The Science of Diabetes Self-Management and Care2023, Vol. 49(2) 150 –162© The Author(s) 2023Article reuse guidelines:sagepub.com/journals-permissionsDOI: 10.1177/26350106231151405journals.sagepub.com/home/tde
AbstractPurpose: The purpose of this study was to assess the feasibility of delivering the Diabetes Tune-Up Group (DTU), a cognitive-behavioral-therapy-based (CBT) multidisciplinary intervention for adults with diabetes distress and elevated A1C using a group in-person delivery format.
Methods: The DTU intervention consisted of 6 weekly group sessions (90 minutes in duration per session). The groups were cofacilitated by a diabetes care and education specialist (DCES) and a master’s-level clinical psychology trainee. The intervention integrated CBT with patient-centered diabetes education. Using a pre/post study design, participants completed assessments at baseline, post-intervention, and 3 months following the intervention.
Results: The sample consisted of 29 adults with type 1 diabetes (N = 8) or type 2 diabetes (N = 21) who were predominantly female (79%), White (59%), and educated (56% with a college degree or greater). Participants attended 131 total sessions out of 174 possible sessions, for an overall attendance rate of 75.3%. At 3-month follow-up, significant improvements were observed in A1C values (mean decrease = 0.39%). Diabetes distress improved significantly from baseline (mean = 3.44, SD = 0.68) to post-intervention (mean = 2.94, SD = 0.68), and 3-month follow-up (mean = 2.55, SD = 0.75). Significant improvements were also observed in diabetes selfefficacy from baseline to post-intervention and at 3-month follow-up.
Conclusions: This group-based, multidisciplinary intervention resulted in improvements in A1C, diabetes distress, and patient self-efficacy in caring for diabetes. Future studies to validate this intervention approach across settings and delivery platforms are needed.
More than 37.3 million individuals are estimated to be currently diagnosed with either type 1 (T1DM) or type 2 diabetes (T2DM), with an additional 89 million at increased risk of developing the disease.1 By 2050, approximately 1 in 3 people will be diagnosed with T1DM or T2DM, with annual costs to individuals, communities, and health care systems exceeding $327 billion.2
Diabetes distress is a highly prevalent psychological condition associated with T1DM and T2DM, with point prevalence rates ranging from 18% to 44%.3-5 Diabetes distress is defined as unrealistic patient expectations for perfection accompanied by feeling overwhelmed by the number and frequency of self-care tasks, limited support, feelings of failure, and lack of influence over glycemic trends when self-care is initiated.6-8 Diabetes distress has been documented to be persistent, with 71% of study participants maintaining high levels of diabetes distress over a 9-month period without intervention, indicating the absence of spontaneous remissions,9,10 and has been consistently shown to be associated with elevated A1C11 values in both cross-sectional and longitudinal studies.9 In longitudinal studies, a reciprocal relationship exists between diabetes distress and A1C and problem-solving skills.10 Fisher et al9 postulated that distress may create an emotional and cognitive barrier to the adoption of education and problem-solving behaviors, which results in elevated A1C. Elevated A1C places patients at increased risk for long-term diabetes complications such as heart disease, nephropathy, neuropathy, proliferative retinopathy, and poor wound healing.5,12
Meta-analyses and systematic reviews have identified multiple interventions and approaches for the treatment of diabetes distress in T2DM samples.13 Characteristics of effective interventions include addressing diabetes education, emotional functioning, cognitive reframing, motivational interviewing, and peer support. Mathiesen and colleagues14 observed a standard mean difference of −0.18 (95% CI, −0.32 to −0.03) improvement in diabetes distress at a 3-month follow-up time period across individual intervention approaches. Combined effects on A1C and health-related quality of life over this same time period were not statistically significant across these studies. Greater levels of impact were observed in intensive and individual interventions.
In the REDEEM trial, which focused on adults with T2DM, Fisher and colleagues3 tested an individualized computer-assisted self-management web intervention used alone and in combination with a diabetes distress problem-solving therapy (CAPS) compared to computer-based minimal support in adults with T2DM drawn from primary care practices. In this pragmatic trial of 392 patients, the authors reported that improvements in diabetes distress were observed in all 3 intervention arms, with larger observed reductions in regimen-related distress in the CAPS condition. Reductions in overall diabetes distress were associated with improvements in healthy eating, physical activity, and medication adherence.
