Disparities in diabetes technology utilization in youth with diabetes

WHAT IS ALREADY KNOWN ON THIS TOPIC Despite the increase in diabetes technology, there are marked disparities in utilization in which non-Hispanic White children have the highest percentage of insulin pump use compared with non-Hispanic Black and Hispanic children.WHAT THIS STUDY ADDS This original research article adds to the existing literature on health disparities with the perspectives of historically minoritized populations and those with limited English proficiency.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE, OR POLICY This study found that clinical providers place more weight on subjective factors rather than objective criteria and perceive high burden to use of technology, though this burden to technology was not expressed by patients. These findings can help in future clinical practices to have more standardized criteria for technology discussion.Introduction There have been marked advances in diabetes care with the introduction of diabetes technologies such as insulin pumps and continuous glucose monitors (CGM). Compared with multiple daily injections or fixed insulin doses, insulin pumps, particularly the more advanced automated insulin delivery systems (AID), can lower hemoglobin A1c (HbA1c), improve metabolic variability, and decrease the incidence of both diabetic ketoacidosis and severe hypoglycemia. 1 2 In addition, CGM use reduces HbA1c across all ages and insulin administration methods and reduces mild to moderate hypoglycemia.3 4 While technology use over the past few years has increased5 and the benefits are clear, there are persistent disparities nationwide in technology utilization.6 7 When controlling for gender, age, and diabetes duration, non-Hispanic White (NHW) children have the highest percentage of insulin pump use compared with non-Hispanic Black (NHB) and Hispanic children with type 1 diabetes (T1D).7 This disparity persists even when corrected for insurance type, household income, and parental education, suggesting the disparities are not only attributable to socioeconomic factors.7The reasons for disparities in diabetes technology utilization are likely multifactorial and include provider and patient perceived barriers to technology uptake. Prior studies have shown provider-implicit bias pertaining to diabetes technology recommendations8 and different levels of clinician readiness in promoting newer diabetes technologies,9 which may unintentionally contribute to disparities. Medical providers are less likely to discuss potentially complex treatment plans with patients who they perceive to be less medically competent, and surveys of physicians indicate that these are often patients from historically marginalized communities.10 Another study showed that limited English proficiency (LEP) in parents predicts lower rates of diabetes technology utilization.11 Additionally, there may be personal (ie, single-parent household, older age at diagnosis), psychological (ie, depressive symptoms, diabetes distress), and structural (ie, insurance coverage, technology costs, lack of standardized patient education) factors that affect uptake and use of diabetes technology.12–15 The aim of this study was to assess racial, ethnic, and socioeconomic disparities in technology utilization and determine provider-, patient-, and caregiver-identified barriers to diabetes device adoption at a large urban pediatric hospital. We hypothesize that disparities in diabetes technology utilization will be present across race/ethnicity and primary language of caregiver but will not be related to health insurance coverage. Additionally, we expect that patients and caregivers from historically marginalized populations will have higher perceived burdens and lower expectations and negative attitudes about diabetes technology. Finally, we anticipate an alignment between patient/caregiver and clinician-perceived barriers to technology utilization.Materials and methods Retrospective electronic medical record-based cohort study Electronic health records (EPIC Systems Corporation, Verona, Wisconsin) of all children attending a diabetes clinic appointment at University of Texas Southwestern Medical Center (UTSW)/Children’s Health (CH) from August 2020 to