BackgroundType 2 diabetes (T2DM) is one of the leading chronic diseases worldwide, requiring effective long-term self-management (SM) to improve health outcomes and reduce complications. Although structured diabetes pathways exist in Belgium, national reports indicate that adherence to the recommended diabetes follow-up and routine monitoring remains suboptimal. Digital health technologies have shown promise in supporting SM; however, little is known about how stakeholders within the Flemish health care system perceive opportunities to improve SM and implement digital health interventions. ObjectiveThis study aimed to explore stakeholders’ perspectives on improving diabetes SM and the potential role of digital tools in supporting people with T2DM in the Flemish health care system. MethodsA qualitative study was conducted using 2 online semistructured focus group discussions with 19 stakeholders, including health care professionals (HCPs) from primary, secondary, and tertiary care settings; researchers; digital health experts; and an individual living with T2DM. Participants were recruited using maximum variation purposive sampling followed by snowball sampling. The discussion guide was informed by the Consolidated Framework for Implementation Research and current evidence on diabetes SM and digital health. Discussions were audio-recorded, transcribed verbatim, translated from Dutch into English, analyzed using Braun and Clarke’s 6-phase reflexive thematic analysis, and supported by NVivo (QSR International). ResultsFive key themes were identified: (1) Empowerment through education, (2) Tailoring education to individual needs, (3) Multidisciplinary and cross-sectoral collaboration, (4) Opportunities for digital technology in self-management, and (5) Regional and national systemic limitations. Participants consistently emphasized that effective diabetes SM may extend beyond patients’ individual behaviors and depends on tailored education, coordinated multidisciplinary care across the health care sectors, supportive organizational structures, and supportive health policies. Digital health technologies were perceived as valuable tools to facilitate patient education, communication, information sharing, and continuity of care. However, participants also highlighted that successful implementation depends on addressing broader organizational and system-level barriers, including fragmented digital infrastructures, limited interoperability, and regulatory constraints within the Belgian health care system. ConclusionsStakeholders viewed improving diabetes SM as a shared responsibility between patients, HCPs, and the health care system. While digital health technologies may strengthen SM support, their value depends on how well they complement existing care processes and address contextual implementation challenges. These findings provide context-specific insights that may inform the development and implementation of digital interventions for T2DM in Belgium and similar health care settings. Further qualitative research involving a larger and more diverse sample of people living with T2DM is warranted to better understand patient experiences and support needs.
BackgroundThe Diabetes Remission Clinical Trial (DiRECT) demonstrated that remission of type 2 diabetes can be achieved through substantial weight loss using a total diet replacement (TDR) intervention combined with structured support. Whether this approach can be delivered exclusively via eHealth has not been well studied. ObjectiveThis pilot study aimed to evaluate whether a TDR intervention delivered solely via eHealth could induce clinically meaningful weight loss and achieve type 2 diabetes remission. MethodsThis was a single-arm, nonblinded, single-center pilot study conducted at the clinical research facilities of Umeå University Hospital. Individuals with type 2 diabetes diagnosed within the past 6 years were included. Participants followed a TDR (~850 kcal/d) for 12 weeks, and then gradually reintroduced food, with a total study duration of 6 months. Diabetes medications were discontinued at the start of the study. Daily body weight and fasting glucose were measured at home and transmitted to the study team via connected devices. All clinical support was delivered through video calls and chat. The 2 prespecified coprimary outcomes were a reduction in weight of 15 kg or more and remission of diabetes, defined as glycated hemoglobin below 6.5% without glucose-lowering medication, from baseline to 6 months. A 75 g oral glucose tolerance test and indirect calorimetry were performed at baseline and at 6 months to assess resting energy expenditure (REE) and respiratory quotient (RQ). ResultsTen participants (mean age 58.5, SD 7.4 y; mean BMI 35.3, SD 5.9 kg/m²; n=9 men and n=1 woman) were enrolled. After 6 months, 6 of 10 participants achieved the prespecified coprimary outcome of ≥15 kg weight loss, and all 10 participants achieved the prespecified coprimary outcome of diabetes remission. Mean weight loss was 16.9 (SD 5.4; P=.005) kg, and glycated hemoglobin decreased by 1.0% (SD 0.8%; P=.008). Daily home measurements showed a marked reduction in fasting glucose within the first week of the intervention. Median REE decreased by 19% (IQR 17%-22%), while median fasting RQ increased by 15% (IQR 5%-27%) and median post–glucose-load RQ increased by 12% (IQR 9%-18%) at 6 months. ConclusionsIn this pilot study, an eHealth-delivered TDR intervention was associated with substantial weight loss and diabetes remission at 6 months without the use of glucose-lowering medication. These findings support further evaluation of fully eHealth-delivered dietary interventions in larger and longer-term randomized controlled trials. Trial RegistrationClinicalTrials.gov NCT04805996; https://clinicaltrials.gov/study/NCT04805996
BackgroundEcological momentary assessment (EMA) is a tool that captures emotional states, experiences, and behaviors in real or near-real time. Concurrent EMA and continuous glucose monitoring (CGM) data can provide an in-depth understanding of the impacts of psychosocial factors on momentary glucose levels. However, study methodology varies widely for EMA and CGM data collection and analysis. ObjectiveThis scoping review aims to summarize the objectives, methodologies, and outcomes of studies analyzing concurrent biopsychosocial EMA and CGM data in diabetes. MethodsThis study was conducted following the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines. Original research studies evaluating participants with diabetes, with concurrent collection and analysis of EMA and CGM data, were included in this review. A total of 782 studies were identified from PubMed, Embase, and EBSCOhost from May 2009 to December 2025. From these, duplicates, non-original research articles, and articles not measuring concurrent EMA and CGM data were excluded, resulting in a total of 19 studies included in this review. Methodological data were abstracted and summarized, including study characteristics, EMA protocols and outcomes, CGM outcomes, and integrated EMA and CGM study objectives. Methodological gaps across the included studies were identified and summarized. ResultsOf the 19 studies, 16 primarily recruited adult populations, and the majority of these studies (n=14, 74%) included participants with type 1 diabetes. The median EMA delivery duration was 14 days (range 3‐18 days), with a median of 5.5 prompts delivered per day (range 1‐28). EMA outcomes included a variety of biopsychosocial factors (emotion, self-care behaviors, interpersonal interactions, symptoms, and cognition). Seventy-four percent analyzed blinded CGM data over a median of 14 (range 3‐18) days. Forty-two percent used nonstandardized CGM glucose outcomes (postprandial glucose, eating patterns, and hypoglycemic events). Integrated EMA and CGM data analysis answered a broad array of study objectives, including the impact of psychosocial factors on momentary glucose metrics; the influence of momentary glucose on emotional states, mood, personal behaviors, sleep, and cognition, or vice versa; and study protocol or mobile app optimization, among others. Methodological gaps included lack of standardization of EMA and CGM outcomes, duration of data capture, reporting of statistical analyses, and study population. ConclusionsThis review provides a detailed summary of the methodology and outcomes of original research articles using combined EMA and blood glucose data measured via CGM. These combined methods allow for an opportunity to elucidate relationships between psychosocial factors and momentary glucose, an understanding of which is imperative for the development of future psychological and behavioral interventions in diabetes. However, standardization of protocols and expansion of participant populations for future EMA and CGM data collection and analysis are needed to ensure data accuracy, reproducibility, and generalizability.
Abstract Nutrient estimation from meal images by multimodal large language models can support diabetes self-management, but its robustness is limited by variability driven by both sensitivity to visual presentation and inherent model instability. Accounting for and mitigating this variability is essential for robust AI-assisted dietary assessment.
