This was an observational study of DDF app users from January 1, 2021, to December 8, 2023. Participants in this study were not recruited but were existing users of the commercially available DDF app. Data were collected from users who voluntarily engaged with the app as part of their normal use, and all data analyzed were deidentified entries made during the first 30 days of app use.

DDF is a freely accessible web-based app with a subscription option [13]. Users could pay a small fee for a structured 30-day program that includes daily lessons, weekly live question and answer sessions, and community support. Users are encouraged to purchase a glucometer for glucose tracking, with the Contour Next (Ascensia Diabetes Care) being the recommended model. DDF has an interactive interface designed to help users follow GGE and provides nutritional guidance based on recorded preprandial (premeal) and postprandial (postmeal) self-monitored glucose levels. Perceived preprandial hunger ratings were recorded in the app using a 5-point scale, with 1 indicating “not hungry” (Figure 1).

During the first 3 days of app use, participants are instructed to eat normally but to test their glucose levels before every eating occasion. Reported preprandial glucose levels during this time are averaged to establish a baseline glucose threshold. After this initial phase, users are encouraged to observe their hunger sensations and test their preprandial glucose levels only when hungry, and to delay eating until their glucose levels fall below their personalized threshold (ie, glucose-guided eating). After 5 days of glucose-guided eating, the threshold is adjusted daily by averaging reported preprandial glucose over the previous 7 days, with a maximum decrease of 0.5% per day. During times marked by higher-than-usual glucose levels due to conditions other than recent eating (eg, high stress, fatigue, pain, and intense exercise) users can opt to use an alternative, higher threshold (+0.6mmol/L or +10 mg/dL above their current threshold).

The DDF app provides automated nutritional guidance based on recorded preprandial and postprandial glucose levels (relative to the current glucose threshold). When preprandial glucose is slightly above the threshold (≤0.4mmol/L or 7mg/dL), users are encouraged to prioritize meals with a higher protein composition. When glucose is just below the threshold, users are encouraged to eat “a nutritious meal.” If preprandial glucose is considerably below the threshold (>0.4mmol/L or 7mg/dL), users are encouraged to prioritize “whole food carbs” to quickly raise glucose into a normal range. Users who choose to also report postprandial glucose levels also receive feedback. Specifically, in the event postprandial glucose levels are more than 1.6 mmol/L or 30 mg/dL above the current threshold, the app suggests they “consider eating less of it (or even avoiding) that meal in the future.”

In addition to monitoring pre- and postprandial glucose levels and perceived hunger levels, the DDF app also allows users to track their fasting glucose levels and body weight. For this study, we extracted all available data entered into the DDP app by users during the first 30 days of use with day 0 representing the first day a preprandial glucose level was recorded. Implausible weight (<30 to >300 kg, and other unlikely values based on neighboring weights), height (<120 to >230 cm), and glucose values (<2.8 to >11.1 mmol/L) were excluded from the dataset.

In an effort to examine the effectiveness of the DDF app, we aimed to assess the changes in body weight (primary outcome) and fasting glucose levels (secondary outcome) between day 0 and day 30 of using the DDF app. The analysis included users who had at least 2 days of recorded entries. A total of 6519 people used the DDF app, with 95.1% (n=6197) using it at least twice in the first 30 days and thus were included in the analytical sample. To ensure the robustness of the findings, a sensitivity analysis was conducted to compare results when including users with 2 or more days of data versus 7 or more days. The findings were consistent across both criteria, confirming that the inclusion of users with fewer days of data did not affect the outcomes. Descriptive results, such as app usage data, were summarized using appropriate measures of central tendency (eg, means or medians) and variability (eg, SDs or IQRs), depending on the data distribution. Participants were categorized based on their baseline BMI and fasting glucose levels using established clinical thresholds [14,15]. Changes in weight and fasting glucose levels between the first and last weeks of app use were assessed using paired comparisons with 95% CIs. For participants who did not have weight or glucose measures in the last week, the last observation carried forward (LOCF) method was applied. A linear mixed-effects model was used to examine the relationship between preprandial glucose and perceived hunger, with preprandial glucose included as a fixed effect and a random intercept for participants to account for repeated measures within individuals. This approach allowed for the estimation of the overall association between preprandial glucose and hunger while accounting for variability in baseline hunger levels across users. Model estimates were reported with 95% CIs, and the significance of the fixed effect was assessed. Analyses were done with R (v4.2.2, R Foundation for Statistical Computing).

This study was approved by the University of Otago Human Ethics Committee (reference: HD22/064). Participants were users of a commercially available app who consented to their anonymized data being used for research purposes through the app’s privacy policy. All data were deidentified before analysis. No compensation was provided for participation.

Previous research indicates that in experimental settings, GGE results in meaningful improvements in body weight, theory-based eating behaviors, and metabolic health outcomes, particularly those related to glucose metabolism. DDF, a commercially available mobile app that supports GGE, represents a scalable, personalized public health intervention with a growing evidence base. This study aimed to evaluate DDF app engagement and the effectiveness of DDF on changes in body weight and fasting glucose levels. The findings indicate that DDF produces modest improvements in body weight, particularly among users with higher weights at baseline, and clinically meaningful reductions in fasting glucose, particularly among users with fasting glucose in the diabetes range at baseline. These results support the effectiveness of DDF despite a steady decline in app use over the 30-day period.

