Type 2 diabetes mellitus (T2DM) is complex, costly, and often preventable; yet, millions still live with this destructive disease, with many more Americans diagnosed annually. Ideally, incidence rates would be on a downward trend; however, this is not the case, with adult incidence rates remaining relatively stable over the past ~20 years, albeit with some variation following 2008 [1]. Thus, with more individuals diagnosed, coupled with this being a lifelong disease upon diagnosis, prevalence rates are on the rise, with an increase from 10% (2001‐2004) to 13% (2017‐2020) [1]. Therefore, this remains a critical national public health issue with additional evidence of differential patterns impacting individuals across both individual-level and place-based characteristics [2].

The United States is home to many residents living outside major metropolitan areas, who have historically faced inequities in accessing critical health care resources [3,4]. Recent estimates suggest that age-adjusted prevalence of diagnosed diabetes was higher among those residing in nonmetropolitan areas, with rates of 8.9% and 7.9% for both men and women, respectively, among large central metro areas, compared to rates of 10.5% and 8.6% in nonmetropolitan areas [1]. Furthermore, in terms of individual-level characteristics, inequities exist across characteristics such as age (with higher rates with increasing age) and sex (with higher rates among males) [1]. Thus, this is a complex and interrelated issue spanning multiple levels. Making matters worse, individuals with diabetes often face potentially preventable medical encounters due to, in part, poor management of diabetes, including higher risks of heart disease, stroke, blindness, kidney disease, amputations, and other items [5]. For many, these complications could potentially be prevented with proper maintenance of diabetes and related comorbidities, as applicable.

Effective self-management involves taking the necessary steps to monitor and maintain optimal glycemic levels and prevent adverse effects [6-8]. Additional components of effective self-management include medication adherence and lifestyle modifications (eg, diet or nutrition modifications, physical activity, and refraining from, or quitting, smoking) [9-11]. Thus, interventions aimed at reducing potentially preventable complications and improving the quality of life of individuals diagnosed with diabetes should consider both individual-level characteristics in combination with place-based measures, consistent with a strong theoretical foundation [12,13].

In terms of interventions, they can take several forms and platforms, including but not limited to online access to critical information; yet, there are some core elements to consider. For example, experts recommend all patients diagnosed with diabetes receive diabetes self-management education and support (DSMES) to provide them with the knowledge, resources, and skills essential to support their self-care journey and manage their condition [8]. Research indicates patients who receive DSMES are more likely to practice recommended self-care behaviors and clinical preventive care practices, such as routine clinical examinations and vaccines [9]. Additionally, positive changes in knowledge [14,15], confidence [8,16], and reduced distress [8,17] have been consistently reported and associated with sociodemographic and health factors, such as age, sex, education, race or ethnicity, rurality, financial stability, insurance coverage, and general health status [18-20].

With the rise in technological development and aftereffects of the COVID-19 pandemic, there is now a greater variety and availability of DSMES in the virtual and digital space. Interventions come in various formats from in-person to virtual (including both synchronous and asynchronous) to hybrid and contain various components or mechanisms of support, including, but not limited to, behavioral tracking and monitoring, peer support, coaching, knowledge-based resources, and skill building [21-23]. As no program is a one-size-fits-all, there are pros and cons to consider for both in-person and virtual DSMES programs.

Consistent with the diffusion of innovations theory [24], many people may be skeptical about the effectiveness of new digital health interventions (which include telehealth, telemedicine, mobile health, and eHealth [25]) and wary to engage in such practices. However, a systematic literature review and meta-analysis revealed that digital health interventions for people with T2DM were associated with significant improvements in hemoglobin A_1c levels [26]. Additionally, digital interventions may fill some gaps of in-person interventions, including accessibility, real-time feedback, and transportation barriers [27,28]. However, it is also important to note that users’ health literacy and/or digital literacy may impact their usage, level of engagement, and outcomes of using digital health interventions [29,30].

