Feasibility and Acceptability of an Agenda-Setting Kit in the Care of People With Type 2 Diabetes: The QBSAFE ASK Feasibility Study

This was a single-arm intervention trial to assess the feasibility and acceptability of the QBSAFE ASK. The study recruited primary care clinicians and diabetes specialists at three study sites: the Mayo Clinic (Rochester, MN), Yale University (New Haven, CT), and Trinity Health of New England (Waterbury, CT). Clinicians were recruited if they cared for adults (≥18 years of age) with diabetes. People with diabetes were recruited if they had a diagnosis of type 2 diabetes and were under the care of an enrolled clinician. Individuals who did not speak English, had severe vision/hearing impairment, or were unable to give consent were not eligible. Individuals were enrolled after learning about the study procedures and providing written informed consent. The Mayo Clinical Institutional Review Board (IRB) approved all procedures (IRB #19-004560, ClinicalTrials.gov identifier NCT04514523).

The full pilot protocol was published previously (6). Once enrolled in the study and before a scheduled appointment, people with diabetes were approached by study personnel in person, by telephone, or via e-mail to review the QBSAFE ASK in a paper (Supplementary Figure S1) or electronic (https://patientrevolution.org/qbsafe) format. The kit was also made available in a paper format during in-person clinical encounters. People who needed to access the kit during telehealth or telephone encounters were directed to the electronic format. Individuals could select up to three cards from the kit to discuss with their clinician during the encounter. Encounters were audio- or video- recorded, although individuals could decline recording and still participate in the study. After encounters, clinicians and patients completed a three-question post- encounter survey to assess acceptability (Supplementary Table S1).

We assessed the feasibility of research procedures to determine the effectiveness of the ASK and to assess the acceptability of using the QBSAFE ASK in practice. Research feasibility was evaluated by assessing the time it took the study team to enroll clinicians and people with diabetes, the percentage of those approached who were willing to participate, and the percentage of participants who completed all study procedures. ASK acceptability was assessed via the three-question survey obtained from participants by study personnel after clinical encounters.

Two coders (S.H. and C.G.-L.) analyzed all recorded clinical encounters to assess the number and type of selected cards, measure the length of the encounter, and characterize the conversations about issues brought up using the QBSAFE ASK. A codebook to define clinician responses and conversations around the QBSAFE ASK was developed by using existing literature and data from prior video analysis. Clinicians’ responses were coded as follows: ignored, addressed conversation card, asked more questions, recommended further/continued monitoring, recommended change in monitoring, recommended referral, suggested new or a change in medication or treatment plan, suggested behavioral change (diet/exercise), and/or made a change in the follow-up plan. The types of conversations elicited by the QBSAFE ASK were developed a priori and categorized into five groups: patient-directed conversations, clinician-directed conversations, collaborative conversations, confrontational conversations, or cards that did not elicit discussion. Coders first had to achieve >90% calibration and then coded subsequent encounters independently. Coders discussed discrepancies or questions in coding to determine appropriate clinician response/conversation, and updates were made to the codebook. Data were recorded within REDCap (Research Electronic Data Caption) software and analyzed via Stata statistical software. We reported simple percentages and averages, when applicable.

Seven clinicians were recruited to participate, including five from the Mayo Clinic, one from Yale, and one from Trinity Health. Three clinicians were female, and most were diabetes specialists. Across the three study sites, 213 people with type 2 diabetes were potentially eligible and contacted for enrollment based on appointment lists of participating clinicians. Of these 213 people, 85 (39.9%) agreed to participate in the study and were enrolled. Of these 85, 70 (82.4%) were ≥50 years of age (mean age 63 ± 14.7 years). Less than half (n = 37 [43.5%]) were female, and 71 (84.5%) self-identified as White. The majority (n = 70 [84.5%]) had an A1C measured 6 months before to 3 months after the recorded clinical encounter. The mean A1C was 7.48 ± 1.17% (Table 1).

