Underserved young adults (YA) with type 1 diabetes (T1D) experience the worst outcomes across the life span. We developed and integrated the Supporting Emerging Adults with Diabetes (SEAD) program into routine endocrinology care to address unmet social and medical challenges.
This study was designed as a longitudinal cohort study, with prospective data collection over 4 years on YA in SEAD compared with usual endocrine care. We used propensity-weighted analysis to account for differences in baseline characteristics, and multivariate regression and Cox proportional hazard models to evaluate change in outcomes over time. Primary outcomes included incidence of hospitalizations, diabetes technology uptake, and annual change in HbA1c levels.
We included 497 YA with T1D in SEAD (n = 332) and usual endocrine care (n = 165); mean age 25 years, 27% non-Hispanic Black, 46% Hispanic, 49% public insurance, mean HbA1c 9.2%. Comparing YA in SEAD versus usual care, 1) incidence of hospitalizations was reduced by 64% for baseline HbA1c >9% (HR 0.36 [0.13, 0.98]) and 74% for publicly insured (HR 0.26 [0.07, 0.90]); 2) automated insulin delivery uptake was 1.5-times higher (HR 1.51 [0.83, 2.77]); and 3) HbA1c improvement was greater (SEAD, −0.37% per year [−0.59, −0.15]; usual care, −0.26% per year [−0.58, 0.05]).
SEAD meaningfully improves clinical outcomes in underserved YA with T1D, especially for publicly insured and high baseline HbA1c levels. Early intervention for at-risk YA with T1D as they enter adult care could reduce inequity in short and long-term outcomes.
Introduction
Young adulthood is a unique phase in life marked by developmentally normative transitions across familial, social, educational, and geographic life domains, while transitioning from pediatric to adult care (1). During this transition, when loss to follow-up is greatest, young adults (YA) with type 1 diabetes (T1D) have 2.5-times greater odds of poorer glycemic control compared with youth who remain in pediatric care (2,3), with only ∼17% achieving American Diabetes Association-recommended glycemic targets (4,5). YA with T1D have higher rates of hospitalizations during this period, with hospitalizations for diabetes-related emergencies in youth <20 years of age with T1D increasing ∼40% from 2006 to 2016 (6,7). In addition, mortality is 1.5-times greater for YA with T1D than for YA without diabetes (8). YA with T1D from underrepresented racial and ethnic groups have even worse outcomes during the transition period due to added negative effects of structural racism in health care and unmet social needs (9,10). As a result, non-Hispanic Black and Hispanic YA have 1–2% higher HbA1c levels, higher incidence of diabetic ketoacidosis (DKA), more frequent severe hypoglycemic events, and lower diabetes technology use compared with non-Hispanic White YA (9–11). There is a critical need to intervene in the YA period to prevent long-term complications and premature mortality in adulthood, and furthermore to provide tailored support to underserved populations in this vulnerable life period.
Health care transition from pediatric to adult care holds several challenges that predispose YA with T1D to deterioration in health outcomes. Lack of YA-centered care delivery, limited support in adult receivership clinics, prolonged transfer time, and competing social and financial constraints pose challenges to diabetes self-management and consistent outpatient care (12–17). Thus, there is impetus to better support the health care transition from a systems perspective to improve downstream health for YA with T1D. Multiple professional societies have published guidelines and strategies to facilitate health care transition (18), but integration into real-world care has been slow. This may be due to fundamental differences in health care delivery in pediatric versus adult care paradigms and an overall lack of resources in adult care to address complex and intersectional YA needs. As a result, there remains a dearth of adult care-based transition receivership programs, despite YA spending most of their lives in adult care.
