The Science of Diabetes Self-Management and Care2023, Vol. 49(6) 426 –437© The Author(s) 2023Article reuse guidelines:sagepub.com/journals-permissionsDOI: 10.1177/26350106231205030journals.sagepub.com/home/tde
Abstract
Purpose: The purpose of this study was to identify factors impacting the acceptability of continuous glucose monitoring (CGM) in adolescents and young adults (AYAs) with type 2 diabetes mellitus (T2DM).
Methods: In this single-center study, semistructured interviews were conducted with AYAs with T2DM and their parents to determine attitudes about CGM, including barriers and facilitators. Interviews were audio-recorded, transcribed, and evaluated using thematic analysis.
Results: Twenty AYAs and 10 parents participated (n = 30 total). AYAs were mean age 16.5 years (SD 2.2, range = 13.7-20.1) and had median diabetes duration of 1.3 years. Most were female (65%) and from minoritized background (40% non-Hispanic Black, 10% Hispanic, 5% Asian). Seven (35%) used CGM. The primary facilitator elicited was convenience over glucose meter use. Important barriers included the impact of physically wearing the device and drawing unwanted attention, desire for AYA privacy, and inadequate education about the device.
Conclusions: In this diverse sample of AYAs with T2DM and their parents, CGM was generally regarded as convenient, although concerns about worsening stigma and conflict with parents were prevalent. These findings can guide the development of patient-centered approaches to CGM for AYAs with T2DM, a critical step toward reducing inequities in diabetes technology uptake.
Youth-onset type 2 diabetes (T2DM) is rising in incidence,1 is associated with high rates of comorbidities at diagnosis,2 and leads to severe complications in early adulthood.3 A central part of management for youth with T2DM is promotion of positive health behaviors,4 including healthy diet and physical activity, but such change is often difficult to achieve. One method to support diabetes self-management and healthy lifestyle change in individuals with T2DM is home glucose monitoring.5,6 For adolescents with T2DM, however, low adherence to glucose monitoring is common. In the Treatment Options for Type 2 Diabetes in Adolescents and Youth study, which enrolled adolescents with recently diagnosed T2DM, nearly one-quarter did not engage in recommended, incentivized, twice-daily home glucose monitoring at all in the first 2 months of participation, and nearly one-half did not at 2-year follow-up.7 Notably, higher glucose monitoring adherence was associated with lower A1C.7 Factors associated with lower adherence included older age and higher body mass index, blood pressure, and number of diabetes- and obesity-related comorbidities, suggesting that burden (ie, medication, emotional) of comorbidities may contribute to poor adherence to glucose monitoring.7
Among youth with type 1 diabetes, in whom glucose monitoring is critical for safe and effective diabetes management, identified barriers include forgetting to check and not having a glucose meter available when needed.8 Thus, continuous glucose monitoring (CGM), via use of small wearable devices that can be used for 10 to 14 days at a time, can overcome logistical barriers and contribute to significant improvements in glycemic control for youth with type 1 diabetes.9 For youth with T2DM, CGM could serve to minimize the burden of frequent blood glucose monitoring, which may ultimately facilitate modifications to medication or health behaviors via improved glycemic awareness. Evidence related to clinical benefits for youth with T2DM has been limited10 but is now emerging,11 following trends in adult T2DM care.12
A first step toward planning for widespread uptake of CGM among youth with T2DM is understanding patient perspectives. To date, attitudes about CGM by youth with T2DM have been explored in only 1 study with 5 participants.11 This is a particularly important knowledge gap given the stark racial/ethnic disparities in diabetes technology use in youth and young adults with type 1 diabetes13,14 and the disproportionate prevalence of T2DM among racial/ethnic minoritized youth.15 Thus, the objective of this qualitative study was to elicit perspectives from a racially diverse group of adolescents and young adults with T2DM about CGM for T2DM management and support of healthy lifestyle.
This qualitative descriptive research design involved semistructured interviews conducted with youth with T2D and their parents. The research team aimed to use everyday language to describe experiences of participants.16,17 Using NVivo software, interview transcripts were evaluated using thematic analysis.
