Feasibility and Effectiveness of an AI-Powered Carbohydrate Counting Educational Platform to Support Parents of Children With Type 1 Diabetes (CARB-AI)

Sultan Qaboos University

Status: Not yet enrolling

Conditions: Type 1 Diabetes Mellitus

Treatments: Behavioral: AI-Powered Carbohydrate Counting Educational Platform

Study type: Interventional

Funder types: Other

Identifiers

NCT07671053

SultanQU-PEDU-002

Details and patient eligibility

About

The goal of this clinical trial is to learn whether an AI-powered carbohydrate counting educational platform can help parents of children with type 1 diabetes improve their carbohydrate counting skills and diabetes management. The study will include parents or primary caregivers of children aged 2-12 years with type 1 diabetes.

The main questions it aims to answer are:

• Is the AI-powered educational platform feasible, acceptable, and easy for parents to use?
• Can the platform improve carbohydrate counting accuracy, parental confidence in diabetes management, and diabetes outcomes compared with usual education alone?

Researchers will compare parents who receive access to the AI-powered carbohydrate counting educational platform plus usual diabetes education with parents who receive usual diabetes education alone to see whether the AI-supported approach provides additional benefits.

Participants will:

• Complete baseline assessments, including questionnaires and a carbohydrate counting test.
• Be randomly assigned to either the AI-supported education group or the usual education group.
• Use the assigned educational resources for 12 weeks.
• Complete a follow-up assessment at 6 weeks and a final assessment at 12 weeks.
• Provide information about their child's diabetes management, including HbA1c and glucose monitoring data.
• Complete questionnaires about confidence, usability, and satisfaction with the educational support they receive.

The AI platform is designed to provide educational support only and does not replace medical advice, insulin dosing decisions, or routine diabetes care provided by healthcare professionals.

Full description

This multicentre randomized controlled feasibility trial will evaluate an AI-powered educational platform designed to support carbohydrate counting education for parents of children with type 1 diabetes (T1D). Accurate carbohydrate counting is an essential component of T1D management because insulin dosing is closely linked to carbohydrate intake. However, many parents experience challenges in estimating carbohydrate content accurately, which may affect glycemic control.

The intervention uses conversational artificial intelligence to provide personalized educational support, interactive learning opportunities, and practical guidance related to carbohydrate counting. The platform is intended as an educational tool and does not provide medical advice or insulin dosing recommendations. Educational content and safety oversight are provided by pediatric endocrinologists, diabetes educators, and registered dietitians.

The primary objective of this feasibility study is to evaluate recruitment, retention, participant engagement, intervention adherence, and data collection procedures to determine whether a future definitive efficacy trial is warranted. Secondary objectives include assessment of participant acceptability and usability, as well as exploration of preliminary effects on carbohydrate counting accuracy, parental diabetes management self-efficacy, and glycemic outcomes.

Participants will be recruited from our pediatric diabetes centers, and randomized to receive either access to the AI-powered educational platform in addition to enhanced usual care or enhanced usual care alone. Study findings will inform the development of larger trials evaluating the role of conversational artificial intelligence in diabetes education and chronic disease self-management.

Enrollment: 80 estimated patients

Sex: All

Volunteers: No Healthy Volunteers

Inclusion criteria

• Primary responsibility for carbohydrate counting and insulin dosing decisions for the child
• English-speaking
• Access to a smartphone (iOS or Android) with internet connectivity
• Willing and able to provide informed consent and complete study procedures
• Diagnosis of type 1 diabetes for at least 1 month
• Receiving intensive insulin therapy (multiple daily injections or insulin pump)
• Using carbohydrate counting for insulin dosing

Exclusion criteria

• Child has significant developmental delay or a medical condition that substantially alters nutritional requirements or carbohydrate metabolism (e.g., celiac disease, cystic fibrosis)
• Parent or caregiver has significant cognitive impairment that would preclude participation
• Family plans to relocate from the study area during the study period
• Participation in another diabetes intervention study

Trial design

Primary purpose: Supportive Care

Allocation: Randomized

Interventional model: Parallel Assignment

Masking: Single Blind

---

80 participants in 2 patient groups

AI-Powered Carbohydrate Counting Educational Platform + Enhanced Usual Care

Experimental group

Description:

Participants receive access to an AI-powered carbohydrate counting educational platform in addition to enhanced usual diabetes care. The platform provides interactive educational support, carbohydrate counting practice, personalized feedback, scenario-based learning, and educational guidance under dietitian and diabetes specialist oversight. Participants also receive standard diabetes education materials and routine clinical care for 12 weeks.

Treatment:

Behavioral: AI-Powered Carbohydrate Counting Educational Platform

Enhanced Usual Care Alone

No Intervention group

Description:

Participants receive enhanced usual diabetes care consisting of standard diabetes education provided by their clinical care team, printed carbohydrate counting educational materials, portion size reference materials, educational PDF resources, and routine clinical care. Participants do not receive access to the AI-powered carbohydrate counting educational platform