A review of the included studies reveals that several behavioral theories were applied in designing digital therapeutics targeting individuals with prediabetes. These included SCT [23,31], theory of planned behavior [23,28], transtheoretical model (TTM) [22,29], delay discounting theory [36], Obesity-Related Behavioral Intervention Trials (ORBIT) model [9], cognitive behavioral theory (CBT) [26], self-determination theory (SDT) [10], and Information-Motivation-Behavioral Skills (IMB) self-care model [25]. The TTM was the most frequently used as a single theoretical framework. Notably, 1 study incorporated 5 distinct theoretical models [23].

Additionally, 11 studies used theory-informed BCTs in their intervention design, yet did not explicitly specify the underlying theoretical framework [11,24,27,30,32-35,37-39]. The frequency of theory application is summarized in Table 2.

To better understand how behavioral science informs digital interventions, it is essential to examine the theoretical models that underpin these interventions and the specific techniques used to operationalize them. While behavioral theories—such as TTM, SCT, and the theory of planned behavior—provide valuable conceptual frameworks for explaining how and why behavior change occurs, they often offer limited guidance on the concrete and replicable components required for intervention design and implementation.

To address this gap between theory and practice, Michie et al developed the BCTTv1 in 2013 [21]. This consensus-based taxonomy classifies 93 distinct BCTs into 16 groups, each with standardized definitions and coding. BCTs are the “active ingredients” of behavioral interventions, observable, replicable, and irreducible strategies intended to modify behavior by targeting specific psychological mechanisms. The BCTTv1 provides a systematic and transparent framework for researchers, practitioners, and intervention designers to describe, implement, and evaluate behavioral interventions, thereby improving reproducibility, comparability, and scientific rigor.

Analysis of the 21 included studies revealed that digital therapeutic interventions commonly incorporated multiple BCTs during their design.

Among the 93 BCTs identified in the BCTTv1 [21], 36 were applied across the included studies. The 5 most frequently used techniques were self-monitoring of behavior, instruction on how to perform the behavior, goal setting (behavior), information about health consequences, and social support (unspecified). Detailed findings are presented in Table 3.

Notably, all studies combined at least 5 BCTs when designing their interventions, highlighting the complexity and multifaceted nature of behavior change strategies in digital therapeutics.

The following section provides examples of how behavioral theories are applied in practice. It shows how specific theoretical constructs are translated into digital components and BCTs. These cases help demonstrate the real-world use of theory-based interventions for individuals with prediabetes.

The TTM was applied in 2 studies to guide diabetes prevention interventions. In Ramachandran et al [22], the intervention was structured around TTM’s 5 stages—precontemplation, contemplation, preparation, action, and maintenance [40]. Stage-specific SMS text messages were sent accordingly—for example, messages about the benefits of exercise for those in the precontemplation stage and actionable instructions such as “take a brisk 30-minute walk daily” for those in the action stage. Messages were adjusted every 6 months based on participants’ stages. The intervention group showed a 36% reduction in diabetes risk and improved dietary adherence, supporting the effectiveness of stage-matched interventions.

Nanditha et al [29] used a similar approach, delivering 75-80 messages per stage and incorporating decisional balance through questionnaires. However, after 2 years, no significant difference in diabetes incidence was observed. This may be due to a lower message frequency in this study (2-3 times per week, approximately 8-12 per month) compared with the higher delivery intensity in the Indian trial. Additionally, the Indian trial used HbA_1c rather than the oral glucose tolerance test, which may be less sensitive to short-term glycemic changes.

Together, these studies demonstrate the feasibility of TTM-based interventions but suggest that effectiveness may depend on factors such as intensity, evaluation methods, and cultural relevance.

Griauzde et al [10] applied SDT to develop a mobile health app that promotes healthy behaviors in individuals with prediabetes by enhancing intrinsic motivation. Participants were encouraged to log daily habits (eg, sleep, activity, and diet) and reflect on how these aligned with their core values. Personalized feedback and predictive tools supported this process, strengthening the link between behavior and internal values and fostering autonomous motivation.

Although quantitative data showed no significant change in motivation, likely due to already high baseline levels, qualitative interviews revealed increased self-reflection and greater awareness of how behavior relates to one’s environment. The results suggest that future interventions should focus on individuals with lower baseline motivation and include interpersonal support to strengthen SDT’s “relatedness” component, thereby better sustaining behavior change.

The IMB model, developed by the Fisher brothers [41], emphasizes that effective behavior change requires 3 elements: accurate information, adequate motivation, and practical behavioral skills. Kim et al [25] applied the IMB model, along with persuasive technology, to design a diabetes management intervention.

Information was delivered via a mobile platform offering Korean-language diabetes education and dietary guidance. Motivation was supported through personalized interviews by community health workers and real-time feedback. Behavioral skills were developed through the use of home monitoring devices and stepwise goal setting.