In response, we developed the Supporting Emerging Adults with Diabetes (SEAD) program, an adult-based health care transition receivership program for YA with T1D between the ages of 18 and 35 years old (19). Preliminary evaluation of the program showed promising improvements in glycemic control and care engagement; however, disparities existed for non-Hispanic Black and Hispanic YA. To address inequity, we worked with underserved YA with T1D, their families, and health care providers to codesign an enhanced version of SEAD (10,20,21). Components of the updated program included new clinical pathways that focused on 1) equal access for all YA with T1D ages 18–35 years old; 2) health-related social needs screening and management; 3) provider training in patient-centered equitable care focused on shared decision making and cultural humility; and 4) enhanced access to diabetes technology. Although original program development involved grant funding, SEAD was integrated into routine endocrinology care at the Fleischer Diabetes Institute at Montefiore Medical Center in 2018, and is the focus of the current analysis.
To evaluate the SEAD program’s performance in routine endocrinology care, we prospectively collected and compared 4-year hospitalization, diabetes technology uptake, and glycemic outcomes of YAs with T1D in SEAD versus usual endocrine care. We chose these outcomes because they are important predictors of short- and long-term health effects of diabetes, have cost implications for health systems, and have been shown to be inequitable among racial and ethnic minority groups. We hypothesized that YA receiving care in SEAD would have greater improvement in HbA1c levels, higher incidence of diabetes technology uptake, and reduced incidence of hospitalizations compared with those in usual endocrine care.
Research Design and Methods
Study Design
To evaluate the SEAD program, we designed a comparative longitudinal cohort study starting in 2019 until 2022 to prospectively study incidence of hospitalizations, glycemic control, and diabetes technology uptake comparing YA in SEAD with usual endocrine care. We decided on the longitudinal real-world cohort design to capture underserved YA who would not have otherwise been possible to recruit into research, given historic underparticipation of YA and underserved communities in randomized controlled trials (22,23). Furthermore, we chose our longitudinal cohort comparison with usual endocrine care because T1D usually requires specialized endocrinology care, thus making comparison with primary care an inappropriate comparison.
Setting
Patients who were referred to the SEAD program included YA with T1D transitioning from pediatric endocrinology, primary care, and obstetrics-gynecology at Montefiore Medical Center, patients with newly diagnosed or established T1D discharged from the hospital, and outside referrals from the New York metropolitan area. The SEAD model encourages care that underscores enhanced YA communication strategies to allow for flexibility and shared decision making in care plans, a focus on continuing T1D diabetes education, diabetes technology support, psychosocial support, pediatric provider partnership, and orientation to adult care.
Patients in the SEAD program were seen by SEAD providers during dedicated clinic sessions every week; however, there were no full-time SEAD providers. SEAD providers included two physicians, one nurse practitioner, one dietitian, and one psychologist. Clinical visits were cohorted into 1 day so that SEAD providers could co-locate in the clinic; however, other days and times were offered for flexible access. In addition, clinic sessions for SEAD patients had “frozen” slots to accommodate acute referrals and closer follow-up for SEAD patients if required.
SEAD used existing resources of the Fleischer Diabetes Institute at Montefiore Medical Center, which included partial time of a clinical diabetes psychologist and use of a licensed practice nurse who was shared with primary care for social needs screening and navigation. Specific to SEAD, financial investment by clinical leadership included allotting five dedicated SEAD clinical sessions per week across three providers at potential lost revenue for missed follow-up visits. Nonfinancial investment by clinical leadership included endorsement and marketing of SEAD at the institution level and flexibility in clinic scheduling to accommodate unique YA needs.
Inclusion Criteria
Participants were included in analyses if they were 18–35 years old, had a clinical diagnosis of T1D defined by ICD-10 codes, and had at least one established in-person or telemedicine visit with an endocrine physician or nurse practitioner from January 2019–December 2022 at Montefiore (SEAD or non-SEAD). Enrollment date in the cohort was defined as the date of the first diabetes visit with a SEAD provider or non-SEAD endocrine provider.
Exclusion Criteria
Participants were excluded from analysis if they had not been seen in the endocrine practice for >1 year before the start date of the study, had a diagnosis by ICD-10 code of steroid-induced diabetes, maturity-onset diabetes of the young, or gestational diabetes mellitus, or a ratio of ICD codes of <50% for T1D to non-T1D. We also excluded data from 48 crossover referrals from usual endocrine care to SEAD to avoid selection bias.