August 2021 were reviewed. Demographic data extracted from the demographics section of the electronic health record included sex, race/ethnicity (recoded as a single variable with values of NHW, NHB, Hispanic, Other), age, and type of health insurance (recoded as commercial (provided by an employer) or non-commercial (ie, government programs, charitable programs, uninsured, self-pay)) at the most recent clinic appointment. Insurance status was considered a surrogate for socioeconomic status. Data were collected using SAP analytics (Walldorf, Germany) to analyze data from Clarity (a Microsoft SQL server database), which were extracted from the EPIC clinical database. Additional clinical data were entered manually in EPIC by providers at each clinic visit. HbA1c was measured as a point-of-care test at the clinic visit. Insulin pump or CGM use was defined as having current pump or CGM related data at time of most recent clinic appointment. The Institutional Review Board of the UTSW approved this study as a retrospective chart review (IRB #STU-2021–1171).Pediatric patients living with diabetes and their caregivers: study population We conducted a cross-sectional, descriptive survey of patients and their caregivers at UTSW/CH to determine their perceived benefits and barriers to technology utilization. Participants were recruited at their standard of care clinical visit with their diabetologist in Dallas, TX (single site) from 2022 to 2023. Eligibility criteria included English and Spanish speaking patients ages 8–17 years with either T1D or T2D receiving insulin therapy (basal or basal/bolus) and caregivers of patients with T1D or T2D ages 1 to 17 years. Patients with secondary diabetes, cystic fibrosis-related diabetes, or monogenic diabetes, and patients not on insulin therapy were excluded. NHB and Hispanic patients and their caregivers were oversampled (goal 35% NHB, 35% Hispanic) to gain perspective of historically marginalized patients and caregivers at our center and equally represent the three largest race/ethnic groups in Dallas, TX. The study was approved by the Institutional Review Board of the UTSW (IRB #STU-2021–1171), and verbal consent was obtained from all the patients’ parents/legal guardians, and assent was obtained from pediatric patients when appropriate.Pediatric patients living with diabetes and their caregivers: survey design and data collection The patient and caregiver surveys (offered in English and Spanish language) included published and validated measures assessing thoughts about diabetes technology as well as questions constructed by the research team evaluating how participants learned about various forms of diabetes technology. Questions evaluating when participants and caregivers learned about diabetes technology (at the time of diagnosis; at first clinic visit; at subsequent clinic visits; at diabetes education class; at diabetes camp; I have not yet learned about this) were limited to pediatric participants diagnosed with diabetes after 2018 when CGMs were accessible and reliable, insulin pumps were proven to be efficacious for improving glycemic variability and available with consistent insurance coverage and the first AID was approved for pediatric use.4 Perceptions of CGM use were determined using Benefit of CGM (BenCGM) and Burdens of CGM (BurCGM) measures, each consisting of 8 Likert-scale items (Cronbach’s alpha 0.89 and 0.87, respectively).16 Using a scale 1 to 5 (1 is strongly disagree, and 5 is strongly agree), higher scores on BenCGM indicate higher positive attitudes, and higher scores on BurCGM indicate higher perceived burden of CGM use. Patient and caregiver thoughts and expectations for AID were assessed using the Insulin delivery Systems: Perceptions, Ideas, Reflections, and Expectations (INSPIRE) measure, which is a 5-level Likert scale (0 strongly disagree to 4 strongly agree and then converted to scaled scores (0 to 100)) (Cronbach alpha 0.95 for patient version, 0.97 parent version).17 The Diabetes-Specific Attitudes about Technology use (DSAT), a 5-level Likert-scale measure (1 strongly disagree to 5 strongly agree) consisting of 5 items, was incorporated to assess general attitudes toward diabetes technology, irrespective of prior diabetes technology use.12 Items are summed to generate a total score (0–25), and higher scores indicate more positive attitudes about diabetes technology (Cronbach’s alpha 0.91).12 All surveys were administered through Research Electronic Data Capture at UTSW.18 19Clinic provider prescribing practices and beliefs We administered questionnaires to pediatric endocrinology physician attendings and fellows, nurse practitioners, and diabetes educators (included certified diabetes care and education specialists and nurses) to determine basic demographics, provider-specific insulin pump prescribing practices such as eligibility criteria, objective, and subjective factors. 