OBJECTIVE

To evaluate the efficacy and safety of a digital therapeutic application (app) delivering cognitive behavioral therapy (CBT) designed to improve glycemic control in patients with type 2 diabetes.

RESEARCH DESIGN AND METHODS

Adults with type 2 diabetes and an HbA1c of 7 to <11% were randomly assigned to receive access to a digital therapeutic app delivering CBT (BT-001) or a control app, both on top of standard of care management. CBT is an established form of psychological treatment that endeavors to identify and change unhelpful thinking patterns. The primary study end point was treatment group difference in mean HbA1c change from baseline to 90 days.

RESULTS

Among 669 randomly assigned subjects who completed app onboarding, the mean age was 58 years, BMI 35 kg/m2, 54% were female, 28% Black, and 16% Latino. Baseline HbA1c was 8.2 and 8.1% in the BT-001 and control groups, respectively. After 90 days of app access, change in HbA1c was −0.28% (95% CI −0.41, −0.15) in the BT-001 group and +0.11% (95% CI −0.02, 0.23) in the control group (treatment group difference 0.39%; P < 0.0001). HbA1c reduction paralleled exposure to the therapeutic intervention, assessed as the number of modules completed on the app (P for trend <0.0001). No adverse events in either group were attributed to app use and no adverse device effects reported.

CONCLUSIONS

Patients randomly assigned to the BT-001 arm relative to the control arm had significantly lower HbA1c at 90 days. The digital therapeutic may provide a scalable treatment option for patients with type 2 diabetes.

In the U.S., about half of adults with diagnosed diabetes attain an HbA1c <7% (1). Factors contributing to suboptimal glycemic control include inadequate self-care, lack of knowledge and ability to implement recommended behavior changes, and psychiatric comorbidities (2,3). Cognitive behavioral therapy (CBT) is an evidence-based treatment grounded in the idea that all behaviors are learned and that dysfunctional core beliefs contribute to errors in processing information, leading to emotional and behavioral dysfunction. Meta-analyses of randomized trials of CBT-based interventions in patients with diabetes have reported HbA1c reductions ranging from 0.3 to 1.6% (4,5). The trials used individual counseling or group interventions delivered in person, by telephone, or internet for periods ranging from 2 weeks to 12 months. The improvement in glycemic control observed with CBT has been attributed to better adherence to lifestyle recommendations and hypoglycemic medication as well as alleviation of depression and anxiety (6,7).

Digital therapeutics delivering behavioral interventions without the need for human intervention have the potential to increase access due to their inherent scalability and reach beyond physical location and scheduling constraints. BT-001 is a digital therapeutic application (app) delivering CBT designed to improve glycemic control in adults with type 2 diabetes via smartphone. An earlier version of the app, when paired with health coaching, reduced self-reported HbA1c by 1.1% after 3 months in a sample of 69 adults with type 2 diabetes (8). A separate pilot study (n = 80) using the BT-001 app alone demonstrated a 23 mg/dL (1.3 mmol/L) fall in fasting blood glucose in a similar time frame (9). We now report the results of a randomized, controlled, open-label trial assessing the efficacy and safety of BT-001 in a larger population.

The trial design and the intervention provided by the BT-001 app have been previously described (10). In brief, adults with type 2 diabetes, with HbA1c 7 to <11% (53–97 mmol/mol) and access to a smartphone, were randomly assigned (1:1) to receive access to BT-001 or a control app for 180 days, both on top of standard of care management. Current smokers were excluded, as were those taking prandial insulin or oral corticosteroids or with active eating disorders. Subjects randomly assigned to both groups were instructed to download their assigned app from the Apple or Android app store and set up an account with username and password; those who did not complete this onboarding activity were excluded from the modified intention-to-treat (mITT) analysis population.

The predominantly automated recruitment process required potential subjects to provide electronic informed consent, complete an online screening questionnaire, and schedule a home visit for phlebotomy. The decentralized study required no in-person clinic visits; a mobile phlebotomy unit collected blood samples and physical measures during home visits using a standardized protocol. Standard of care management was provided via telehealth visits with investigators at days 90 and 180. Both the BT-001 and control groups provided patient-reported outcome data; BT-001–allocated subjects also received access to the CBT content. Antihyperglycemic medications could be adjusted as needed during the study consistent with standard of care, and all concomitant therapies were permitted.

