For the most part, behavioral health lacks the objective measurement tools used in physical health care, making behavioral conditions comparatively harder to detect, track and manage. But Ellipsis Health is trying to change that.
Founded in 2017 and based in San Francisco, the company uses machine-learning techniques to analyze speech in order to detect and measure depression and anxiety symptoms. That’s as opposed to traditional questionnaires such as the PHQ-9 and the GAD-7, which Ellipsis CEO and co-founder Mainul Mondal calls subjective and “not enough.”
“You can’t manage what you can’t measure,” Mondal told Behavioral Health Business. “Mental health has been a squishy field, where there are not great measurement tools. Now, the thing we’re doing is [figuring out:] How do we measure something objective, like the way you have your A1C to measure your blood sugar level?”
To do that, Ellipsis settled on speech. Its app uses machine-learning algorithms to assess patient speech for tone of voice, acoustics and language. Deep-learning models then generate an automated assessment of anxiety and depression symptoms to help providers and payers identify and evaluate high-risk patient populations.
The goal is to help people in an affordable, scalable way. Currently, neither adjective is commonly used to describe the behavioral health industry, which is more often characterized by clinician shortages and high, out-of-pocket costs.
“We want to create something that’s going to give the ability for everyone on planet Earth to be able to measure and manage [behavioral health] using voice,” Mondal said. “It’s a super cheap tool … [that] can measure and manage mental health wellness with smart care pathways that never existed before.”
Ellipsis has raised $14 million to date, according to the fundraising tracker Crunchbase.
Rather than work directly with consumers, as many digital behavioral health companies opt to do, the company partners with major payers and health systems nationwide, according to Mondal. He said the company has more than 10 such partnerships to date; however, he was unable to name them due to reasons related to permissions.
Meanwhile, the company’s research partners include heavy hitters such as the Mayo Clinic and Vanderbilt University, among others.
Ellipsis’ partners use the company’s technology to securely monitor patients’ speech before, after and during care. The only prerequisite is that patients consent to having their speech monitored. Mondal detailed a pre-visit use case to explain how the process works.
“You get the link, and then do a sort of voice journaling: So it’d be like, ‘How are you? What are your problems? What brings you here?’” Mondal said. “Then with that two minutes of speech, by the time you see your provider, they know where you are for mental health and wellness.”
The app can also listen to conversations patients have with providers in real time — again, only if given permissions to do so. In both use cases, Ellipsis analyzes the speech, then sends that analysis to the appropriate EHR or case management software for clinical purposes.
The goal is to give providers information to best inform patients’ behavioral health treatment.
“We want to give them the ability to say: ‘Here’s where you are. Here are the recommended actions. We can connect with a therapist. And by the way, here’s the intervention that Ellipsis itself can provide,’” Mondal said. “That needs to be in the hands of every single individual. Health care doesn’t have to be expensive to be good.”