For this qualitative study, adolescents and young adults (AYAs) with T2DM, ages 10 to 22 years old, were recruited from the UPMC Children’s Hospital of Diabetes Clinic, a large, academic, pediatric hospital-based clinic in Pittsburgh, Pennsylvania, USA. Purposive sampling was used to ensure (1) representation of roughly equal proportions of younger (10-15 years) and older (16-22 years) adolescents, (2) representation of the racial composition of patients with T2DM who attend our clinic, and (3) enrollment of at least 25% of the final study sample who used CGM for diabetes management to elicit perspectives from those with lived experience with the technology. Due to the important role caregivers play in pediatric diabetes management,18,19 parents of participants younger than 16 years were invited to participate in a joint interview with their child. Informed consent and assent were obtained, as applicable based on participant age. The study received approval from the University of Pittsburgh Institutional Review Board. Our study results are reported according to the Consolidated Criteria for Reporting Qualitative Research reporting guideline for qualitative research.20
Semistructured individual youth or parent-youth dyad interviews were conducted via HIPAA-compliant video conferencing software (Zoom Video Communications Inc). The research team developed an interview guide that drew on the technology acceptance model (Table 1),21 which includes perceived usefulness and ease of use as key determinants of acceptance of digital technologies, such as CGM use among patients with type 1 diabetes.22 The interview guide was initially developed by the principal investigator (MEV), a pediatric endocrinologist with formal training in qualitative research methods. The guide was iteratively refined based on pilot testing with non-medicalprofessional study team members. Interview questions covered the AYA’s lived experience with T2DM and baseline knowledge and attitudes about or experience with CGM use. To overcome any lack of prior knowledge or experience about CGM, the session included an explanatory slideshow about CGM, including examples of the data provided by the devices.
Interviews were conducted by VP, a clinical research assistant trained in conducting semistructured interviews. The interview guide was structured to be adaptable such that interview and follow-up questions were flexible based on the participant’s experience. The interviews took place between June and August 2022 and lasted 30 to 60 minutes. For joint interviews, which were performed with caregivers and younger adolescents, nearly all interview questions were directed to the adolescent to elicit their responses first; caregivers were invited to respond as needed to support their child due to any uncertainties and to state agreement or share differing opinions. Caregivers were also separately asked for their additional perspectives and insights beyond their responses to interview questions asked to their child. Interviews were audio-recorded and transcribed verbatim by a professional transcription service, after which they were reviewed by VP and corrected as needed. Transcripts were deidentified before analysis. Thematic saturation, with no additional concepts emerging, was reached with 20 AYA participants.
Medical record review was performed to obtain medical history and the most recent (within 1 year) laboratory and anthropometric data obtained at a clinic visit. Additional survey questions were asked of participants or parents directly as needed, including self-identified race, ethnicity, and gender.
Transcripts were analyzed using thematic analysis.23 First, an initial set of codes was created based on the interview guide (eg, CGM baseline knowledge/experience, concerns about use, attitudes about data sharing). Second, 2 study team members (VP, MEV) coded the first 3 transcripts together to minimize any discrepancies. Third, the same study team members individually coded all remaining transcripts, meeting frequently to discuss new codes and to resolve coding discrepancies. To enhance trustworthiness of our findings,24 we used member checking within the diverse study team (undergraduate and medical students, public-health-trained study coordinator, medical trainees, attending clinicians and researchers) and with peers, including through presentation of preliminary findings to clinicians, clinical staff, and researchers within pediatric endocrinology. Dependability of findings was addressed via double coding of all transcripts by 2 members of the research team. The multidisciplinary study team met to develop themes based on similar finalized codes, and select quotations were chosen to illustrate each theme. To facilitate data coding and retrieval, NVivo software version 12 (QSR International) was used. Summary statistics were used to describe and compare continuous (mean, standard deviation, t test) and categorical (proportion, chi-square test) demographic and clinical variables.
AYA participants (n = 20) had a mean age of 16.5 years (SD 2.2; range 13.7-20.1) and median diabetes duration of 1.3 (IQR 0.6-2.6; range 0.3-7.1) years. All had obesity, with mean body mass index of 38.9 (SD 6.6) kg/m2. Eight participants (40%) were non-Hispanic Black, 2 were Hispanic (10%), 1 (5%) was non-Hispanic Asian, and 9 (45%) were non-Hispanic White. Twelve participants (61.5%) identified as female, 7 (30.8%) as male, and 1 (7.7%) as nonbinary gender. Eight (40%) participants had only public insurance. Nearly all (17/20, 85%) were prescribed metformin, 8 (40%) were prescribed insulin, and 3 (15%) were prescribed a glucagon-like peptide 1 receptor agonist. Seven (35%) were actively using CGM (n = 4 Freestyle Libre 2; n = 3 Dexcom G6), with 1 additional planning to start use soon. The most recent A1C was a median of 46.4 mmol/mol (6.4%; IQR 37.7-48.6 mmol/mol, 5.6%-6.6%). Body mass index was lower among youth who used CGM (mean 34.4 vs 40.8 kg/m2, P = .03; Table 2). Caregiver participants (n = 10) included 8 mothers and 2 fathers who had varying levels of education (n = 3 high school, GED, or technical degree; n = 4 some college or associate degree; n = 1 bachelor degree; n = 2 graduate degree). T2DM was common among parents, reported in 7 or 10, and prediabetes was reported in 1 of the 3 parents without T2DM.