Each week, BT-001 would ask subjects to complete a new behavioral module along with one or more related skill-based exercises. Modules addressed topics such as: personal beliefs and barriers (e.g., those related to a subject’s ability to change and control his or her behaviors); beliefs about macronutrients and the importance of various food types; hedonic-related beliefs about pleasant or unpleasant sensations experienced by eating or exercising; and beliefs about exercise.

BT-001 is intended to be prescribed for use between clinic visits within 90-day treatment cycles, each of which includes 13 modules. The CBT content for the second treatment cycle was available during the first treatment cycle if subjects chose to work ahead. They were not required to complete any specific number of modules. BT-001–allocated subjects were asked to record information about their diet and exercise behaviors, perceptions, and beliefs, self-measured blood pressure, and glucose levels within the app. Based on these factors, the app presented a treatment plan summarizing daily and weekly goals to improve their glycemic control.

The primary efficacy end point was the difference in mean HbA1c change from baseline to day 90 among BT-001–allocated subjects compared with those assigned to the control app. Safety assessments included adverse events and adverse device effects. An independent Data Safety Monitoring Committee reviewed data throughout the study.

Statistical Considerations

The planned sample size based on the primary end point was 648 subjects, 324 in each treatment group. The sample size (90% power; α = 0.05) was selected to detect a 0.4% difference in the change from baseline HbA1c between the intervention and control arms assuming a dropout rate of 20%, SD of 1.4% based on prior studies conducted by the sponsor (8,9), and a common SD for the two treatment groups.

The primary end point was assessed in the mITT population by ANCOVA with baseline HbA1c as a covariate. Treatment group differences in biomarker change from baseline were assessed in a similar manner. Change in antihyperglycemic medication was compared by χ2. Subject-reported fasting blood glucose values were assessed as weekly change from baseline with baseline defined as the first three values after completion of onboarding. Weekly values were the mean of all values entered during the week. Fasting glucose trend was assessed by linear regression. The mITT population included all randomly assigned subjects who completed the onboarding process for their assigned app. Safety assessments were also conducted on the mITT population. Analyses for this study were independently conducted by CPC Clinical Research, an academic research organization affiliated with the University of Colorado, using the academic statistical analysis plan and SAS version 9.4.

Between April and December 2021, 725 subjects were randomly assigned at 12 sites in the U.S. App onboarding, which involved downloading the assigned app and setting up an account with an e-mail address and password, was completed by 669 subjects (BT-001, 326; control app, 343) who constituted the mITT population (Supplementary Fig. 1). Of these, 610 subjects (91%) had paired baseline and day 90 HbA1c values (BT-001, 291; control app, 319). During the 90-day treatment period, 14 subjects in the BT-001 group (4%) withdrew consent and 12 (4%) were lost to follow-up; in the control app group, 9 (3%) withdrew consent and 6 (2%) were lost to follow-up.

Demographics and baseline characteristics are summarized (Table 1); no statistically significant differences between treatment groups were observed. Mean age was 58 ± 9 years in both treatment groups (range 31–75 years), and BMI was 35 ± 7 kg/m2. The cohort included 56% women and 28% Black, 5% Asian, and 16% Latino subjects with similar proportions in the two treatment groups. Baseline HbA1c was 8.2 ± 1.0% (66 mmol/mol) in the BT-001 group and 8.1 ± 1.0% (65 mmol/mol) in the control group. The mean number of antihyperglycemic medications at baseline was 2.1 ± 1.1 in both groups. Metformin use was reported by 90%, sulfonylurea by 42%, basal insulin by 22%, glucagon-like peptide 1 agonists by 35%, and sodium–glucose cotransporter 2 inhibitors by 31% (Supplementary Table 1).

