Greater Telehealth Use Results in Increased Visit Frequency and Lower Physician Related-Distress in Adolescents and Young Adults With Type 1 Diabetes

Abstract

Background:

Type one diabetes (T1D) management is challenging for adolescents and young adults (AYAs) due to physiological changes, psychosocial challenges, and increasing independence, resulting in increased diabetes distress and hemoglobin A1c (HbA1c). Alternative care models that engage AYAs and improve diabetes-related health outcomes are needed.

Methods:

A 15-month study evaluated an adaptation of the Colorado Young Adults with T1D (CoYoT1) Care model. CoYoT1 Care includes person-centered care, virtual peer groups, and physician training delivered via telehealth. AYAs (aged 16-25 years) were partially randomized to CoYoT1 or standard care, delivered via telehealth or in-person. As the study was ending, the COVID-19 pandemic forced all AYAs to transition to primarily telehealth appointments. This secondary analysis compares changes in clinic attendance, T1D-related distress, HbA1c, and device use between those who attended more than 50% of diabetes clinic visits via telehealth and those who attended more sessions in-person throughout the course of the study.

Results:

Out of 68 AYA participants, individuals (n = 39, 57%) who attended most (>50%) study visits by telehealth completed more diabetes care visits (3.3 visits) than those (n = 29, 43%) who primarily attended visits in-person (2.5 visits; P = .007). AYAs who primarily attended visits via telehealth maintained stable physician-related distress, while those who attended more in-person visits reported increases in physician-related distress (P = .03).

Conclusions:

Greater usage of telehealth improved AYA engagement with their care, resulting in increased clinic attendance and reduced physician-related diabetes distress. A person-centered care model delivered via telehealth effectively meets the needs of AYAs with T1D.

Keywords

type one diabetes adolescents young adults telehealth diabetes distress

Introduction

Type one diabetes (T1D) management is challenging during the transition from adolescence to young adulthood due to psychological and physiological changes, increasing independence, and competing life events (eg, college, work, relationships).¹ Consequently, fewer than 17% of adolescents and young adults (AYAs) with T1D in the United States achieve recommended blood glucose levels as determined by the American Diabetes Association’s (ADA) guidelines (<7.5% hemoglobin A1c [HbA1c] for adolescents).¹ AYAs living with T1D who experience persistently elevated blood glucose levels and diabetes-related adverse events (eg, diabetic ketoacidosis, severe hypoglycemia) are at risk of developing diabetes-related complications in their twenties.^2,3 Physical complications are compounded by T1D-related distress, which is associated with increased HbA1c, less engagement with self-care regimens, and reductions in quality of life.^4,5 Systemic barriers to care and discrimination also uniquely impact racially and ethnically diverse AYAs, leading to worse outcomes.^6-8

Since the start of the COVID-19 pandemic, telehealth has emerged as a promising alternative approach to delivering diabetes care to AYAs with T1D. Telehealth studies have led to improvements in HbA1c, self-efficacy, and other health-related outcomes compared with in-person care among medically underserved populations from rural settings and with lower socioeconomic status.^9-15 If deployed effectively, telehealth can reduce health inequities by reducing time missed from work and school. However, while general digital access has steadily improved in recent decades, there is still a digital divide in access to internet, computer literacy, and ability to access culturally and linguistically appropriate digital health content.^16,17

In order to meet the needs of young adults (YAs; aged 18-25 years) with T1D, the Colorado Young Adults with T1D (CoYoT1 Clinic) model was developed. CoYoT1 Clinic is an innovative T1D care model that involves telehealth, peer groups, and person-centered care.¹⁸ In pilot studies, YAs with T1D receiving CoYoT1 Care reported clinically and statistically significant benefits in engagement, satisfaction,¹⁸ and diabetes distress compared to YAs in standard care.¹⁹

Our interdisciplinary team adapted the CoYoT1 Clinic model to address the T1D care needs of a racially and ethnically diverse population of AYAs with public insurance in a large, urban pediatric hospital in California.²⁰ The CoYoT1 Clinic model was adapted using pilot study feedback, as well as input from stakeholders (eg, AYAs, clinic administrators, clinicians), an AYA advisory board, and a multidisciplinary team (eg, occupational therapist, psychologists, health economist, physicians, social workers, registered nurses). The revised model, CoYoT1 to California, included video post-visit summaries, virtual peer group sessions, caregiver inclusion at the AYA's discretion, shared decision-making tools for both the AYA and physicians, and a centralized diabetes data platform.²⁰ The care model was recently evaluated in a randomized controlled trial (RCT) comparing telehealth to in-person care and CoYoT1 Care to standard care practices. The current study is a secondary analysis that evaluates diabetes health-related outcomes associated with higher telehealth usage (>50% of visits via telehealth) compared to higher usage of in-person care (>50% of in-person visits).

