Justification for Change - Revised

Request for Approval of a Non-Substantive Change:

Data Collection for the Residential Care Community and Adult Day Services Center Components of the National Study of Long-Term Care Providers

OMB No. 0920-0943 Exp. Date: 07/31/2015

March 11, 2014

Lauren Harris-Kojetin

Chief, Long-Term Care Statistics Branch

Division of Health Care Statistics

National Center for Health Statistics

Phone: 301.458.4369

Fax: 301.458.4693

Email: [email protected]

Data Collection for the Residential Care Community and Adult Day Services Center Components of the National Study of Long-Term Care Providers

A1. Justification Circumstances Making the Collection of Information Necessary

This request is for a nonsubstantive change to an approved project (OMB No. 0920-0943 Exp. Date: 07/31/2015), data collection for the residential care community (RCC) and adult day services center (ADSC) components for the National Study of Long-Term Care Providers (NSLTCP), including the estimated sample size and estimated annual burden. We conducted data collection in 2012, and based on what we learned from this experience, we would like to make a few nonsubstantive changes. As we did in 2012, the data to be collected from RCCs and ADSCs will include basic characteristics, services offered, staffing, and practices of providers, as well as distributions of the demographics, physical functioning, and cognitive functioning of users (RCC residents and ADSC participants) aggregated to the RCC/ADSC level. The 2014 NSLTCP survey will be administered by mail, web, and telephone, as we did in 2012. As we did in 2012, data will be collected from a sample of 11,690 RCCs and a census of 5,000 ADSCs in the 50 states and the District of Columbia.

Changes for 2014 based on 2012 experience

• Revise RCC sampling

• Drop and revise select questionnaire items on RCCs and ADSCs that were fielded in 2012.

• Add questionnaire items on number of hours worked by staff, resident diagnoses, and falls

• Revise NCHS cover letters that will be sent to survey respondents, streamline questionnaire mailing materials and change mode of mailings from FedEX to USPS

• Lengthen field period

• Drop three methods experiments

• Add follow-up calls to 1,000 respondents after mailings to provide technical assistance and answer questions.

• Revise approach to determining which cases go to computer-assisted telephone interviewing (CATI)

A comparison of the proposed 2014 questions and the 2012 questions that were fielded is in Attachment A, and the 2014 NSLTCP questionnaire items are in Attachment B. The revised NCHS letters are presented in Attachment C. The questionnaire emailing follow-up is in Attachment D.

A2: Purpose and use of information collection

In 2012, we aimed to achieve a 65% weighted response rate for both ADSCs and RCCs. We achieved a 67% response rate for ADSCs a 55% weighted response rate for RCCs. We also had an ineligibility rate for RCCs that was higher than we expected. In this multi-mode survey, we had fewer completions by mail and web than anticipated at the point in the field period when CATI non-response follow-up started, and fewer completions by CATI than expected, both of which resulted in higher than expected survey costs. For these reasons, we are proposing to make a few nonsubstantive changes for the 2014 NSLTCP wave, in order to try to improve RCC response rates, maintain ADSC response rates, and better control and predict field costs.

First, we would like to make some revisions to our RCC sampling strategy. The 2012 NSLTCP sampling design assumed the RCC strata-specific (4 bedsize strata) response rates and eligibility rates achieved in the 2010 National Survey of Residential Care Facilities, which was designed only to produce national estimates (not state estimates). To benefit from the more precise state-level experience in 2012, the 2014 NSLTCP sampling design assumes 2012 NSLTCP response rates and eligibility rates, by state and RCC size. For the 2012 design, we assumed that estimates will be made with a 15% RSE for a 30% estimate, therefore requiring a minimum of 104 completes for all states where possible and a census in states with insufficient sample to meet the 104 minimum.  For the 2014 wave, we assumed a 17% RSE for a 30% estimate requiring 81 completes per state and a census for states with insufficient sample to meet the minimum of 81 completes.  These assumptions have resulted in a sample size of 10,245. Since NCHS budgeted for a sample size of 11,690 (the same as the 2012 sample size), we reallocated the remaining cases to states that were designated to be sample states. Our goals in reallocating the additional sample were to have as many census states as possible, and have the same 30 census states in 2014 as we had in 2012.  This allocation resulted in a total of 29 census states (all 2012 census states included with the exception of Kansas for which there were not enough extra sample cases to make it a census state).  The 2014 design has a total sample size of 11,663.