Table 1

Baseline characteristics

BT-001Control app
N 326 343 
Age, years, mean (SD) 57 (9) 58 (8) 
Female, n (%) 176 (54) 190 (55) 
Race, n (%)   
 White 189 (59) 201 (59) 
 Black or African American 89 (28) 98 (29) 
 Asian 13 (4) 15 (4) 
 American Indian or Alaskan Native 3 (1) 3 (1) 
 Native Hawaiian or other Pacific Islander 1 (0.3) 1 (0.3) 
 Other (includes multiple races) or not reported 26 (8) 25 (7) 
Hispanic or Latino ethnicity, n (%) 50 (17) 41 (13) 
BMI, kg/m2, mean (SD) 35 (7) 35 (7) 
Weight, kg, mean (SD) 99 (20) 101 (24) 
Systolic blood pressure, mmHg, mean (SD) 127 (15) 126 (14) 
Diastolic blood pressure, mmHg, mean (SD) 78 (10) 78 (9) 
Using antihypertensive medication, n (%) 202 (63) 233 (68) 
Total cholesterol, mmol/L, mean (SD) 4.4 (1.1) 4.4 (1.2) 
HDL-C, mmol/L, mean (SD) 1.2 (0.3) 1.2 (0.3) 
LDL-C, mmol/L, mean (SD) 2.4 (0.9) 2.4 (0.9) 
Triglycerides, mmol/L, mean (SD) 2.2 (1.5) 2.3 (4.3) 
HbA1c, %, mean (SD), mmol/mol, mean 8.2 (0.1), 66 8.1 (0.1), 65 
Fasting blood glucose, mmol/L, mean (SD) 9.4 (3.2) 9.3 (3.0) 
Years since diagnosis of diabetes, mean (SD) 11 (8) 11 (8) 
Number of antihyperglycemic medications, mean (SD) 2.1 (1.1) 2.1 (1.1) 
BT-001Control app
N 326 343 
Age, years, mean (SD) 57 (9) 58 (8) 
Female, n (%) 176 (54) 190 (55) 
Race, n (%)   
 White 189 (59) 201 (59) 
 Black or African American 89 (28) 98 (29) 
 Asian 13 (4) 15 (4) 
 American Indian or Alaskan Native 3 (1) 3 (1) 
 Native Hawaiian or other Pacific Islander 1 (0.3) 1 (0.3) 
 Other (includes multiple races) or not reported 26 (8) 25 (7) 
Hispanic or Latino ethnicity, n (%) 50 (17) 41 (13) 
BMI, kg/m2, mean (SD) 35 (7) 35 (7) 
Weight, kg, mean (SD) 99 (20) 101 (24) 
Systolic blood pressure, mmHg, mean (SD) 127 (15) 126 (14) 
Diastolic blood pressure, mmHg, mean (SD) 78 (10) 78 (9) 
Using antihypertensive medication, n (%) 202 (63) 233 (68) 
Total cholesterol, mmol/L, mean (SD) 4.4 (1.1) 4.4 (1.2) 
HDL-C, mmol/L, mean (SD) 1.2 (0.3) 1.2 (0.3) 
LDL-C, mmol/L, mean (SD) 2.4 (0.9) 2.4 (0.9) 
Triglycerides, mmol/L, mean (SD) 2.2 (1.5) 2.3 (4.3) 
HbA1c, %, mean (SD), mmol/mol, mean 8.2 (0.1), 66 8.1 (0.1), 65 
Fasting blood glucose, mmol/L, mean (SD) 9.4 (3.2) 9.3 (3.0) 
Years since diagnosis of diabetes, mean (SD) 11 (8) 11 (8) 
Number of antihyperglycemic medications, mean (SD) 2.1 (1.1) 2.1 (1.1) 

P > 0.05 for all treatment group comparisons.

HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol.

From baseline to day 90, HbA1c decreased 0.28% (95% CI −0.41, −0.15) in the BT-001 group and increased 0.11% (95% CI −0.02, 0.23) in the control group (treatment group difference −0.39 [95% CI −0.57, −0.20]; P < 0.0001) (Fig. 1, left panel). The magnitude of HbA1c reduction in the BT-001 group increased in parallel with exposure to the therapy, as indicated by the number of modules completed (Fig. 1, right panel) (P for trend <0.0001). Subjects completed an average of eight modules during the 90-day treatment period.

Figure 1

Treatment group difference in HbA1c change from baseline to day 90. HbA1c (mean, 95% CI) change from baseline to day 90 by randomized treatment group, adjusted for baseline HbA1c (left panel). HbA1c change from baseline to day 90, adjusted for baseline HbA1c, by tertile of CBT modules completed in the BT-001 app (right panel). Subjects assigned to the control app completed no modules (blue bars; n = 319); orange bars show HbA1c change among BT-001–allocated subjects by tertile of modules completed (n = 91, 99, and 101 for <5, 5–10, and >10 modules, respectively).

