A qualitative analysis of family experiences across diabetes-specific family functioning types during a family-focused intervention for adults with type 2 diabetes

Interviews were audio-recorded, transcribed verbatim, and then uploaded in NVivo 12 for coding and analysis (Ltd., 2018). No repeat interviews were conducted, and transcripts were not returned to participants. First, interviews were structurally coded based on the main sections of the interview guide (Table 1). During the structural coding processes, coders reviewed the entire interview transcript for all interviews and allocated data by topic discussed in the interview. After structural coding, we reviewed the data within each structural code and agreed the Family Interactions Regarding Diabetes code contained content with the most generalizable information and most relevant for our questions about experiences by type, rather than information specific to FAMS. If a participant spoke about family interactions outside of the interview questions designated in Table 1, that content was included during the structural coding process. Content in other structural codes was focused largely on feedback about the specific intervention and RCT procedures. Therefore, we decided to focus on data about Family Interactions Regarding Diabetes for thematic analysis. Next, a team of three raters [EB, SM, CG] distributed and reviewed the data coded to Family Interactions Regarding Diabetes, and independently generated notes about emergent themes. The raters then met and discussed their notes to develop an initial thematic codebook for the data pertaining to Family Interactions Regarding Diabetes (Braun & Clarke, 2006). This coding scheme was refined by study authors who had participated in the structural coding and were familiar with the interviews. Then, the final coding scheme was applied iteratively to first three (κ = 90) and then later five (κ = 80) transcripts to establish and confirm inter-rater reliability across 10% of the transcripts. All types were represented across 10% of the transcripts used to evaluate inter-rater reliability. During this process, updates to the codebook were made as needed. Raters then independently thematically coded the rest of the data relevant to Family Interactions Regarding Diabetes, meeting regularly to discuss any questions or resolve any uncertainties.

Next, [MKR & EB] conducted constant comparison analyses by diabetes-specific family functioning types on thematic codes appearing within the topic of Family Interactions Regarding Diabetes (Strauss & Corbin, 1990). Constant comparison procedures compare all data relevant to a particular thematic code with other occurrences of the same thematic code within and across groups (in this case, diabetes-specific family functioning type) to uncover similarities and differences in prevalence as an indicator of salience and dimensions within the themes. We reviewed participants’ narratives about their family and friends’ involvement in their diabetes to identify similarities and differences in thematic codes across and between participants by type. The results presented are based on these constant comparative analyses, whereby similarities and differences in the prevalence and dimensions of the thematic codes by type are summarized.

To enhance the transparency and trustworthiness of our results, we used investigator triangulation (multiple authors were involved in the reading and interpreting of codes and themes) and audit trails (a thorough accounting of each step of the process we undertook) during the coding process. We also triangulated our qualitative results with quantitative research on the typology–both descriptions of the types (Aim 1; Mayberry et al., 2021, 2023) and experiences with the intervention (Aim 2; Mayberry et al., 2021, 2023).

There was evidence of the defining characteristics of types in participants’ descriptions of their experiences with family. Collaborative & Helpful, notable for high levels of helpful involvement, satisfaction, and collaboration, had a participant who described, “I like the fact that–I mean, I know I'm accountable for my own actions, but that little extra push from somebody, especially somebody that cares about you that loves you and doesn’t want to see anything bad happen to you, you’re going to always try and do what’s right versus what’s wrong” (54 year-old, non-Hispanic Black female). Another Collaborative & Helpful participant shared how his grandson, who has type 1 diabetes, taught him how to use a continuous glucose monitor, stating, “[my grandson] talked me through it, and he told me where to put it and how to place it and, you know, how to prep the site and when to push the button and said, ‘Don’t be scared, Granddaddy. You know it’s going to… It might hurt just a little bit, but it’s okay’… It was good for our family,” (70 year-old non-Hispanic white male).

Critically Involved, notable for high levels of helpful and harmful involvement, high criticism, and low effectiveness had a participant who said, “Well, my wife and daughter… I told my doc that they were getting on my nerves, harassing me, basically” (46 year-old, non-Hispanic Black male). Another Critically Involved participant stated their family member, “acts like she’s my mom, so she asks me like my mom would ask me. You know, you’re not–you’re not my mom… it was her tone of voice and the way she used it, not–maybe not necessarily the way she asked it. Just her tone of voice, it was like my mom. ‘How come you didn’t do it that way? You did great, but you could have done better’ (61 year-old non-Hispanic white female).