20 We used a modified version of a previously published self-reported survey administered to Pediatric Endocrine Society members and included questions regarding AID to remain consistent with recent advances in technology. Providers were also asked to report clinician-perceived barriers to CGM use by patients using a previously published list of barriers (including selecting multiple options and free text) created from literature review and market research.21 Unfinished questionnaires were included in the final analysis. For statistical analysis, only survey responses from attending physicians and diabetes educators were compared, as these two groups had the largest sample sizes (16 and 13, respectively).Statistical analysis Logistic regression analyses were performed with treatment modalities (insulin pump use, CGM use) as separate dependent variables and race/ethnicity, recent HbA1c, and insurance status as independent variables. With NHW children as the reference group, ORs were calculated for NHB versus NHW and for Hispanic versus NHW. Normally distributed data were described as mean±SD and non-normally distributed data as median (IQR). χ2 or Fisher-Freeman-Halton exact tests, as appropriate, were used to compare the responses at time of diagnosis versus not at time of diagnosis for knowledge of the use of insulin pump, AID, or CGM for the three groups (Hispanic, NHW, and NHB). If the χ2 test was significant, a pairwise, post hoc, comparison of the groups was examined to determine which groups were significantly different using Tukey-type multiple comparison testing among proportions. 22Survey responses were analyzed using analysis of variance and non-parametric Kruskal-Wallis. Data analysis was performed using SAS 9.4 software (Cary, NC: SAS Institute, Inc 2024).Results Retrospective cohort study Insulin pump and CGM usage rates were assessed in pediatric patients with T1D (n=1742) ( table 1). NHW patients with commercial insurance had the highest percentage of insulin pump and CGM use, whereas NHB and Hispanic patients with non-commercial insurance had the lowest percentages of insulin pump use (table 1). NHB and Hispanic children were less likely to use insulin pumps (OR: 0.4 (95% CI 0.27 to 0.59) and OR: 0.35 (95% CI 0.25 to 0.50), respectively) and CGM (OR: 0.74 (95% CI 0.54 to 1.0) and OR: 0.54 (95% CI 0.41 to 0.71), respectively) independent of insurance status. This disparity persisted after adjusting for insurance and HbA1c (figure 1).Table 1Clinical and demographic data for patients with type 1 diabetes (T1D) seen at UTSW between August 2020 and August 2021TotalNHWNHBHispanicAsiann=1742n=988n=338n=361n=55T1DMale sex914 (53%)514 (52%)178 (52.7%)198 (54.8%)24 (43.6%)Language English1585 (91%)982 (99.4%)332 (98.2%)218 (60.4%)53 (96.4%) Spanish143 (8%)1 (0.1%)0 (0%)142 (39.3%)0 (0%) Other14 (1%)5 (0.8%)6 (1.8%)1 (0.3%)2 (3.6%)Commercial Insurance1039 (60%)740 (74.1%)134 (39.6%)120 (33.2%)45 (81.8%) CGM729 (70%)556 (75.1%)72 (53.7%)67 (55.8%)34 (75.6%) Pump389 (37%)309 (41.8%)36 (26.9%)27 (22.5%)17 (37.8%)Non-commercial703 (40%)248 (26.1%)204 (60.4%)241 (66.8%)10 (18.2%) CGM352 (50%)146 (58.9%)80 (39.2%)120 (49.8%)6 (60%) Pump105 (14.9%)62 (25%)21 (10.3%)19 (7.9%)3 (30%)CGM, continuous glucose monitors; NHB, non-Hispanic Black; NHW, non-Hispanic White; UTSW, University of Texas Southwestern Medical Center.Figure 1(A) Logistic regression models of pump use based on race/ethnicity, insurance status, and HbA1c. (B) Logistic regression models of CGM use based on race/ethnicity, insurance status, and HbA1c. CGM, continuous glucose monitors; HbA1c, hemoglobin A1c.Survey recruitment Of the 299 