Figure 1

Treatment group difference in HbA1c change from baseline to day 90. HbA1c (mean, 95% CI) change from baseline to day 90 by randomized treatment group, adjusted for baseline HbA1c (left panel). HbA1c change from baseline to day 90, adjusted for baseline HbA1c, by tertile of CBT modules completed in the BT-001 app (right panel). Subjects assigned to the control app completed no modules (blue bars; n = 319); orange bars show HbA1c change among BT-001–allocated subjects by tertile of modules completed (n = 91, 99, and 101 for <5, 5–10, and >10 modules, respectively).

Close modal

Subjects in the BT-001 group recorded physical activity, plant-based meals, and fasting glucose values an average of 5, 5, and 6 days/week, respectively. Fasting glucose recorded in the app trended downward (Fig. 2) (P for trend <0.0001), consistent with central laboratory glucose (Table 2) and HbA1c results. During the 90-day treatment period, 48 subjects (7%) increased and 10 (1%) decreased the dose or number of their antihyperglycemic medications (increase: control, 28 and BT-001, 20; decrease: control, 7 and BT-001, 3; P = 0.29).

Figure 2

Fasting glucose levels change from baseline (mean, 95% CI). Fasting glucose levels reported in the BT-001 app are shown as the mean of each subject’s change from baseline by week. Baseline glucose was defined as the mean of the first three values recorded. Weekly glucose levels are the mean of values recorded by individual subjects each week. Subjects were not required to record any particular number of glucose values; on average, subjects recorded six values per week.

Figure 2

Fasting glucose levels change from baseline (mean, 95% CI). Fasting glucose levels reported in the BT-001 app are shown as the mean of each subject’s change from baseline by week. Baseline glucose was defined as the mean of the first three values recorded. Weekly glucose levels are the mean of values recorded by individual subjects each week. Subjects were not required to record any particular number of glucose values; on average, subjects recorded six values per week.

Close modal
Table 2

Weight, blood pressure, and plasma lipids: change from baseline to day 90 by treatment group

BT-001Control app
nChange from baselinenChange from baseline
Weight, kg 200 −1.5 (6.3) 230 −1.0 (5.5) 
Blood pressure, mmHg 200  228  
 Systolic  −3.0 (15.4)  −1.0 (14.2) 
 Diastolic  −1.5 (9.7)  −1.0 (9.8) 
Lipids, mmol/L 183  205  
 Total cholesterol  −0.13 (0.74)  −0.08 (0.87) 
 HDL-C  −0.01 (0.15)  0.02 (0.15) 
 LDL-C  −0.11 (0.58)  −0.06 (0.70) 
 Triglycerides  −0.15 (0.90)  −0.05 (1.31) 
Fasting glucose, mmol/La 217 −0.68 (3.49) 232 −0.07 (3.28) 
BT-001Control app
nChange from baselinenChange from baseline
Weight, kg 200 −1.5 (6.3) 230 −1.0 (5.5) 
Blood pressure, mmHg 200  228  
 Systolic  −3.0 (15.4)  −1.0 (14.2) 
 Diastolic  −1.5 (9.7)  −1.0 (9.8) 
Lipids, mmol/L 183  205  
 Total cholesterol  −0.13 (0.74)  −0.08 (0.87) 
 HDL-C  −0.01 (0.15)  0.02 (0.15) 
 LDL-C  −0.11 (0.58)  −0.06 (0.70) 
 Triglycerides  −0.15 (0.90)  −0.05 (1.31) 
Fasting glucose, mmol/La 217 −0.68 (3.49) 232 −0.07 (3.28) 

HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol.

a

P < 0.05.

After 90 days of exposure to their assigned app, weight, blood pressure, and plasma lipids were lower in the BT-001 group compared with the control group, although none of the treatment group differences were statistically significant (Table 2).

