Skip to content

Wearable Sensors and Artificial Intelligence for Carbohydrate Counting

Feasibility of Using Wearable Sensors and Artificial Intelligence for Carbohydrate Counting in Chinese Americans With Type 2 Diabetes

Status
Completed
Phases
Unknown
Study type
Observational
Source
ClinicalTrials.gov
Registry ID
NCT05335889
Enrollment
12
Registered
2022-04-20
Start date
2022-07-18
Completion date
2023-09-22
Last updated
2024-04-12

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Type 2 Diabetes

Brief summary

This is a one-group pilot study where Chinese immigrants who are English speaking with T2D from NYU Langone Health and NYU Brooklyn Family Health Center (Sunset Park) will be recruited.

Detailed description

To evaluate the estimation accuracy using eButton, researchers will collect carbohydrate data via weighing food by registered dietitian nutritionist (RDN) (gold standard) (2 days/week at research labs) and food diaries by participants (2 days/week at research labs and 3 days/ week at participant home). Then, the estimated carb grams will be compared head to head among each other. Assessment will be at 0 and 2 weeks, including surveys and qualitative audio interview.

Interventions

DEVICEeButton

The eButton is a wearable camera that takes pictures every 6 seconds of whatever is in front of participants. The recorded data are processed by algorithms to determine food names, volumes, and nutrient value of the consumed food (e.g., grams of carbohydrates). The eButton is a wearable device containing a multicore microprocessor, a rechargeable battery capable of 10-15 hours of continuous operation (upon a flexible choice of battery capacity), a miniSD card for data storage.

DEVICEContinuous Glucose Monitoring (CGM)

The use of this device provides ambulatory glucose profiles, giving graphic and quantitative information on 24-hour glucose patterns. It does not require finger-prick testing for calibration. The system consists of a reader and a sensor (35 mm x 5 mm). The sensor is applied to the back of a person's arm. The sensor automatically measures interstitial glucose at 15-minute intervals during daily activities like work, sleep, eating, and exercise. It is able to store blocks of glucose data for 14 days.

Sponsors

NYU Langone Health
Lead SponsorOTHER

Study design

Observational model
CASE_ONLY
Time perspective
PROSPECTIVE

Eligibility

Sex/Gender
ALL
Age
18 Years to 99 Years
Healthy volunteers
No

Inclusion criteria

* ≥18 years old * Diagnosed with T2D at least one year prior * Self-identified as first- or second-generation Chinese immigrants * Feel comfortable communicating in English, the reason is that the questionnaires/surveys are validated in English * Have a computer and internet connection

Exclusion criteria

* Plan frequent travel or vacations or plan to relocate in the next five weeks * Have serious diabetes-related complications, physical illness, or mental illness (e.g., schizophrenia, bipolar disorder, or substance abuse) that would preclude participation * Have severe cognitive impairments (e.g., dementia, intellectual disability)

Design outcomes

Primary

MeasureTime frameDescription
Accuracy of Carbohydrate Counting using eButton (absolute error)Day 14The estimated carb grams using the eButton, gold standard, and participants' food diaries will be compared head to head among each other. The absolute error will be computed: the difference between the estimated value and the gold standard (estimated - gold standard). Bland-Altman plots will be used to examine the level of agreement between eButton and gold standard measurements.
Accuracy of Carbohydrate Counting using eButton (relative error)Day 14The estimated carb grams using the eButton, gold standard, and participants' food diaries will be compared head to head among each other. The relative error will be computed: the percentage difference between the estimated value relative to the gold standard. Relative errors will be reported using boxplots to allow visual comparison of the distribution and variability in errors across all methods. Bland-Altman plots will be used to examine the level of agreement between eButton and gold standard measurements.
Proportion of participants who are fully compliant with eButton useDay 14Proportion of participants who are fully compliant with eButton use and proportion of meals evaluated using eButton, with 95% confidence intervals will be qualitatively reported. Perceived usefulness and perceived ease of use will be summarized as mean and standard deviation.

Countries

United States

Outcome results

None listed

Source: ClinicalTrials.gov · Data processed: Feb 4, 2026