DDF produced an average weight loss of 2.2 kg in the completer analysis, comparable with the results of studies using GGE in more controlled research settings of different durations. Average weight loss in the published clinical trials ranges from 1.5 kg in a 2-week feasibility trial for participants with overweight or obesity [2] to 4‐5 kg over 6 months [7,10]. Despite its simplicity and reliance on intermittent glucose monitoring, DDF achieved 30-day weight losses comparable with that reported in a study evaluating the January AI app, which uses continuous glucose monitoring and artificial intelligence–driven personalized dietary feedback [16]. This comparison underscores the effectiveness of GGE, which relies on intermittent preprandial glucose measurements and simpler personalized thresholds, in achieving comparable real-world weight loss. The modest inverse association between hunger and glucose observed in this study aligns closely with findings from previous work conducted with participants across a range of BMI categories [2]. Results from this commercial implementation of GGE are reassuring, confirming the real-world potential of this approach to improve body weight.

In contrast, this is the first GGE study to report a clinically meaningful effect on fasting glucose [17]. The results here are similar to a recent 8-week clinical trial of time-restricted eating in a sample of adults with obesity, which resulted in a reduction in fasting blood glucose of 0.4 mmol/L with greater improvements among participants who had elevated blood glucose levels at baseline [18]. Interestingly, among users who had normal fasting blood glucose at baseline, use of the DDF app resulted in a slight increase in fasting glucose levels over the 30-day period (0.11 mmol/L to 0.23 mmol/L or 2 mg/dL to 4 mg/dL), although still within the normal range (<5.6 mmol/L or <100 mg/dL). Future studies should aim to explore this phenomenon, if replicated. Despite subtle fluctuations of fasting glucose among normoglycemic users of the DDF app, the majority of users experienced clinically meaningful improvements in glycemia. Future clinical trials should aim to replicate the beneficial effects of GGE on markers of glucose metabolism (and insulin resistance), particularly among people with prediabetes or diabetes.

The strengths of this study include the real-world application of GGE and a very large sample size that included a diverse range of baseline weights and fasting glucose levels. The results demonstrate the outcomes of the DDF method when implemented under real-world conditions. Data collected from people in free-living conditions are more representative of effectiveness and can be better generalized to the wider population. The DDF app provides an accessible and structured method for glucose-guided eating, leveraging user-friendly tools like glucometers to promote self-awareness and behavioral changes.

Despite these strengths, there was a lack of available data describing users in terms of age, ethnicity, and health status; however, the app could be easily modified to collect these data for future research purposes. In addition, our analyses examining the effect of DDF on body weight and fasting glucose levels were limited by the use of self-reported data. Adherence to the program could only be inferred from interaction with the DDF app and the absence of granular tracking tools, such as Google Analytics, limits the ability to fully evaluate user interaction with the app. It is possible that users continued to follow GGE without recording all measures or used the app incorrectly. While the median app usage of 19 days reflects real-world engagement patterns, it may not capture the long-term impact of sustained use. However, previous research has demonstrated that the benefits of GGE can persist for up to 5 months following the initial training period [10]. The DDF app focuses on helping users establish effective self-regulation strategies within the first 30 days, which may contribute to lasting benefits. Future research should investigate the durability of these outcomes over extended periods. The study design did not include a control group, which limits our ability to definitively conclude that changes in our outcomes were due to the use of DDF. While the use of the glucometer alone (and their companion apps) might have induced added benefits, research to date suggests that these benefits are likely to be minor without a structured program [19,20]. Compared with gold-standard measures like laboratory fasting glucose or HbA_1c, preprandial glucose monitoring with a glucometer offers a more accessible yet less precise approach [21]. This trade-off highlights the practicality of the DDF app for real-world settings, particularly for individuals seeking to manage weight and glucose levels without requiring expensive or complex equipment.

For users, the DDF app provides a simple and structured framework to improve metabolic health by promoting self-awareness of hunger and glucose patterns. For health care professionals, the app provides a unique opportunity to observe real-world glucose patterns and weight trends, offering insights that could inform personalized dietary and metabolic health interventions. Adherence and outcomes in the real world are likely to be affected by a range of external factors, such as social support, food security, family structure, and food preferences, which were not possible to measure in this setting. Nevertheless, these preliminary findings are positive, particularly in light of the real-world research setting, and suggest a more rigorous examination of the long-term potential benefit of DDF is warranted.

GGE, supported by the DDF app, is a simple intervention that could be easily implemented at scale with minimal support services. Glucometers are widely available and inexpensive, and continuous glucose monitors are available over-the-counter in several countries. Results from this study suggest that the DDF app may be most beneficial for people with overweight or obesity and those who have dysglycemia. The DDF app could be used alongside other lifestyle interventions to support behavior change among these populations. Future randomized controlled trials are warranted to confirm the metabolic benefits of DDF and GGE in specific populations, such as individuals with prediabetes or diabetes, and to assess the sustainability of the behavioral and metabolic improvements achieved through the app over the long term.