participants approached to participate in the survey study, 18 declined to participate, and six participants only completed the demographics portion of the survey. The remaining 275 participants who completed the surveys included 130 patients and 145 caregivers ( table 2). Twelve partial responses were included. Data for patients with T2DM (n=16) and caregivers (n=14) were not analyzed due to limited sample size. There were five patients with T1D who chose to complete the survey in Spanish; however, one survey was incomplete. Due to incomplete data and small sample size, these participants were excluded. There were 14 caregivers who completed the survey in Spanish (LEP), and one survey was incomplete and excluded. These 13 caregivers with LEP were analyzed separately. The final total of English-speaking patients with T1D who completed the survey was 109, and the total number of caregivers was 117 (table 2).Table 2Demographics and technology utilization of patients and parents who completed surveysCharacteristicsType 1 diabetesYouth (n=109)Caregivers (n=117)Male55 (50%)N/AAge, years: mean (SD)13 (2.5)N/ADuration of Diabetes Mellitus (DM): mean (SD)6.4 (3.8)N/APatient race/ethnicity  Non-Hispanic White30 (28%)40 (34%) Non-Hispanic Black34 (31%)41 (35%) Hispanic41 (37%)31 (26%) Other2 (2%)2 (2%) Multiracial2 (2%)3 (3%)Pump utilization  Non-Hispanic White14 (47%)18 (45%) Used to use1 (3%)1 (3%) Non-Hispanic Black10 (29%)15 (37%) Used to use1 (3%)2 (5%) Hispanic16 (39%)12 (39%)CGM  Non-Hispanic White24 (80%)32 (80%) Used to use3 (10%)3 (8%) Non-Hispanic Black25 (74%)32 (80%) Used to use4 (12%)4 (10%) Hispanic34 (83%)27 (87%) Used to use1 (2%)AID Non-Hispanic White7 (23%)10 (25%) Non-Hispanic Black4 (12%)5 (13%) Used to use1 (3%) Hispanic5 (13%)5 (17%) Used to use2 (5%)1 (3%)CGM, continuous glucose monitor.Survey results from pediatric patients with T1D While most pediatric patients (91 out of 109 surveyed) used CGM, the usage of insulin pumps differed by race and ethnicity ( table 2). Patients diagnosed with diabetes in 2018 onward (n=50) first heard about diabetes technology at variable times. More NHW patients stated they first learned about insulin pump technology at time of diagnosis (71%) compared with the NHB (22%) and Hispanic (38%) patients (p<0.025 between NHW and NHB). Similarly, NHW patients (71%) remember first learning about CGM at time of diagnosis, compared with NHB patients (28%) and Hispanic patients (47%) (p<0.05 between NHW and NHB). Most NHW patients (71%) first heard about AID at time of diagnosis, compared with NHB and Hispanic patients (23% and 28% respectively) (p<0.025 between NHW and NHB; p<0.05 between NHW and H). There were no differences in perceived benefits of CGM among race/ethnicity groups (table 3A). However, Hispanic patients perceived more barriers compared with NHW and NHB patients including cost (“CGM is too expensive to wear”), worry (“CGM causes too much worry”), and difficulty in interpreting CGM data (“it is too hard to understand CGM information”) (table 3B). There were no differences based on race/ethnicity in expectations of AID technology using the INSPIRE measure or attitudes toward technology using the DSAT measure (data not shown).Table 3Benefits (A) of CGM use and Burdens (B) of CGM use for patients with T1D3A. Median BenCGM scores of T1D patients by race/ethnicityNHWNHBHispanicn=30n=32n=41I think…Median (IQR)…CGM makes taking care of diabetes easier4.5 (4–5)5 (4–5)5 (4–5)…CGM helps take care of low blood sugars4 (4–5)5 (4–5)4 (3–5)…CGM alarms are helpful4 (4–5)5 (4–5)4 (4–5)…CGM makes me/would make me feel more secure5 (4–5)4 (4–5)4 (3–5)…CGM lets me/would let me do less finger sticks5 (4–5)5 (4–5)4 (4–5)…my family wants me to wear a CGM5 (4–5)5 (4–5)4 (3–5)…I take/would take better care of my diabetes with a CGM4 (4–5)5 (4–5)4 (4–5)…CGM helps with managing blood sugar during exercise5 (3–5)5 (4–5)4 (3–5)3B. BurCGM scores of T1D patients by race/ethnicityNHWNHBHispanicn=30n=32n=41I think…Median (IQR)…CGM sensor readings cannot be trusted2 (2–2.75)2 (1–3)2 (2–3)…CGM takes too much time to use2 (1–2)1 (1–2)2 (1–2)…CGM is not helpful1 (1–2)1 (1–2)2 (1–2)…CGM is painful to wear2 (1–3)1 (1–2)2 (1–2)…the CGM is too expensive to wear regularly*2 (1–3)2 (1–3)3 (2–3)…CGM causes too much worry about blood sugars†2 (1–2)1 (1–2)2 (2–3)…I feel/would feel embarrassed about wearing CGM1 (1–2)1 (1–2)2 (1–2)…it is too hard to understand CGM information*1 (1–2)1 (1–2)2 (1–2)Median score scale 1 to 5: 1 is strongly disagree, and 5 is strongly agree; the higher the median score, the greater is the benefit.Median score scale 1 to 5: 1 is strongly disagree, and 5 is strongly agree; the higher the median score, the greater the burden.