Data Collection
Baseline age, sex, BMI, race and ethnicity, preferred language, and primary insurance status were extracted from the electronic health record (EHR). ICD-10 codes were extracted for baseline psychiatric disease given the high incidence in YA with T1D (24,25), and as a potential confounder of hospitalizations. HbA1c levels were extracted as laboratory or point-of-care measurements, as available. Diabetes technology prescription data were collected using visit and nonvisit prescription data in the EHR, and used to identify participants who had an active continuous glucose monitor (CGM) or insulin pump prescription in their medical record from the prior month or as a new prescription with no preceding prescription in the past 3 years. CGM prescriptions included Dexcom G4/G5/G6, FreeStyle Libre 1/2/3, and Guardian 3. Pump prescriptions included Medtronic 670G/770G, Tandem t:slim Basal IQ or Control IQ, Omnipod Eros, Omnipod DASH, and Omnipod 5. Automated insulin delivery (AID) system was defined as Dexcom G6-Tandem t:slim Control IQ, Dexcom G6-Omnipod 5, and Guardian-Medtronic 670G/770G systems. Hospitalization data were extracted directly from the EHR and included data from Montefiore Health System as well as from New York area hospitals participating in Epic Care Everywhere.
Primary Outcomes
Primary outcomes were defined a priori and measured longitudinally over the 4-year study time frame, including incidence of annual hospitalizations, diabetes technology uptake (CGM alone vs. AID), and mean annual change in HbA1c levels.
Statistical Analysis
Baseline characteristics were tabulated as mean ± SD for continuous variables, and frequency (n) and percentage (%) for discrete variables. We used propensity-weighted analysis to account for differences in baseline characteristics between the SEAD and usual endocrine care groups. A logistic model was constructed predicting group membership conditional on date of study accrual, age at study accrual, sex, race and ethnicity, history of hyperglycemia, ketoacidosis, nephropathy, retinopathy, neuropathy, gastroparesis, cardiovascular disease, thyroid disease, obesity, kidney transplant, psychiatric hospitalization, and insurance category. In adjusted analyses contrasting the two groups, each observation was weighted by the reciprocal of its probability of membership in its group. In analyses relying on Cox proportional hazard models, loss to follow-up was handled by censoring that patient's observations as of the date of their last activity in our data set. Analyses of the rate of HbA1c change over time include all HbA1c measurements available between cohort enrollment and the end of follow-up. The resulting estimated rates of change are not applicable beyond the end of follow-up. To compare hospitalization outcomes between groups, we created Kaplan-Meier plots and used a Cox proportional hazards model to compare incidence of hospitalization, adjusting for age, sex, race and ethnicity, insurance, and baseline rate of DKA. To compare diabetes technology outcomes between groups, the time interval from joining SEAD or usual care to the first orders, if any, for CGM alone and AID were extracted for use in time-to-event analysis. Time, in years, from entry into the study group to diabetes technology uptake was contrasted across study groups using the log-rank test, and Kaplan-Meier plots were obtained. To compare change in HbA1c between groups, the trajectory of A1C was modeled as a linear time trend in a fixed-effects regression, with time between cohort join date and HbA1c measurement (latency), and its interaction with study group (SEAD vs. usual care) as predictors.
Results
The analysis included 497 YA with T1D who received SEAD (n = 332) or usual endocrine care (n = 165) care from January 2019–December 2022. Table 1 presents the demographic and clinical characteristics of the participants, comparing YA in SEAD with usual endocrine care. Given that the SEAD group had a higher proportion from the White racial and ethnic group and commercially insured, we used propensity-weighted analysis to account for differences at baseline between the two groups. Despite potential differences in demographic risk factors for severity, however, baseline HbA1c levels were elevated and similar between groups, suggesting equal baseline risk of poor outcomes.
