Attachment D_CMS-10816 Supporting Statement B_Non sub request

Supporting Justification for the Clearance of the Medicare Part C and Medicare Part D Enrollment Form Interviews

Supporting Statement B: Statistical Methods for the Collection

(CMS-10816; OMB 0938-1440)

Revised December 4, 2023

Submitted by:

US Department of Health and Human Services (HHS)

Centers for Medicare and Medicaid Services (CMS)

7500 Security Boulevard

Baltimore, MD 21244

Part C and D Enrollment Form Interview

Supporting Statement – Part B

PART B INTRODUCTION

CMS has a long-term goal of collecting race and ethnicity data from all Medicare enrollees. CMS received OMB approval of the revised collection of information based on the inclusion of race and ethnicity data on the Medicare Advantage (Part C) and/or Prescription Drug Plans (Part D) enrollment form. CMS will initially focus efforts on enrollees who newly elect or change coverage in the Medicare Part C and D programs. CMS OMH has contracted with NORC at the University of Chicago (NORC) to plan for and conduct interviews to understand why enrollees do not answer the race/ethnicity questions on the Part C and Part D enrollment form. The data collected through this study will be used to understand among enrollees who did not elect to respond to the race and ethnicity questions, how they perceive the addition of those questions on the form and to inform strategies to improve race and ethnicity data collection among enrollees.

With this submission, CMS is requesting OMB approval for a new collection and analysis of qualitative data on reasons for not electing to respond to race and ethnicity questions on the Part C and Part D enrollment form (see Instruments 1 & 2: Cognitive Interview Protocol – English/Spanish).

Each participant (up to 130) for whom verbal consent are secured will participate in one interview. The interview guide was developed by NORC and informed by NORC experts in conducting cognitive testing interviews.

B1. Respondent Universe and Sampling Methods

The respondent universe will include Medicare enrollees newly enrolling or switching from Part C and/or Part D plans in the 2023 calendar year that do not elect to answer the race/ethnicity questions on the enrollment form that has been in place starting January 1, 2023 (OMB No. 0938-1378). In the 2022 cycle, there were about 4.28 million people that switched or newly enrolled in MA plans and another 2.52 million people that switched or newly enrolled in Part D plans. We assume a similar number of people will also switch or newly enroll in the 2023 calendar year during the MA Open Enrollment Period (Jan 1 – Mar 31), Open Enrollment Period (Oct 15 – Dec 7), and as people become eligible for Medicare. NORC will ask CMS for a list of new and existing enrollees who switched plans in April 2023 (MA Open Enrollment Period) and December 2023 (Medicare Open Enrollment Period) as the two different enrollment populations may differ. NORC will look at how respondents that do and do not answer the race/ethnicity question differ on certain characteristics (e.g., age, gender, plan name (i.e., parent organization), region). Based upon the identified characteristics, NORC statisticians will stratify participants.

The goal of this data collection effort is to complete a total of 130 cognitive testing interviews, with the number of interviews for Part C (i.e., MA only) and for Part D (i.e., MA-PDP, and PDP-only) relative to the number of people who did not respond to race and ethnicity questions (or until saturation is reached) with a varied convenience sample pulled from the respondent universe. Interviews will also conclude by June 2024 to allow for timely review of the results to inform enrollment form design. From previous experience conducting cognitive testing interviews for the Medicare Current Beneficiary Survey (MCBS), 20-30 interviews have been sufficient to identify patterns. Conducting 130 interviews for Part C and Part D should account for the stratification variables that may be used (Phase 1 variables included CMS Region, age group category, large vs. small plans, and plan type). We will reserve 40 interview slots across Part C and Part D to be completed for Wave 2 (enrollment during July – Dec 2023).