*p value <0.05.†p value <0.01.BenCGM, benefits of continuous glucose monitoring; BurCGM, burdens of continuous glucose monitoring; NHB, non-Hispanic Black; NHW, non-Hispanic White.Survey results from caregivers of pediatric patients with T1D Comparable to the patient responses, caregivers (n=58) to patients diagnosed after the year 2018 onward first learned about technology at different times. More NHW caregivers (67%) first learned about insulin pumps at diagnosis compared with NHB caregivers (24%) and Hispanic caregivers (47%) (p <0.025 between NHW and NHB). Similar trends were noted for CGM and AID, but they did not reach statistical significance. There were no significant differences in perceived benefits or burdens to CGM use by caregivers based on race/ethnicity. Caregivers in general had a positive attitude toward technology based on the DSAT measure.Survey results from Hispanic English-speaking caregivers compared with Hispanic caregivers with LEP We compared survey responses from Hispanic caregivers who were English-speaking (n=30) to those with LEP (n=13) to better understand the role language plays in disparities related to diabetes technology utilization.Due to the smaller sample size, as well as different primary language, these 13 caregivers were not included in the previous analyses. Participants were provided a choice of whether they preferred to complete the surveys in English or Spanish. Among caregivers with LEP, there were four caregivers who had children utilizing insulin pumps and seven caregivers who had children using CGM. Due to small sample size, we were unable to make meaningful comparisons of when caregivers with LEP learned about diabetes technology. Hispanic English-speaking parents had higher perceived benefits of CGM compared with LEP parents, particularly for simplifying diabetes care (“CGM makes taking care of child’s diabetes easier”), security (“CGM makes me feel more secure”), and exercise (“CGM helps with management of blood sugar during exercise”) (table 4A). Conversely, Hispanic LEP parents had higher perceived burdens to CGM compared with Hispanic English-speaking parents, identifying lack of trust as barriers (“CGM sensor reading cannot be trusted”), difficult to understand, and time (“CGM takes too much time to use”) (table 4B). There were no differences in expectations of AID technology of caregivers based on English proficiency.Table 4Benefits (A) of CGM use and Burdens (B) of CGM use for Hispanic Caregivers by Language4A. Median BenCGM scores for Hispanic caregivers by languageEnglish speakingLEPn=30n=11I think…Median (IQR)…CGM makes taking care of diabetes easier*5 (5-5)4 (3.5–5)…CGM helps take care of low blood sugars5 (4-5)4 (3.5–4.5)…CGM alarms are helpful5 (4.25–5)4 (4-5)…CGM makes me/would make me feel more secure*5 (4.25–5)4 (4-5)…CGM lets me/would let me do less finger sticks5 (5-5)5 (4.5–5)…my family wants me to wear a CGM5 (5-5)5 (4-5)…I take/would take better care of my diabetes with a CGM5 (4.25–5)4 (4-5)…CGM helps with managing blood sugar during exercise*5 (4-5)4 (3.5–4.5)4B. Median BurCGM scores in Hispanic caregivers by languageEnglish speakingLEPn=30n=11I think…Median (IQR)…CGM sensor readings cannot be trusted*2 (1–3)3 (2–4)…CGM takes too much time to use†1 (1–2)3 (2–3)…CGM is not helpful1 (1–1)1 (1–2)…CGM is painful to wear2 (1–2)2 (1.5–3)…the CGM is too expensive to wear regularly3 (2–3)3 (1.5–3.5)…CGM causes too much worry about blood sugars1 (1–2)2 (1.5–2.5)…I feel/would feel embarrassed about wearing CGM1 (1–2)2 (1.5–3)…it is too hard to understand CGM information†1 (1–2)2 (2–2.5)Median score scale 1 to 5: 1 is strongly disagree, and 5 is strongly agree; the higher the median score, the greater is the benefit.Median score scale 1 to 5: 1 is strongly disagree, and 5 is strongly agree; the higher the median score, the greater the burden.