Adverse events were reported by 76 (22%) of subjects in the control and 68 (21%) in the BT-001 group with most mild or moderate in severity. The most commonly reported adverse events were coronavirus disease 2019 (13 subjects in each group) and headache (BT-001, 3; control, 9). Hypoglycemia was reported by two subjects in the BT-001 group and none in the control group. No patient reported treatment-emergent depression; anxiety was reported by four and six subjects in the BT-001 and control groups, respectively. Serious adverse events were reported by 10 (3%) and 4 (1%) of patients in the control and BT-001 groups, respectively. The only category with more than one serious adverse event was for injury, poisoning, and procedural complications. None of the adverse events were attributed to app use, and no adverse device effects were reported.

In adults with type 2 diabetes, CBT delivered via a digital therapeutic app lowered HbA1c by 0.39% (P < 0.0001) at 90 days compared with a control app, both on top of standard of care. A dose response was observed; subjects with greater exposure to the therapy demonstrated greater HbA1c reduction (P for trend <0.0001). In parallel, glucose levels fell with exposure to BT-001 (P for trend <0.0001). Fewer adverse events were reported with BT-001 compared with the control app, and no adverse device effects were reported.

A strength of the study is the demographic profile of the cohort, which is generally representative of adults with diabetes in the U.S. (11). In particular, the cohort included 28% Black, 4% Asian, and 14% Latino subjects as well as 55% women. Another strength is the real-world setting of the study conduct, facilitated by the decentralized, pragmatic study design, including the lack of restrictions on concurrent diabetes management and entirely remote study conduct. The latter enabled participation of rural residents as well as persons with mobility, transportation, or childcare constraints. Despite the entirely remote study conduct, rates of withdrawal and loss to follow-up were low. Although permitted, adjustments to antihyperglycemic medication were infrequent and similar between treatment groups during the 90-day treatment period. The impact of BT-001 on medication use may become clearer with longer follow-up.

A limitation of the study is the decentralized enrollment process, which may have selected a more motivated cohort than expected in clinical practice, potentially overestimating the efficacy of BT-001. The observation that subjects completing more modules had greater HbA1c reduction supports the importance of engagement, consistent with in-person CBT (12) and behavioral interventions in general (3). Other limitations include the exclusion of non-English speakers, as the app is currently only available in English, and the exclusion of current smokers, a group comprising 14% of adults with diabetes (1), who might be expected to benefit from improved glycemic control. This gap will be addressed in ongoing studies of BT-001 (NCT05266625, NCT05302050, and NCT05094401), which include current smokers. Allowing adjustment of antihyperglycemic medication during the 90-day treatment period incurred the risk of attenuating the treatment group difference in HbA1c, although medication changes were similar between groups. The fasting blood glucose levels reported within BT-001 reflect intrinsic limitations of glucometer measurements as well as potential reporting bias, although they are consistent with the central laboratory glucose results, suggesting that selective reporting was not an issue.

The observed HbA1c reduction of 0.4% at 90 days with BT-001 can be assessed alongside randomized trials of in-person CBT for glycemic control. In a meta-analysis of 23 trials with 2,619 patients, HbA1c change from baseline in patients with type 2 diabetes ranged from +0.2% to −1.09% with an overall difference of −0.3% (95% CI −0.4, −0.1; P < 0.01) (4). Another meta-analysis of 20 nonoverlapping trials with 2,900 patients found treatment group differences ranging from +0.74 to −1.66% with an overall difference of −0.97 (95% CI −1.37, −0.57; P < 0.0001) (5). The duration of CBT in these studies ranged from 2 weeks to 12 months and was delivered individually or in group settings. Baseline HbA1c levels varied, as did management of the control group. The magnitude of HbA1c reduction observed with BT-001 appears generally in line with studies of CBT delivered in person. The relationship between improved glycemic control and increased consumption of plant-based meals as well as minutes of physical activity with CBT will be evaluated, as this trial completes 180 days of follow-up and in conjunction with ongoing studies of BT-001.

Cardiovascular risk reduction is one goal of glycemic control; HbA1c reductions >0.3% have been associated with reductions in major cardiovascular events, with the magnitude of fall in HbA1c paralleling risk reduction (13). The least squares mean HbA1c 0.39% reduction observed with BT-001 was for the overall cohort. In the current study, patients who engaged more with the app, using modules completed as a surrogate for engagement, achieved greater HbA1c reduction. As data accrue from ongoing real-world clinical studies of the app, characteristics of patients likely to engage with the app will be sought with the potential aim of targeting those likely to achieve the greatest benefit. Cardiovascular risk reduction occurs over a lengthy period. Whether the digital therapeutic app has a durable effect on behavior is currently being assessed. If HbA1c reduction persists after completion of one or more cycles of treatment, this would contrast with medications that are only effective while being taken. The current study is assessing use of the app through 180 days, whereas ongoing real-world clinical studies will collect follow-up data for up to 2 years.