Baseline characteristics of participants comparing SEAD program with usual endocrine care
. | Total (N = 497) . | SEAD (n = 332) . | Usual care (n = 165) . |
---|---|---|---|
Demographics | |||
Age (years) | 25.3 ± 4.5 | 23.9 ± 4.0 | 28.0 ± 4.0 |
Female sex | 48 (240) | 48.5 (161) | 47.9 (79) |
Race and ethnicity | |||
Non-Hispanic Black | 27 (135) | 23 (76) | 36 (59) |
Hispanic | 46 (230) | 48 (158) | 44 (72) |
Non-Hispanic White | 12 (59) | 15 (48) | 7 (11) |
Other | 15 (72) | 15 (49) | 14 (23) |
Primary language | |||
English | 92 (458) | 93 (308) | 92 (150) |
Spanish | 6 (31) | 6 (19) | 7 (12) |
Other | 0.2 (1) | 0 | 0.6 (1) |
Primary insurance | |||
Private | 51 (252) | 56 (186) | 40 (66) |
Public | 49 (243) | 43 (144) | 60 (99) |
Clinical characteristics | |||
HbA1c (%) | 9.2 ± 2.3 | 9.2 ± 2.4 | 9.1 ± 2.2 |
Baseline DKA hospitalization rate | 9 (46) | 7 (24) | 13 (22) |
. | Total (N = 497) . | SEAD (n = 332) . | Usual care (n = 165) . |
---|---|---|---|
Demographics | |||
Age (years) | 25.3 ± 4.5 | 23.9 ± 4.0 | 28.0 ± 4.0 |
Female sex | 48 (240) | 48.5 (161) | 47.9 (79) |
Race and ethnicity | |||
Non-Hispanic Black | 27 (135) | 23 (76) | 36 (59) |
Hispanic | 46 (230) | 48 (158) | 44 (72) |
Non-Hispanic White | 12 (59) | 15 (48) | 7 (11) |
Other | 15 (72) | 15 (49) | 14 (23) |
Primary language | |||
English | 92 (458) | 93 (308) | 92 (150) |
Spanish | 6 (31) | 6 (19) | 7 (12) |
Other | 0.2 (1) | 0 | 0.6 (1) |
Primary insurance | |||
Private | 51 (252) | 56 (186) | 40 (66) |
Public | 49 (243) | 43 (144) | 60 (99) |
Clinical characteristics | |||
HbA1c (%) | 9.2 ± 2.3 | 9.2 ± 2.4 | 9.1 ± 2.2 |
Baseline DKA hospitalization rate | 9 (46) | 7 (24) | 13 (22) |
Data are presented as % (n) or mean ± SD.
Incidence of Hospitalizations
Overall, SEAD YA had a lower incidence of hospitalizations compared with usual endocrine care (P < 0.011) (Fig. 1). In adjusted analyses stratified by baseline HbA1c level, SEAD YA with HbA1c >9% had a statistically significant reduction in hospitalizations by 64%, compared with usual endocrine care (hazard ratio [HR] 0.36 [95% CI 0.13, 0.98]). When stratified by insurance, YA in SEAD who were publicly insured and had a baseline HbA1c >9% had even greater benefit, with 74% reduction in hospitalizations compared with usual endocrine care (HR 0.26 [0.07, 0.90]) (Table 2).
Incidence of hospitalization (left) and automated insulin delivery uptake (right) comparing SEAD program with usual endocrine care.
Incidence of hospitalization (left) and automated insulin delivery uptake (right) comparing SEAD program with usual endocrine care.
