Hackensack Meridien Health partners with AI-driven speech analysis start-up

Health system Hackensack Meridian Health is partnering with Canary Health, which uses proprietary machine learning modeling to assess digital biomarkers in the human voice to explain or predict health-related outcomes. This will enable Hackensack Meridien to accelerate its artificial intelligence technology for wellness.


Canary Speech’s patented technology monitors health factors “in a fast, non-invasive and accurate way,” according to Thursday’s announcement.

Developed by the neurology and speech AI team behind Amazon Alexa, Canary Speech aims to provide proactive screening in real time with algorithms built to augment standardized clinical assessments like GAD7 for anxiety and PHQ8 for depression.

The company uses machine learning to model both acoustic and linguistic features of voice and human speech and aims to provide clinical-grade screening to detect changing emotional wellness states and mental health deterioration in real time.

Hackensack Meridian says the tool’s vocal scores can replace subjective measurements with “objective, actionable” care plans.

Canary Speech is also researching voice assessments for diseases like Huntington’s, Alzheimer’s and Parkinson’s as well as PTSD and other conditions.

“Investment of this kind is a smart strategic move, with an eye toward the future,” said Dr. Ihor Sawczuk, Hackensack Meridian Health’s president of academics, research and innovation, and founding chair of the Hackensack Meridian Health Research Institute.


Even in daily and seemingly unemotional conversation, the human voice can reveal clues – within the patterns, inflections and timing of those words – to both an individual’s emotional and physiological states.

In December, Grace Chang, cofounder and CEO of Kintsugi, explained how novel machine learning- and deep learning-based voice biomarker technology is deployed in health systems and payer organizations to identify signs of clinical depression and anxiety and improve access to mental health services.

Thousands of experiments during development revealed that what somebody says is not as important as how they say it, Chang said.

In a demo on HIMSSTV, Chang showed how the software tool – which protects privacy, provides real-time results and is language-agnostic – and API architecture can be integrated into telehealth systems.

At the beginning of a virtual care session, patients are asked to give consent to be recorded for assessment. Through the course of the practitioner’s visit, the technology identifies that there may be a risk of more severe depression or anxiety and flags the clinician.

“This then gives them the opportunity to offer mental health services to the patient that the patient may not know is necessarily provided as under their plan, and then were able to route accordingly based on the severity of their depression or anxiety,” Chang explained.

The demo showed a visit to discuss a patient’s back pain. The clinician and patient discussed her physical experience as the tool provided GAD7 and PHQ8 scores, which then encouraged her clinician to ask further questions about how she is feeling about it and offer her a referral.


“This is a fascinating innovation, and we’re eager to support a promising way to leverage technology for wellness,” said Robert C. Garrett, CEO of Hackensack Meridian Health. 

“Artificial intelligence is impacting every aspect of modern life, and our health network wants to help accelerate the efforts of companies like Canary Speech.”

Andrea Fox is senior editor of Healthcare IT News.
Email: [email protected]

Healthcare IT News is a HIMSS Media publication.

Source: Read Full Article