Van der Ven and colleagues15 conducted a randomized controlled trial of a group-based cognitive behavioral therapy (CBT) intervention compared to Blood Glucose Awareness Training in a sample of 107 adults with T1D.16 Diabetes distress outcomes improved in both groups with no significant differences between groups.
In the T1-REDEEM trial, Fisher et al3 compared the effectiveness of 2 intervention approaches on diabetes distress and A1C outcomes in a sample of 301 adults with T1DM. An emotion-focused intervention (OnTrack) and the KnowIt educational intervention both demonstrated significant within-group improvements in both diabetes distress and A1C, but no significant differences were observed between groups. Post hoc comparison of changes in A1C to a nonintervention sample indicated greater improvement in diabetes distress and A1C outcomes compared to a comparable separate sample of adults who did not receive an intervention.17
Taken together, this literature speaks to the need to improve diabetes distress interventions on both direct (diabetes distress) and indirect (A1C) outcomes. Such improvement is most likely to be successful when all available tools are leveraged to enhance outcomes across domains, including cognitive, educational, emotional, and motivational interventions. Previous interventions have used subsets of these domains but not all combined. One of the challenges to fully integrated interventions across these domains is the scope of expertise among diabetes care and education specialists (DCESs) and behavioral health professionals. DCESs are expert in education and motivational interviewing tools, whereas cognitive and emotional interventions are typically delivered by behavioral health professionals. There remains a need for a scalable intervention that draws on all of these domains simultaneously that can be easily adopted and implemented within the existing structure of health care systems. The purpose of the current study was to assess the feasibility and efficacy of the Diabetes Tune-Up Group (DTU), a multidisciplinary CBT-based group intervention for diabetes distress and glycemic outcomes. It was designed to deliver integrated psychoeducational, cognitive, motivational, and emotional interventions in a scalable manner. The aims of the study were (1) to assess the feasibility and acceptability of the intervention and (2) to assess changes in diabetes distress and glycemic outcomes from baseline to post-intervention and 3-month follow-up.
A pilot and feasibility study of DTU was conducted in adults with T1DM and T2DM using a multidisciplinary approach. DCESs and graduate-level clinical health psychology students were trained to use the CBT-based intervention to address diabetes distress using both didactic and experiential training approaches. The DTU intervention was then tested as a single-arm, pre-post intervention design with assessments occurring at baseline (T1), post-intervention (T2), and 3 months following completion of the intervention (T3). Adults with T1DM or T2DM of 1 year or greater duration, experiencing moderate or higher levels of diabetes distress and with an A1C ≥ 8.0% within 12 months of study entry were recruited from endocrinology and primary care clinics (see Figure 1 below).
The DTU intervention consisted of a 6-session, CBT-based curriculum delivered in an in-person group format (see Table 1 for an intervention outline). Groups were cofacilitated by a DCES and a clinical health psychology PhD student who had attained at least a master’s degree at the time of the intervention. Group sessions were held at multiple locations in the metropolitan area to reduce barriers to attendance.
Training for DCESs. DCESs and master’s-level clinical health psychology trainees attended a 3-hour classroom training course conducted by a licensed clinical health psychologist (the first author). The training included 3 primary components: (1) overview to the literature on diabetes distress and adherence to diabetes self-care and the purpose of the study; (2) an introduction to cognitive behavioral principles, practice, and their applicability to patients with diabetes; and (3) an overview of the DTU study protocol. Following the didactic training, each DCES cofacilitated the DTU intervention with a psychology trainee.
