
Canary Speech Vocal Biomarker Model for Behavioral Health
Contact Information
info@canaryspeech.com
Release Stage
Release Date
Version
Regulatory Approval
HIPAA compliant, HITRUST E-1 Certified, ISO/IEC 27001:2022 and 42001:2023 Certified
Global Availability
USA; available for pilot and integration in select international markets
Summary
Canary Speech’s Behavioral Health model is a standardized instrument that provides clinical decision support to screen for Major Depressive Disorder and Generalized Anxiety Disorder in patients aged 18 and older. Our software utilizes an artificial intelligence algorithm to analyze speech patterns, providing clinicians with vocal biomarkers that provide actionable insights for early detection and ongoing management. Built for seamless integration, it functions within digital health workflows on any smart device. Ultimately, clinicians are responsible for the diagnosis of Major Depressive Disorder and Generalized Anxiety Disorder per the standard of care.
Uses and Directions
Intended Use
o Designed to inform and augment (not replace) clinical management of adult patients at risk or being treated for behavioral health conditions such as Major Depressive Disorder and Generalized Anxiety Disorder in ambulatory, telehealth, acute, or remote monitoring settings. o Operates as a clinical decision support tool: clinicians collect ≥40 seconds of patient speech via app/device; AI analyzes speech for vocal biomarker patterns; clinician receives results in near real-time. o Requires human review and clinical correlation—clinicians maintain oversight.
Primary Users
Primary Care Providers, Psychiatrists, behavioral health professionals, credentialed telehealth clinicians. Requires basic digital literacy and familiarity with depression and anxiety screening.
How to Use
All configuration settings are documented and designed specific to the customer use case. These settings are defined in the configuration file hosted in the secure Canary Speech cloud (Azure) environment.
Canary Speech’s vocal biomarker technology is designed to be used either in-person or remotely. In either setting, patient assessments require capturing at least 40 seconds of conversational speech in the system. o Clinician and patient meet in an environment that is quiet and where background noise is limited. o Clinician obtains informed consent from patient regarding the recording of their voice for purposes of conducting an assessment. o Clinician logs into the system via a Canary Speech-compatible audio capture device (such as a tablet, smartphone app, or telehealth software). o Clinician places the audio capture device in a location within 3 feet of the patient’s mouth ensuring that the device microphone is unobstructed, if in person, or conducts their usual telehealth visit with the patient remotely. o During the patient visit, once 40 seconds of the patient’s voice has been obtained Canary Speech will analyze the speech sample in real-time. o The analysis typically takes about 10-15 seconds and then the system returns a result to the clinician’s device. o The software will return results for Major Depressive Disorder and Generalized Anxiety Disorder.
Target Population
Patients aged 18 or older
Cautioned Out-of-Scope Settings
Cautioned Out-of-Scope Settings
Audio Capture Requirements The minimum requirements for audio capture from a Canary Speech-compatible voice capture device are at least a 16kHz sampling rate with 16-bit depth with a single channel recording. This provides compatibility with most modern voice recording devices (from smart devices to bluetooth headphones to telephony systems).
EHR Compatibility
Compatible EHRs:
- Epic
- Cerner / Oracle Health
- Meditech