How Big Data (Artificial Intelligence) Could Improve Your Next Therapy Session

Why my experience in therapy and working at The Crisis Text Line has taught me to trust the ARTIFICIAL INTELLIGENCE (BIG DATA).

About eight months ago, I sat in a counselor’s office at my university and we closed the therapy session with her asking me how I was doing.

“I’m doing a lot better,” I said. “Thanks for all the help you’ve been. It means a lot.”

But in truth, I was lying. I was actually doing worse. Life had gotten harder and I was faced with unorthodox challenges and grief that inevitably made me barely able to sleep. I was drinking too much, taking out my anger on my friends, and behaving in a multitude of other self-destructive ways.

I was more depressed than I was the first time I saw the counselor, but I didn’t want her to know it. I didn’t want her to see it as an affront to her competence or aptitude as a social worker. The truth was she cared a lot, and I knew it. She was very competent, asked all the right questions, gave me great coping mechanisms, and it showed.

But even if you have the best social worker in the world, you can still get worse. And I did, but I didn’t want her to take it personally.

I remembered all this when I read, “What Your Therapist Doesn’t Know” by psychologist Tony Rousmaniere in The Atlantic. In the article, Rousmaniere explains his initial strong resistance to big data in his practice:

Psychotherapy is unlike any other field, I’d thought, with the arrogance that comes from being untested. We work in a human relationship. What we do can’t be measured.

All of these ideas changed, however, when Rousmaniere, as a 34-year-old young therapist, had a patient named Grace. Grace was a recovering heroin addict who had been clean for six months. She was an unemployed single mother who’d been in a plethora of abusive relationships, and she was in the process of putting her life together and retaining custody of her son.

Everything seemed to be going well. She had a boyfriend who respected her, and she attended NA meetings regularly. When prompted for feedback, Grace assured Rousmaniere that their sessions were productive, but her responses were often rushed with a forced smile. When Rousmaniere told his supervisor about Grace’s process, his supervisor was cautious:

“Getting clean is hard,” she told me, “but staying clean is harder.”

It became clear that Rousmaniere’s supervisor was right: Grace no-showed three appointments and relapsed on heroin when she reappeared. Everything in her life was falling apart. She had no jobs, no boyfriend, and was using regularly.

A couple of months later after relapsing, Grace died, and her son was in foster care. The events led to a crisis-like episode of introspection for Rousmaniere. He asked himself, “What could I have done differently? How could I become a more effective therapist?” He began to recall a talk about whether psychotherapy could benefit from data and analytics. After Grace died, he found himself open-minded to infusing more data into his practice.

Rousmaniere goes on in the article to explore whether performance feedback, which gives therapists awareness of how well they’re doing their job, would improve psychotherapy’s effectiveness. In a highly challenging and sensitive field like psychotherapy, incorporating objective feedback and data is extremely difficult. Especially in confidential and sensitive settings, therapists work in private and sheltered environments that don’t allow for objective feedback.

Clients also might sugarcoat their responses in a misguided effort to protect their therapists. Matt Blanchard and Barry Farber of Columbia University asked 547 patients in an online therapy survey whether they were dishonest about the effectiveness of their therapy and the therapeutic relationship.

What they found was alarming: 93% of respondents reported having lied to their therapist. 72.6% reported lying about at least one therapy-related topic. The most common motives for lying included the following rationales:

Romantic and sexual feelings for a therapist.

“I wanted to be polite.”

“I wanted to avoid upsetting my therapist.”

“This topic was uncomfortable for me.”

To counter the dishonesty that often accompanies therapy feedback, Margarita Tartakovsky describes an approach known as feedback-informed treatment or FIT. The approach uses data from formal feedback to enhance client well-being and decrease dropout rates and no-shows. The scale allows clients to be honest about bad negative feedback to their therapists, and for therapists to collect accurate data to help determine how their clients are really doing.

Rousmaniere himself used FIT to great effect: He had a client named June who reported that the skills Rousmaniere was teaching her were good. He had a patient that he thought was doing a lot better, but was actually doing worse. She told him she was feeling better, but once she put in feedback on an iPad, he learned that she had been doing worse, despite his best efforts. He incorporated video feedback of his sessions to help him improve as a therapist, and started acting more like an “equal partner” and less like an authority figure.

Here, big data helped Rousmaniere’s patient get better.

I’ve seen first-hand how challenging it is to use metrics and data when I first started out as a volunteer in my school’s suicide hotline and a crisis counselor at Crisis Text Line. I’ve had conversations with people on the phone that made it seem like they’ve gotten a lot better, only to have them call back several times later as “chronic callers.” My supervisors would make management plans to try to get them to seek more sustainable and effective help because we weren’t therapists, we were trained volunteers.

At the time, I took my failures as an insult. I worked on the suicide hotline on the phone and had trouble believing that I was being helpful at all.

It wasn’t until I worked on the Crisis Text Line that my opinion on management plans and data indicating ineffectiveness changed. Once I started working with the anonymity behind the Crisis Text Line, texters seemed to be a lot more honest about how they were feeling after each conversation. Some said they felt exactly the same, if not worse. Some told me that they felt a lot better acknowledging their strengths.

The Crisis Text Line also employs the use of extensive data that helps me improve. The research shows that the best counselors need to adapt their language to the people they’re talking to. Using the language of the texter themselves has been an essential point of feedback.

The Crisis Text Line also uses algorithms to boost the effectiveness of text messaging for crisis counseling. As the Crisis Text Line mostly targets young people, 75% of whom are under 25, the anonymity and confidentiality from social judgment are important for younger texters. Many texters have told me that this is the first time they’ve ever told people about their problems.

Instead of being put on the spot, texters have expressed to me the importance of being given time to think and feel before sending a text message.

Even though National Suicide Prevention Awareness Month is observed in September, suicidal ideation peaks in April and May, according to data from the Crisis Text Line.

As a formula, an effective crisis counseling situation involves building rapport, exploring support systems, identifying a goal, and collaboratively problem-solving and making a plan. Crisis counseling isn’t therapy, but it helps de-escalate patients.

Through the Crisis Text Line, big data has also shown trends about which issues spike and when. For example, even though National Suicide Prevention Awareness Month is observed in September, suicidal ideation peaks in April and May, according to data from the Crisis Text Line. The data has also revealed trends that one in five users under the age of 13 mention self-harm.

Made with Visme Presentation Maker

Made with Visme Infographic Maker