DTU intervention. The DTU consisted of 6 group sessions stratified by diabetes type (ie, separate groups for T1DM and T2DM). Each series was offered at clinical and community settings. Following eligibility screening, participants were assigned to the next available intervention series specific to matching type of diabetes. Enrollment in each series was capped at 10 participants to allow for all members to effectively engage in the group discussion. Intervention meetings were scheduled for a predetermined day and time for 6 weeks over the course of 2 months. The topics for each session are outlined in Table 1. At the end of session 1, each group selected 2 topics from a menu of diabetes education topics for presentation by the DCES cofacilitator in sessions 2 and 5. The list of topics that participants could choose from is presented in Table 2. Use of the selection menu provides patients with empowerment to tailor the content to the needs of the group. At the time the interventions were conducted (2018-2019), generally accepted practice was to utilize the terms “distress” and “burnout” interchangeably. As a result, we have used these terms interchangeably throughout this article.
The DTU was designed to (1) reduce diabetes burnout in patients with T1DM and T2DM, (2) improve adherence to self-care behaviors in patients with T1DM and T2DM, and (3) enable patients to identify and alter cognitions and behaviors that contribute to poor adherence. Each intervention session included psychoeducation, group discussion, and take-home assignments to be completed between sessions. The intervention components integrate key strategies that have been demonstrated to be effective in previous studies: motivational interviewing and introduction to stages of change; diabetes education; cognitive behavioral therapy with a particular emphasis on identification of cognitive distortions, thought stopping, and cognitive reframing; emotional regulation; social support and miscarried helping; and behavioral strategies to support diabetes self-management. Take-home assignments include setting goals for improving diabetes self-care, identifying negative thoughts and reframing as positive thoughts, and problem-solving barriers to achieving goals.
Recruitment. Recruitment took place from March 2018 through May 2019. Participants were recruited primarily via 3 methods: flyers, direct calls, and electronic medical record (EMR) messaging. Flyers were placed in endocrinology offices in the local health system. For direct calls, a list was received from the diabetes clinics of patients who had met with a DCES in the previous year. Study staff reviewed the EMRs for those patients on the list to assess initial eligibility by reviewing A1C results. Those with an A1C value of 8.0% or greater in the previous year were contacted, either by phone or, for those patients who utilize the online EMR patient portal, direct message through the patient portal. Individuals who were interested in learning more were directed to call the study coordinator at the number listed on the flyer.
Once reached by phone, interested patients were provided with detailed study information and screened for eligibility. Patient eligibility criteria included: (1) age ≥ 21 years, (2) diagnosis of T1DM or T2DM for at least 1 year, (3) an A1C value of at least 8.0% in the last year, and (4) a Diabetes Distress Scale18 (DDS) score of at least 2.5 at the time of screening. Exclusion criteria included (1) limited English proficiency, (2) presence of a serious mental illness that would preclude group participation (eg, psychotic disorders or severe mood disorders with suicidal ideation), (3) diagnosis of an acute medical disorder (eg, myocardial infarction, stroke, cardiac rehabilitation) within the last 3 months, and (4) initiation of new medical treatment regimens (eg, chemotherapy) for medical diagnoses that would require the patient’s primary attention.
Patients were screened for eligibility by the study coordinator using standardized medical history questions, the Structured Clinical Interview for the DSM-IV Screener, and the DDS. Individuals with a DDS mean item score of 2.5 (moderate severity) or greater and who met other eligibility criteria were invited to participate.
Data collection. Baseline assessments took place 1 hour prior to the first session of the DTU. Post-intervention assessments occurred at session 6. The 3-month follow-up assessment took place approximately 3 months following completion of the intervention.
All measures were assessed at baseline (T1), post-intervention (T2), and 3-month follow-up assessment (T3), with the exception of demographics (T1 only) and the patient satisfaction questionnaire (T2 and T3 only). Participants received $20 for completion of the baseline assessment, $30 for completion of the post-intervention assessment, and $40 for completion of the 3-month follow-up assessment. All questionnaires were completed by participants on paper and entered into a REDCap database by the study team.
Demographic information (ie, age, race/ethnicity, education, marital status, diabetes type, duration of diabetes) was assessed at T1 using a self-administered demographic questionnaire.