Target accrual was reached in 145 days (seven clinicians and 85 people with diabetes). Most people with diabetes (n = 66 [77.7%]) agreed to have their clinical visit recorded, and all study procedures were completed for 58 patients (68.2%). Across 69 encounters for which clinicians completed surveys, clinicians usually agreed that they felt confident responding to issues raised through the use of the tool (in 62 encounters [89.9%]), that they would use the tool again (in 54 encounters [78.2%]), and that they would recommend the tool to their colleagues (in 56 encounters [82.4%]). Among people with diabetes who completed study procedures, 37 (63.8%) agreed that the QBSAFE ASK helped them discuss their situation, 45 (77.6%) agreed that others could benefit from the tool, but only 22 (37.9%) said they would use the tool at their next visit.

Clinical encounters were either in-person (n = 36 [54.6%]) or conducted via telehealth video (n = 19 [28.8%]) or telephone (n = 11 [16.7%]). The mean length of clinical encounters was 29.4 ± 12.1 minutes. The mean length of encounters was not significantly different based on the use of the ASK (29.9 minutes for encounters with ASK cards selected vs. 27.7 minutes for those with no cards selected, P = 0.59).

The QBSAFE ASK was introduced in almost every recorded encounter (n = 65 [98.5%]), and in most cases the conversation cards were brought up by the clinician (n = 46 [69.7%]). Half of the participating people with diabetes with recorded encounters (n = 33) viewed the conversation cards via the electronic format. Fifty-five people with diabetes selected one or more cards (83.3%) across 66 recorded encounters. Every conversation card within the QBSAFE ASK was selected at least once. When people with diabetes used the QBSAFE ASK tool, they most often selected one card (median 1 [interquartile range 1–2]).

The most frequently selected card was “I find it hard to follow your suggestions about diet and exercise” (n = 19 [34.5%]), followed by “I struggle with monitoring my blood sugar” (n = 14 [25.5%]), “I struggle with remembering, taking, or managing my medications” (n = 11 [20.0%]), and “I have another issue related to my diabetes that I’d like to talk about” (n = 11) (Figure 1).

In four of the 55 encounters where cards were selected, people with diabetes chose a positive conversation card (“Diabetes has had some positive impacts on my life” or “I have felt moments of pride while managing my diabetes”). The cards exclusively elicited a person-directed conversation, with clinicians asking elaborating questions.

The most common clinician response to the selected conversation cards was to ask elaborating questions (n = 45 [81.8%]), followed by making a change in medication/treatment (n = 19 [34.5%]), recommending behavioral change (n = 17 [30.9%]), continued monitoring (n = 12 [21.8%]), and recommending referral (n = 12 [21.8%]) (Figure 2).

The QBSAFE ASK most often elicited collaborative conversation between clinicians and people with diabetes (n = 29 [52.7%]), followed by person-directed conversations (n = 10 [18.2%]) and clinician-directed conversations (n = 6 [10.9%]) with no confrontational or conflicted conversations observed. We did not capture the types of conversations elicited when the ASK was not used.

In this single-arm intervention pilot and feasibility study, study procedures were feasible, and the QBSAFE ASK tool was generally acceptable to people with diabetes and their clinicians. Conversations elicited when the QBSAFE ASK tool was used were often collaborative and involved brainstorming to find a solution, often in response to challenges participants with diabetes experienced with their diabetes self-management. This approach led to clinicians asking elaborating questions and to occasional changes in treatment, with no significant difference in encounter length between those using and not using the ASK. These findings suggest that the QBSAFE ASK holds promise in helping people with diabetes discuss their situation with clinicians and eliciting collaborative problem-solving.