As with 2012, the data to be collected in 2014 from RCCs and ADSCs will include basic characteristics, services offered, staffing, and a profile of the demographics, physical functioning, and health status of residents/participants aggregated to the RCC/ADSC level. However, we would like to make some revisions to the questionnaires for 2014, mainly to (a) shorten the questionnaire in hopes of increasing response rates and (b) revise some items to improve their measurement validity. Our data contractor for the 2012 NSLTCP, RTI International, and NCHS’ QDRL (OMB No. 0920-0222) each conducted rounds of cognitive interviewing using the 2012 NSLTCP questions, and identified problematic questions within each of the questionnaire sections. Based on their recommendations, we are proposing to drop, revise and add items for 2014. We plan to rotate back in some of the dropped items in the 2016 wave. We propose the following changes for 2014, by questionnaire section (unless specified, changes refer to both ADSC and RCC):

Background Information

• Drop items on continuing care communities, number of respite care residents/participants, other sources of funding (ADSC question), organization affiliation, and other type of provider license (ADSC question).

Services Offered

• Drop items on dementia care staffing (RCC questions), case management services, and person-centered care.

• Revise items on depression screening and services.

Staff Profile

• Drop number of full-time equivalents (FTEs) for all staffing categories.

• Add items on number of hours worked by full- and part-time staff.

Resident/Participant Profile

• Drop items on living arrangements (ADSC question), cognitive impairment screening, needing assistance with bed transferring (ADSC only), use of a wheelchair or scooter, hospital readmissions, movement in/out of RCCs and ADSCs, and deaths.

• Add cardiovascular disease and diabetes to diagnoses and an item on falls.

Record Keeping

• Drop capabilities for tracking resident/participant information.

• Add hospital to electronic health information exchange item.

A comparison of the fielded 2012 questions and proposed 2014 questions are in Attachment A. The 2014 questionnaire items are in Attachment B.

During the 2012 field period, some directors/administrators from sampled providers (via the survey help line, feedback to CATI interviewers) noted that they were receiving too many FedEx packages too close in time to each other. In response to this feedback, in 2014 we are sending the questionnaire mailings by USPS instead of FedEx, and lengthening the overall field period from less than 6 months to 7 months, which allows us to lengthen the time between mailings. We will also use project-specific mailing envelopes, in hopes that they may appear more appealing than generic FedEx envelopes, resulting in respondents being more likely to open them. Compared to 2012, for 2014 we are also (a) streamlining mailing materials, to include the minimal set of documents needed to enable the resident to complete and mail back the questionnaire; and (b) cutting text from the cover letters for the questionnaire mailings, so that the font can be enlarged to increase readability. Now that we have 2012 data available, we also plan to include a link to 2012 results in the 2014 letters. One of the 2012 experiments showed that providing a deadline and consequence results in getting questionnaires back sooner, so we are adding deadlines in the letters (e.g., submit your questionnaire by DATE in order to avoid receiving another questionnaire).

We conducted three methods experiments in 2012 (drive to web, deadline with consequence, promise of non-financial incentive). Because the deadline with consequence results showed a positive effect, we are incorporating this into the 2014 contact materials. The promise of a tailored report made no difference in response rates, so we are not offering this in 2014. The drive to web experiment results showed that cases that received materials that highlighted only the web option in the first two mailings had a lower response rate than cases whose materials highlighted both mail and web. As a result, we are not using a drive to web strategy in 2014; the contact materials for all mailing to all cses will highlight both mail and web as options for responding.

In the 2014 wave, we would like to add follow-up calls to 1,000 ADSC and RCC cases after the first questionnaire mailing to confirm receipt of the questionnaire, provide technical assistance, and answer any questions. We hypothesize that these calls will increase the number of web and mail completes earlier in the field period, resulting in fewer casese needing CATI follow-up in 2014 compared to 2012.