Methods

Participants

Children’s Hospital Los Angeles (CHLA) serves a predominantly publicly insured, racially and ethnically diverse pediatric population (63% publicly insured or uninsured; 87% non-white). English-speaking AYAs aged 16 to 25 years, with T1D for at least six months, receiving or pending care at CHLA were eligible to participate. AYAs with public insurance, self-pay, and/or private insurance (with the exception of United Healthcare or Healthnet due to lack of telehealth nonreimbursement prior to the COVID-19 pandemic) were eligible for inclusion. Pregnant and non–English-speaking AYAs were excluded, as were AYAs with cognitive challenges that may preclude survey completion or virtual peer group participation.

Recruitment

Electronic medical record (EMR) queries were conducted to screen AYAs prior to being recruited in the clinic, electronically via phone or email or by direct physician referral. After reviewing study information and indicating interest, research coordinators reviewed consent forms and obtained signatures from all AYAs and their families. AYAs were provided electronic gift cards for their participation in appointments and REDCap/Qualtrics^21,22 surveys—$140 for visits one to four and $20 for end-of-study activities.

Study Design and Pandemic Adjustments

Originally designed as an RCT, the CoYoT1 to California study used pragmatic care assignment and partial randomization to evaluate the intervention over a 15-month period. Three physicians were assigned to deliver standard care, and three were assigned and trained to provide CoYoT1 Care. Participants whose physicians were participating in the study were given the choice of retaining their physicians, and if they chose to retain, they indirectly selected their treatment group (CoYoT1 Care or standard care). Those who did not retain their prestudy physician were randomly assigned a treatment group. All AYAs were randomly assigned to receive treatment via telehealth or in-person visits at enrollment, resulting in four study groups (Figure 1). AYAs randomized into telehealth groups were expected to attend three telehealth visits and one in-person visit over 12 to 15 months (approximately one visit per quarter); those randomized to in-person groups were expected to attend four in-person appointments over 12 to 15 months. All study procedures and recruitment materials were approved by the CHLA Institutional Review Board.

Figure 1.

Study and intervention design. Abbreviations: PCC, person-centered care; VGP, virtual peer groups; CoYoT1, Colorado Young Adults with T1D; T1D, type one diabetes.

On March 15, 2020, the state of California entered lockdown due to the COVID-19 pandemic, which required modification of the study design. Using the technology implemented for the telehealth study groups, all AYAs still participating in the study were transitioned to receive appointments primarily via telehealth, regardless of initial study assignment. Onboarding materials and procedures established for the telehealth study groups were used to train all AYAs who transitioned to telehealth during the pandemic. All AYAs still completed one in-person visit, following the protocol for those assigned to telehealth. The overall timing of appointments was not delayed because of the pandemic and did not differ between telehealth and in-person participants. The present study considers differences based on actual telehealth use, which was more common in those assigned to a telehealth study group at enrollment.

Study Treatment Groups

Standard care

For standard care, physicians delivered routine care, and no intervention was provided. AYAs also received routine care and received copies of their treatment plan summaries as usual.

CoYoT1 care

CoYoT1 Care is a person-centered care approach that uses shared decision-making tools that support patient autonomy. Training sessions were provided to CoYoT1 Care physicians, and debriefing sessions were set up for additional support throughout the course of the study. AYAs in CoYoT1 Care rank their diabetes-related discussion topics prior to their care appointments and then review these issues with their physicians at the beginning of the visit. Both AYAs and physicians collaborate on the visit agenda and treatment plan, which is then summarized in videos that are sent to participants. Additionally, AYAs had access to eight 30- to 60-minute virtual peer group sessions led by a YA facilitator with T1D. Each session covered a specific diabetes-related topic (eg, diabetes distress, burnout). During these sessions, a research coordinator was available for technical support.

Telehealth

To participate in telehealth appointments, AYAs needed an internet-connected computer, tablet, or smartphone that was equipped with a microphone and camera. Only one AYA did not have a smartphone; the device was provided for the study. AYAs were instructed to obtain HbA1c tests from a nearby laboratory before their telehealth appointments. AYAs in the telehealth cohort uploaded their diabetes technology to Tidepool (Palo Alto, CA, USA), a free virtual diabetes management software.²³ When technical issues arose, data were downloaded through the diabetes devices' proprietary software or hand-written in a blood glucose log.