Persuasive techniques—such as tailored prompts and social support—were integrated throughout. For example, abnormal readings triggered human follow-up. After 6 months, 56% of participants achieved normal HbA_1c levels, demonstrating the IMB model’s effectiveness in chronic disease management.

Han et al [9] used the ORBIT model [42] to design a mobile health intervention. This study applied the full 4-phase ORBIT framework.

In phase I (define), the team identified poor diet quality and self-management barriers in people with prediabetes. Phase II (refine) tested the feasibility of core app functions. Phase III (optimize) involved iterative improvements, including the addition of real-time nutrition advice and social support. In phase IV, a randomized trial demonstrated a significant improvement in dietary quality, with a 6.2-point increase in Alternative Healthy Eating Index–2010 scores in the intervention group (P<.01).

This phased approach ensured scientific rigor and facilitated the successful translation of theory into practice.

CBT [43] served as the foundation of the BetaMe or Melon digital health program, evaluated by McLeod et al [26], which aimed to improve diabetes self-management through structured modules. CBT principles were applied in health coaching (via regular video or audio sessions to reframe negative thinking), goal tracking (using behavioral experiments to monitor diet and activity), and a peer support forum (to promote adaptive coping).

The intervention targeted unhelpful beliefs (eg, “change is impossible”) and reinforced behavior to boost self-efficacy. Although HbA_1c improvements were not statistically significant, the intervention group demonstrated gains in secondary outcomes, including self-management confidence. These results support the role of CBT in promoting behavior change and highlight the need for optimized dosage and personalization.

Block et al [23] developed the Alive-PD (a fully automated diabetes prevention behavioral intervention program from Turnaround Health, a division of NutritionQuest) by integrating multiple behavioral theories, showcasing an innovative and synergistic approach. Based on SCT [44], the program used digital goal-tracking tools and web-based social support to enhance self-efficacy and reshape environmental influences. The theory of planned behavior [45] guided a personalized goal-setting algorithm by translating attitudes and social norms into actionable steps.

A point-based reward system rooted in behavioral economics [46] provided immediate positive feedback to boost intrinsic motivation. Learning theory supported incremental goal setting with automated reminders as habit cues [47], while positive psychology [48] enhanced achievement and identity through visual progress tracking.

This multi-theory design surpassed the limits of single-model approaches, allowing cognitive, emotional, and environmental pathways to shape behavior through automated, personalized delivery. The results showed that 84% of participants remained highly engaged, and 35% achieved a weight loss of 5% or more, highlighting the benefits of integrated theory in promoting adherence and effective behavior change.

The intervention formats could be categorized into single-format and multi-format approaches. Three studies used a single-format intervention, primarily relying on mobile apps or communication software. In contrast, 18 studies used a combination of digital therapeutic formats. See Table 4 for details.

Current digital interventions are predominantly mobile-based, with smartphone apps serving as the primary platform. These apps commonly integrate features such as data synchronization and remote health coaching. Communication software and hardware devices serve as supporting components. Communication tools are used for reminders, education, and social support, while hardware devices facilitate data collection.

Common app-connected devices include glucometers, pedometers, wireless scales, and Bluetooth wristbands. As shown in Table 4, digital therapeutic interventions exhibit considerable diversity in format, with most studies adopting a combination of 2 or more delivery modes. The most frequently used configuration involves integrating mobile apps with health coach guidance.

The content of digital therapeutic interventions can be summarized into 6 core modules.

Goal setting reduces psychological barriers to behavior change by breaking long-term objectives into actionable short-term tasks. Various strategies were used across studies. Some used personalized micro-goals based on questionnaire analysis, with systems automatically generating weekly targets in physical activity, diet, and psychosocial areas according to individual profiles [23]. Others combined web-based skills training with motivational interviewing to provide participants with diabetes self-management education and individualized goal counseling, supporting healthy lifestyle behaviors [25]. Stage-matched SMS text messaging interventions grounded in the TTM delivered periodic messages offering prompts, advice, and reinforcement strategies related to goal setting, physical activity, diet, and overall lifestyle change [29,31]. Web- and app-based goal tracking was also common: for example, Lee et al [34] used a website to tailor weight loss or maintenance goals and daily caloric intake based on participants’ body weight status; Epstein et al [36] used the “MyFitnessPal” (MyFitnessPal, Inc) app, requiring participants to set calorie goals and adhere to a traffic light-based dietary system (<2 servings of red-light foods and >5 servings of green-light foods daily), with exercise goals of at least 150 minutes of moderate or 75 minutes of vigorous activity per week. Lim et al [37] implemented a step-counting app that automatically increased daily step targets, progressing from 3000 steps per day in week 1 to 10,000 steps by week 3.

