Table 1.

Total Quality Framework (TQF) Criteria for Planning, Conducting, Interpreting, and Evaluating Applied Qualitative Research^*

Criterion	Methodological Decisions	Rationale
Credibility	Clearly defined target population	See “Methods – Participants” in report
	Assessment of extent to which population sample is incomplete and implications of this omission for study conclusions	See “Discussion (para 7–8)” in report
	Sample design and size	For this pilot study that was limited in resources for recruitment and focus group facilitation logistics, we elected to recruit older adults with type 1 diabetes from a hospital-based outpatient endocrinology clinic. Recruitment was conducted on a rolling basis. Saturation was defined as the point when no new or original themes emerged.9 Because the same moderator and assistant led the groups, ARK and CS periodically assessed saturation of themes that emerged from discussion and group model building activities. Verification of saturation achievement was sought out by presenting preliminary themes to research mentors LAY and KHL who have extensive experience and training in clinical diabetes care for older adults and group model building. Recruitment was halted when consensus was reached that saturation had been achieved. We sought to conduct groups with 4–5 participants to maintain required COVID social distancing spacing in the room where the groups were able to be conducted, minimize intimidation of speaking in a larger group and because groups with <10 participants have been noted to promote depth and focus of discussion.9 We also conducted groups of variable sizes to accommodate participant availability and because dyads and triads have been shown to be particularly useful for rich data not too dissimilar from in-depth interviews. Groups were conducted separately with those who currently used CGM and those who formerly or never used a CGM, which was initially identified via medical record screening and confirmed by participant self-report. Because understanding perspectives and experiences about CGM was a prominent interest to our study and potentially a sensitive topic, we sought group homogeneity on this behavior to increase participant comfort in sharing thoughts with people who have lived through similar experiences and increase the nuance of our understanding of the reasons for use and non-use. Older adults vary in age, physical/cognitive limitations, relationships, and resources, and we therefore anticipated variability in the extent to which family members or other caregivers would be involved in older adult diabetes management. In order to limit introducing bias to our sample, we thus let all prospective participants choose whether to invite a caregiver to attend the groups with them.
	Strategies to gain access and gain cooperation from participants	Selection of facilitated group model building process, which accommodates different forms of expression and support multiple ways for older adults to share their perspectives (See Methods – Data Collection (para 1) in report). In order to promote participant cooperation and honesty and minimize risk of participants disclosing what they thought were the ‘right answers’ that the moderator wanted to hear, the moderator prefaced each group by explaining the objectives of the research, as well as emphasizing that the participants were experts from whom the researchers sought to learn, and that all thoughts related to the research objectives were valuable. As part of efforts to build rapport with and promote uninhibited disclosure by participants, the moderator disclosed her own type 1 diabetes status to participants in her introduction given the possibility that participants might be less forthcoming with a moderator who was markedly younger and appeared unrelatable. An icebreaker prompted participants to describe their favorite thanksgiving food or candy and an open-ended prompt asked the participants to share their relationship with diabetes in order to naturally initiate group rapport building and discussion about the research objectives. All participants were prompted to share in order to prevent against dominance by extroverted participants. At the end of each group, participants were also given the opportunity to write down any additional information they felt was relevant but did not raise in the course of the discussion or felt was too private to share with the group. In order to limit bias introduced by the moderator, attention was paid to use neutral language in facilitating the group so as not to give participants indication that some responses were more right than others. The same moderator, the principal investigator (ARK), led the discussions and activities to promote consistency in quality of data gathered across groups and prevent against erroneous claims that variations in data exist because of variation in discussion guide coverage across groups.9
	Selection of constructs and attributes that map to research objectives	See “Methods – Data Collection (para 1)” in report
	Selection of data collection mode	See “Methods – Data Collection (para 1)” in report
	Operationalization of constructs and attributes in the form of a data collection tool	See “Methods – Data Collection (para 1)” in report
	Evaluation of data collectors as sources of bias and inconsistency	See “Strategies to gain access and gain cooperation from participants” in present table
Analyzability	Context and subject matter guided transcription	See “Methods – Data Processing and Analysis (para 1–3)” in report
	Coding format that standardizes way data are processed	The principal investigator (ARK) along with two analysts (ACS, CS) with training and research experience in type 1 diabetes and qualitative methods read through all transcripts from the perspective of the research objectives and constructs of interest. Each analyst then independently inductively coded a different subset of all transcripts before meeting to discuss a preliminary set of codes to ensure that the codes were representative of the full range of data. Once analysts reached consensus about the draft codebook, each analyst independently applied the codebook to the same subset of transcripts, after which they met to compare code applications, resolve discrepancies, and revise the codebook. This process of codebook revision was repeated until all transcripts were coded. Two of the analysts (ACS, ARK) independently applied the final codebook to all transcripts and the third analyst (CS) compared code application and compiled any remaining divergences. All analysts then met to achieve consensus on final coding of all transcripts.
	Systematic identification of themes interpretations from the data	Once coding consensus was achieved by all analysts (ACS, ARK, CS), two analysts (ACS, ARK) independently identified categories across codes and identified themes within and across categories and summarized them independently.
	Verification of accuracy of the data, themes, and interpretations	A set of final results, interpretations, and implications were reached by analysts by a) comparing and reconciling those that they had independently drawn b) looking for data that contradicted analysts’ claims (i.e., deviant case analysis) and c) seeking verification by researchers with subject matter expertise and who had been involved in supporting research conceptualization, data collection and processing (CS, LS, RM).
Transparency	Complete disclosure of all aspects related to study design, data collection, and analysis	Throughout report and the present supplement, by explaining rationale for methodological choices according to the Total Quality Framework (TQF) and ensuring reporting conforms to the TQF and Standards for Reporting Qualitative Research (SRQR).9,10
	Discussion of specific aspects of design, data collection, and analysis processes that impact outcomes and interpretations of data
Usefulness	Evaluation of relevant questions for researchers and users of research	Answers to the following questions can be found in “Discussion” in the report: • What can and should be done with the study now that it has been completed? (para 2–6) • Has the study identified important knowledge gaps that future research should try to help close? (para 7–8) • Has the study offered recommendations for action that are worthy of further testing or worthy of actionable next steps? (para 2–6)
	Inclusion of “Usefulness of the study” section to articulate and justify how study should be interpreted, acted upon, or applied in other research contexts in the real world	See “Discussion” in report

^*Roller, Margaret R., and Paul J. Lavrakas. Applied qualitative research design: A total quality framework approach. Guilford Publications, 2015.