Ieso has focused on depression and generalized anxiety disorder, and has used data-driven techniques to boost recovery rates. The recovery rate for depression is 62% in 2021, compared to a national average of 50% and a national average of 73% for generalized anxiety disorder.
Ieso says it has focused on anxiety and depression because they are two of the most common conditions. They respond better to CBT than other disorders. It is not clear how far the clinic can go, but it will focus on more conditions.
Ieso has not yet studied the direct impact of NLP on the efficiency of care, but in theory, using artificial intelligence to monitor quality frees up clinicians to see more clients because better therapy means fewer unproductive sessions.
Freer says they can treat between 80 and 90 clients with 1,000 hours of therapy time. We want to move that needle and ask if you can treat 200, 300, or 400 clients with the same amount of therapy hours.
Unlike Ieso, Lyssn does not offer therapy. Quality control and training are what it provides to other clinics and universities in the UK and the US.
In the US, a program in California wants to monitor the quality of care given by its providers. The company is working with the University of Pennsylvania to set up a network of therapists with its technology.
Trent Psychological Therapies Service is an independent clinic that is commissioned by the National Health Service to provide mental-health care. Trent is still using the software. The clinic had to work with Lyssn to make it recognize British regional accents because the model was built in the US.
Dean Repper believes that the software could help therapists standardize best practices. He thinks therapists who have been doing it for a long time would get the best outcomes. They don't necessarily. Repper compares it to learning to drive a car, where you are taught to do a number of safe things. After a while, you stop doing some of those safe things and maybe get fined for speeding.
Improving, not replacing.
The purpose of the artificial intelligence is to improve human care. The lack of quality mental-health care is not going to be fixed by short-term fixes. Reducing stigma, increasing funding, and improving education are some of the things that will be needed to address that problem. Many of the claims for artificial intelligence are dismissed by Blackwell. He says there is a lot of hype.
There is a lot of buzz about things like chatbot therapists and round-the-clock monitoring by apps. Most of the tech is years away and never going to happen.
It is not about well-being apps or anything like that. Putting an app in someone's hand that says it's going to treat their depression probably serves only to inoculate them against seeking help.
One problem with making psychotherapy more evidence-based is that it requires therapists and clients to open up their private conversations. Will therapists object to having their performance monitored?
Repper expects some reluctance. He says that the technology represents a challenge for therapists. It is as if they have someone else in the room for the first time. Trent PTS is only using Lyssn's software with trainees who expect to be monitored. Repper thinks that therapists may accept the monitoring if they are used to it. More experienced therapists might need to be convinced of its benefits.
Imel, who used to be a therapist himself, says that the technology is not used as a stick but as support. He thinks people will like the extra information. He says it is hard to be alone with your clients. It can be difficult to improve if you sit in a private room with another person for 20 or 30 hours a week without getting feedback from your colleagues.
Freer agrees. At Ieso, therapists discuss feedback with their supervisors. The idea is to let therapists take control of their professional development, showing them what they are good at, and not so good at, things that other therapists can learn from.
There is a clear potential for learning things about therapy that are revealed only by mining large data sets. A meta-analysis that pulled together over 1,000 hours of therapy without the help of artificial intelligence was published in the year of 2019. He says thatLyssn processes that in a day. New studies analyzing tens of thousands of sessions have been published.
In a paper published in JAMA Psychiatry, Ieso researchers described a deep-learning model that was trained to categorize utterances from therapists in more than 90,000 hours of CBT sessions with around 14,000 clients. It was learned that different phrases and short sections of conversation were instances of specific types of CBT-based conversation, such as checking the client's mood, setting and reviewing homework, discussing methods of change, planning for the future, and so.