*p value <0.05.†p value <0.01.BenCGM, benefits of continuous glucose monitoring; BurCGM, burdens of continuous glucose monitoring; LEP, limited English proficiency.Survey results from clinical diabetes providers Forty providers (16 attending physicians, five fellow physicians, five nurse practitioners, 13 diabetes educators completed the survey on technology prescribing practices. Provider demographics revealed that although race and ethnicity varied greatly, the largest professional category was attending physicians (40%) and was female (82%), largest race category was White (42%), and largest ethnicity category non-Hispanic (80%). Roughly one-third of providers were more than 20 years post-graduation from classroom-based medical education. While most providers (92.5%) stated that they routinely discuss insulin pump therapy with patients with T1D at any follow-up visits, a little over half (57.5%) routinely discuss insulin pump therapy at the time of diagnosis. Nearly half of providers required patients to meet a minimum number of daily blood glucose checks prior to initiation of insulin pump therapy. Of those, most (88%) required four to five blood glucose checks per day. Most providers did not require a minimum duration of time after diagnosis to initiate an insulin pump (64.1%) or a minimum number of visits per year to receive insulin pump therapy (79.5%) or a HbA1c criteria to initiate an insulin pump prescription (89.7%). Most attending physicians and diabetes educators rated diabetes-related factors such as recurrent hypoglycemia, fluctuations in glucose levels, and insulin sensitivity as important or very important considerations for prescribing an insulin pump. Compared with attending physicians, a higher percentage of diabetes educators placed importance on a patient’s elevated HbA1c level ( table 5). Significant differences were detected between the importance that diabetes educators and attending physicians placed on subjective patient/family factors (table 5). Compared with attending physicians, the majority of the diabetes educators assigned more importance to subjective factors such as psychosocial stability of family, health literacy, prior adherence to diabetes self-care management tasks, and parental education (table 5). Clinic providers were asked to select potential patient barriers to CGM use (option to select multiple answers and free text). The highest-rated barrier was patients not wanting a diabetes device on their body (84.6%), followed by insurance coverage (76.9%), cost of device (76.9%), and cost of supplies (76.9%).Table 5Clinician-reported factors in decision to initiate insulin pump by provider type: median scoresAttending physicianDiabetes educatorsn=16n=13Diabetes-related factorsMedian (IQR)Current regimen compromises lifestyle4.0 (4–5)4.0 (3–4)Early complications such as neuropathy or nephropathy3.5 (2.75–4)3.0 (3–4)Elevated HbA1c*3.0 (3–3.25)4.0 (3–4)Extreme insulin sensitivity4.0 (3–5)4.0 (4–5)Frequent exercise/competitive athletes4.0 (4–4.25)4.0 (3–4)Recurrent, severe, or unpredictable hypoglycemia4.0 (4–5)4.0 (4–5)Wide fluctuations in blood glucose levels4.0 (4–5)4.0 (4–5)Patient and family-related factorsAbility to demonstrate carbohydrate counting4.0 (3–4)4.0 (4–5)Financial stability (ability to pay medical costs)4.0 (2.75–4)3.0 (3–4)Health literacy*3.5 (3–4)4.0 (4–5)Insurance coverage of insulin pump5.0 (4–5)4.0 (3–5)Number of parents/guardians in household2.0 (1–3)2.0 (2–3)Parental education†3.0 (2–3.25)4.0 (4–5)Parent request†4.5 (4–5)4.0 (3–4)Patient and families’ ability to communicate with medical team4.0 (4–4)4.0 (4–4)Patient and family motivation5.0 (4–5)5.0 (4–5)Patient and/or family psychological wellness4.0 (3–5)4.0 (4–5)Patient request5.0 (4–5)4.0 (4–5)Physical ability to use pump4.0 (4–5)4.0 (4–5)Prior adherence with diabetes self-care measures*3.0 (3–4)4.0 (4–5)Psychosocial stability of the family†3.0 (3–4)4.0 (4–5)Realistic expectations of insulin pump use4.0 (4–4.25)4.0 (4–5)Median score scale 1 to 5: the higher the median score, the greater is the importance.*p value <0.05†p value <0.01CDCES, certified diabetes care and education specialists; HbA1c, hemoglobin A1c.Discussion Our study of pediatric patients with T1DM found that Black and Hispanic patients were less likely to use diabetes technology such as insulin pumps and CGM even when adjusted for insurance status. 