. Methods to Maximize Response 3

B4. Tests of Procedures or Methods 3

B5. Statistical and Data Collection Consultants 4

References 5

SUPPORTING STATEMENT:

Field Test of Survey of End-of-Life Care

B. Collection of Information Employing Statistical Methods

B1. Respondent Universe, Sampling and Respondent Selection

The respondent universe for this field test will be caregivers of adult patients (i.e., age 18 or older) who died while receiving care from a large integrated healthcare system. The sample will be selected at random by RAND staff experienced in sample selection for CAHPS surveys, using the universe of caregivers provided by the integrated healthcare system. The integrated healthcare system will provide RAND with a sample file drawn from their administrative records. This sample file will contain a list of patients who have died in the previous 8 months, along with data elements that will allow RAND to identify patients eligible for the survey and to administer surveys to their caregivers.

Determining Eligibility

Pre-Sampling Eligibility. Caregivers of adult patients (i.e., age 18 or older) who died in the last 4 to 8 months, and who had at least two health care visits in the year prior to death, will be eligible to take the survey. Caregivers will be identified as the primary healthcare decision maker, primary caregiver, or emergency contact for the patient; if no such individuals are known, the survey will be sent to the next of kin at the decedents’ last known address. 

Sampling

We propose to draw a sample of 1,700. We assume a response rate of 30%, based on the reported response rate to a similar survey conducted among bereaved caregivers in an integrated health care system (Wang et al., 2019). At this response rate, the proposed sample would yield approximately 510 completed surveys.

The proposed sample size will allow us to make empirical comparisons between the care experiences of (a) those who received palliative or hospice care and those who did not and (b) those who completed an advance directive and those who did not.

Table 1 describes the expected distribution of the proposed sample of 1,700 across these subgroups of interest. A recent study of end-of-life care within an integrated health system reported that slightly more than half of caregivers were responding for patients who received palliative care or for those enrolled in hospice (56% and 52%, respectively), slightly more than one quarter (27%) for those who received neither palliative nor hospice care, and 44% for those with an advance directive on record (Wang et al., 2019).

Table 1. Sampling Plan by Subgroups of Interest

	Expected Sample Sizea	Completed Surveys (Assumes 30% Response Rate)	Completed Surveys Eligible for Analysisb
Overall	1,700	510	485
By Subgroups of Interest:
Health Care at End of Life
Palliative or hospice care	1,241	372	372
Neither palliative nor hospice care	459	138	113
Documented Care Plan
Advance directive on record	748	224	184
No advance directive on record	952	286	234

^a Number of sampled cases that meet pre-sampling eligibility criteria: age 18 or older at time of death, and two health care visits in the year prior to death.

^b Number of completed surveys that are eligible for inclusion in survey data analysis, after removal of those who are ineligible post-sampling because their survey responses indicated that: they never oversaw or were involved in the patient’s care; that the patient did not received health care in the last month of life; and that the patient’s cause of death was an injury or accident. The expected post-sampling ineligibility rate is 5% overall (0% for those in palliative or hospice care and 18% for all other subgroups).

Since some information regarding eligibility for the survey will not be available at the time of sampling, we will use survey responses to apply the following post-sampling eligibility criteria. Cases will be excluded from analysis if the caregiver indicates that:

• They “never” oversaw or took part in the patient’s care

• The patient did not receive care from a health care provider in the last month of life

• The patient died of accident or injury AND did not have a chronic underlying condition. Those who died of accident or injury but did have a chronic underlying condition will be included.

We assume that all caregivers for those receiving palliative or hospice care will be deemed eligible for the survey. For all other groups (i.e., caregivers for patients who did not receive palliative or hospice care), we assume that 18% of cases will be deemed ineligible post-sampling.

The proposed sample size supports overall assessments of ceiling and floor effects, overall estimation of internal consistency reliability, and assessments of the correlations between reports of care experiences and overall ratings of care. When comparing responses to survey items, the sample design will allow us 80% power to detect small differences (Cohen’s d=0.27 to d=0.31 SD) between subgroups of interest within a given category (e.g., those who received palliative or hospice care versus those who did not).

B2. Data Collection Procedures

Data will be collected using a mixed mode (mail-telephone) survey protocol. After sampling, data collection will begin with a survey packet sent to sampled caregivers. The packet will contain a personalized cover letter, a self-administered survey, and a pre-addressed business reply envelope. The letter will include a statement informing caregivers that Spanish-language materials are available and instructions to follow to request those materials. Three weeks after the survey packet mailing, follow-up calls will begin to sample caregivers who have not previously completed the mail survey, refused to participate, or been determined to be ineligible. The telephone survey will be available in both English and Spanish. We will close data collection three weeks after telephone data begins, six weeks after the survey packet mailing.

B3. Methods to Maximize Response

We will strive to maximize the survey response rate while retaining the voluntary nature of the field test. Several aspects of our field test procedures are associated with increasing response to mail surveys (Dillman et al., 2014):

• Using mixed-mode data collection,

• Personalizing letters and survey packets,

• Using first class postage, and

• Providing a survey translation.

B4. Tests of Procedures or Methods

This data collection effort includes:

1. A test of the newly designed Survey of End-of-Life Care. This survey is being fielded for the first time under this data collection effort.

2. A test of care experiences across patient subgroups of interest. This includes empirical comparisons of responses between those who received palliative or hospice care and those who did not and those who had an advance directive on record and those who did not.

B5. Statistical and Data Collection Consultants

The survey, sampling approach, and data collection procedures were designed by the RAND Corporation under the leadership of:

Ron D. Hays, PhD

Senior Policy Researcher

RAND Corporation

1776 Main Street

Santa Monica, CA 90407

Marc N. Elliott, PhD

Senior Principal Researcher

RAND Corporation

1776 Main Street

Santa Monica, CA 90407

References

Dillman, D. A., Smyth, J. D., Christian, L. M., & Dillman, D. A. (2014). Internet, mail, and mixed-mode surveys: The tailored design method. Hoboken, N.J: Wiley & Sons.

Wang, M.C., Johnson, D., Glass, D., Martel, H., Mularski, R.A., Bellows, J. (2020). Assessing End-of-Life Care across Settings in an Integrated

Health Care Delivery System. J Palliat Med. 23(6):825-828.