Mental health assessment tool Deliberate AI joins FDA's ISTAND program

Steven Loeb · January 26, 2024 · Short URL:

Its product uses words, vocal and visual cues, and other modalities to measure mental health

While physical health problems can often be measured through biological markers, mental health assessment is often made through self assessment and questionnaires. Actually measuring mental illness has been an area that's been extremely tough to crack.

That might be changing thanks to the proliferation of artificial intelligence: that's the idea behind Deliberate AI. a digital health technology company that uses AI to deliver more precise measurement of mental health. Its AI-generated Clinical Outcome Assessment (AI-COA) solution functions similarly to human perception but with enhanced perception and analytical capabilities, combining data that a person normally processes through sight and hearing, assimilating it in a way analogous to the human brain.

"We've developed a highly transparent machine learning algorithm, designed to identify the severity of depression and anxiety using audiovisual signals captured through an ordinary smartphone or computer. This involves fusing a multitude of audiovisual features in a multimodal algorithm that infers clinical outcome assessments," Marc Aafjes, founder and CEO of Deliberate AI, told VatorNews.

"These assessment scores, which our models infer are FDA-approved for use in clinical trials, showcasing our technology's ability to bridge the gap between subjective observation and objective, quantifiable data in mental health evaluation."

On Friday the company announced that it has been selected by the U.S. Food and Drug Administration (FDA) to participate in the Innovative Science and Technology Approaches for New Drugs (ISTAND) Pilot Program, which was designed to encouraging development of Drug Development Tools (DDTs) that are out of scope for existing DDT qualification programs but may still be beneficial for drug development. 

Under the program, AI-COA will receive support to become a qualified DDT.

"The ISTAND announcement focuses on our advanced work with depression and anxiety AI-COATMs. These tools are specifically designed to monitor symptoms of depression and anxiety across various diagnoses, including major depressive disorder, generalized anxiety disorder, and Bipolar Disorder," Aafjes explained.

"Additionally, our technology shows promise for measuring depression and anxiety as transdiagnostic symptoms in individuals who have not been formally diagnosed with a mood disorder."

Founded in 2020, the New York City-based Deliberate operates its tool seamlessly in the background so, from the patient's perspective, using its assessment tool is straightforward and unobtrusive, especially during direct patient-clinician interactions. That can be in assessment interviews in clinical trials or during clinical encounters in healthcare; the key requirement with AI-COA is to ensure that the video camera is positioned to clearly capture the patient's face and neck without any obstruction and in good lighting. 

"Patients have expressed a strong desire for data-driven insights into their mental health, in line with what they receive from blood tests and radiological results in other areas of healthcare. They appreciate having tangible, objective data for their mental health, which our tools provide. They also want safe, efficacious and affordable therapies fast – which requires significant improvement in clinical trials for those," said Aafjes.

Currently, AI-COA is not a diagnostic; rather, the AI model infers a severity measure similar to clinician rating scales. While the company is looking to develop diagnostic and prognostic models, this will require shifting to a different approach with current diagnostic criteria.

In addition, the company is also working to extend the AI-COA for use in clinical practice to enable effortless measurement-based care through ambient tracking and monitoring of clinical encounters, and was as other indications, with R&D underway for PTSD, suicide risk, OCD, and schizophrenia.

Deliberate AI's current customers include the Defense Advanced Research Project Agency (DARPA), for a project to improve suicide risk assessment, clinical trial sponsors, and mental health clinics. It’s also currently running a pilot with one of New York’s largest healthcare systems.

In terms of ROI, the company is optimistic about the potential in drug development, said Aafjes.

"By quantitatively capturing mental health symptoms, our AI-COA models offer a consistent, bias-free assessment, enhancing the precision and reliability of clinical trials. This means significantly more powerful, efficient, and faster drug development processes, leading to improved patient care and significant public health advancements. Early indications suggest opportunities to reduce trial sizes and reduce the duration of Phase 3 trials by months," he explained.

Now that the company is part of ISTAND, it will have more rapid interaction with the FDA towards ultimate qualification at a DDT.

For Deliberate AI, inclusion in the program shows the regulatory support for the approach of AI-COA in clinical trials, which is an important consideration for sponsors considering including this technology in their late-stage clinical trials, Aafjes said, and the company expects this will lead to faster adoption in the industry.'

"It shows the regulatory support for the approach of AI-COA in clinical trials, which validates our technology implementation approach. The FDA’s decision 'is based on several factors, including the scientific merit of the submission, the ability of the DDT to address a specified drug development need, the availability of information and resources that support the proposed qualification effort and demonstration that the DDT is feasible and practical within the proposed context of use.' To obtain qualification we will generate the required additional evidence validation through a prospective confirmatory trial," he said.

(Image source:

Support VatorNews by Donating

Read more from our "Trends and news" series

More episodes