As new programs are developed, researchers have analyzed the acceptability, usability, and impacts of digital DSMES [7]. However, most studies analyzing the impacts and outcomes of DSMES compare the use of a digital program to a control group [7,31-35]. Additionally, studies tend to recruit from one community, limiting the generalizability of findings and the inability to observe the effectiveness of different digital health DSMES programs on behavioral outcomes across different geographic regions.

This study, part of the Texas A&M University Rural Health Moonshot Initiative [6], attempted to address these research gaps by comparing the impact of 2 digital DSMES programs delivered separately and in combination to adults with T2DM across many settings in Texas. The objectives were to examine (1) the sociodemographic and general health characteristics of adults living with diabetes who agreed to participate in a randomized clinical trial and whether any of these baseline characteristics were significantly associated with group assignment which served as a test of our randomization protocol; (2) mean changes and effect size differences in knowledge, confidence, distress, and self-care activities by intervention and rurality status; and (3) whether knowledge, confidence, and distress, controlling for sociodemographic factors and general health status predicted changes in self-care activities over time.

As described in detail in an earlier study [6], we recruited 188 adults aged 25 years and older living in Texas from November 2020 to March 2022 who reported having T2DM with baseline hemoglobin A_1c (HbA_1c) levels of 7.5 or higher, indicating poorly managed diabetes and potential benefit from an intervention [36,37] targeted toward improving one’s management of diabetes. Recruitment efforts were geared toward participants from both rural and urban communities in Texas. Interested participants were considered eligible if they had access to a smartphone with internet connectivity and were able to read and speak English.

This trial was generated based on the research team’s experiences of implementing diabetes management programming in rural and underserved communities [6,38,39]. Feedback was obtained from program implementers and participants in relation to program content, access, and attendance. Using this information obtained in the years preceding the trial, we updated the intervention content and delivery format accordingly. While originally attempting to test the in-person version of the intervention, the COVID-19 pandemic required us to pivot to virtual delivery. Representatives from community-based organizations, health care organizations, program facilitators, and past participants were engaged in the material development and testing phase of the digital translation. Furthermore, these representatives were engaged when attempting to identify the best recruitment and retention strategies, given high levels of attrition in past initiatives. Community and clinical partners were instrumental in participant recruitment efforts. While participants were not directly involved in the selection of measures, previously used subjective measures were included based on their success in related interventions and trials [6,38]. Findings were shared with the funder and community partners in accordance with their preferences. These took the form of informal meetings and discussions with stakeholders, flyers, virtual correspondence (eg, websites and social media), conference presentations, and publications.

Knowing that multiple strategies would be needed to reach community-dwelling participants throughout the state, a variety of recruitment strategies were used, including radio advertisements, radio segments, flyers, bulletin boards, social media campaigns, directed letter campaigns from in-network health care providers, newspapers, email listserves, in-person recruitment, and the study’s website. We also used multiple retention strategies: participant incentives, distribution of study information, and frequent communication via text messaging, emails, and phone calls to limit attrition.

This study was reviewed and approved by the Texas A&M University Institutional Review Board (IRB 2019-0804D) and registered at ClinicalTrials.gov (ID NCT06370494). All participants provided written consent to participate in this randomized controlled trial (RCT). Participation was voluntary, and participants could withdraw from the study at any time without penalty. Deidentified data on major study outcomes were placed in the Texas A&M data repository upon completion of data analyses and report writing. Only the study team could access study data. Depending on their participation in the study, each participant could receive upward of US $375 for completing data throughout the study period.

Participants were randomized into 1 of 3 intervention arms, as described in the following paragraph. Each intervention aligned with American Diabetes Association national standards for diabetes self-management and best practices for behavioral change [40]. DSMES is a continuous process designed to help individuals with diabetes gain the knowledge, skills, and confidence needed to manage their condition effectively [41]. Each DSMES may vary in components, objectives, and mechanisms, but there is consensus that goals of DSMES are to promote informed decision-making, encourage healthy self-care habits, enhance problem-solving abilities, foster collaboration with health care providers, and ultimately improve health outcomes, overall well-being, and quality of life [41,42]. Furthermore, the American Diabetes Association and International Diabetes Federation established guidelines for DSMES programs to not only be reimbursable but also recognized and accredited as evidence-based and effective programs.

