A Study to Train a Machine Learning Algorithm for an Evaluation of the Use of Biometric Data Captured at the Wrist for the Identification of Acute Opioid Use Events and the Quantification of Opioid Withdrawal in Opioid Dependent Individuals

OpiAID

Status: Enrolling

Conditions: Treatment for Opioid Use Disorder

Treatments: Device: Train and evaluate the accuracy and reliability of the Strength Band Platform in identifying acute opioid dosing events from time-stamped biometric data collected from wrist-worn devices.

Study type: Interventional

Funder types: Industry; NIH

Identifiers

NCT07405398

4R44DA058474-02 (U.S. NIH Grant/Contract)

OA-SBIR-25-01

Details and patient eligibility

About

To train a machine learning model/algorithm for an evaluation of the use of biometric data captured at the wrist for the identification of acute opioid use events and the quantification of opioid withdrawal in opioid dependent individuals.

Full description

The goal of this real-world, multi-center, outpatient study is to train a machine learning model/algorithm utilizing patient-specific physiological parameters from the OpiAID Strength Band Platform™ can accurately detect MOUD events during the induction phase with an 80% classification success when comparing the True Positive Rate against the False Positive Rate as plotted on a Receiver Operator Curve. In addition to MOUD detection, machine learning will be used to quantify participant withdrawal level from physiological parameters. To demonstrate that withdrawal quantification performs as well or better than current measures used for this purpose the correlation between quantified withdrawal and time since last opioid dose (TSLD) will be computed and compared against the association between SOWS and TSLD in a non-inferiority analysis.

Enrollment: 420 estimated patients

Sex: All

Ages: 22+ years old

Volunteers: Accepts Healthy Volunteers

Inclusion criteria

• Male or female
• Age ≥22 years at signing of informed consent
• Patients with a DSM-5 diagnosis of OUD who are eligible for MOUD induction with methadone or buprenorphine

Exclusion criteria

• Sleeve tattoo covering the wrist
• Subject unable to independently navigate and operate smartwatch applications
• Subject not proficient with written and spoken English
• Subject determined likely to be non-compliant by physician/HCP
• Subject likely to not be available to complete all protocol-required study visits or procedures, and/or to comply with all required study procedures to the best of the subject and investigator's knowledge.
• History or evidence of any other clinically significant disorder, condition, or disease that, in the opinion of the investigator, would pose a risk to subject safety or interfere with the study evaluation, procedures or completion.
• Subject has diminished decision making capability

Trial design

Primary purpose: Supportive Care

Allocation: N/A

Interventional model: Single Group Assignment

Masking: None (Open label)

---

420 participants in 1 patient group

Single arm 14 day monitoring period

Experimental group

Description:

The goal of this real-world, multi-center, outpatient study is to train a machine learning model/algorithm utilizing patient-specific physiological parameters from the OpiAID Strength Band Platform™ can accurately detect MOUD events during the induction phase with a predefined classification success when comparing the True Positive Rate against the False Positive Rate as plotted on a Receiver Operator Curve. In addition to MOUD detection, machine learning will be used to quantify participant withdrawal level from physiological parameters. To demonstrate that withdrawal quantification performs as well or better than current measures used for this purpose the correlation between quantified withdrawal and time since last opioid dose (TSLD) will be computed and compared against the association between SOWS and TSLD in a non-inferiority analysis. Prescribing physician must determine appropriate starting dose (titration expected over 2-6 weeks)

Treatment:

Device: Train and evaluate the accuracy and reliability of the Strength Band Platform in identifying acute opioid dosing events from time-stamped biometric data collected from wrist-worn devices.