Book Review: All Therapy Books | Slate Star Codex
Published by ☕,
Book Review: All Therapy Books
Posted on November 20, 2019 by Scott Alexander
[Related: CBT In The Water Supply, Scientific Freud, Book Review: Method Of Levels, Different Worlds]
I.
All therapy books start with a claim that their form of therapy will change everything. Previous forms of therapy have required years or even decades to produce ambiguous results. Our form of therapy can produce total transformation in five to ten sessions! Previous forms of therapy have only helped ameliorate the stress of symptoms. Our form of therapy destroys symptoms at the root!
All psychotherapy books bring up the Dodo Bird Verdict – the observation, confirmed in study after study, that all psychotherapies are about equally good, and the only things that matters are "nonspecific factors" like how much patients like their therapist. Some people might think this suggests our form of therapy will only be about as good as other forms. This, all therapy books agree, would be a foolish and perverse interpretation of these findings. The correct interpretation is that all previous forms of therapy must be equally wrong. The only reason they ever produce good results at all is because sometimes therapists accidentally stumble into using our form of therapy, without even knowing it. Since every form of therapy is about equally likely to stumble into using our form of therapy, every other form is equally good. But now that our form of therapy has been formalized and written up, there is no longer any need to stumble blindly! Everyone can just use our form of therapy all the time, for everything! Nobody has ever done a study of our form of therapy. But when they do, it’s going to be amazing! Nobody has even invented numbers high enough to express how big the effect size of our form of therapy is going to be!
Consider the case of Bob. Bob had some standard-issue psychological problem. He had been in and out of therapy for years, tried dozens of different medications, none of them had helped at all. Then he decided to try our form of therapy. In his first session, the therapist asked him "Have you ever considered that your problems might be because of [the kind of thing our form of therapy says all problems are because of]?" Bob started laughing and crying simultaneously, eventually breaking into a convulsive fit. After three minutes, he recovered and proceeded to tell a story of how [everything in his life was exactly in accordance with our form of therapy’s predictions] and he had always reacted by [doing exactly the kind of thing our form of therapy predicts that he would]. Now that all of this was out in consciousness, he no longer felt any desire to have psychological problems. In a followup session two weeks later, the therapist confirmed that he no longer had any psychological problems, and had become the CEO of a Fortune 500 company and a renowned pentathlete.
Not every case goes this smoothly. Consider the case of Sarah. Sarah also has some standard-issue psychological problem. She had also been in and out of therapy for years, tried dozens of different medications, none of them had helped at all. Then she decided to try our form of therapy. In her first session, the therapist asked her "Have you ever considered that your problems might be because of [the kind of thing our form of therapy says all problems are because of]?" Sarah said "No, I don’t think they are." The therapist asked "Are you sure you’re not just repressing the fact that they totally definitely are, for sure?" As soon as Sarah heard this, she gasped, and her eyes seemed to light up with an inner fire. Then she proceeded to tell a story of how [everything in her life was exactly in accordance with our form of therapy’s predictions] and she had always reacted by [doing exactly the kind of thing our form of therapy predicts that she would], only she was repressing this because she was scared of how powerful she would be if she recovered. Now that all of this was out in consciousness, she no longer felt any desire to have psychological problems. In a followup session two weeks later, the therapist confirmed that she no longer had any psychological problems, and had become the hand-picked successor to the Dalai Lama and the mother of five healthy children.
Previous forms of therapy have failed because they were ungrounded. They were ridiculous mental castles built in the clouds by armchair speculators. But our form of therapy is based on hard science! For example, it probably acts on synapses or the hippocampus or something. Here are three neuroscience papers which vaguely remind us of our form of therapy. One day, neuroscience will catch up to us and realize that the principles of our form of therapy are the principles that govern the organization of the entire brain – if not all of multicellular life.
II.
Maybe I’m being unfair here. I’m basing this off a small sample of therapy books (five textbooks I can think of, plus scattered papers on psychodynamic and psychedelic therapies), and only a subset are quite this bad.
But my basic confusion is this: I work in a clinic with about ten therapists. Some are better than others, but all of them are competent. I send my patients to them. In a few hundred patients I’ve worked with, zero have had the sudden, extraordinary, long-lasting change that the therapy books promise. Many have benefited a little. A few would say that, over the course of years, their lives have been turned around. But sudden complete transformations? Not that much.
Of course, this fits with the therapy books’ perspective. My colleagues practice normal therapy. Sometimes it’s from a boring old school like CBT; other times it’s "eclectic" or "supportive" or any of the other words we use to describe what we’re doing when we don’t know what we’re doing. So maybe there are two sets of therapies: boring old therapies that ordinary people practice, and exciting new therapies that people write glowing books about. And maybe the first set really don’t work (or work only a little), and the second set really is that good.