In-person

AYAs assigned to in-person care downloaded diabetes device data and completed HbA1c testing in the clinic during routine diabetes care appointments or reviewed data on their diabetes device with their physician. During the pandemic, many of these participants were moved to Tidepool or hand-written data logs.

Participant-Reported Outcomes

Demographics and clinical data were collected from the EMR. AYAs were also asked to complete self-reported measures. Questionnaires were completed at baseline (prior to visit one) and after every quarterly visit (visits one-four). Due to delays in scheduling, some appointments were more than three months apart, and participants were in the study for up to 15 months. For participants who did not attend a visit, research staff attempted to collect data within six weeks of the missed quarterly appointment date. Participants completed three validated measures at baseline and after visit four, as follows:

The Diabetes Distress Scale (DDS)²⁴ uses 17 items to assess distress related to living with T1D across four subscales: emotional burden, regimen-related, physician-related, and interpersonal distress. Items are scored on a six-point scale (1, not a problem, through 6-a, very serious problem). An average total DDS score or average subscale score of three or greater indicates high distress.²⁵

The Diabetes Empowerment Scale-Short Form²⁶ uses eight items rated on a five-point scale (1, strongly disagree, to 5, strongly agree) to assess diabetes-related psychosocial self-efficacy, including managing the psychosocial aspect of diabetes, assessing dissatisfaction and readiness to change, and setting and achieving goals.

The Center for Epidemiologic Studies Depression Scale (CES-D)²⁷ uses 20 items to assess depressive symptoms, each rated on a four-point scale (0, none, to 3, most of the time). A total CES-D score ≥16 suggests clinically significant depressive symptoms.

Statistical Analysis

For this secondary analysis, patients were classified based on the proportion of visits completed via telehealth, regardless of the number of visits completed during the study. AYAs who attended more than half of their study visits via telehealth were compared to AYAs who completed more than half of their visits in-person (ie, fewer than 50% via telehealth). Demographic and clinical characteristics were compared between these groups using independent samples t tests for continuous measures or Fisher’s exact tests for categorical measures. Changes in participant-reported outcomes from baseline to the study end were compared between groups using linear regression models adjusted for age, gender, retention of prestudy physician, participation in CoYoT1 Care, and baseline values for each outcome; fit via maximum likelihood²⁶ estimation with robust standard errors; and weighted based on the inverse probability of treatment selection to account for partial self-selection of treatment group. To include measurements obtained throughout the study, changes in HbA1c were evaluated using linear mixed-effects models adjusted with the same covariates. Changes in binary outcomes (continuous glucose monitoring [CGM] and insulin pump use) were evaluated using mixed-effects logistic regression models adjusted with the same covariates.

A modified intent-to-treat analysis approach was used to include all available data from participants, including those with missing follow-up data, without imputation of missing values. Given the data from n = 68 participants across study groups, the study was powered to detect a significant treatment effect of R² = 0.22 or greater using an limited mixed-effects model or f² = 0.09 and greater using a linear regression model, assuming a type I error rate α = 0.05 and a type II error rate β = 0.20 (or 80% power). Even with models utilizing incomplete data (with as few as n = 48 cases), the study was powered to detect medium-to-large effects across outcomes. All treatment effect model terms were considered significant at P values <.05 and marginally significant at values <.10. Actual treatment effects obtained from regression models are reported without standardization. Effect sizes for binary outcomes are reported as percentage change in use. All analyses were conducted using Stata/SE 14.2 (StataCorp LLC, College Station, TX, USA).

Results

Demographics

Eighty individuals initially enrolled in the study. Eleven individuals did not complete baseline assessments and did not follow up with research staff, and one participant withdrew and requested that their data not be used (15% attrition). Although more male participants dropped out of the study than female participants, there were no significant differences in terms of racial, ethnic, or gender identities between study participants and those who dropped out of the study. Data on the remaining 68 participants, who completed baseline surveys and then either completed or partially completed their assessments throughout the study, were analyzed and included here (Figure 2). Demographic and baseline clinical characteristics for those who completed most visits via telehealth and those who completed most visits in-person are shown in Table 1. Racial, ethnic, and gender identities were similar across all study groups.

Table 1.

Demographic and Clinical Data by Study Group.

