Diabetes Telemedicine Mediterranean Diet (DiaTeleMed) Study: study protocol for a fully remote randomized clinical trial evaluating personalized dietary management in individuals with type 2 diabetes

Lauren Berube, Collin Popp, Margaret Curran, Lu Hu, Mary Lou Pompeii, Souptik Barua, Emma Bernstein, Vanessa Salcedo, Huilin Li, David E. St‐Jules, Eran Segal, Michael Bergman, Natasha Williams, Mary Ann SevickTrials 2024 · DOI

Abstract: ClinicalTrials.gov NCT05046886. Registered on September 16, 2021.

Introduction

The DiaTeleMed Study represents a significant shift toward precision medicine in diabetes management. Traditional "one-size-fits-all" dietary advice often fails due to the substantial inter-individual variability in postprandial glycemic responses (PPGR). This study evaluates whether a machine-learning-personalized nutrition plan is more effective than a standard Mediterranean diet for managing Type 2 Diabetes (T2D) in a fully remote clinical trial setting.

The Challenge of Glycemic Variability

Standard dietary guidelines often overlook the fact that two individuals can consume the same carbohydrate source yet experience vastly different blood sugar spikes. These spikes are driven by a complex interplay of genetics, lifestyle, and the gut microbiome. The DiaTeleMed trial aims to address this by predicting individual spikes and tailoring dietary choices accordingly to improve overall glycemic control.

Methods and Study Design

The trial is a parallel-group, three-arm randomized clinical trial (RCT) involving 255 participants. A defining feature of this study is its fully remote nature, leveraging telemedicine, continuous glucose monitors (CGM), and mobile health applications to reach participants in their home environments.

DiaTeleMed Study Flow: From virtual screening to 6-month follow-up.

Participants are randomized into three distinct intervention arms:

  1. Usual Care Control (UCC): Provides basic Mediterranean diet education via instructional videos.
  2. Standardized Behavioral Intervention: Combines the Mediterranean diet with behavioral counseling based on Social Cognitive Theory and real-time macronutrient tracking.
  3. Personalized Behavioral Intervention: Includes all standardized elements plus real-time feedback on predicted PPGR for every meal, generated by a proprietary machine-learning algorithm.
Comparison of intervention components across the three study arms.

The Personalized Intervention: "Getting to Green"

In the personalized arm, participants use the DayTwo app, which employs an algorithm trained on the Personal Nutrition Project (PNP) cohort. This model predicts blood sugar responses based on microbiome data, blood tests, and lifestyle factors. Participants are coached on a "Getting to Green" strategy: if a planned meal is predicted to cause a "red" or "yellow" spike, they are taught to modify the meal's composition or portion size to achieve a "green" rating before consumption.

Outcome Measures

The primary outcome is the Mean Amplitude of Glycemic Excursions (MAGE), a measure of glycemic variability (GV). While HbA1c remains the gold standard for long-term monitoring, the researchers argue that GV—the "peaks and valleys" of blood glucose—is a more sensitive indicator of the risk for diabetes-related complications.

Pilot Data and Rationale

The trial is supported by pilot data from Israel, where an algorithm-guided diet significantly reduced glycemic exposure compared to a standard Mediterranean diet. Notably, 61% of pilot participants with an initial HbA1c ≥ 6.5% reached a level below 6.5% after six months of the personalized intervention, providing a strong rationale for this larger-scale trial.

Limitations and Future Outlook

The study faces challenges regarding generalizability, as it is centered within the NYU Langone Health system and currently limited to English-speaking participants with sufficient digital literacy. However, if successful, this remote, scalable model for precision nutrition could redefine the standard of care for T2D management, potentially being adapted for diverse geographic and socioeconomic settings.

Conclusion

By integrating machine learning with telemedicine, the DiaTeleMed Study seeks to validate a scalable approach to personalized nutrition that moves beyond generalized dietary guidelines toward truly individualized care.

Summary updated

Abstract

ClinicalTrials.gov NCT05046886. Registered on September 16, 2021.
BibTeX
@inproceedings{berube2024diabetes,
  author = {Lauren Berube and Collin Popp and Margaret Curran and Lu Hu and Mary Lou Pompeii and Souptik Barua and Emma Bernstein and Vanessa Salcedo and Huilin Li and David E. St‐Jules and Eran Segal and Michael Bergman and Natasha Williams and Mary Ann Sevick},
  title = {{Diabetes Telemedicine Mediterranean Diet (DiaTeleMed) Study: study protocol for a fully remote randomized clinical trial evaluating personalized dietary management in individuals with type 2 diabetes}},
  booktitle = {Trials},
  year = {2024},
  doi = {10.1186/s13063-024-08337-w},
}
Figure