Four themes were identified, consisting of benefits (Theme 1) and potential barriers (Themes 2-4; Figure 1).
Theme 1: convenience and favorability over finger sticks. Most AYA participants were interested in using CGM due to perceived convenience and favorability over finger sticks. Caregivers shared a similar perspective and believed that using a CGM would be reassuring and helpful in their child’s diabetes management.
AYA participants reflected on the challenge of remembering to check blood glucose and to take medications (Table 3, Quote 1 [Q1]). Several who had not previously used CGM expressed interest in use because they believed it could make glucose monitoring more convenient. The ability to check glucose without pricking a finger, which was described as painful and burdensome, was appealing for many (Q2).
All CGM-using participants emphasized the convenience of CGM over glucose meter (Q3). Some also described changing their diabetes management due to the device, including dietary changes (Q4). One participant explained that her analysis of the CGM data, thanks in part to its alerts, helped her to make lifestyle changes, stating she made these changes “[so I] won’t hear as much of the alarms because [my sugar] won’t be as high” (17, female, CGM user).
Theme 2: physical and auditory barriers. Although most AYAs were interested in the practicality of the device, some shared their concerns about the appearance of the device on their skin and potentially disruptive alarms. AYAs considered these barriers as important factors in affecting their decision to begin using CGM.
AYAs noted that device size and visibility were important considerations (Q5). Of the 13 participants not using CGM, 6 (46%) were unsure of whether they would be willing to use CGM due to wearability concerns (Q6). Participants also reflected on the stigma surrounding T2DM and the ways in which CGM, which may be visible to others, could draw unwanted attention to their diagnosis (Q7). A CGM-using adolescent highlighted how the curiosity of others can make use uncomfortable and potentially limit their social interactions with others (Q8).
Audible alerts also contributed to disruptions in daily life. CGM-using adolescents expressed that alarms can “get annoying” in public (Q9). Audible alerts also disturbed sleep. These were felt to be particularly frustrating when reflecting inaccurate data (Q10).
Theme 3: desire for privacy and autonomy, in conflict with parents. AYAs were concerned about the lack of privacy due to the sharing feature of the device. They feared losing autonomy because this sharing feature allows their caregivers to monitor their blood glucose levels at any time. However, caregivers positively viewed this feature and believed it would be useful in their child’s diabetes care.
When asked about their views on the option for parents to track their child’s glucose in real time, many adolescents raised concerns, including hesitancy about being continuously monitored by their parent while in public (Q11). One CGM-using adolescent described difficulty at school that resulted from his parent’s real-time monitoring of the CGM and response to hyperglycemia (Q12). However, others acknowledged the added parental support that was made possible through real-time sharing (Q13, Q14).
All parents were in support of their child’s use of CGM for several reasons. Some felt a CGM could provide overall reassurance via alerts when glucose is out of range, while others emphasized the benefits of the sharing feature, which could improve their ability to be involved with their child’s diabetes care (Q15). Parents of CGM-using adolescents expressed positive views toward the device because they felt it provided them with a better understanding of their child’s diabetes and how to manage the condition (Q16).
Theme 4: need for adequate education about CGM use. CGM users shared difficulties experienced when troubleshooting the device and emphasized the importance of proper education about CGM. Caregivers were also interested in being part of this learning process and requested various forms of instruction, including hands-on training.
AYAs were concerned about practical issues related to device use, including engaging in troubleshooting (Q17). Preferred methods to learn about CGM use varied. Most preferred videos, but additional preferred methods included slideshows, handouts (Q18), and hands-on instruction (Q19). Family and friends using CGM also played an important role in AYA decision-making about CGM, providing comfort (Q20) and support in getting started (Q21). Those who did not know anyone who had used CGM also noted the potential value of learning through the experiences of others when making their own decision on use (Q22).
The potential to use CGM data to inform lifestyle change or diabetes medication management was a novel concept to many participants and several parents (Q23). However, information from CGM alone was not always enough to drive behavior change (Q24). Some parents also expressed concerns that their child would need very clear explanation of how to use CGM to support health behavior change.