DDS. Diabetes distress was assessed using a 17-item version of the Diabetes Distress Scale (DDS-17). The DDS- 17 presents patients with single-sentence phrases describing psychosocial experiences consisting of 4 subscales: emotional burden, regimen distress, relationship with health care providers, and interpersonal relationships (eg, “Feeling that diabetes is taking up too much of my mental and physical energy everyday”). Patients rated the degree to which each item is currently problematic for them on a Likert scale from 1 (no problem) to 6 (serious problem). A total mean item score (range 1-6) was calculated. Cronbach’s alpha for the DDS-17 is 0.93.18 The measure was used for screening for eligibility (mean item score ≥ 2.5) and administered at each of the assessments. A mean item score of between 2.0 and 2.9 is indicative of moderate distress, and 3.0 and above indicates high distress.
Diabetes Knowledge Test19 was assessed using an adapted version of the Michigan Diabetes Knowledge Test, a 23-item knowledge test battery developed by the Michigan Diabetes Research Training Center. These 23 items represent a test of general knowledge of diabetes. The first 14 items are appropriate for people who do not use insulin. All 23 items can be administered to people who do use insulin. The total score is reported as the percentage of items correctly answered by the participants. The full test has a reliability coefficient of 0.77, and the insulinuse subscale has a coefficient of 0.84.19
Patient health questionnaire. Depressive symptoms were measured at each time point using the Patient Health Questionnaire (PHQ-9).20 The PHQ-9 presents the patient with statements of symptoms of depression based on the diagnostic criteria listed in the Diagnostic and Statistical Manual of Mental Disorders. Participants indicated how often they have been bothered by each symptom over the past 2 weeks, responding with a number from 0 (not at all) to 3 (nearly every day) to indicate the frequency of their symptoms. Higher scores indicate more severe levels of depressive symptoms. In people with diabetes, scores of at least 7 on the measure have been shown to have a sensitivity of 81.0% and specificity of 87.7% at detecting depressive disorders.21
Short Form-12. The Short Form-1222 (SF-12) is a 12-item measure of general health status. Items assess physical functioning, pain, role limitations, and mental health functioning. Response formats vary by item but are either on a scale from 1 to 3 (1 = limited a lot, 2 = limited a little, 3 = not limited at all) or from 1 to 5 (all of the time, most of the time, some of the time, a little of the time, none of the time). Items are divided into 2 subscales, the Physical Component and Mental Component scores. Total item and subscale scores are converted to a 100-point scale, with higher scores indicating better functioning. The SF-12 has been validated for use in multiple age groups, diseases, and conditions, including diabetes and depression.22
Confidence in Diabetes Self-Care. The Confidence in Diabetes Self-Care23 (CIDS) scale is a 20-item questionnaire assessing diabetes-specific self-efficacy. All items begin with the stem, “I believe I can . . . .” Items relate to various tasks specific to diabetes self-care such as “Treat a high blood glucose correctly” and “Plan my meals and snacks according to dietary guidelines.” Responses are on a 1 to 5 Likert scale, with 1 indicating “No, I am sure I cannot” and 5 indicating “Yes, I am sure I can.” Total scores range from 0 to 100, and higher scores indicate higher perceived self-efficacy. The CIDS was found to be highly internally consistent (0.90) and negatively correlated with A1C (r = −0.25) in the scale’s original test population.
Satisfaction questionnaire. A brief self-report satisfaction form was created to capture participant experience of each element of the intervention, including intervention content, format of delivery, perceived usefulness of the material, and overall experience. Participants were presented with components of the intervention and asked to rate how satisfied they were with each component and how useful they found each component. Responses were on a Likert scale from 1 to 5 (very dissatisfied to very satisfied and not at all useful to very useful). Higher scores indicate greater satisfaction and usefulness of the intervention.
A1C. A1C was measured using the DCA 2000 Analyzer (Bayer Diagnostics, Inc, 1998). The DCA 2000 Analyzer provides glycated hemoglobin data from whole blood samples using the measurement of glycated fractions of Hemoglobin A, which reflects the glucose level in the blood over approximately a 3-month time span. The DCA 2000 Analyzer has demonstrated acceptable rates of precision and clinical utility in multiple clinical trials.24-26
Statistical analyses were conducted using SAS 9.5 (SAS Institute, 2017). Data were analyzed for normality, skewness, and kurtosis. Paired t tests and descriptive statistics were used to assess change from T1 to T2 and T3.