In 2012, we included CATI follow-up for non-responders. We sent all cases to the CATI system that had not responded via mail or web by the time the CATI phase started. In 2014, we would like to revise our CATI strategy by delaying the start of CATI and sending fewer cases to the CATI system. We will target the cases we send to CATI by looking at response rates by strata and focusing on characteristics among RCCs and ADSCs with low response, such as size and state.

We intend that these changes to our sampling; questionnaires; respondent contact materials; field period, type of mailing, and spacing between mailings; and approach to CATI will help increase response rates, decrease eligibility issues, lessen missing data, result in lower and more reliable field costs, and increase the reliability and validity of NSLTCP.

A12. Estimates of Annualized Burden Hours and Costs

A. Burden Hours

Table 1 includes the average annual burden for data collection over the three year clearance with our 2014 changes. We learned from the 2012 NSLTCP that on average it took more time to gather records than we thought. However, for 2014 we estimate the burden for each response to the questionnaire will be the same as 2012 (30 minutes), because we shortened the questionnaire and revised problematic questions. In 2014, 1,000 ADSC and RCC Directors will receive a follow-up call after the first questionnaire mailing. We expect this call to take on average 10 minutes with a total burden of 167 hours. The new total estimate of annualized burden is 9,044 hours.

Table 1: Estimated Annualized Burden Hours

Type of Respondent	Form Name	Number of Respondents	Number of Responses	Average Burden/ Response (in minutes)	Response Burden (in hours)
RCC Director	RCC Questionnaire	11,663	1	30/60	5,832
ADSC Director	ADSC Questionnaire	5,254	1	30/60	2,627
RCC and ADSC Directors	Data Retrieval	1,670	1	15/60	418
RCC and ADSC Directors	Follow-Up Call	1,000	1	10/60	167
Total	9,044

A15. Explanations for Program Changes or Adjustments

This package was originally approved for 8,769 annualized burden hours for three years, of which 8,769 was indeed utilized for year one. In 2014, 1,000 ADSC and RCC Directors will receive a follow-up call after the first questionnaire mailing. We expect this call to take on average 10 minutes with a total burden of 167 hours. The new total estimate of annualized burden is 9,044 hours. Over the remaining two years of burden (representing a balance of 17,538 hours), only one additional wave of data collection is planned, for which 9,044 hours of anticipated burden will be utilized.

B2. Procedures for the Collection of Information

As with the 2012 wave, the 2014 wave includes NCHS letters that will be included in a series of survey mailings. For 2014, we would like to make minor revisions to the letters so that they are shorter, updated, and include a deadline for respondents. We would like to minimize redundant text between the letters, and revise them so they emphasize the 2014 wave of data collection. The proposed letters are in Attachment C.

B3. Methods to Maximize Response Rates and Deal with Nonresponse

In the 2012 wave of NSLTCP, we conducted three methods experiments to maximize response rates; we would like to delete these for the 2014 wave. Findings from these experiments will be disseminated at conferences and in journals (see the AAPOR 2013 conference as an example https://www.aapor.org/Recent_Conferences1/6230.htm).

In the 2014 wave, we would like to add follow-up calls to 1,000 ADSC and RCC cases after the first questionnaire mailing to confirm receipt of the questionnaire, provide technical assistance, and answer any questions. We hypothesize that these calls will maximize our response rates.

Attachments

Attachment A-1: Comparison of 2012 and 2014 ADSC Questionnaire Items

Attachment A-2: Comparison of 2012 and 2014 RCC Questionnaire Items

Attachment B-1:  2014 NSLTCP ADSC Questionnaire Items

Attachment B-2:  2014 NSLTCP RCC Questionnaire Items

Attachment C-1:  2014 NSLTCP NCHS Advance Notification Letters

Attachment C-2:  2014 NSLTCP NCHS First Mailing Letters

Attachment C-3:  2014 NSLTCP NCHS Follow-up #1 Letters

Attachment C-4:  2014 NSLTCP NCHS Follow-up #2 Letters

Attachment C-5:  2014 NSLTCP NCHS Thank You Reminder Letters

Attachment C-6:  2014 NSLTCP NCHS Trouble Reaching You Letters

Attachment C-7:  2014 NSLTCP NCHS Chain Packet Letters

Attachment D: Questionnaire Mailing Follow-Up Call

Attachment I: Data Retrieval Call

7