Clinician-directed conversations during routine diabetes care often center around treatment targets, reinforced by guidelines (4). A treat-to-target approach narrowly focuses on metrics such as A1C goals without incorporating individuals’ treatment burden or other quality-of-life issues related to living with diabetes. Particularly in people not meeting treatment goals and experiencing significant treatment burdens, this approach may lead to further treatment intensification, which can be ineffective and even harmful (1). Furthermore, as seen in our study, people who are at or near their goal A1C can still experience a significant burden of disease and treatment. With the QBSAFE ASK, we propose shifting the focus of clinical encounters from the treat-to-target approach that is entrenched in clinical care to one that is collaborative and responsive to problems identified by individuals at that specific point in their life. Our findings suggest that the QBSAFE ASK may be an effective intervention that prioritizes individuals’ concerns and leads to productive conversations about how to solve them (i.e., may help care fit) (7).

In recent years, professional societies and clinical guidelines have emphasized the need to engage people with diabetes in shared decision-making (4). Numerous tools have been developed to facilitate shared decision-making between individuals and clinicians (8–11). These tools have ranged from handouts outlining guideline recommendations (4) to interactive, Web-based resources (10). One study described the development of agenda-setting cards similar to the QBSAFE ASK to facilitate conversations between people with diabetes and clinicians (12). Most tools have demonstrated high levels of acceptability by both people with diabetes and clinicians (10–12). Tools involving conversation cards have focused on themes similar to those in the QBSAFE ASK, such as treatment burden and quality of life (12). People with diabetes found conversation cards useful in describing their experience, although one study described a subset of individuals who thought the cards would benefit others but not themselves (12). However, prior studies have not evaluated responses to conversation cards in a clinical setting.

The QBSAFE ASK often elicited conversations related to diet/exercise, glucose monitoring, and medications. Prior studies have demonstrated that diabetes medications and monitoring are common concerns (13,–,15). Similarly, a prior qualitative study found that tasks such as implementing complex medication regimens and making lifestyle changes were burdensome for some people (16). Clinicians most often responded to conversation cards by asking elaborating questions, similar to results from our prior study assessing clinicians’ responses to patients’ concerns (5). These forms of problem-solving conversations demonstrate purposeful shared decision-making (17) and illustrate how shared decision-making can be an effective method for care (18).

There are some limitations to our study. This was a single-arm study trial, which means we could not compare the impact of the QBSAFE ASK with usual care versus usual care without the kit. Both clinicians and people with diabetes who participated in this study may have been more likely to use and positively respond to the conversation cards than the general population of clinicians and people with diabetes. Most participating clinicians were specialists and may have felt more comfortable addressing individuals’ diabetes-specific concerns. Most participating people with diabetes were established patients and tended to be in an older age-group. The coronavirus disease 2019 pandemic occurred at the beginning of the trial, which limited and delayed recruitment. Finally, because this was a proof-of-concept feasibility trial, future studies are needed to assess whether the findings are generalizable. Nevertheless, our findings support the subsequent evaluation of the QBSAFE ASK tool in a randomized trial to compare its effectiveness to that of usual care without the kit and to estimate the extent to which its use can reduce treatment burden and improve clinical outcomes.

The QBSAFE ASK received a positive response from clinicians and people with diabetes. The tool was used often in clinical encounters and elicited collaborative problem-solving conversations between individuals and their clinicians. Study procedures were feasible in the setting in which they were studied. Further research is needed to assess the tool’s feasibility in other clinical settings and to determine its efficacy.

This study was supported by a National Institutes of Health T32 grant (DK007058), National Institute on Aging R21 grant (1R21AG061427-01, application no. 9650947), and National Institute of Diabetes and Digestive and Kidney Diseases R01 grant (R01DK129616). The funding bodies did not have a role in the study design and did not take part in data collection, analysis, and interpretation or in writing the article.

No potential conflicts of interest relevant to this article were reported.

S.H. analyzed data and wrote the manuscript. C.G.-L., M.B., V.M.M., and K.J.L. analyzed data and reviewed and edited the manuscript. J.C., D.L.G., S.A.H., and B.S. collected data and reviewed and edited the manuscript. K.R.B. reviewed and edited the manuscript. S.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.