6 7 These findings are consistent with other studies involving young patients and adults.6 7 23 Variability in timing when first learning about different forms of diabetes technology from their providers was consistent across patient and caregiver responses. Therefore, in addition to relying on objective criteria, as observed in the current study findings and previous reports,20 a clinical implication could be to implement a systematic and standardized approach to diabetes technology-related discussions and uptake. We learned that caregivers of patients who were not using CGM reported higher perceived burdens than CGM users, which may dissuade families from pursuing CGM for their children. One theory worth investigating in subsequent studies may be negative caregiver experiences with CGM in their own diabetes management.NHB and Hispanic patients and caregivers had positive attitudes toward diabetes technology and positive expectations regarding AID use. This suggests that the trend of less technology utilization in historically marginalized populations is not due to disinterest in technology. Hispanic caregivers with LEP had higher perceived burdens to CGM compared with Hispanic English-speaking caregivers. Similarly, Hispanic patients had higher perceived burden to CGM use compared with NHW and NHB patients. Given that LEP caregivers of youth with complex medical conditions face many challenges in communicating with healthcare providers, it is likely that communication barriers play a significant role in accounting for this disparity.24 It is possible that clinicians have a lower inclination to offer insulin pumps to LEP families due to anticipated challenges in communication or a biased perception that they might not be good candidates for pump use. One qualitative study found that Spanish-speaking parents of children and adolescents with T1D cited reluctance to employ insulin pumps because of concerns that their child/adolescent would experience discomfort during sports or have less freedom in their dietary choices. The authors noted how these concerns were directly contradictory to the stated benefits of insulin pumps, suggesting possible miscommunication with medical providers or a general misunderstanding of the pump technology.25 This suggests an area of improvement to provide more culturally and linguistically appropriate resources for LEP families.Limitations of this study include small sample sizes for caregivers with LEP. There may have been an element of recall bias, as participants were asked to think back to the time when they or their child was diagnosed with diabetes. Additionally, we are unable to measure the patient/caregiver knowledge of diabetes technology and were unable to assess whether discussions regarding diabetes technology were patient/family-initiated or provider-initiated. Race and ethnicity data were obtained from our electronic health record which may be discordant from how patients self-report their race/ethnicity. It is important to note that the goal of this study was to gain the perspectives of historically marginalized patients and the race/ethnicity demographics of the participants in this study are not representative of our clinic patient’s demographics.Decisions to use diabetes technologies are multifactorial and influenced by provider prescribing practices, as well as patient and/or caregiver perceptions toward technologies. This study shows that clinic providers may place more importance on subjective factors rather than objective criteria, which is consistent with prior studies.20 Additionally, though clinic providers perceive many barriers to technology use by patients, the patient/caregiver responses suggest low burdens to certain forms of technology and overall positive attitudes.The findings of our study provide an opportunity to increase awareness of potential bias and encourage providers to be more mindful of introducing diabetes technology to all patients using a standardized method. Future interventions could incorporate use of a shared decision-making model to allow for a systematic approach to discuss diabetes technology with patients and families.