SUPPORTING STATEMENT NH SOPS Part B 2-21-07

SUPPORTING STATEMENT

Voluntary Questionnaire and Data Collection Testing

to Pilot Test Proposed Nursing Home Survey on Resident Safety

B. STATISTICAL METHODS

1. Potential Respondent Universe and Sample Selection Method

The potential respondent universe will be all employees, contractors and agency staff of all job classes/levels in each nursing home. A purposive sample of 40 nursing homes will be recruited, distributed across bed size and urbanicity (Table 2). This number of sites is needed to allow for multi-level confirmatory factor analyses of the survey data at both the individual and nursing home levels. Efforts will be made to recruit nursing homes distributed across geographic region and type of ownership (proprietary, non-profit, government, and other). Recruited nursing homes will be allocated to each category in numbers roughly proportionate to the national distribution of nursing homes in those categories.

TABLE 2. PURPOSIVE SAMPLE OF FACILITIES FOR PILOT OF NURSING HOME SAFETY CULTURE SURVEY

Group	Bed Size	Urban Setting	Number of Facilities Needed
A	< 50	Rural	2
B	< 50	Urban	3
C	50 – 99	Rural	7
D	50 – 99	Urban	8
E	100 – 199	Urban	9
F	100 – 199	Rural	7
G	200+	Urban	4
Total			40

In nursing homes with 200 employees or less, a census of employees will be surveyed. In nursing homes with more than 200 employees, a random sample of 200 employees will be surveyed. It is estimated that approximately 5,500 nursing home staff will be surveyed for the pilot test. The targeted response rate is 70%. Job titles of each employee surveyed will be obtained to examine survey non-response by job title at the conclusion of data collection to determine if there are differences in response rates by job title.

The survey being field tested is an expansion of the Hospital Survey on Patient Safety Culture, which has been in use for approximately two years. Therefore, the purposive sample of nursing homes with variation in bed size, urbanicity, geographic region, and type of ownership, provides for a meaningful field test.

2. Information Collection Procedures

The data collection will include these steps:

• Mailing of survey packets (which includes a cover letter) to the nursing home contact

• Nursing home point of contact distributes survey packets to nursing home staff

• Reminder postcard distributed to nonrespondents

• Distribution of a second questionnaire to nonrespondents

Attachment B contains example data collection materials, including the cover letter, reminder postcard, and the draft pilot survey.

3. Methods to Maximize Response Rate

We are using incentives as noted in #9 on page 3, and we are following up with a reminder postcard and second questionnaire to nonrespondents.

4. Tests of Procedures

AHRQ and Westat have taken practical steps to consult with nursing home employees to help minimize the burden of the collection of information and to develop a better, more reliable instrument. The steps taken have primarily been in the form of cognitive interviews that serve as survey pretests. In-person cognitive pretests were conducted with persons employed at a rural nursing home, an urban nursing home, and a suburban nursing home. Cognitive testing participants were selected from a variety of staff categories, including administrative, maintenance, food service, linen, and nursing. Revisions to the survey were made based on feedback provided during this pretesting, in addition to input from nursing home researchers and experts in long term care to arrive at the version of the survey instrument that will be pilot tested.

5. Statistical Consultation and Independent Review

Input from statistical analysts was used to develop the study design and plans for data analysis, including:

Veronica Nieva, PhD, Vice President, Westat, 301-251-8271

Joann Sorra, PhD, Senior Study Director, Westat, 301-294-3933

Naomi Dyer, PhD, Research Analyst, Westat, 301-610-8842

Paul Hanges, PhD, University of Maryland College Park, 301-405-5930