Modest, nonstatistically significant reductions in body weight, blood pressure, and lipids were observed with BT-001 compared with the control app. The magnitude of change in these biomarkers may reflect the limited duration of the intervention as well as the fact that average values for most exploratory markers were near normal at baseline, likely due to significant concurrent medication use, and thus, the study was not powered to detect significant changes in these biomarkers at 90 days. These findings do provide reassurance that BT-001 is not having an unwished-for impact.

In contrast to diabetes-focused smartphone apps that catalog foods’ nutritional content or organize individuals’ finger-stick glucose or physical activity data (14), digital therapeutic apps delivering CBT are therapeutic. The intervention is complementary to concurrent pharmacotherapy and has the potential advantage of leveraging the close relationship people have with their smartphones (15). Engagement is built and reinforced by manual input of biomarker values, reporting of plant-based meals (10) and minutes of physical activity, completion of modules and skill exercises, and by receiving individualized encouragement and feedback from the app.

In contrast to in-person CBT, the digital therapeutic is available at any time in the user’s pocket. Delivery of a cognitive behavioral intervention via smartphone app can provide a scalable option for improving glycemic control.

Clinical trial reg. no. NCT04886388, clinicaltrials.gov

This article contains supplementary material online at https://doi.org/10.2337/figshare.20814727.

M.A.B. and M.P.B. contributed equally as co-senior authors.

Funding. M.P.B. was supported by the American Heart Association Strategically Focused Research Network in Vascular Disease under award numbers 18SFRN3390085 (BWH-DH SFRN Center) and 18SFRN33960262 (BWH-DH Clinical Project). The apps used in this study were provided by Better Therapeutics, Inc.

Better Therapeutics, Inc., reviewed the manuscript before submission, but the co-authors maintained the authority of final approval of the manuscript to submit for publication. Statistical analyses were independently conducted by CPC Clinical Research.

Duality of Interest. Study funding was provided by Better Therapeutics, Inc. J.H. owns AstraZeneca stock. N.L.G. and M.A.B. are employees of Better Therapeutics, Inc., and own stock in the company. P.L. is a consultant for Better Therapeutics, Inc. D.D. is an employee of Clinical Trials of Texas, which receives research funding from Better Therapeutics, Inc. J.H., A.G., and M.P.B. are employees of CPC Clinical Research, a nonprofit academic research organization affiliated with the University of Colorado that receives research grant/consulting funding from Abbott Laboratories, Agios, Alexion Pharmaceuticals, Alnylam Pharmaceuticals, Amgen, Angionetics, Anthos Therapeutics, ARCA Biopharma, Array BioPharma, AstraZeneca, Atentiv Health, Audentes, Bayer, Better Therapeutics, Inc, Bristol-Myers Squibb, Cardiol Therapeutics, CellResearch Corp., Cook Medical, Cook Regentec, CSL Behring, Eidos Therapeutics, EP Trading Co., Esperion Therapeutics, Everly Health, Faraday Pharmaceuticals, Fortress Biotech, Inc, HDL Therapeutics, Inc., HeartFlow, Hummingbird Bioscience, Insmed, Janssen, Kowa Research Institute, Inc., Lexicon Pharmaceuticals, Inc., Merck, Medtronic, Moderna, Novate Medical, NovoNordisk, Pfizer, PhaseBio, PPD Development, Prothena Biosciences, Regeneron Pharmaceuticals, Regio Biosciences, Sanifit Therapeutics, Sanofi, Smith+Nephew, Stealth BioTherapeutics, and WraSer Pharmaceuticals. M.P.B. receives consulting fees from Audentes and owns stock in Medtronic and Pfizer. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. J.H. and M.P.B. wrote the first draft of the manuscript, and all of the other authors contributed to the final version. P.L. and A.G. conducted statistical analyses. J.H. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Data from this study were presented at the 2022 Vascular Scientific Sessions of the Society for Vascular Medicine, Denver, CO, 1 October 2022.