Satisfied with Low Involvement, notable for low levels of involvement and collaboration without dissatisfaction had a participant who described, “family involved in [diabetes]… while I understand how important that could be in certain circumstances, in my particular circumstance it wasn’t quite that important at all” (67 year-old, non-Hispanic white male). Another Satisfied with Low Involvement participant shared, “I really didn’t [interact with family/friends about diabetes]. If my kids asked me about it, I would tell them, but I was just kind of private about it, and I just didn’t talk about it with them. I didn’t share my A1c numbers. I didn’t share my blood sugar numbers when I was taking it. I just sort of went about my business” (62 year-old non-Hispanic white female).

Finally, Want More Involvement, notable for low involvement and collaboration with dissatisfaction had a participant who explained, “a large part of the reason I got into the study was because of my support person and because she hadn’t grown up around diabetics and didn’t realize on a day-to-day basis what was involved and what the challenges were” (55 year-old, non-Hispanic white male). Another Want More Involvement participant described how their family was aware of their diabetes but did not pay as much attention to it, sharing, “they knew it was there, but nobody knew how serious it was. So, it really wasn’t there. It was there, but it wasn’t. It was in the room, but it was on the–in the corner of the room,” (53 year-old, non-Hispanic white male).

We found themes Presence of Harmful Involvement, Increases in Communication, and Received Helpful Involvement were present within each type’s description of their experiences with family during FAMS, but there was notable variability across types in prevalence or dimensions (detailed in the paragraphs below). Presence of Harmful Involvement captured participants’ descriptions of specific harmful behaviors family engaged in such as nagging or making negative comments about their diabetes self-management choices or bringing certain foods into the home despite requests not to. When directly questioned, participants denied involvement that was “harmful;” however, upon reading and coding transcripts, we found examples of harmful behaviors described. Participants also noted Increases in Communication with their family about their diabetes and health across their intervention experience. This included both new conversations that had not been had before as well as inviting family into deeper conversations about diabetes. Finally, participants described Received Helpful Involvement, including both instrumental support (help with specific tasks or solving specific problems) and emotional support (having someone to non-judgmentally listen, discuss their diabetes, and encourage them). Specific similarities and differences by type are described below with quotations in Table 4.

When Collaborative & Helpful discussed harmful involvement, their statements were qualified to give their family the benefit of the doubt by assuming this was due to family not being able to remember all the specifics of managing diabetes. Collaborative & Helpful described feeling closer to family during FAMS. Their descriptions of interacting with family were very collaborative, high in warmth, and suggested they trusted their family. They felt the topic of diabetes was less taboo to discuss or normalized through FAMS, allowing for specific conversations about their self-care and dietary needs with their families.

Critically Involved provided examples of harmful involvement included family nagging, bringing ‘sweets’ into the home, and policing what they were eating. Some Critically Involved reported that their conversations with family had a deeper ‘level’ or better ‘tone’ than before FAMS; however, several examples provided suggested the support was still more controlling than autonomy supportive throughout the intervention. Participants in this type also noted that it was easier to talk about diabetes due to FAMS.

Satisfied with Low Involvement, like other types, were hesitant to directly discuss harmful involvement from family. The minority who did described their family bringing home ‘sweets’ and being aware of their dietary needs but not changing their behavior in response. Throughout FAMS, Satisfied with Low Involvement described sharing their experience with diabetes with family more frequently and in return receiving encouragement and support. Additionally, Satisfied with Low Involvement found that as they opened up to their family about their diabetes, the shame and stigma they felt around the disease decreased and the support they were able to access increased. Participants described these new conversations as a safe place to land that historically may not have felt as welcoming.

Finally, Satisfied with Low Involvement emphasized emotional support over instrumental support. Emotional support was described as opening new conversations and using family as ‘sounding boards.’ Similar changes in family behaviors were noted (e.g. food choices, walking together) but the person with diabetes seemed less enthusiastic about these changes than Critically Involved.

Want More Involvement described their family as being harmful by critiquing them and embarrassing them for how they were managing their diabetes. Want More Involvement described sharing more information with their families and having conversations they had never had before during FAMS. One described developing an accountability partner relationship where they shared goals and kept each other motivated. Want More Involvement reported equally liking emotional and instrumental support. There were different types of emotional support described including heightened awareness of their diabetes-related needs, checking in, and accountability. For instrumental support, reminders were very common, walking together was mentioned, and a few expressed desires for more instrumental support than experienced.