In this qualitative study including a racially diverse group of AYAs with T2DM and parents, the primary perceived benefit of CGM was convenience, with some AYAs and parents also recognizing the potential to support changes in diabetes management or health behaviors. Several barriers to use arose, including potential worsening of preexisting diabetes-related stigma due to calling attention to the diagnosis, impingement on privacy and autonomy, and the need for appropriate education about use. These barriers were recognized by both CGM-using and nonusing AYAs, highlighting the value of proactively addressing these concepts when engaging in shared decision-making about CGM initiation and continuation among AYAs with T2DM. The authors acknowledge race as a social construct.
Although glycemic benefits of CGM use for youth with T2DM are not yet sufficiently demonstrated,10 this technology will likely become an increasingly utilized and valuable tool for several reasons. First, CGM use by adults with T2DM on insulin is now endorsed by the American Diabetes Association due to its beneficial effect on glycemic control.25-28 Second, due to the frequent and rapid progression to insulin dependence among youth with T2DM,29 glycemic control and ongoing monitoring is critical to reduce risk of severe complications in early adulthood.3 Third, the rising prevalence of T2DM in youth15 necessitates a flexible approach to glycemic monitoring and diabetes management that incorporates use of remote monitoring and interaction, particularly given the barriers to clinic visit attendance experienced by many.
To our knowledge, only 1 other study has elicited perspectives about CGM use among AYAs with T2DM; this study assessed changes in quality of life and acceptability in AYAs who completed a 12-week period of real-time CGM use.11 AYAs reported improved quality of life (n = 7 evaluated) and general convenience and acceptability of CGM use (1 focus group of 5 AYAs). However, studies of adolescents with type 1 diabetes have shown that perceptions of CGM use may be divided. For some, CGM is viewed as a useful tool that can lower diabetes distress due to easy access to blood glucose data and the ability to enhance understanding of diabetes management; for others, it is perceived to be more of a burden.30,31 The more negative views about CGM use were associated with more pessimistic attitudes about diabetes technology and lower self-efficacy.30 Among AYAs with type 1 diabetes, CGM uptake is significantly lower among non-Hispanic Black and Hispanic individuals than among non-Hispanic White individuals.13,32 These disparities are driven by both lower initiation and higher discontinuation among minority youth,14,32 which may be due in part to provider inadvertent withholding of information or access to technology and failure to address AYAs’ concerns about technology use.33 The disproportionate impact of youth-onset T2DM on racial/ethnic minority youth make inequities in diabetes technology use especially critical to address proactively.
This study highlights the danger of worsening perception of diabetes stigma through the use of CGM. In adults with T2DM, diabetes-related stigma is associated with greater depressive and anxiety symptoms,34 which are common in youth with T2DM.35 In our sample, 7 of 20 (35%) AYA participants had a history of anxiety, depression, or both. Our findings are consistent with a study of Australian youth with T2DM, who were also from a minority, lower-income population.36 This study found evidence of diabetes-related shame, with participants reporting hesitancy about disclosing their diagnosis to peers.36 Due to the tight link between T2DM and obesity, youth with T2DM may also already face weight-related stigma, which increases the risk of disordered eating behaviors, anxiety and depression, and health-related quality of life.37 Thus, the risk of worsening diabetes-related stigma through CGM use should be carefully weighed using a shared decision-making approach with youth and their parents. Notably, the authors did find that CGM-using participants had lower BMI than non-CGM participants; however, our study was not designed to assess causal relationships between glycemic or health outcomes and CGM use.
An important consideration at the time of this study was the lack of insurance coverage for CGM use for youth with T2DM, particularly those not on multiple daily injections of insulin.38 Thus, our focus was not on elucidating barriers to prescription or initiation of CGM but, rather, to evaluate patient-level factors associated with decisionmaking about use if it were available. However, our identified theme of need for adequate education to support CGM use and the value of speaking with other users are similar to findings from a qualitative study with youth with type 1 diabetes.39 In addition, the importance of learning about CGM use through video and hands-on demonstration was noted by AYAs in our study and youth with type 1 diabetes.39 These findings highlight the need for implementation strategies that incorporate patient-preferred methods of exposure to the technology and education.