Variables	>50% Visits in-person (n = 29)	>50% Visits via telehealth (n = 39)	Total (N = 68)	P value
Demographic data
Gender, n (%)				.58
Female	10 (34)	17 (44)	27 (40)
Male	10 (34)	14 (36)	24 (35)
Unknown/declined	9 (31)	8 (21)	17 (25)
Race, n (%)				.15
African American	5 (17)	1 (3)	6 (9)
Asian	1 (3)	3 (8)	4 (6)
Native North and South American	2 (7)	6 (16)	8 (12)
White	10 (34)	19 (49)	29 (42)
Multiracial	3 (10)	5 (13)	8 (12)
Unknown	8 (28)	5 (13)	13 (19)
Ethnicity, n (%)				1.00
Latinx	15 (52)	19 (49)	34 (50)
Non-Latinx	14 (48)	20 (51)	34 (50)
Insurance, n (%)				.34
Public (Medicaid)	22 (76)	29 (74)	51 (75)
Private	7 (24)	7 (18)	14 (21)
Both public and private	0	3 (8)	3 (4)
Clinical data, mean (SD)
Age at enrollment (years)	18.38 (1.21)	17.97 (1.20)	18.15 (1.21)	.18
T1D duration (years)	8.13 (3.58)	8.16 (4.26)	8.15 (3.98)	.98
Unadjusted baseline HbA1c (%)	9.25 (2.19)	9.01 (2.28)	9.11 (2.23)	.66
Annual prestudy attendance	2.51 (0.94)	2.47 (0.91)	2.49 (0.23)	.87
Study group assignment, n (%)
CoYoT1 Care	17 (59)	23 (59)	40 (59)	1.00
Randomized to in-person care	15 (52)	3 (8)	18 (26)
Randomized to telehealth	2 (7)	20 (51)	22 (32)
Standard care	12 (41)	16 (41)	28 (41)
Randomized to in-person care	8 (28)	10 (26)	18 (26)
Randomized to telehealth	4 (14)	6 (15)	10 (15)

Abbreviations: SD, standard deviation; T1D, type one diabetes; HbA1c, hemoglobin A1c; CoYoT1, Colorado young adults with T1D.

Figure 2.

Consolidated Standards of Reporting Trials (CONSORT) diagram. Abbreviations: CoYoT1, Colorado Young Adults with T1D; T1D, type one diabetes.

Appointment Attendance

Study visit attendance was strongly related to study group assignment. Among those participants who attended more than half of the clinic visits by telehealth (n = 39), 67% were assigned to the telehealth group at enrollment; among those who attended more than half of their sessions in-person (n = 29), 79% were assigned to in-person care at enrollment. Participation in telehealth was equal across care groups (59% were in CoYoT1 Care, and 41% were in standard care; P = 1.00, see Table 1).

Participation in the study primarily via telehealth was associated with significantly higher attendance. AYAs who attended more than half of their visits via telehealth were seen 3.33 times over the study period on average, compared to 2.47 times over the study period for participants who attended more sessions in-person (P = .007; see Table 2 and Figure 3).

Table 2.

Adjusted Mean Changes in Health-Related Outcomes Over Study by Treatment Group.

Measure/subscale, mean (SD)	>50% Visits in-person (n = 29)				>50% Visits via telehealth (n = 39)				Tx effect	P value
Measure/subscale, mean (SD)	Baseline	Study end	Δ	P value for Δ	Baseline	Study end	Δ	P value for Δ	Tx effect	P value
Study attendance	2.51 (0.94)	2.47 (1.72)	−0.04	.95	2.47 (0.91)	3.33 (1.03)	0.86	.0001	0.86	.007
DDS
Average score	2.21 (1.00)	2.58 (1.28)	0.37	.20	1.87 (0.82)	2.02 (0.84)	0.15	.90	−0.57	.13
Emotional distress	2.13 (1.04)	2.56 (1.34)	0.43	.16	1.79 (0.75)	1.97 (0.84)	0.18	.83	−0.59	.12
Interpersonal distress	2.39 (1.36)	2.65 (1.46)	0.26	.41	2.05 (1.26)	2.17 (1.13)	0.12	.90	−0.48	.22
Physician distress	1.98 (0.83)	2.62 (1.19)	0.64	.05	1.71 (0.71)	1.84 (0.78)	0.13	.93	−0.77	.03
Regimen distress	2.36 (1.11)	2.56 (1.38)	0.20	.37	1.97 (0.96)	2.09 (0.96)	0.12	.99	−0.47	.23
DES-SF	4.11 (0.71)	4.40 (0.64)	0.29	.34	3.93 (1.24)	4.26 (0.84)	0.33	.57	−0.14	.46
CES-D	20.68 (12.99)	22.52 (10.19)	1.84	.56	17.86 (9.71)	21.81 (13.15)	3.95	.11	−0.71	.84
HbA1c %	9.25 (2.19)	8.36 (2.18)	−0.89	.46	9.01 (2.28)	8.48 (1.62)	−0.53	.69	−0.28	.40
CGM Use, n (%)	7 (24%)	5 (17%)	−2 (−7%)	.48	12 (31%)	15 (38%)	3 (+7%)	.19	+14%	.22
Pump Use, n (%)	8 (28%)	3 (9%)	−5 (−19%)	.04	19 (49%)	20 (51%)	1 (+2%)	.66	+21%	.06

