Supporting B

Supporting Statement B: Data Collection Procedures

OWH Postpartum Depression Campaign Focus Group Research Plan Overview

To test creative concepts and draft message frames, the LTG postpartum depression (PPD) formative research team consisting of LTG and NORC will conduct 30 focus groups across the nation. Online virtual focus groups will be conducted using the Zoom video conferencing platform. Materials designed and reviewed as part of this package are a moderator’s guide and two screeners.

A two-stage screening process will be used for these focus groups to ensure the closest matching of participants to the characteristics of the segments of women at risk for PPD identified in the initial formative research. Active consent, as illustrated in the Phase 1 and Phase 2 screeners, will be obtained during and passive consent, as illustrated in the focus group guide, will occur at the beginning of the focus group sessions.

Phase 1 Recruitment: Initial qualification will be based on age (18-44) and birth of baby in the past year. Women who work as social workers, counselors or in a mental health care field will be excluded in order to capture unbiased reactions to the creative concepts. Other characteristics will be collected to balance geographic residence, demographics and risk factors for PPD. Items collected will include race-ethnicity, number and ages of children living in the home, education level, marital status, and income. We also plan to collect data to assess the likelihood of active or passive participation in the group sessions.

We anticipate 2-3 waves of review of the data collected during Phase 1 recruitment to balance the sample by geography, demographics, and risk factors. We will use the characteristics from the segmentation survey to match the sample characteristics as much as possible during the recruitment process. The geographic distribution for the sampling is provided below by moderator with a map provided of the Census Regions and Divisions in Figure 1. To maximize time and efficiency, we will have 3 moderators conducting groups over 2 weeks simultaneously with a goal of scheduling 5 groups per week.

Moderator 1

• Eastern Standard Time

• Northeast Census Region and East South Central Census Division and South Atlantic Census Division

• 5 Groups from Northeast (Women from PA, NY, RI, MA, CT, VT, NH, ME – will be based on recruiter availability)

• 5 Groups from South Central/Atlantic (Women from KY, WV, VA, MD, DE, DC, TN, NC, SC, GA, AL, MS - will be based on recruiter availability)

Moderator 2

• Mainly Central Time Zone, with a portion in Eastern Standard Time

• Midwest Census Region and the West South-Central Census Division

• 5 Groups from Midwest Region (Women from OH, MI, IN, Il, WI, MO, IA, MN, ND, SD, NE, KS – will be based on recruiter availability)

• 5 Groups from West South-Central Division (Women from TX, OK, AR, LA – will be based on recruiter availability)

Moderator 3

• Pacific and Mountain Standard Time

• West Census Region

• 5 Groups from Pacific Census Division (Women from WA, OR, CA, AK, HI – will be based on recruiter availability)

• 5 Groups from Mountain Census Division (Women from ID, MT, WY, CO NV, UT, AZ, NM – will be based on recruiter availability)

Figure 1.

Phase 2 Recruitment: Once we have a balanced sample of women from the Phase 1 initial qualification review process, we will send out a link to a 5–7-minute survey to the participants identified from Phase 1 recruitment. This online set of questions will be used to assign women to the audience groups/segments from the segmentation study. Key questions from the larger survey will be used to identify their assignments. Once women have been assigned to groups, any gaps in the coverage among the audience groups/segments will be examined and a final recruitment strategy may occur. Twelve women will be recruited for each group to ensure an optimum number of 10 women and a minimum of 8 women per group. Each moderator will conduct sessions with 2 of each of the 5 audience groups/segments for a total of 10 groups per moderator.

Logistics and Moderation: Professional moderators will facilitate the group sessions. Professional recruiters will be engaged who will identify, manage, and pay incentives for all respondents. As noted, professional moderators will be recruited and trained to conduct the focus groups. Data will be managed from the screening process, and respondents placed into groups and scheduled for sessions. LTG will host the sessions and provide access to the virtual platform for participants and those requesting to view any of the sessions. The sessions will be recorded through the Zoom video conferencing platform and all respondents will be made aware of the recording during screening and asked to provide a response to the Zoom recording consent at the beginning of the session. Recruitment and attendance progress/tracking reports will be shared throughout the fielding period to ensure coverage of sample specifications including audience segments and respondents with key risk factors.

Materials for Testing:

• 1-2 visuals that represent each of the three overarching campaign concepts. The same visuals will be presented to all the groups in a random order. The visual concepts will also be compared to each other at each group.

• One set of 3-5 text-based messages for each of the audience groups/segments that reflect the “think, believe, do” messages tailored to their audience group/segment. We will get feedback on each of the messages but will not be comparing any of the messages against the other.

Analysis and Reporting: Transcription will be conducted by a professional transcription service and will begin immediately after each focus group. Data cleaning will occur as the transcripts are ready. Data coding will be conducted by two analysts with inter-rater reliability checks, and then a thematic analysis using an inductive process to identify emerging themes.

We plan to use a two-phase process of data analysis. In the first phase, data from each focus group will be analyzed and outcomes summarized. In the second phase, all groups from each segment will be compared and those outcomes summarized.

LTG uses NVivo software to code qualitative data, tagging recurrent themes and issues in order to develop a thematic framework. A preliminary list of codes will be generated prior to data collection and refined as transcripts are reviewed. Each code will have a written definition that is improved or focused further as the study proceeds. During the initial coding, the text will be read line by line in order to gain a broad comprehension of the data. Initial coding involves tagging blocks of texts with descriptive codes. As the data accumulate and are coded, and the reviewers gain additional understanding of the content, the data will be revisited and newly coded to reflect the emerging patterns and themes. The last step is focused coding, in which the themes are further operationalized by clustering similar tags or breaking up existing tags into sub-categories. Two researchers will code the same data set and inter-coder reliability will be established at 70%.

After the LTG qualitative analysis team have reached consensus on the coding of the data, text will be sorted by tag and inspected for themes and patterns. These themes and patterns will provide the context for understanding the perspective of the respondents and interpreting the findings. As noted, each focus group analysis will be discrete with comparison within segment being conducted after all within-segment groups have been analyzed. From this analysis, detailed segments will be described with means to target specific beliefs and behaviors.

By qualitative feedback we mean information that provides useful insights on perceptions and opinions but are not statistical surveys that yield quantitative results that can be generalized to the population of study. This feedback will provide insights into customer or stakeholder perceptions, experiences and expectations, provide an early warning of issues with service, or focus attention on areas where communication, training or changes in operations might improve delivery of products or services.

A presentation of the findings will be provided two weeks after the coding is complete and a final report will be provided two weeks after the presentation.