Assuming 50% have missing race/ethnicity data and 50% of the 6.8 million enroll in the first part of the year, we estimate a universe of 1.7 million people for the first set of interviews and another 1.7 million for the second set. The samples will draw from their respective universe. The expected response rate is based on the 2020 baseline MCBS panel. In Fall 2020 the panel had a 41.9% response rate. Given that our target population elected to not respond to the race/ethnicity questions and are likely to not want to discuss their reasonings, we estimate a more conservative expected response rate of 35%. The MCBS response rate includes only 35% of phone numbers connecting interviewers to respondents as the MCBS uses a batch locating service to identify phone numbers. This suggests pulling an initial sample of approximately 250 for the first wave of interviews to account for a 35% response rate. The proposed plan for the first set of interviews is to release a sample of 40 enrollees for each Part (80 total) to assess the response rate, with the remaining 170 sampled enrollees to be released as needed during the interview period (i.e., as an enrollee is changed from active to unreachable or unwilling to participate, the interviewer will receive a new enrollee to start contacting).

Based on the screening and eligibility rates observed during Wave 1, CMS anticipates that an estimated 350 enrollees will be recruited and screened, resulting in an estimated 130 completed interviews.

B2. Procedures for Collection of Information

NORC staff will reach out to the selected enrollees via phone. NORC staff will contact the enrollees by telephone and record the interviews with the participants’ consent. NORC staff will attempt up to three calls to connect with the sampled enrollee and invite them to participate in the interview (Attachment A). If the enrollee agrees to participate in the interview, the NORC staff will screen the participant for eligibility (Attachment B) and if eligible, attempt to conduct the interview if the enrollee is available to complete it at that time. If the enrollee is not available, the NORC staff will schedule a callback time to conduct the interview. Prior to beginning the interview, the NORC staff will provide information on the study, answer any questions (Attachment C) and obtain verbal informed consent (Attachment D) from the participant including consent to record the discussion. NORC will be recording interviews in order to use them as a reference in developing notes of the interviews. During the interview, the NORC staff member will conduct the interview and take notes. A unique identifier will be assigned to interview transcripts.

The process of recruiting potential study participants is the same across respondents and reflects sensitivity to issues of efficiency, accuracy, and respondent burden. Different questions will be asked depending on whether the respondent recalls seeing the race/ethnicity questions (see Instruments 1 and 2). The interviews will be conducted in both English and Spanish.

During the data collection period, interviewers will meet weekly to discuss key findings from interviews. The study lead will determine whether data saturation is being reached when the same comments are being expressed repeatedly from respondents.

Best practices will be used for the analysis of the qualitative data collected from the interviews. Analysis of qualitative data will not require statistical analysis, as it is not appropriate. A kappa will be calculated to assess reliability of qualitative coding between two coders. NORC will use NVivo 12 Pro to analyze the qualitative data collected from the interviews.

B3. Methods to Maximize Response Rates and Deal with Nonresponse

To ensure a sufficient number of interviews, Wave 1 will be pulled from the pool of all new and switched Part C and D enrollees through March 31 2023, with a pool of sampled replicates. And then similarly for Wave 2, the starting sample will be pulled from the pool of all new and switched Part C and D enrollees through Dec 31 2023. Sampled replicates will be released based on the reachability, cooperation, and availability of sampled people. If the sampled enrollee cannot be reached for an interview after three attempts or does not agree to be interviewed, alternates (sample replicates) will be released. The use of replicates ensures that the target sample size is achieved without exceeding the target number of interviews. CMS expects to achieve a response rate of 35% for the interviews. The sampling plan is expected to provide information that can be generalized to other Part C and D enrollees who elected to not respond to the race/ethnicity questions.

NORC will reach out to sampled enrollees up to three times to allow an opportunity to participate in the interviews. Additionally, in order to encourage people to participate in the interviews, prospective participants will be told about the $40 incentive that is available to eligible participants after completion of the 30-minute interview.

B4. Test of Procedures or Methods to be Undertaken

The interview guide was developed following best practices for cognitive interviewing of survey questions and using lessons learned from NORC’s group of experienced interviewers.

B5. Individuals Consulted on Statistical Aspects and Individual Collecting and/or Analyzing Data

Interview data will be collected and analyzed by NORC. See Attachment E for a list of individuals CMS consulted on instrument development, the sampling plan, and/or analyzing data. See Attachment F for the analysis plan.