Δ—mean change from baseline to study end; Tx effect—treatment effects, obtained from mixed models with age, gender, participation, and baseline values of outcomes as covariates, weighted using the inverse probability of treatment selection (IPTW). Baseline values for study attendance are based on attendance in the year prior to study enrollment. Baseline values are unadjusted, and study end values are obtained from adjusted mixed models. Significant effects and P values are marked in bold.

Abbreviations: Tx effect, treatment effects; SD, standard deviation; DDS, Diabetes Distress Scale; DES-SF, Diabetes Empowerment Scale-Short Form; CES-D, Center for Epidemiologic Studies Depression Scale; CGM, continuous glucose monitor; HbA1c, hemoglobin A1c.

Figure 3.

Clinic attendance by telehealth usage.

Diabetes Distress Scale

Participants who attended more than half of their clinic visits via telehealth reported little change in physician-related T1D distress, whereas those who attended more in-person visits reported increased distress over the course of the study (P = .03; Table 2; Figure 4). Other domains that showed similar benefits were those who attended 50% or more sessions via telehealth reported less severe increases in distress than those who attended more sessions in-person, although these effects did not reach statistical significance.

Figure 4.

Changes in DDS physician-related distress over the study period by telehealth usage (>50% of visits in-person vs >50% visits via telehealth). Abbreviation: DDS, Diabetes Distress Scale.

When participants were examined relative to their treatment group assignment and telehealth usage, AYAs who participated in the full intervention (CoYoT1 Care primarily via telehealth) reported significant benefits for physician-related distress (Δ = −1.03; P = .009) and marginal benefits for emotional burden (P = .05), interpersonal distress (P = .09), and regimen-related distress (P = .07) in comparison to those in standard care who attended more in-person care visits, who reported large increases in distress over the study period (DDS physician distress Δ = 0.62; P = .05). This additional benefit was not found in the group that primarily attended CoYoT1 Care sessions in-person when they were compared to standard care in-person attendees (Δ = −0.26; P = .24) but was observed in standard care participants who attended more sessions via telehealth (Δ = −0.77; P = .03) when they were compared to standard care in-person attendees.

Diabetes Technology Device Use

CGM use did not significantly change over the study period. However, those who attended fewer sessions via telehealth reported significant decreases in insulin pump use over the study (Δ = −19%; P = .04), while those who attended more than half of their sessions via telehealth reported small increases in pump use (Table 2).

Hemoglobin A1c

Changes in HbA1c were similar across telehealth usage groups (Table 2). Notably, when participants were examined relative to both telehealth usage and care group assignment, those in the full intervention (CoYoT1 Care primarily via telehealth) reported significant reductions in HbA1c over the course of the study period (ΔA1c = −0.82%, P = .048) in comparison to those in standard care who primarily attended in-person visits and reported increases over the study (ΔA1c = 0.42%, P = .29). Those who attended CoYoT1 Care sessions in-person showed smaller, nonsignificant changes (ΔA1c = −0.43%, P = .39) than standard care in-person attendees, while those who attended standard care sessions via telehealth showed little difference in A1c compared to standard care in-person attendees (ΔA1c = −0.08%, P = .89).

Other Measures

Participants did not report any significant differences in terms of diabetes empowerment or depressive affect over the course of the study (Table 2).