Differing in modality, the 3 intervention arms (further described in Table 1) included (1) Virtual Making Moves with Diabetes (vMMWD) Group, (2) Technology-Based Education and Support (TBES) Group, and (3) a combined approach of vMMWD and TBES. Participants from all groups received an HbA_1c monitoring kit, a glucose meter, and supplies to test their blood sugar at home. On the basis of the prior literature documenting positive effects of DSMES [8,9,16,17,21-23,43-46], we hypothesized that all intervention strategies would have a beneficial effect on diabetes-related study outcomes, but that effects would be stronger for vMMWD over TBES, given the greater personalization, and that the combined intervention strategy (representing the strengths of each intervention modality) would have the strongest effects. We also hypothesized that participants from traditionally underserved and resource-limited settings (eg, minority identity, less education, lower financial security, no health insurance, and rural residence) who typically have less access to diabetes self-care programs might have the greatest improvements.

We selected standardized measures that have been previously used in diabetes education studies, as described below.

This study examined the comparative effectiveness of 3 digital DSMES interventions among adults with T2DM, with a particular focus on differences by intervention modality and sociodemographic factors. Findings demonstrated that (1) randomization was effective in creating 3 distinct intervention groups (objective 1); (2) all intervention modalities significantly improved diabetes-related confidence, distress, and self-care behaviors over time, with sustained effects (objective 2); and (3) confidence and distress, controlling for sociodemographic factors and general health status, predicted significant changes in self-care activities over time (objective 3). Importantly, no statistically significant differences were observed between intervention modalities, and improvements were largely consistent across rural and urban participants, a major covariate.

To our knowledge, this study is among the first to rigorously evaluate and directly compare different digital diabetes self-management programs using an RCT design in this way. All 3 interventions—vMMWD, TBES, and the combined approach—led to significant improvements in diabetes-related confidence, distress, and self-care behaviors, with effects through 6 months and supported by moderate to strong effect sizes. These findings generally reflect positive findings reported in other in-person and digital health diabetes interventions [8,9,16,17,21-23,43-46]. However, the TBES intervention did not result in significant improvements in diabetes-related knowledge from baseline to follow-up, contrary to other research showing that coaching-based apps can significantly increase self-management diabetes knowledge.

While all interventions demonstrated overall positive outcomes, no statistically significant superiority was observed among the intervention modalities, consistent with HbA_1c outcome findings reported elsewhere [6,55]. These results offer partial support for the study’s hypotheses, confirming the effectiveness of each intervention but not the expected added benefit of the combined approach. This suggests the observed improvements were likely due to shared attributes (ie, DSMES components) rather than the intervention modality. Thus, whichever modality is easily accessible to participants is likely to benefit their self-care behaviors.

The study findings also partially supported our expectations that those from underserved or low-resourced settings with greater need would experience stronger positive effects, providing evidence for the heterogeneous treatment effect [56]. However, with regard to rurality, participants in both rural and urban settings experienced significant improvements across outcomes, with no statistically significant differences observed between groups. While rural participants showed comparable—if not slightly larger—gains in some outcomes (eg, knowledge and self-care), the study was not sufficiently powered to detect interaction effects. These findings suggest that digital DSMES may help mitigate geographic disparities in access to diabetes education.

