How Artificial Intelligence Can Disrupt The Mental Health Industry

Updated: Feb 25

Each year, millions of Americans with mental illness struggle to find care. It’s estimated that at least half of the 36 million people at educational institutions and workplaces struggling with mental health conditions in the United States go without any treatment.

In 2016, NAMI, the National Alliance on Mental Illness, found that people with mental illness in the US experience significant barriers to finding affordable, accessible mental health care. Barriers include high rates of denials of care by insurers, high out-of-pocket costs for mental health care, difficulties accessing psychiatric medications and problems finding psychiatrists, therapists, and other mental health services. These barriers exist whether the person is covered by private insurance or by a public insurance plan.

The parity law does not require insurers to provide mental health benefits—rather, the law states that if mental health benefits are offered, they cannot have more restrictive requirements than those that apply to physical health benefits. A person’s eligibility to claim mental health insurance coverage demands to be evaluated on a case-by-case basis. Depending on the particular insurance plan's deductible, in most cases, people have to pay out of pocket before the insurance company will pay any claims.

Globally, the total number of people suffering from mental health disorders is believed to be over 600m, according to the World Health Organization (WHO). A survey in 2018 by the American College Health Association found that 63 percent of US college students reported feeling overwhelming anxiety over the previous year, 42 percent said they felt so depressed it was difficult to function over the previous year, and 12 percent seriously considered suicide. Among U.S. adults aged 18–44, mood disorders such as depression are now the third most common reason for hospitalizations. Considering these facts, it is not surprising that the Google searches for anxiety, depression, and similar mental disorders are way up.

According to the National Institute of Mental Health (NIMH), one in five adults in the United States, that is 18 percent of the population, experiences some type of mental health disorder every year. As stated by the consulting firm McKinsey & Company, one in two will experience a mental illness at some point in their lives. Mental illness not only reduces an individual’s quality of life, but it also links with increased health spending.

“How are you doing today?” This might seem like a simple question a caring friend would ask. Yet, it can also be the start of a conversation with your virtual therapist. Advancements in artificial intelligence (AI) are bringing therapy services to more people who need it. AI will facilitate individualized treatment planning and improve counseling and therapy outcomes.

The University of Southern California’s Institute for Creative Technologies (ICT) launched a virtual therapist named Ellie. It was designed to treat veterans experiencing depression and post-traumatic stress syndrome. Ellie can detect not only words, but also nonverbal cues and the software automatically tracks and analyzes in real-time facial expressions, body posture, acoustic features, linguistic patterns and higher-level behavior descriptors (e.g. attention and fidgeting). Nonverbal signs are very important in therapy, yet it can be subtle and difficult to pick up even for trained and experienced therapists. The team led by Albert “Skip” Rizzo and Louis-Philippe Morency developed their virtual therapist in order to gather and analyze multisensory data to offer more precise treatment responses.

Studies show that we react to avatars as if they were real humans. For example, Mel Slater and his team observed this conduct when they orchestrated experiments where people were aware that they were interacting with androids, yet they related to them as if they were real. However, the Turing test, developed by Alan Turing in 1950—a test of a machine's ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human—is still to be passed by a machine.

Healthynox, a healthtech startup located in Cambridge, MA is developing software to provide high-quality, affordable mental health services, created for educational institutions like colleges and modern workplaces that care about their stakeholders. The startup created a mental health platform that provides the full spectrum of therapy services by combining virtual therapy powered by artificial intelligence (AI) with digital counseling sessions with certified best-in-class therapists.

Modern Natural-Language Understanding (NLU) and Processing (NLP)

To have practical clinical utility in mental health care, an AI software must encompass machine-learning systems capable of processing gigantic volumes of structured data, Natural Language Understanding (NLU) and Natural Language Processing (NLP) and software capable of mining unstructured data such as narrative text in electronic health records and medical imaging data.

Natural Language Processing (NLP) allows for the tagging, parsing, and extracting of information from text. To give some context to the technology available, it should be emphasized that with the help of machine learning tools it is possible to utilize data as a source of insights into the epidemiology (study & analysis of distribution, patterns, and determinants of health and disease determinants) of emotions and mental illness. Natural Language Processing (NLP) and Natural-Language Understanding (NLU) can be, for instance, used to identify if the author of a text message showed indications for a mental health disorder by comparing their words to a lexicon of terms that were more common among depressed people, such as mentions of medications, depressive symptoms and words related to disclosure.

By extracting meaningful information from large datasets, NLP and NLU tools can provide therapists the information they need to detect complex patients. In the future, NLP, NLU and other machine learning tools are the key to better clinical decision support and patient health outcomes.

By applying techniques such as sequence-based conversation models, language model comparisons, message clustering, and psycholinguistics-inspired word frequency analyses, AI can develop personalized conversation strategies that are associated with better treatment.

Market Sizing with a Focus on Educational Institutions and Workplaces

The same study that found that over half (63 percent) of college students are feeling ‘overwhelming anxiety’ during their studies reports that almost 30 percent of students experience feelings of hopelessness, and 42 percent report feeling so depressed that it's difficult to function.

There were 4,298 degree-granting postsecondary institutions in the U.S. as of the 2017-2018 school year, according to the National Center for Education Statistics. U.S. college enrollment statistics for public and private colleges show that in 2020 there are 14.69 million students enrolled in public universities and 5.24 million enrolled in private institutions, resulting in 19.93 million students overall in the U.S. According to NCES, between 2017 and 2028, undergraduate enrollment in 4-year institutions is projected to increase by 2 percent (from 10.8 to 11.1 million students).

63 percent of US college students reported feeling overwhelming anxiety over the previous year, which means that at least 12.56 million students suffer from anxiety alone. According to the National Institute of Mental Health (NIMH), one in five adults in the United States, that is 18 percent of the population, experiences some type of mental health disorder every year.

In the education system, at least 12.56 million students would benefit from a virtual mental health assistant. According to the National Institute of Mental Health (NIMH), one in five adults in the United States, that is 18 percent of the population, experiences some type of mental health disorder every year.

In 2019, about 130.6 million people were employed on a full-time basis. That means that in the US, 23.6 million workers would benefit from a virtual mental health assistant; thus, the total addressable market (people suffering from a mental illness at educational institutions and workplaces) consists of at least 36.16 million people who should be classified as potential consumers/end-customer.

Artificial Intelligence Will Disrupt The Mental Health Industry

A virtual mental health assistant can completely transform the way we treat mental illness, therapy in general, and how people live. By leveraging AI & Machine Learning programs including NLU and NLP, the technology has the potential to find its application at every educational institution, workplace, and maybe even every home in the world.

In terms of the risk factors perceived for the use of digital software and AI in mental health care, we should reflect on the fact that the first steam engines also caused mass-feelings of unwantedness, as the machines were capable of assisting or even replacing humans, but no one today would imagine turning back the clock. Artificial intelligence in mental health care makes everyday lives significantly better and brings many benefits to educational institutions and workplaces alike. Change and advancement are inevitable.