top of page

What are upcoming advances for speech-based depression and anxiety screening?

At Ellipsis Health, we’ve been looking back at some of the breakthrough applications of vocal biomarker technology that we’ve unlocked alongside our health plan, health system, academic medical center, and digital health partners.

Today, we’re going to focus on current studies and future ambitions that get us closer to our goal of putting mental health assessment on the same level as physical health. These projects all have the same ultimate aim in common — improving the quality and accessibility of depression and anxiety screening for the greatest number of people. So join us as we explore the different ways of achieving that aim and the positive impacts on our partners and the people they serve.

Tackling a growing rise in adolescent depression

At Ellipsis Health, we have already made significant strides toward improving screening accessibility for adolescents, but more work is needed to help this particularly vulnerable age group.

Even before the surge in demand for mental healthcare due to the COVID-19 pandemic, rates of adolescent depression almost doubled[1] between 2009 and 2019. The downstream effects of depression also have wider implications for society as a whole. For example, a 2018 study showed that students with notable signs of depression are more than twice as likely to drop out compared to their peers[2]. Similarly, the increase in depression coincided with an alarming rise in suicide between 2007 and 2018, climbing 57%, from 6.8 per 100,000 to 10.7 per 100,000. While it is vitally important to acknowledge the effects of depression on society, we can’t ignore the need for innovative action to promote early intervention and prevent escalation.

According to a study in the Archives of General Psychiatry[3], 50% of overall lifetime mental illness takes shape by the age of 14, and 75% by the age of 24. Close monitoring is critical in these important stages of development if we are to manage or arrest the development of mental health conditions.

At Ellipsis Health, we are currently working with the University of Denver Graduate School of Social Work to validate the efficacy of our technology as a depression and anxiety screening tool for 700 adolescents aged 11 to 17. The study focuses on a range of factors, from user engagement with the app to the acceptability of a mental health resource page for school mental health clinicians. With all of these elements working in tandem, we expect this pilot to yield positive outcomes that can be scaled up to more extensive cohorts. 


As opposed to traditional methods that rely on self-reporting or infrequent one-to-one sessions, patients invariably receive treatment that is too little or too late. Using secure, validated AI to monitor patients at scale, healthcare providers receive real-time data and can channel resources to intervene in serious cases at a much earlier stage. Ellipsis is hoping to shift the paradigm toward prevention not only for adults but whole families.

The benefits are not just for the patients themselves – healthcare partners also see efficiency gains. As Dr. Michael Aratow, MD, Co-Founder and Chief Medical Officer of Ellipsis Health, says: “The Ellipsis Health voice technology addresses two critical care gaps in the behavioral health patient journey — upstream screening and downstream monitoring."

Improving mental health screening across demographics

One of Ellipsis Health’s core principles is to break down the structural inequities patients face when accessing care in terms of race, language, and cultural background. For example, a Blue Cross Blue Shield Association study suggests that millennials from majority Black and Hispanic communities are diagnosed with major depression at a lower rate than majority white communities, at 31% and 55% lower, respectively.

To help redress the balance, the Ellipsis Health team is continuing to develop models that reduce systemic biases as much as possible. By ensuring our databases of voice recordings feature strong representation from diverse ages, races, and backgrounds, we can level the playing field for assessments and effective triage. 

Since the US population is so wonderfully diverse, healthcare providers - of which we have a national shortage - cannot possibly deliver culturally competent care to each person. To bridge that gap, we train our machine-learning models on diverse datasets that incorporate differences and nuances in language and ways of speaking.

Providing access to remote care for underserved populations

Cultural demographics aren’t the only barriers to quality medical care. Social determinants of health like inadequate transportation, and financial and time restraints often block pathways to care, in addition to the well-documented shortage of behavioral health clinicians. A study in Health Affairs [4] found that 55% of counties in the continental US do not have a single psychiatrist.

At Ellipsis Health, we envisage a future free of these barriers at the screening and triaging stage. By providing remote solutions through, a digital front door, or by layering our technology onto a telehealth visit or case management call, we can reach patients in underserved areas and those who are unwilling or unable to attend in-person assessments. Through this innovative approach, we hope to offer patients more feasible access to mental health services while reducing the significant burden on overworked physicians.

Inviting patients to take an active role in their mental health monitoring

Traditional screening methods for depression and anxiety usually involve one-to-one sessions with a provider, where the patient fills out a GAD or a PHQ survey to detect signs of anxiety or depression. Even without taking into account the inherent unreliability of self-reported methods, two-thirds of American primary care practices struggle to find mental health specialists they can refer patients to.

These inefficiencies leave patients waiting in limbo for an available session, and that wait is often dangerously long. The National Comorbidity Survey found an average delay of eleven years between the onset of behavioral disorder symptoms and intervention.[1]  During the lag time, mental states can worsen, leading to more difficult and costly treatment that places a greater burden on health systems and payers.

Ellipsis Health has already begun to challenge this deadlock by involving patients in their own treatment. With our simple and engaging interfaces, people can run initial self-assessments using just their voice and monitor their progress over time.  While providers can only see one patient at a time, our models can analyze vast amounts of voice recordings simultaneously, flagging cases of depression and anxiety so healthcare providers can prioritize appointments according to severity.

Test a Use Case with Ellipsis Health

At Ellipsis Health, we know we’re at the beginning of something big. Our mission from day one has been to give mental health the same priority as physical health, and our success so far shows that we’re on the right track.

As we prove the efficacy of our clinical-grade mental health assessments in an ever-increasing number of clinical cases, we are always on the lookout for the next challenge.

So if you represent a healthcare company looking to streamline your workflows while improving outcomes for your patients, we’d love to hear from you!


[1] Wilson, S., & Dumornay, N. M. (2022). Rising Rates of Adolescent Depression in the United States: Challenges and Opportunities in the 2020s. The Journal of Adolescent Health: Official Publication of the Society for Adolescent Medicine, 70(3), 354.

[2] Dupéré, V., Dion, E., Nault-Brière, F., Archambault, I., Leventhal, T., & Lesage, A. (2018). Revisiting the Link Between Depression Symptoms and High School Dropout: Timing of Exposure Matters. The Journal of adolescent health : official publication of the Society for Adolescent Medicine, 62(2), 205–211.

[3] Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of general psychiatry, 62(6), 593–602.

[4] Bishop T., et al. (2016). Population of US practicing psychiatrists declined, 2003-13, which may help explain poor access to mental health care. Health Affairs, 35(7), 1271-1277.


bottom of page