The study findings can be interpreted through the lens of access, cost, and quality in diabetes self-management. From a quality perspective, all 3 interventions demonstrated encouraging effectiveness. Our prior research [6] demonstrated improvements in HbA_1c values, whereas this follow-up research revealed positive impacts on psychosocial factors linked to clinical outcomes [57,58]. Specifically, our DSMES-based interventions significantly improved psychological outcomes—such as confidence and distress—as well as actual self-care behaviors. These results suggest that the interventions enhanced participants’ understanding and confidence in managing their diabetes, reduced diabetes-related distress, and facilitated real behavioral changes. As emphasized in classic models of chronic disease management, sustained behavioral improvements often arise from underlying psychological shifts—particularly in self-efficacy and emotional readiness [59]—which are foundational in effective chronic illness care [60].

These concepts are further supported by the Chronic Care Model [61], which underscores the importance of informed, activated patients and prepared, proactive health care teams. Prior research supports the notion that self-management programs are most effective when they improve knowledge and confidence while alleviating distress [8,41,62]. Importantly, this study also offers evidence of sustainability, with program effects maintained from baseline through 3 and 6 months. This is especially encouraging, given previously mixed findings in the literature regarding the long-term effectiveness of digital diabetes self-management programs [63,64].

The longitudinal regression findings can also be interpreted as supporting the social cognitive theory. The positive association between confidence and self-care behaviors is consistent with evidence that self-efficacy is a key driver of sustained behavior change [59,60], while the negative effect of diabetes-related distress reflects well-documented emotional barriers that undermine adherence and self-management capacity [50,62]. Together, these results reinforce that behavior change is not driven by knowledge alone, but by psychosocial mechanisms that enable individuals to translate knowledge into action. By demonstrating that confidence and distress—but not knowledge—predict longitudinal improvements in self-care, this study strengthens the theoretical understanding of how DSMES interventions influence behavior over time.

From an access perspective, this study highlights the feasibility and value of digital diabetes self-management programs. Accelerated by advancements in digital technology and the societal shifts prompted by the COVID-19 pandemic, there is increasing acceptance of accessing health care services via digital platforms [65]. Given persistent barriers to in-person care—particularly in rural areas where geographic isolation, limited transportation options, and inadequate health care infrastructure prevail—these findings support the use of digital programs as a viable alternative to traditional in-person services [66,67].

A prior meta-analysis further supports the potential of digital diabetes self-management education to improve clinical and behavioral outcomes while also highlighting significant variability in program implementation, user engagement, and access across populations [63]. Policymakers, public health professionals, and individuals with diabetes should consider DSMES as a practical and scalable solution, provided that participants have access to the necessary devices and internet connectivity. Notably, leveraging community resources—such as public libraries with Wi-Fi and private telehealth spaces—could bridge the digital divide and extend digital health access in underserved communities [68]. From a cost perspective, self-management remains a critical pillar of diabetes care, with substantial evidence supporting both the effectiveness and cost-efficiency of DSMES [8,69]. Although this study did not directly assess the cost of digital DSMES, its findings contribute to the growing literature supporting the economic and practical value of digital delivery methods. Digital interventions can reduce or eliminate common logistical barriers—such as travel time, facility utilization, and provider burden—making them particularly advantageous for underserved and rural populations. Prior studies have shown that telehealth-based diabetes interventions significantly enhance access to care and treatment adherence in rural settings by minimizing travel burden and expanding provider reach [70-72].

Future research steps include examining the mechanisms through which digital interventions improve diabetes outcomes—particularly the interrelationships between knowledge, confidence, emotional distress, and self-care behaviors. A deeper understanding of these pathways can inform the design of more targeted, effective interventions. Ultimately, digital DSMES programs hold promise not only for improving health outcomes (quality) but also for increasing accessibility to care and reducing the economic burden associated with diabetes management—advancing the triple aim of better care, improved population health, and lower costs [73].

This study has several limitations typical in single-site randomized clinical trial studies. First, the sample was limited to one state (Texas), which may affect the generalizability of findings to other geographic regions. However, this state includes a large and geographically diverse population (including an international border, many rural counties, and several major metropolitan areas). Furthermore, Texas is located within the South census region with a history of high rates of diabetes [2,4], with other work even pointing to higher diabetes mortality rates relative to other census regions [74]. Thus, Texas is a unique area, yet also representative of various sociodemographic groups.