Discussion

Transitioning from adolescence to young adulthood is challenging, and AYAs with T1D must follow intensive diabetes regimens while also navigating a complex and expensive health care system, often leading to decreased engagement with clinical care.^28-30 Telehealth presents a means of increasing access and engagement in diabetes care for AYAs with T1D, which may result in improved health outcomes and cost savings. In this study, having access to telehealth increased visit frequency, compared with those who primarily attended in-person visits. While the COVID-19 pandemic was unprecedented and the design of the study was altered, the benefits of telehealth were still significant among these individuals, and they were often still able to meet the ADA recommendation of quarterly clinic visits. This outcome is similar to that of the CoYoT1 Clinic pilot study, where an increase in clinic appointment attendance also occurred.¹⁸ Additionally, telehealth and interventions focusing on peer connection have been shown in prior studies to increase engagement.^18,31 The support of a research coordinator and their availability via telehealth to answer questions or connect them to needed services may also increase engagement. In other studies, AYAs attended more diabetes care visits when transition care programs had a clinic coordinator.^32,33

Physician-related distress was stable in those who received the majority of their care through telehealth, while those who received less than half of their care via telehealth saw a significant increase in physician-related distress. Those exposed to the full intervention (CoYoT1 Care via telehealth) showed the most stability over the study, while those who received some component of the intervention still benefited relative to those in standard care who attended sessions primarily in-person. The lack of changes in diabetes distress, even during a global pandemic, may be due to the compound effect of having access to care via telehealth and having a person-centered care approach in CoYoT1 Care. Their engagement with their physician was altered from standard care to a more collaborative care approach, where AYAs felt more comfortable in expressing their unique diabetes concerns and goals at regular sessions that were uninterrupted during the pandemic.²⁰

The CoYoT1 model was developed in a standalone diabetes center and examined in a predominantly non-Latinx white AYA population with T1D and private insurance, and participants demonstrated positive diabetes-related outcomes. When it was adapted in a culturally appropriate and person-centered manner with stakeholder input and subsequently examined in an urban hospital, serving a predominantly racially and ethnically diverse AYA population with T1D, the intervention continued to show benefits. These outcomes indicate that providing access to care via telehealth and utilizing an alternative care model (CoYoT1 Care) can engage diverse groups of AYAs with the greatest need, regardless of their location, background, or insurance status.

Although telehealth usage increased visit attendance and reduced physician-related distress, no significant differences in HbA1c were observed when telehealth usage alone was evaluated. However, individuals who participated in CoYoT1 Care via telehealth (ie, the full intervention) did show significant improvements in glycemic control, while those who participated in CoYoT1 Care in-person showed smaller gains that did not reach significance. Taken together, these results suggest that AYAs can benefit from more than one-on-one T1D care if they are given sufficient access to a tailored intervention via telehealth.

Individuals who used telehealth more frequently did report modest increases in CGM use, while those who attended more in-person visits reported decreases in insulin pump use. Telehealth use itself may be insufficient to encourage device use, as individuals who adopt telehealth more already report higher rates of device use, reflecting their own comfort with technology. Person-centered care or peer support is more likely to generate interest in device use, as participants can share experiences and benefits with others.

Other diabetes-related outcomes, including empowerment and depression, also did not show significant changes over the study period when telehealth usage was considered. Empowerment is more likely to be impacted by a person-centered intervention that exposes patients to language and ideas that are associated with care autonomy. Furthermore, telehealth may be useful for ameliorating depressive symptoms in some individuals, but its usefulness for addressing feelings in AYAs with T1D requires evaluation in a longer study.

Limitations

While the results from this intervention were positive, some limitations exist. Because this study was conducted at an urban pediatric hospital with a unique population, it may not be generalizable in other settings (eg, hospitals in rural locations). Allowing participants to retain their care provider led to uneven distribution of study groups, although the primary focus of this analysis was telehealth use, which was fully randomized prior to the pandemic. Although most participants had completed or nearly completed the study at the onset of the COVID-19 pandemic, participants were all moved to predominantly telehealth visits to comply with safety and travel restrictions. Those originally assigned to telehealth were equipped to handle the pandemic more effectively than those originally assigned to in-person care. Additionally, AYAs that enrolled in the study may differ from those that declined, although we have minimal data on patients who declined participation.

Conclusion

Greater telehealth usage during the CoYoT1 to California trial was associated with increased attendance and stable diabetes distress, compared to increased distress seen in those who attended clinic visits in-person. Notably, this effect was observed in a diverse AYA population, who frequently encounter challenges to accessing care. These results suggest alternative care models can assist in engaging AYAs in their diabetes care.