Feasibility. Feasibility of the DTU was examined using descriptive statistics. Enrollment rates were calculated as the ratio of patients who consented to receive the intervention to the total number of patients who screened eligible to participate. Attendance was calculated as the ratio of the number of group sessions attended to the 6 potential sessions each patient could attend. Retention rates were calculated as the ratio of number completing the postintervention assessment to the number who completed the baseline assessment.
Outcomes. Paired t tests were used to assess the changes in each questionnaire from T1 to T2 and T1 to T3 time points. The primary outcome of interest was change in A1C at T2 and T3 compared to T1. The change in participant psychosocial measures at T2 and T3 from T1 was also calculated. The changes in scores from T1 to T2 and T1to T3 were analyzed using t tests. To be included in these analyses, an individual participant must have had a score on the individual measure at each time point being analyzed, T1 and T2 or T1 and T3. Cohen’s d was calculated for change in outcome measures at T2 and T3. For Cohen’s d calculations, effect sizes of 0.2, 0.5, and 0.8 are interpreted as small, moderate, and large, respectively.27
Of the 41 people who screened eligible for study participation, 29 enrolled, for an enrollment rate of 70.7%. Table 3 presents demographic information for the study participants. Overall, participants were female (79%), married (60%), White (59%), and highly educated (56% had at least a college degree). All participants had health insurance and a primary care provider. Participants attended 131 total sessions out of 174 possible sessions, for an overall attendance rate of 75.3%. Retention at the post-intervention and 3-month follow-up assessments was 86.2% and 65.5%, respectively.
The group means for outcomes measures are presented in Table 4. For the group as a whole, significant improvements were observed in A1C values at T3 (a mean decrease of 0.44%; P < .05) Diabetes distress (DDS) and depression symptom (PHQ-9) scores improved at both T2 and T3 assessments (decreased scores on the DDS and PHQ-9 indicate improvements on the measures). DDS scores decreased from a mean of 3.44 (SD = 0.68) at T1 to 2.94 at T2 (SD = 0.81; P < .001) and 2.55 at T3 (SD = 0.75; P < .001). PHQ-9 scores also decreased from T1 (M = 8.20, SD = 5.55) to T2 (M = 6.75, SD = 5.14) and T3 (M = 5.37, SD = 5.31).
Using paired data, Cohen’s effect size data were calculated for all measures. Change in A1C was observed to have a moderate effect size (Cohen’s d between 0.50 and 0.79). Effect sizes in change of diabetes distress (DDS) scores were large in magnitude at both T2 (d = 0.86) and T3 (d = 1.18). Effect sizes for changes in depressive symptoms (PHQ-9) were observed to be small at T2 (d = 0.39) and approximately moderate at T3 (d = 0.46).
CIDS scores increased across all groups (T1: M = 58.90, SD = 5.55; T2: M = 69.62, SD = 10.98, P = .016; T3: M = 77.11, SD = 14.90, P < .001), indicating improvements in diabetes-related self-efficacy. Significant improvements were also observed in health-related quality of life on the Physical Component Score of the SF-12 (T1 M = 44.19, SD = 10.94), indicating that participants felt better about their physical health at the post-intervention (T2: M = 44.16, SD = 9.75, P = .014) and 3-month follow-up (T3: M = 47.67, SD = 7.34, P = .018) assessments compared to baseline.
The satisfaction survey used a 5-point Likert response format ranging from 1 (very dissatisfied) to 5 (very satisfied) and calculated the average item score. Results indicated that all participants were somewhat to very satisfied with all aspects of the intervention (M = 4.59, SD = 0.40).
Evaluation of outcomes by diabetes type (see Table 5) indicated similar trends to those observed in the whole sample. Both T1DM and T2DM groups had lower mean A1C values at T3, although these reductions were not statistically significant, likely due to the small sample sizes. The effect size for the T2DM group was moderate (d = 0.56) and small for the T1DM group (d = 0.37). Similar patterns were seen in the T1DM group compared to the whole sample. For the T2DM group, significant changes were seen in DDS from T1 (M = 3.49, SD = 0.71) to T2 (M = 2.82, SD = 0.71, P < .002) and T3 (M = 2.59, SD = 0.74, P < .004). Fewer significant results were seen for the T1DM group, likely due to the small sample size. Diabetes distress scores improved significantly in people with T1DM from T1 (M = 3.30, SD = 0.60) to T3 (M = 2.45, SD = 0.81, P = .015).