Second, we were not able to recruit as many participants from rural areas as we had projected, constraining our ability to conduct formal subgroup analyses. While preimplementation sample size estimates indicated adequate power for testing main effects of DSMES programs with moderate or larger effect sizes, the relatively small sample size within each intervention arm may have limited the ability to detect differences between rural and urban subgroups. Hence, we were not able to distinguish relative intervention benefits for rural versus urban populations.

Third, all participants were required to have internet access and a smartphone, potentially introducing selection bias and limiting generalizability to those with lower digital literacy or access. However, according to a Pew Research Survey, the percentage of Americans who “rely on their smartphones for going online” (percentage of adults with a smartphone, but no broadband at home) has seen increases across all income categories from 2013 to 2021 [75]. In fact, the largest gains were for those with household incomes of less than US $30,000 (the lowest income level in the study) from 12% to 27% from 2013 to 2021 [75]. Thus, this is actually a strength, in that this method (eg, use of a smartphone to access diabetes-related training) is likely to gain more momentum over time with more Americans becoming more familiar with their smartphones as a major means of accessing information, etc. In fact, more than 90% of study participants reported being comfortable with smartphones, minimizing concerns about digital literacy in this group of participants. Yet, as indicated earlier, our study participants may not be generalizable to the broader population of people living with diabetes who may have lower digital literacy and underscore equity concerns needing policy attention [76,77].

Fourth, all outcome measures were based on self-report, which may be subject to recall or social desirability bias. However, this is not uncommon in survey research and is applicable to even large nationally representative surveys such as the BRFSS [78].

Fifth, the study did not collect data on participants’ actual engagement with the digital interventions, limiting insights into usage patterns and dose-response relationships. However, the introduction of random assignment was a major strength that likely led to greater internal validity [54].

Finally, although our study had a 6-month follow-up period revealing short-term maintenance effects, a longer follow-up period, while not feasible during our study period, would have addressed the extent to which benefits from different digital interventions could be maintained over time.

This RCT demonstrates the effectiveness of 3 digital DSMES interventions—vMMWD, TBES, and a combined modality—in improving diabetes-related confidence, reducing distress, and enhancing self-care behaviors among adults with T2DM. While most intervention groups showed improvements through 6 months on all diabetes-related outcomes, no single modality proved statistically superior. One notable exception emerged; knowledge gains were not significant in the TBES-only group. These findings underscore the value of digital DSMES as a flexible and scalable approach to support behavior change. Importantly, the study highlights how digital interventions can promote health equity by expanding access to self-care education for individuals in underserved and rural settings. Although the study did not directly assess costs, the digital format’s ability to reduce logistical barriers, such as travel, provider burden, and infrastructure demands, suggests strong potential for cost-efficiency.

These findings have important implications for public health practice and health care delivery. Digital DSMES programs represent a scalable and adaptable approach to supporting individuals with diabetes, particularly in underserved and resource-limited settings. Future research should explore the mechanisms linking knowledge, confidence, and distress to behavioral outcomes to optimize digital delivery strategies. Furthermore, given recent advancements in artificial intelligence, we also expect additional improvements in the user interface (including how individuals access digital data) on smartphones with potential opportunities for subject-specific (where applicable) tailoring of knowledge and other outcomes to improve the user experience in the future. For example, digital DSMES programs and platforms could incorporate adaptive messaging that responds to users’ ongoing data, such as increasingly motivational prompts and action steps when confidence scores decline [79]. Future research should focus on identifying the specific components and mechanisms that drive intervention effectiveness, including the roles of engagement, personalization, and behavioral support. Taken together, this work supports the promise of digital DSMES in advancing the goals of improved care quality, enhanced access, and reduced costs—aligned with the broader aim of transforming chronic disease management.