Footnotes

Acknowledgements

The authors would like to thank the adolescents, young adults, and their families for participating in research. This work would not be possible without their trust, commitment, and willingness to participate in research. The authors thank the grant funders, The Patrick and Catherine Weldon Donaghue Medical Research Foundation, for their support and entrusting the authors to improve diabetes health outcomes across communities of need. The authors would also like to express their deepest gratitude to the team (eg, clinicians, administrators, coordinators) as this work is only possible with their support.

Correction (May 2023): Article updated; for further details please see for more details.

Abbreviations

ADA, American Diabetes Association; AYA, adolescents and young adults; CES-D, Center for Epidemiologic Studies Depression Scale; CGM, continuous glucose monitor; CHLA, Children’s Hospital Los Angeles; DDS, Diabetes Distress Scale; CoYoT1, Colorado Young Adults with T1D; EMR, electronic medical record; HbA1c, hemoglobin A1c; RCT, randomized controlled trial; T1D, type one diabetes; YA, young adult.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by The Patrick and Catherine Weldon Donaghue Medical Research Foundation.

ORCID iDs

Jaquelin Flores Garcia

Ethan Faye

Mark W. Reid

Daniel I. Bisno

Alejandra Torres Sanchez

Sarah Hiyari

Jennifer K. Raymond

References

Foster

Beck

Miller

, et al. State of type 1 diabetes management and outcomes from the T1D Exchange in 2016-2018. Diabetes Technol Ther. 2019;21(2):66-72. doi:10.1089/dia.2018.0384.

Park

Scott

Adams

Brindis

Irwin

Jr.

Adolescent and young adult health in the united states in the past decade: little improvement and young adults remain worse off than adolescents. J Adolesc Health. 2014;55(1):3-16. doi:10.1016/j.jadohealth.2014.04.003.

Monaghan

Helgeson

Wiebe

Type 1 diabetes in young adulthood. Curr Diabetes Rev. 2015;11(4):239-250. doi:10.2174/1573399811666150421114957.

Buysse

de Moor

de Maeseneer

Introducing a telemonitoring platform for diabetic patients in primary care: will it increase the socio-digital divide?

Prim Care Diabetes. 2013;7(2):119-127. doi:10.1016/j.pcd.2012.10.085.

Ramirez

Ojeaga

Espinoza

Hensler

Honrubia

Telemedicine in minority and socioeconomically disadvantaged communities amidst COVID-19 pandemic. 2020;164(1):91-92.doi:10.1177/0194599820947667.

Lipman

Smith

Patil

Willi

Hawkes

CP.

Racial disparities in treatment and outcomes of children with type 1 diabetes. Pediatr Diabetes. 2021;22(2):241-248. doi:10.1111/pedi.13139.

Walker

Hood

Gurka

, et al. Barriers to technology use and endocrinology care for underserved communities with type 1 diabetes. Diabetes Care. 2021;44(7):1480-1490. doi:10.2337/dc20-2753.

Fantasia

Wirunsawanya

Lee

Rizo

Racial disparities in diabetes technology use and outcomes in type 1 diabetes in a safety-net hospital. J Diabetes Sci Technol. 2021;15(5):1010-1017. doi:10.1177/1932296821995810.

Shea

Kothari

Teresi

, et al. Social impact analysis of the effects of a telemedicine intervention to improve diabetes outcomes in an ethnically diverse, medically underserved population: findings from the IDEATel study. Am J Public Health. 2013;103(10):1888-1894. doi:10.2105/AJPH.2012.300909.

10.

Khairat

Haithcoat

Liu

, et al. Advancing health equity and access using telemedicine: a geospatial assessment. J Am Med Inform Assoc. 2019;26(8-9):796-805. doi:10.1093/JAMIA/OCZ108.

11.

Cuttriss

Bouchonville

Maahs

Walker

AF.

Tele-rounds and case-based training: project ECHO telementoring model applied to complex diabetes care. Pediatr Clin North Am. 2020;67(4):759-772. doi:10.1016/j.pcl.2020.04.017.

12.

Predieri

Leo

Candia

, et al. Glycemic control improvement in Italian children and adolescents with type 1 diabetes followed through telemedicine during lockdown due to the COVID-19 pandemic. Front Endocrinol (Lausanne). 2020;11:595735. doi:10.3389/fendo.2020.595735.

13.

Guljas

Ahmed

Chang

Whitlock

Impact of telemedicine in managing type 1 diabetes among school-age children and adolescents: an integrative review. J Pediatr Nurs. 2014;29(3):198-204. doi:10.1016/j.pedn.2013.10.013.

