AI Could Help Alleviate America’s Approaching Mental Health Crisis

The U.S. is experiencing a chronic shortage of medical professionals. Of course, this includes physicians. But psychiatrists will also be in scant supply. In five years, the U.S.’s supply of psychiatrists will be 15,600 short of the demand from patients, according to a 2017 study from the National Council for Behavioral Health.

AI Can Lend a Helping Hand

Although the technology is relatively new, AI applications have already been implemented in medical settings to help diagnose diseases, clarify treatment plans, and study radiology images.

These applications are quite straightforward, while the field of psychiatry is as human and personal as it gets. With mental health, many things don’t “make sense”, and we employ humans to help clarify what may be going on in our brain. Still, psychiatry-based AI does not need to have the emotional intelligence, experience, or perspective of a human being to assist psychiatrists. Experts say that AI can analyze data and find patterns that signal an impending problem.

The fact of the matter is that doctors and providers these days don’t get to spend much time with their patients. We all have experienced a long waiting room time, only to be seen by the doctor for fifteen minutes. On top of that, Peter Foltz says, “Patients … may be seen by a clinician [only] once every three months or six months.” Foltz is a research professor at the University of Colorado Boulder who is excited about the prospect of AI in psychiatry.

With AI, Foltz says, clinicians can spend quality time with their patients and direct them to additional resources as needed. Clinicians can use the insights from AI-enabled data analysis to make diagnoses, which helps the patient start treatment sooner.

The Potential of Mobile Apps in Mental Health

Foltz is even more excited about the potential of mobile apps and other software to use AI to monitor patients remotely. These medical applications can alert providers immediately while keeping an eye on the patient in between appointments and doses. This is an area of major impact because research shows that patients need regular check-ins, especially if they’re in mental distress or suicidal.

According to psychiatrists, it’ll be another five to ten years before clinics start routinely using AI algorithms, so mobile apps offer a great testing ground for AI. Woebot is one such AI-enabled mental health app; it reminds you to track your mood and offers a chatbot that utilizes AI and cognitive behavioral therapy.

Henry Nasrallah is a psychiatrist at the University of Cincinnati Medical Center who’s interested in AI’s place in psychiatry. Nasrallah says speech and language are two areas that AI can tackle well. He says speech and mental health are closely related; for example, fast speech can be a sign of mania, while monotonous speech can indicate depression. On the other hand, a psychiatrist may want to look into schizophrenia when their patient uses disjointed words.

Often, these speech patterns may not be picked up by clinicians, but AI isn’t an overworked medical professional. An algorithm can possibly pick up on these irregular speech patterns much faster than a psychiatrist.

Foltz and his research team are working on a mobile app that journeys patients through a series of repeatable phrases. Patients tell a story or they answer questions about their mental and emotional state. These responses are fed to an AI algorithm that analyzes the speech and verbiage for mental distress. The algorithm also compares the newest responses to the patient’s older responses, and it compares these responses against the responses from the rest of the users.

The researchers tested the app’s performance on 225 patients who live in rural Louisiana or Northern Norway. These two areas suffer from a lack of access to adequate mental health care. Results showed that the app was at least as accurate as psychiatrists in identifying the speech-based qualities of mental distress.

Technology’s Effect on Mental Health

John Torous is the director of digital psychiatry at Boston-based Beth Israel Deaconess Medical Center. He’s also serving as the chair of the American Psychiatric Association’s Committee on Mental Health Information Technology.

Torous warns that AI is only as good as the data it’s trained on. This is known to be one major caveat of AI algorithms; they perform extremely poorly when they’re trained on low-quality data, and psychiatry is one field that cannot suffer from any mistakes caused by AI. The results could be disastrous for patients.

Torous also points out that mental health professionals have not been acquiring or generating data long enough or well enough to use as training data for an algorithm. In fact, programming an AI algorithm is going “to be an uphill challenge,” according to Torous.

But ophthalmologist Alastair Denniston sees AI in a different light. Denniston says that technology can alleviate many monotonous and technological responsibilities from providers, giving them time and energy back to focus on the human aspects of medicine.

Making Medicine More Human

Ultimately, Denniston says, AI can free up time for providers to spend communicating more effectively and more humanely towards their patients.”Rather than being diagnostic machines… [doctors can] provide some of that empathy that can get swallowed up by the business of what we do,” adds Denniston.

And sometimes, all a mental health patient needs is to see and feel that someone cares.