Self-monitoring enhances individuals’ awareness of their behaviors through the collection and analysis of multidimensional health data. Common strategies included app-based tracking of sleep, diet, and physical activity [10,27,28,30,35-39], as well as the use of home-based monitoring devices to record weight, blood glucose, blood pressure, dietary intake, and exercise [11,25,31,34]. Several interventions provided personalized recommendations based on data from wearable devices [37-39]. For example, Lim et al [37] used an app with automated response features that assessed dietary choices and generated culturally tailored food substitution lists. Two studies [38,39] used AI-driven algorithms to deliver personalized suggestions on diet and physical activity, including reducing calorie intake, increasing dietary fiber and protein, and promoting physical activity. Notably, 1 study [39] incorporated principles of traditional Chinese medicine, using body constitution analysis and meridian energy profiling to tailor lifestyle recommendations further.

Grounded in behavioral economics principles, real-time feedback and reward systems were implemented to enhance motivation and reinforce behavior. Strategies included point-based rewards and team competitions, where participants earned points by completing weekly goals that could be redeemed for incentives or used to support team challenges [23]. Automated feedback mechanisms were integrated into behavior change programs, such as the “Noom program” (Noom, Inc), which delivered daily tasks (eg, food selection) and evaluated responses in real-time to provide immediate feedback [27]. Some studies also used in-app reward points (eg, “virtual gold”), allowing participants to earn rewards by completing questionnaires, quizzes, or personal goals, which could then be exchanged for tangible incentives [39]. Feedback was further delivered through SMS text messaging and app-based notifications [25,35,37], as well as real-time data syncing from wearable devices [11,28,31]. For example, Kim et al [25] provided participants with Bluetooth-enabled home monitoring devices (eg, glucometers, blood pressure monitors, pedometers, and scales), which automatically transmitted data to a smartphone app. Similarly, in studies by Katula et al [11] and Batten et al [35], wearable trackers and smart scales synced automatically with the intervention platform, facilitating continuous behavioral feedback.

Social networks—both digital and offline—were leveraged to enhance adherence to interventions and behavioral engagement. Strategies included web-based team-based competitions to foster peer accountability [23], as well as mobile app-based coach-participant interactions through mobile platforms. For instance, Toro-Ramos et al [27] enabled participants to access coach and group messages via mobile devices and communicate directly with coaches as needed. Similarly, in studies by Katula et al [11] and Batten et al [35], participants engaged in web-based group discussions and in-app community interactions to ask questions, responded to peer feedback, and shared personal progress. Additional approaches included follow-up phone calls to maintain participant engagement [11,31,32] and the provision of gym memberships to promote physical activity through environmental and social reinforcement [33]. In that study, Al-Hamdan et al [33] offered free gym access to all participants in the intervention group. Remote dietary counseling was also provided in several studies [9,30,37], where participants recorded data on their diet, glucose levels, sleep patterns, weight, and physical activity using an app, and nutritionists delivered tailored feedback. In the study by Lim et al [37], nutritionists reviewed submitted data regularly. They used motivational interviewing techniques to provide real-time feedback and help participants overcome behavioral barriers to sustainable change.

Multichannel reminder systems were used to counteract memory decay and promote adherence. Scheduled text message delivery was one of the most common strategies, with messages sent at fixed times according to participant preferences [10,22-26,28,29,31]. For example, Ramachandran et al [22] and Nanditha et al [29] structured messages around the 5 stages of the TTM (precontemplation, contemplation, preparation, action, and maintenance), delivering 60-80 stage-matched messages in cycles. These messages were tailored based on participants’ physical activity and dietary data. In the study by Nanditha et al [29], bilingual texts were used to accommodate diverse racial and cultural backgrounds. App-based motivational messaging was also used to support behavior change. In a study by Staite et al [28], motivational interviewing principles were integrated into a smartphone app that sent automated messages to facilitate health intention formation, encourage self-monitoring, and enhance social support. Additionally, graphical progress dashboards were used to provide visually enriched feedback and reinforce engagement throughout the intervention period [32].

A structured knowledge delivery chain—from basic awareness to applied practice—was embedded into many interventions. Common formats include general health information messaging covering healthy lifestyles, the benefits of physical activity and nutrition, and diabetes prevention strategies [22-25,28-30,33,35]. One study incorporated the Diabetes Prevention Program–certified educational content [27]. Video-based courses were also used in 2 studies [32,37], both of which delivered 3-month curricula. In the study by Muralidharan et al [32], the content was presented in a reality television format, accompanied by weekly feedback calls from health coaches. In contrast, Lim et al [37] provided videos on weight management, diabetes prevention, healthy meal planning, and food selection. Web-based courses and toolkits were provided in another study [34], featuring an 18-session, 24-week program that guided participants through lifestyle modification, diet, physical activity, and self-monitoring, with supplementary materials and offline coaching. Additionally, Katula et al [11] implemented a year-long stepped curriculum, including 16 weeks of intensive weight-loss content followed by 36 weeks of maintenance-focused education.