Multi-mode screening description

Attachment 2a

National Health and Nutrition Examination Survey (NHANES)

Multi-Mode Screening Feasibility Study Description

NHANES Multi-Mode Screening Feasibility Study

Eligibility: All households in NHANES PSUs are eligible. The maximum number of respondents would be 2060.

Informed Consent: Informed consent will be obtained as part of the regular National Health

and Nutrition Examination Survey (NHANES) consent process.

Exclusion Criteria: There are no exclusion criteria for this project.

Data Collection: NHANES screener data will be collected using an online questionnaire for this feasibility study. This web screener collects the same information as what is currently approved for use in NHANES and done in-person with an field interviewer. The feasibility study will be conducted in six NHANES locations. The in-person screening procedures used in NHANES are not changing during this feasibility study.  All households from released sample will still be screened in person in accordance with the existing NHANES protocol.  During this project, a randomly selected subset of households, in the selected locations, will also complete the web based screener. Households randomly selected for web screening will be offered the web screener before the in-person screening. On-line screening results will be validated in-person by an NHANES field interviewer. Success would be indicated by a high-level of agreement between the web screener information and the in-person screener information, ideally 80% for SP selection (i.e. both the web screener and the in-person screener would include or exclude the same participants 80% of the time).

Report of Findings: This is a screening project. As such, there are no findings to report.

National Health and Nutrition Examination Survey (NHANES)

Multi-Mode Screener Feasibility Study ERB Protocol

Overview:

Currently NHANES only conducts in-person screenings of potential participants. For several reasons, NHANES would like to test, and if successful, implement a multi-mode (online and in-person) screening approach. The primary reason would be to identify and more quickly screen out households which do not contain persons eligible to be in NHANES. Interviewers who spend less time screening could instead spend more time with individuals who screened into the survey, thereby potentially saving time and money. The language of the screener would not change.

Pilot Overview:

The Multi-Mode Screener Pilot Test will be conducted at six primary sampling units (PSUs), also known as stand locations, for roughly 2,060 households. All dwelling units (DUs) in the initially released subsample(s), typically one half to two thirds of the total DUs in PSU sample, will be included in the Pilot Test and the related analysis. Pilot Test DUs in each segment will be randomly split into two groups: Group 1 receives the Web Screener mailings; and Group 2 receives the currently approved NHANES Advance Letter and does not receive Web Screener mailings. The current NHANES In-person field interviewer protocol will be the same for both groups. No incentives for completing the screener will be offered to respondents for this pilot. If successful, a future pilot of screening incentives may be conducted.

Research questions we hope to answer by conducting the Pilot Test:

• Does web screening result in data quality equivalent to in-person screening? Does either mode result in misrepresentation that would result in the misclassification of household eligibility?

• Does web screening result in efficiency gains by reducing the level of in-person screening contact attempts required by field interviewers?

• Is there item nonresponse for key screening variables when conducted via the web?

Protocol for Group 1 Households (randomly assigned to receive the Web Screener Mailings)

1. Roughly half of all DUs in each segment will be randomly assigned into a multi-mode Web Screener Group. The remaining half will be assigned into an In-Person screener only group and treated as under the current NHANES screener protocol.

2. DUs assigned into the Web Screener Group will receive up to three mail contact attempts, as follows:

   1. Six weeks before the stand opens, Web Screener households will be mailed a packet that includes the Web Screener advance letter, magnet, and confidentiality brochure.

      1. The advance letter will be similar to the current advance letter except that it will include instructions on how to access the Web Screener and a unique login ID and password.

      2. The address, login ID, and password will be printed directly onto the letter and mailed in a window envelope to avoid matching errors with envelope labels and inserting the correct letter into the correct envelope.

      3. The password does not get updated by the user. The password remains the same as printed directly on the letter throughout the Web Screener.

