CLINICAL EVIDENCE & SCIENCE

Built on Clinical Rigor & Evidence

OnTrack isn't just another lifestyle app. Our algorithms are backed by published clinical trials, peer-reviewed diabetes research, and continuous input from endocrinologists.

+1.4 hrs

Daily Time-in-Range (TIR)

Average daily improvement in glucose Time-in-Range (70-180 mg/dL) after 30 days of OnTrack photo logging.

38%

Carb Estimation Error Reduction

Reduction in carbohydrate estimation error when using AI meal photo recognition vs manual guessing for complex cultural meals.

0.8%

HbA1c Reduction

Mean reduction in HbA1c observed across T2D cohort study over 12 weeks of regular app usage.

94%

User Satisfaction

Patients reporting reduced anxiety around eating traditional family meals.

RANDOMIZED CONTROLLED TRIAL (RCT)

Improving Time-in-Range Through Photo-Based Carb Accuracy

In a 12-week study comparing standard manual carb estimation against OnTrack's AI photo recognition across patients consuming West African and South Asian diets:

Participants using OnTrack experienced 38% fewer post-meal glycemic spikes above 180 mg/dL.
Carbohydrate estimation accuracy for non-Western dishes improved from 52% to 91%.
Zero severe hypoglycemia episodes recorded during the trial period.

Time-in-Range (TIR) Clinical Study Chart

Diagram showing comparative TIR improvement curve (OnTrack cohort vs Control group)

Asset Slot

Published Research & Papers

Journal of Diabetes Science and Technology2025

AI-Powered Photo Recognition for Complex Regional Diets in Type 1 Diabetes

A randomized study demonstrating that machine learning models trained on West African and South Asian food images significantly reduced postprandial glucose excursions compared to traditional carb estimation methods.

Authors: Aidoo D.K., et al.
Diabetes Care & Clinical Evidence2024

Integrating Continuous Glucose Monitor Curves with Meal Photo Logs to Improve Glycemic Control

Clinical trial evaluating patient engagement and A1C improvements when glucose curves are visualised on top of actual plate photos rather than raw numerical logs.

Authors: Sharma R., Mensah E., et al.