1.
Fang
M
,
Wang
D
,
Coresh
J
,
Selvin
E
.
Trends in diabetes treatment and control in U.S. adults, 1999–2018
.
N Engl J Med
2021
;
384
:
2219
2228
2.
Young-Hyman
D
,
de Groot
M
,
Hill-Briggs
F
,
Gonzalez
JS
,
Hood
K
,
Peyrot
M
.
Psychosocial care for people with diabetes: a position statement of the American Diabetes Association
.
Diabetes Care
2016
;
39
:
2126
2140
3.
Draznin
B
,
Aroda
VR
,
Bakris
G
, et al.;
American Diabetes Association Professional Practice Committee
;
American Diabetes Association Professional Practice Committee
.
5. Facilitating behavior change and well-being to improve health outcomes: Standards of Medical Care in Diabetes—2022
.
Diabetes Care
2022
;
45
(
Suppl. 1
):
S60
S82
4.
Yang
X
,
Li
Z
,
Sun
J
.
Effects of cognitive behavioral therapy-based intervention on improving glycaemic, psychological, and physiological outcomes in adult patients with diabetes mellitus: a meta-analysis of randomized controlled trials
.
Front Psychiatry
2020
;
11
:
711
5.
Chapman
A
,
Liu
S
,
Merkouris
S
, et al
.
Psychological interventions for the management of glycemic and psychological outcomes of type 2 diabetes mellitus in China: a systematic review and meta-analyses of randomized controlled trials
.
Front Public Health
2015
;
3
:
252
6.
Safren
SA
,
Gonzalez
JS
,
Wexler
DJ
, et al
.
A randomized controlled trial of cognitive behavioral therapy for adherence and depression (CBT-AD) in patients with uncontrolled type 2 diabetes
.
Diabetes Care
2014
;
37
:
625
633
7.
Uchendu
C
,
Blake
H
.
Effectiveness of cognitive-behavioural therapy on glycaemic control and psychological outcomes in adults with diabetes mellitus: a systematic review and meta-analysis of randomized controlled trials
.
Diabet Med
2017
;
34
:
328
339
8.
Berman
MA
,
Guthrie
NL
,
Edwards
KL
, et al
.
Change in glycemic control with use of a digital therapeutic in adults with type 2 diabetes: cohort study
.
JMIR Diabetes
2018
;
3
:
e4
9.
Guthrie
NL
,
Appelbaum
KJ
,
Berman
MA
.
A Software Application Delivering Behavioral Therapy Improved Glycemic Control in Adults With Type 2 Diabetes
.
Presented at the 2020 Endocrine Society Scientific Sessions, 28–31 March 2020, at the Moscone Convention Center, San Francisco, CA. Available from https://www.abstractsonline.com/pp8/#!/7897/presentation/6610
10.
Canonico
ME
,
Hsia
J
,
Guthrie
NL
, et al
.
Cognitive behavioral therapy delivered via digital mobile application for the treatment of type 2 diabetes: rationale, design, and baseline characteristics of a randomized, controlled trial
.
Clin Cardiol
2022
;
45
:
850
856
11.
Cheng
YJ
,
Kanaya
AM
,
Araneta
MRG
, et al
.
Prevalence of diabetes by race and ethnicity in the United States, 2011–2016
.
JAMA
2019
;
322
:
2389
2398
12.
Harrison
P
,
Hardy
GE
,
Barkham
M
.
The relationship between expected engagement and talking therapy outcome
.
Clin Psychol Psychother
2019
;
26
:
492
501
13.
Ambrosi
P
,
Daumas
A
,
Villani
P
,
Giorgi
R
.
Glycosylated hemoglobin as a surrogate for the prevention of cardiovascular events in cardiovascular outcome trials comparing new antidiabetic drugs to placebo
.
Cardiology
2020
;
145
:
370
374
14.
Doyle-Delgado
K
,
Chamberlain
JJ
.
Use of diabetes-related applications and digital health tools by people with diabetes and their health care providers
.
Clin Diabetes
2020
;
38
:
449
461
15.
Alderman
L
.
The Phones We Love Too Much
.
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.