      4. Segments anticipated to be highly Hispanic will be mailed both an English and Spanish version of the letter for all mailings. The pilot will not be conducted in Asian languages.

   2. A week after the initial mailing, a reminder postcard will be sent, if the Web Screener was not completed. This reminder postcard is modeled after the advance post card previously used by NHANES.

   3. Two weeks after the reminder postcard mailing (three weeks before the stand opens), if the Web Screener was not completed, another Web Screener Second Letter will be sent to invite households to complete the Web Screener. The letter will contain instructions on how to access the Web Screener as well as the unique login ID and password.

3. Field interviewers will begin screening in-person once the stand opens. The In-person screener will be completed for all households in both samples, even if the screener was previously completed on the web. The In-Person screener conducted by the NHANES field interviewer is the screener of record.

Protocol for Group 2 Households (randomly assigned to the current NHANES protocol and will not receive the Web Screener Mailings)

1. In-person screener only group households, i.e. households not randomly selected to receive the Web Screener, are the “control group” and will be treated under the current NHANES screening protocol as outlined in Protocol #2017-11.

Web Screener Details

1. Any listed DU which has undeliverable mail will not receive mailings. This is true for the Web Screener DUs and the In-person screener only DUs, and is in accordance with existing protocol.

2. The Web Screener will be 508 compliant, and the confidentiality and burden statement and OMB number will be displayed on the first screen of the Web Screener.

3. A response must be answered for each question before the Web Screener will advance. Don’t Know and Refusal responses will be accepted.

4. Respondents will have the option to save their answers and come back later to complete the Screener. Upon reopening the Web Screener, the respondent will log in and be returned to where s/he left off. Respondents will be able to back up and change answers, including after logging back in after saving and exiting. After a Web Screener is complete and submitted, respondents will no longer be able to view or change answers.

5. The Web Screener will terminate with a display message saying:

“Thank you for taking the time to complete this important survey. One of our representatives will visit your home to talk to you about the opportunity to continue to take part in the National Health and Nutrition Examination Survey (NHANES). If you have any questions, please call 1-800-452-6115 or visit www.cdc.gov/NHANES.”

6. An incomplete web Screener will be kept open until the end of the stand and flagged as Non-Response.

7. If a NHANES survey participant asks why he/she has to provide information to the field interviewer that was already provided via the Web Screener, the Field Interviewer will respond that it is needed for quality control purposes. The Field Interviewer will provide this information only if asked.

Proposed interviewer response to the question, “I already answered these questions when I completed the questionnaire online. Why are you asking me again?”

NHANES participants are selected randomly by the computer. To make sure that all members of your household have an opportunity to be selected, I am required to independently re-ask these questions to confirm who meets the study requirements.

Verification

1. The screener completed by the Field Interviewer in-person is the screener of record. The data collected from the Web Screener will be compared to the data collected by the Field Interviewer in-person with the latter being considered the true value.

2. Field interviewers will go to all released addresses to complete the screener, whether the DU received the Web Screener mailings or not, in accordance with the existing NHANES field interviewer protocol. Information from the Web Screener will not be available to the field interviewer. This process ensures that information is collected by the Field Interviewer without a preconceived notion of household composition and also ensures independent verification.

3. If at the end of the stand, contact cannot be made with a DU that has completed the Web Screener; no verification will be made. These DUs will be classified as screener non-response for the purposes of the NHANES data collection for the stand. These cases will be included in the analysis of the pilot.

4. Validation of Web Screeners will be performed by the Westat Home Office by reviewing overall Web Screener completion statistics and by comparing the information collected from the Web Screener with the information collected by the field interviewers. Specifically, it will involve percentage of Web Screeners started and not submitted, submitted but without all data required to determine eligibility, submitted with data required to determine eligibility, percentage of fields with identical data across modes, and percentage of fields that are “generally” accurate (e.g. different spelling of names, etc.). The decision to perform a second pilot or implement multi-mode screening will be based on an analysis of the data.