Incidence of hospitalization and diabetes technology uptake comparing SEAD program with usual endocrine care
. | Incidence of hospitalization . | Diabetes technology uptake . | |||
---|---|---|---|---|---|
. | Overall . | Baseline HbA1c ≤9% . | Baseline HbA1c >9% . | CGM alone . | AID system . |
SEAD vs. usual care | |||||
Crude | 0.75 (0.6, 0.93) | 0.91 (0.67, 1.24) | 1.73 (1.10, 2.73) | ||
Adjusteda | 0.52 (0.24, 1.13) | 0.88 (0.42, 1.84) | 0.36 (0.13, 0.98) | 0.86 (0.59, 1.25) | 1.51 (0.83, 2.77) |
Stratified by insurance type | |||||
Private | 0.72 (0.2, 2.5) | 0.54 (0.14, 2.03) | 0.96 (0.53, 1.74) | 1.47 (0.62, 3.47) | |
Public | 1.16 (0.47, 2.88) | 0.26 (0.07, 0.90) | 0.81 (0.51, 1.30) | 1.45 (0.62, 3.39) |
. | Incidence of hospitalization . | Diabetes technology uptake . | |||
---|---|---|---|---|---|
. | Overall . | Baseline HbA1c ≤9% . | Baseline HbA1c >9% . | CGM alone . | AID system . |
SEAD vs. usual care | |||||
Crude | 0.75 (0.6, 0.93) | 0.91 (0.67, 1.24) | 1.73 (1.10, 2.73) | ||
Adjusteda | 0.52 (0.24, 1.13) | 0.88 (0.42, 1.84) | 0.36 (0.13, 0.98) | 0.86 (0.59, 1.25) | 1.51 (0.83, 2.77) |
Stratified by insurance type | |||||
Private | 0.72 (0.2, 2.5) | 0.54 (0.14, 2.03) | 0.96 (0.53, 1.74) | 1.47 (0.62, 3.47) | |
Public | 1.16 (0.47, 2.88) | 0.26 (0.07, 0.90) | 0.81 (0.51, 1.30) | 1.45 (0.62, 3.39) |
Data are presented as the HR (95% CI).
aAdjusted by inverse probability of treatment weighting, propensity score calculated from all demographic and comorbidity baseline variables.
Diabetes Technology Uptake
Overall, SEAD YA had more rapid uptake of AID compared with usual endocrine care (P = 0.017) (Fig. 1), with no difference in CGM uptake alone between the groups. Adjusted analyses revealed that SEAD YA were 1.5-times more likely to adopt AID technology compared with usual endocrine care (HR 1.51 [95% CI 0.83, 2.77]) (Table 2). When stratified by insurance, SEAD YA remained more likely to adopt AID technology, regardless of having private versus public insurance.
HbA1c Levels
YA in SEAD had an improvement in mean annual HbA1c of −0.37% [95% CI −0.59, −0.15] vs. −0.26% [−0.59, 0.05] in usual endocrine care, for a difference of −0.11% between groups [−0.49, 0.28], although higher variability was seen in the usual care group (Table 3). When stratified by insurance, improvement in HbA1c was greater for SEAD YA with private insurance (−0.52% [−0.73, −0.31] versus public insurance.
HbA1c trajectories comparing SEAD program with usual care
. | SEAD . | Usual care . | Difference betweenSEAD and usual care . |
---|---|---|---|
Mean change in HbA1c per year (%) | |||
Crude | −0.49 (−0.67, −0.31) | −0.12 (−0.35, 0.11) | −0.37 (−0.67, 0.07) |
Adjusteda | −0.37 (−0.59, −0.15) | −0.26 (−0.58, 0.05) | −0.11 (−0.49, 0.28) |
Stratified by insurance type (%) | |||
Private | −0.52 (−0.73, −0.31) | −0.23 (−0.59, 0.14) | −0.29 (−0.71, −0.13) |
Public | −0.21 (−0.61, 0.18) | −0.28 (−0.73, 0.17) | 0.07 (−0.53, 0.67) |
. | SEAD . | Usual care . | Difference betweenSEAD and usual care . |
---|---|---|---|
Mean change in HbA1c per year (%) | |||
Crude | −0.49 (−0.67, −0.31) | −0.12 (−0.35, 0.11) | −0.37 (−0.67, 0.07) |
Adjusteda | −0.37 (−0.59, −0.15) | −0.26 (−0.58, 0.05) | −0.11 (−0.49, 0.28) |
Stratified by insurance type (%) | |||
Private | −0.52 (−0.73, −0.31) | −0.23 (−0.59, 0.14) | −0.29 (−0.71, −0.13) |
Public | −0.21 (−0.61, 0.18) | −0.28 (−0.73, 0.17) | 0.07 (−0.53, 0.67) |
Data are presented with the 95% CI.
aAdjusted by inverse probability of treatment weighting, propensity score calculated from all demographic and comorbidity baseline variables.
