Novel AI-enabled wearable sensor for scratch monitoring

The Food and Drug Administration awarded Sibel Health Inc. a $499,998 cooperative agreement on September 24, 2025, to validate and finalize qualification of the ADAM (Acoustic Detection and Motion) wearable sensor for objective scratch monitoring as a behavioral biomarker of itch severity in atopic dermatitis.

The ADAM device uses acoustic and motion sensors with artificial intelligence to detect scratching behavior with high temporal resolution, addressing limitations of patient-reported outcome measures—particularly in pediatric populations and during sleep. Under the award, Sibel Health will execute three aims: validate scratch intensity by comparing AI-generated outputs against force plate measurements in controlled conditions; develop and validate algorithms for estimating body surface area scratched and body location of scratching using supervised videography; and conduct a one-week clinical validation study in 30 pediatric and adult atopic dermatitis patients using infrared nighttime videography and daily diaries as gold standards. The work responds to FDA feedback received following the agency's 2021 acceptance of the initial letter of intent, which enabled development and clinical testing of the first ADAM version, achieving 99 percent AI classification accuracy and IEC 60601 safety standard compliance.

Performance occurs in Chicago, Illinois, with a period of performance extending through June 30, 2027. This award falls under FDA Research (CFDA 93.103), which supports research, demonstration, education, and information dissemination activities for medical device and therapeutic development.

Abstract

Itch is a distressing and highly prevalent symptom across a wide range of dermatological and systemic diseases, profoundly impacting quality of life, particularly through sleep disturbance and skin damage. In atopic dermatitis (ad), itch is the hallmark symptom and drives both disease burden and therapeutic response. However, the current standard for measuring itch relies heavily on patient-reported outcomes (PROS), which suffer from limited reliability, especially in pediatric populations and during unconscious periods such as sleep. Scratching is the primary behavioral response to itch and offers a unique, objective, and quantifiable proxy for itch severity. Our team has developed and iteratively refined a novel, hand-mounted wearable device (ADAM) that uses both acoustic and motion sensors to detect scratching behavior with high temporal resolution. In 2021, the FDA accepted our COA letter of intent which allowed US to develop, validate, and clinically test the initial version of ADAM. We demonstrated high AI classification accuracy (99%), compliance with IEC 60601 safety standards, and achieved key milestones including battery life extension and cybersecurity improvements. In this proposal, we aim to respond to recent FDA feedback and finalize qualification of the updated ADAM sensor towards full qualification. In aim 1, we will validate scratch intensity by comparing AI-generated outputs against force plate measurements under controlled conditions. In aim 2, we will develop and validate new algorithms for estimating body surface area (BSA) scratched and body location of scratching using supervised videography. In aim 3, we will conduct a 1-week clinical validation in 30 pediatric and adult ad patients using infrared nighttime videography and daily diaries as gold standards. This work will close the GAP between subjective PROS and objective behavioral biomarkers of itch, fulfilling a critical need in dermatology and drug development. Our validated digital endpoint will directly support regulatory-grade applications in ad and form the foundation for broader deployment in other pruritic disorders. The proposed research aligns with the FDA's clinical outcome assessment qualification program and the NIH's interest in advancing digital health technologies for real-world patient monitoring.