This study tested the feasibility and psychosocial outcomes of the DTU intervention, a group-based CBT approach to addressing diabetes distress in people with T1DM and T2DM using a manualized approach for use by a multidisciplinary team. High levels of retention and intervention satisfaction were observed from participants, suggesting that the group delivery of this content was both feasible and acceptable to participants. Qualitative data from participants indicated that the availability of flexible meeting locations in the community was preferable to meeting within the health care system. Additional work is underway to adapt to the intervention to virtual formats that have emerged since the onset of the COVID-19 pandemic. This study also demonstrated the feasibility of cofacilitated groups comprised of DCESs and behavioral health providers. The acceptability of this format suggests that the addition of behavioral health professionals to an intervention targeting diabetes distress extends the breadth and depth of interventions that can be offered to patients.
Evaluation of psychosocial and medical outcomes observed improvements in A1C, diabetes distress, and other psychosocial measures at post-intervention and 3-month follow-up time points. The effect sizes observed in changes from baseline to follow-up time points far exceed those reported in the meta-analysis by Mathiesen et al,14 suggesting that the multipronged approach of this intervention and the multidisciplinary nature of the intervention may have favorable synergistic effects on primary outcomes and quality of life. An effective intervention that addresses emotional, cognitive, and behavioral aspects of diabetes is consistent with Fisher et al’s9 hypothesis that addressing emotional aspects of diabetes facilitates use of behavioral strategies that in turn improve medical outcomes. Further research is needed to evaluate the long-term impact on diabetes distress and A1C in larger samples of adults with type 1 and type 2 diabetes.
Limitations of the current study include the small sample size, which limited evaluation of differences by diabetes type, and the absence of a control group. While this study and sample size was designed to test feasibility, significant and robust changes in outcome variables were observed that suggest that further evaluation of this intervention shows promise for impact on patient outcomes when delivered on a larger scale.
This study demonstrated the feasibility and positive outcomes of a diabetes distress and adherence support intervention delivered by DCESs and behavioral health professionals that is designed for scalability in diabetes centers to support the delivery of integrated patient care. Future studies will focus on adapting and providing the intervention to larger groups of patients with diabetes using flexible virtual formats to demonstrate efficacy and generalizability on a larger scale.
The current study’s findings suggest that diabetes-related distress can be effectively treated by combining diabetes education with cognitive behavioral therapy using a manualized group treatment approach. While DCES professionals are the frontline workers in addressing diabetes distress, education alone is not sufficient to address diabetes distress that is endemic to the current treatment environment. An approach that leverages the expertise of a multidisciplinary team and equips DCES professionals with the full range of intervention tools is important to improve diabetes distress and A1C. Future research is needed to establish the comparative effectiveness of this approach to treatment approaches currently in practice.
We would like to thank the participants and interventionists who helped to make this study a success. The diabetes care and education specialists included: Holli Bowser, RD, CDE; Lisa Gold, RN, CDE; Kate Haynes, RN; Ariane Sturdy, RN, CDE; Deb Vine, RD, CDE; and Kerry Warner, RD, CDE. The psychology trainees were: Loretta Hsueh, PhD; Brittanny Polanka, PhD; and Ekin Secinti, MS. In addition, we thank Tonya Somers, RD, CDE, administrator for the IU Health, Inc. Diabetes Center, for her collaboration.
This study was funded by an Indiana University Health Values Fund for Education grant.
Mary de Groot https://orcid.org/0000-0003-4021-8824
From Indiana University School of Medicine, Indianapolis, Indiana (Dr de Groot, Ms Myers, D. Baker, Dr Cavaghan), and Indiana University Health, Indianapolis, Indiana (Dr de Groot, Dr Baker, Ms Daily, Dr Cavaghan).
Corresponding Author:Mary de Groot, Indiana University School of Medicine, 410 W. 10th St., Suite 3100, Indianapolis, IN 46202, USA.Email: mdegroot@iu.edu