Conclusions
In this longitudinal cohort study, compared with YA with T1D in usual endocrine care, YA receiving care in SEAD had meaningful improvement in hospitalizations, uptake of automated insulin delivery systems, and HbA1c levels. For the most vulnerable YA in SEAD, with baseline HbA1c >9% and public insurance, even greater improvements occurred, including 74% reduction in hospitalizations and 1.5-times higher likelihood of adopting AID. These study results demonstrate that YA with T1D, especially those most vulnerable, derive great benefit from a real-world adult care-based receivership program.
Our study showed that SEAD resulted in 64% reduction in the incidence of hospitalizations in YA with baseline HbA1c >9% despite similar baseline HbA1c levels between groups, and an even more pronounced benefit of 74% in YA with baseline HbA1c levels >9% and public insurance. Although other studies focusing on reducing hospitalizations in YA with T1D have yielded promising results, many required additional resources that may not be achievable in routine care paradigms. An Australian study used a multipronged approach with a diabetes educator as a transition clinic coordinator and phone outreach for sick day management, demonstrating a 33% reduction in DKA admission rates and a 3.6-day reduction in length of stay for readmissions (26). Other programs have focused on using extra resources for YA-specific care, such as transition care navigators, resulting in milder DKA presentation and lower health care costs driven by a reduction in length of stay (26,27). In contrast, SEAD used existing resources while changing the approach to care, promoting diabetes technology and shared decision making with YA, such that better care engagement may have resulted in a reduction in hospitalizations. Given that diabetes-related hospitalizations have increased by 40% from 2006 to 2016 in YAs with T1D in the U.S. (6) and that hospitalizations are a major driver of health care costs, the SEAD model has implications for cost savings that are immediate, potentially able to be achieved with existing resources, and could help health systems care for their high-utilizer populations.
Our study also demonstrated that SEAD YA were significantly more likely to adopt AID technology compared with usual endocrine care, which provides great potential to reduce inequity in outcomes. Established benefits of diabetes technology have led to CGM and insulin pumps becoming standard of care for people with T1D (28). Nevertheless, the T1D Exchange has shown that YA have lower rates of diabetes technology use compared with other age-groups (5) and that youth 13–25 years were more likely to discontinue insulin pump use compared with younger or older age-groups (29). In addition, there has been widespread documentation of racial and ethnic inequity in diabetes technology use (9,10). Our pivotal qualitative studies in minoritized YA with T1D outlined barriers at each step in adopting diabetes technology, including not being universally offered technology with high HbA1c levels or public insurance, lack of shared decision making when choosing technology, challenges in insurance navigation, troubleshooting technology failures and connectivity, and interruptions in supplies (10,20,21). Increased diabetes technology use has benefits for YA in transition, with some studies showing that pump users are more likely to have better clinic attendance after transition (30), improvement in glycemia, and overall treatment satisfaction compared with those treated with multiple daily injections (31).
Given mounting evidence of technology benefits and exclusion of underserved YA in recent trends of use, the SEAD model focused on technology equity by training providers to increase information and prescription access to all technologies for YA, using shared decision making to incorporate YA preferences, providing closer oversight after technology initiation, and performing social needs management to enhance use of technology. SEAD also provided better technology equity by leveraging portal messaging with YA as opposed to using extra visits or staff for follow-up after starting technology, providing a model for other programs to integrate real-world care approaches that increase diabetes technology uptake.