This study has several strengths, most notably the direct assessment of perspectives of a diverse sample of AYAs with T2DM and their parents to guide future strategies to address CGM use in this population. The study included an initial educational session, leading to a more informed discussions among a group of individuals who may not otherwise have been aware of CGM due to either clinical indications or differential technology prescription by race or insurance type.40,41 Inclusion of parents of younger adolescents gave voice to individuals whose preferences must be addressed when making recommendations about diabetes management. Parent interview data were coded similarly to AYA interview data and provided important insight on perspectives of those who are involved in their child’s diabetes management.17 The study also has important limitations. As is true for many qualitative studies, given the small sample, findings are not necessarily generalizable. Specifically, the mean glycemic control of the participants was at target,42 which could correlate with higher self-efficacy and more positive attitudes about diabetes technology. However, CGM was not universally favored, and perspectives reflecting hesitancy about technology use did emerge. Additionally, although we sought to maximize unique, youth-specific responses by requesting that youth participants respond first to questions during dyad interviews, the presence of parents during interviews (or in the room during an individual youth interview) may have biased their responses toward perspectives they felt comfortable sharing with their parents. Another important limitation is exclusion of participants with limited English proficiency, which may reduce applicability of results to Hispanic youth, who have a higher prevalence of T2DM than non-Hispanic White youth.43 Although 2 out of 20 AYA participants in our study identified as Hispanic, to truly address barriers of CGM uptake among Hispanic youth, future studies should include the option of interviews conducted in Spanish. Finally, our sample of AYAs who used CGM was small; larger qualitative studies of youth with T2DM who have used CGM would provide greater insight into lived experiences with the devices, including how and whether they use the data to address diabetes-related health behavior change and relevant barriers.
In conclusion, in this sample of AYAs with T2DM, CGM was perceived to have several benefits, most notably convenience. In light of the frequent poor adherence to glucose self-monitoring in adolescents with T2DM, the detailed real-time data made available by CGM could be a valuable tool for improving glycemic control in AYAs with T2DM. However, CGM was also noted to carry the risk of increasing T2DM stigma and limiting privacy and autonomy. Careful approaches to education at the time of CGM initiation are needed to increase likelihood of successful uptake and continuation. As CGM becomes more widespread in AYAs with T2DM, efforts must be made to reduce the potential development of racial/ethnic disparities in technology use. These findings highlight important benefits and barriers to discuss with patients, but ongoing research to uncover and address barriers to equitable implementation will be needed.
We sincerely thank the participants for generously sharing their perspectives with us.
Ms Vaishnavi Peyyety designed the study, developed the interview guide, conducted interviews, performed qualitative and quantitative analysis, and drafted the initial manuscript. Dr Mary Ellen Vajravelu designed the study, developed the interview guide, conducted interviews, performed qualitative and quantitative analysis, and revised the initial manuscript draft. Ms Brianna Hewitt, Dr Adriana Rodriguez Gonzalez, and Dr Iswariya Mani piloted the interview and provided critical review of the manuscript. Dr Margaret F. Zupa, Dr Temiloluwa Prioleau, Dr Jessica McCurley, and Dr Yu Kuei Lin provided critical review of the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
Interests The authors have no relevant conflicts of interest.
ARG was supported by an Endocrine Fellows Foundation Grant. JM was supported by NIH K23HL157763. YKL was supported by NIH K23DK129724. MZ was supported by NCATS KL2TR001856. MEV was supported by NIH K23DK125719. Quantitative data collection was supported through the Clinical and Translational Sciences Institute at the University of Pittsburgh through NIH UL1-TR-001857. The NIH had no role in the design and conduct of the study.
Margaret F. Zupa https://orcid.org/0000-0002-8244-9295
Yu Kuei Lin https://orcid.org/0000-0003-1988-7046
Mary Ellen Vajravelu https://orcid.org/0000-0002-5839-7923
From Division of Pediatric Endocrinology, Diabetes, and Metabolism and Center for Pediatric Research in Obesity and Metabolism, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania (Ms Peyyety, Ms Hewitt, Dr Mani, Dr Vajravelu); UPMC Children’s Hospital of Pittsburgh, Pittsburgh, Pennsylvania (Ms Peyyety, Ms Hewitt, Dr Rodriguez Gonzalez, Dr Mani, Dr Vajravelu); Division of Endocrinology and Metabolism, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania (Dr Zupa); Department of Computer Science, Dartmouth College, Hanover, New Hampshire (Dr Prioleau); Department of Psychology, San Diego State University, San Diego, California (Dr McCurley); and Division of Metabolism, Endocrinology and Diabetes, University of Michigan, Ann Arbor, Michigan (Dr Lin).
Corresponding Author:Mary Ellen Vajravelu, UPMC Children’s Hospital of Pittsburgh, 4401 Penn Ave, Faculty Pavilion 6th Floor, Pittsburgh, PA 15224, USA.Email: vajravelume@upmc.edu