14.

Tchero

Kangambega

Briatte

Brunet-Houdard

Retali

Rusch

Clinical effectiveness of telemedicine in diabetes mellitus: a meta-analysis of 42 randomized controlled trials. Telemed J E Health. 2019;25(7):569-583. doi:10.1089/tmj.2018.0128.

15.

Pujara

Sutton

Rhee

Peer Reviewed: telemedicine in the management of type 1 diabetes. Prev Chronic Dis. 2018;15(1):170168. doi:10.5888/PCD15.170168.

16.

Latulippe

Hamel

Giroux

Social health inequalities and eHealth: a literature review with qualitative synthesis of theoretical and empirical studies. J Med Intern Res. 2017;19(4):e136.

17.

Feng

Xie

Digital divide 2.0: the role of social networking sites in seeking health information online from a longitudinal perspective. J Health Commun. 2015;20(1):60-68. doi:10.1080/10810730.2014.906522.

18.

Reid

Krishnan

Berget

, et al. CoYoT1 Clinic: home telemedicine increases young adult engagement in diabetes care. Diabetes Technol Ther. 2018;20(5):370-379. doi:10.1089/dia.2017.0450.

19.

Bakhach

Reid

Pyatak

, et al. Home telemedicine (CoYoT1 Clinic): a novel approach to improve psychosocial outcomes in young adults with diabetes. Diabetes Educ. 2019;45(4):420-430. doi:10.1177/0145721719858080.

20.

Raymond

Reid

Fox

, et al. Adapting home telehealth group appointment model (CoYoT1 clinic) for a low SES, publicly insured, minority young adult population with type 1 diabetes. Contemp Clin Trials. 2020;88:105896. doi:10.1016/j.cct.2019.105896.

21.

Research Electronic Data Capture. https://www.project-redcap.org/. Date unknown. Accessed September 19, 2022.

22.

Qualtrics. https://www.qualtrics.com/. Date unknown. Accessed September 19, 2022.

23.

Tidepool. https://www.tidepool.org/. Date unknown. Accessed May 4, 2021.

24.

Polonsky

Fisher

Earles

, et al. Assessing psychosocial distress in diabetes development of the Diabetes Distress Scale. Diabetes Care. 2005;28(3):626-631. doi:10.2337/DIACARE.28.3.626.

25.

Fisher

Hessler

Polonsky

Mullan

When is diabetes distress clinically meaningful? establishing cut points for the Diabetes Distress Scale. Diabetes Care. 2012;35(2):259-264.

26.

Anderson

Fitzgerald

Gruppen

Funnell

MS.

The diabetes empowerment scale-short form (DES-SF). Diabetes Care. 2003;26(5):1641-1642.

27.

Radloff

LS.

The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas. 1997;1(3):385-401.

28.

Bryden

Dunger

Mayou

Peveler

Neil

HA.

Poor prognosis of young adults with type 1 diabetes a longitudinal study. Diabetes Care. 2003;26(4):1052-1057. doi:10.2337/diacare.26.4.1052.

29.

Hilliard

Perlus

Clark

, et al. Perspectives from before and after the pediatric to adult care transition: a mixed-methods study in type 1 diabetes. Diabetes Care. 2014;37(2):346-354. doi:10.2337/dc13-1346.

30.

Garvey

Beste

Luff

Atakov-Castillo

Wolpert

Ritholz

Experiences of health care transition voiced by young adults with type 1 diabetes: a qualitative study. Adolesc Health Med Ther. 2014;5:191-198. doi:10.2147/ahmt.s67943.

31.

Wood

Clements

McFann

Slover

Thomas

Wadwa

RP.

Use of telemedicine to improve adherence to American Diabetes Association standards in pediatric type 1 diabetes. Diabetes Technol Ther. 2016;18(1):7-14. doi:10.1089/dia.2015.0123.

32.

Svoren

Butler

Levine

Anderson

Laffel

LM.

Reducing acute adverse outcomes in youths with type 1 diabetes: a randomized, controlled trial. Pediatrics. 2003;112(4):914-922. doi:10.1542/PEDS.112.4.914.

33.

Holmes-Walker

Llewellyn

Farrell

A transition care programme which improves diabetes control and reduces hospital admission rates in young adults with type 1 diabetes aged 15-25 years. Diabet Med. 2007;24(7):764-769. doi:10.1111/J.1464-5491.2007.02152.