Lastly, our study showed significant and durable improvement in glycemic control in YA in SEAD compared with usual endocrine care, despite similar baseline HbA1c levels in both groups. Improvement in HbA1c was slightly greater in SEAD YA with private compared with public insurance. Longitudinal studies in T1D health care transition have shown a 2.5-times greater risk of deterioration in glycemic control among youth with T1D transitioning to adult care compared with those who remain in pediatric care (32–34). Although structured transition programs have shown positive results in care engagement, only a few studies have shown improvement in HbA1c after transfer, and few have examined improvements by insurance status (26,27,35–37). A multicenter randomized controlled trial in Canada evaluating a transition program using a transition coordinator demonstrated improvements in HbA1c trends (8.3% vs. 8.8%, P = 0.057) compared with usual care in 205 YA with T1D; however, benefits were not maintained at 12 months after the intervention (38), likely because continued support of a coordinator was needed and was limited by funding. That study was also conducted in a higher-resourced cohort in Canada, therefore making glycemic benefits unclear in underresourced populations. Lastly, a study in the U.S. testing a transition program in majority publicly insured YA with T1D that incorporated tailored diabetes education, group education, and access to a young adult diabetes clinic showed improvement in glycemia at 12 months (−0.4% vs. 0.42%, P = 0.01) compared with usual care (39), but glycemia was not measured beyond the first adult visit, and whether most privately insured YA would have benefitted even more is unclear. SEAD has provided durable benefits for YA that may be replicable given its integration into routine care and its focus on care engagement using preferred provider styles and social needs management rather than additional transition-related staff.
Our study has several limitations. Owing to sample size, we were not powered to directly assess the effect of diabetes technology uptake on improvement in HbA1c and reduction in hospitalizations. Nevertheless, this is one of the largest and longest-studied underserved YA cohorts using a real-world transition intervention, adding unique value to the field. Secondly, our HbA1c trajectory analysis could have been affected by participants with varied follow-up during the study period and with insufficient HbA1c assessments. In addition, the study was designed as a longitudinal cohort study rather than as a randomized controlled trial, which can have inherent limitations. However, recruiting such a large cohort of YA in real-world care would have been challenging if not for our chosen design. Lastly, future analysis should include longitudinal analysis of psychosocial outcomes, which would have been important to measure quality-of-life benefits of the SEAD model.
Our study adds value to the field in several ways. First, we report a longer duration of follow-up of YA outcomes in adult care compared with other studies. Second, our results are bolstered by the fact that SEAD is a real-world program, signaling its higher likelihood of feasibility and sustained effect. Third, we were able to retain a diverse and very underserved population of YA which allowed us to examine long-term trends over time.
In conclusion, YA with T1D in SEAD demonstrated meaningful improvement in hospitalizations, automated insulin delivery uptake, and annual HbA1c levels over 4 years compared with routine endocrinology care. For the most vulnerable YA, with baseline HbA1c >9% and public insurance, SEAD maintained equal access to technology uptake while achieving greater reduction in hospitalizations. Our results suggest that a real-world receivership program dedicated to YA with T1D, tailored using an equity lens, has meaningful and durable benefits for T1D outcomes. Moreover, given that SEAD is a YA-centered model of care, it could potentially be tailored to YA with type 2 diabetes, which is a growing and concerning demographic in diabetes. As the prevalence of youth with diabetes from racial-ethnic minority groups increases in the U.S. (40), it is critical that programs like SEAD are developed and integrated into real-world care.
Received 28 June 2024 and accepted 22 August 2024
Article Information
Acknowledgments. The authors thank the patients, diabetes providers, clinical/support staff, and clinical and practice leadership (Y. Tomer, J. Crandall, E. Epstein, V. Tabatabaie, and C. Barrett) at the Fleischer Institute for Diabetes and Metabolism for their role in the SEAD clinic.
Funding. Financial support for this study was provided by the National Institutes of Health, National Institute of Diabetes and Digestive Kidney Diseases (5K23-DK115896, P30DK111022) for the corresponding author’s time to publish but not care for the SEAD patients.
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. S.A., P.M., C.S., and J.A.L. conducted and analyzed the study and wrote and edited the manuscript. M.F., M.G., S.L.L., and S.M. contributed to the work presented in the study and edited the manuscript. S.A. 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.
Prior Presentation. Parts of this study were presented at an oral presentation at the 83rd Scientific Sessions of the American Diabetes Association, San Diego, CA, 23–26 June 2023.
Handling Editors. The journal editors responsible for overseeing the review of the manuscript were Elizabeth Selvin and Stephanie L. Fitzpatrick.