The problem is, the boring old therapies that everybody uses nowadays inspired equal excitement when they first arose. This is the point that I make in CBT In The Water Supply, and that Oliver Burkeman makes more cogently in Why CBT Is Falling Out Of Favor. Look at therapy books from the 1990s, and they were all about how CBT was a new miracle therapy that would cure your anxiety forever in a few sessions. From a cognitive therapy book:
[When I first learned about cognitive-behavioral therapy, I thought] depression and anxiety seemed far too serious and severe for such a simplistic approach. But when I tried these methods with some of my more difficult patients, my perceptions changed. Patients who’d felt hopeless, worthless, and desperate began to recover. At first, it was hard to believe that the techniques were working, but I could not deny the fact that when my patients learned to put the lie to their negative thoughts, they began to improve. Sometimes they recovered right before my eyes during sessions. Patients who’d felt demoralized and hopeless for years suddenly turned the corner on their problems. I can still recall an elderly French woman who’d been bitterly depressed for more than fifty years, with three nearly-successful suicide attempts, who started shouting "Joie de vivre! Joie de vivre!" ("joy of living") one day in my office. These experiences made such a strong impact on me that I decided my calling was in clinical work rather than brain research. After considerable soul-searching, I decided to give up my research career and become a full-time clinician. Over the years, I’ve had more than 35,000 psychotherapy sessions with depressed and anxious patients, and I’m every bit as enthusiastic about CBT as when I first began learning about it.
But look at therapy books now, and they’re all people saying "Sure, CBT barely outperforms placebo…but what about this exciting new therapy which blows CBT out of the water?"
Studies reflect this decline:
…with the average studied effect size of CBT shrinking from 2.5 to 1.0 over the course of a generation. People have come up with various explanations for this. Maybe therapist quality is falling – when CBT was the hot new thing, you had to be a really plugged-in up-to-date therapist to have heard about it and to make the effort to retrain in it, so only the best therapists would practice it, but now it’s the default therapy used by everyone who’s just clocking it in. Maybe placebo effect is falling – when people viewed it as an astounding miracle therapy, it got astounding miracle results, but now that it’s lost its luster nobody takes it seriously anymore. Maybe its ideas are spreading, so that patients come into their first session already aware of CBT insights and inoculated against them. Or maybe it’s like all science, where the first studies are done quickly by true believers, and the later studies are done carefully by the Cochrane Collaboration, and so the level of hype naturally goes down.
These explanations have different practical implications. If it’s all about therapist quality and placebo expectations, then you should go get the exciting new therapies described in therapy books, since their unusually-qualified therapists and unusually-high expectations will deliver you the miracle cure you’re looking for.
If it’s just that study quality gets better and better until we realize how crappy the exciting new therapies really are, you might as well get the boring old therapies. At least insurance probably covers them.
And they also have different philosophical implications. If it’s all about therapist quality and placebo expectations, then even if it’s hard to deliver high-quality therapy consistently at scale, it means high-quality therapy is a thing. It means that if enough factors go right at once, therapy can be the kind of powerful tool that cures someone’s life-long psychiatric issues in a few sessions with a high success rate. If this is true it would be fascinating. It would be like saying that bananas cure cancer, but only if they’re really fresh bananas. Even if there are practical issues in getting every cancer patient a banana that’s fresh enough, you still want to take a step back and think "Whoa, what’s up with this?"
I can only say that I’ve had a few patients try the exciting new therapies, and none of them have reported miracle cures. They’ve all maybe gotten a little better over long periods, same as the boring old therapies. This makes me think it’s more likely that early results from the exciting new therapies get oversold, not that some combination of therapist skill and excitement makes them go shockingly well. And the Efficient Market agrees with my low estimation, given that therapists aren’t rushing to learn these new strategies and patients aren’t rushing to use them.
But the therapy books still confuse me. They’re full of stories of incredible instant cures, with the authors assuring us that these are all real and typical of their experience. How can you get this from merely "stretching the truth", as opposed to outright data falsification? Are therapy book authors blatantly lying? I try to have a really low prior on this sort of thing, but I’m not sure.
Therapy books are often written by the researcher who invented the therapy. I imagine if you invent a therapy yourself, then it perfectly fits your personality and communication style, you believe in it wholeheartedly, and you understand every piece of it from the ground up. You’re also probably a really exceptional and talented person who’s obsessed with psychotherapy and how to make it better. So maybe they get results nobody else can replicate?
But that still raises the philosophical implication of it being possible, for somebody, to consistently produce dramatic change through therapy. This still bothers me a lot.
III.
Most therapy books share some assumptions, so deep as to be unspoken: current problems serve some purpose related to past traumas.
Different therapies take this in different directions. Some view problems as a passive residue of past traumas: for example you were abused as a child, that filled you with stress and rage, and now you take that out on other people and yourself. Others view them as maladaptive learning from past trauma: for example, you were abused as a child, that taught you that other people would hurt you if you opened up to them, so you never open up to anybody. I don’t know the official name for this, but let’s call it historicism: symptoms are the result of something that happened in a patient’s life history.
Some weak forms of historicism are obviously true. Many (though not all) phobias began with a clear incident where the patient was endangered by the phobic object; someone mauled by a dog as a child who then has cynophobia as an adult is hardly a medical mystery. Many (though not all) depressions are precipitated by some depressing event. And post-traumatic stress disorder has the historical perspective right there in the name; at the very least, going through trauma dysregulates something inside you. But it’s a long way from there to saying that a patient’s psychosomatic blindness is caused by persistent shame at having seen their parents having sex thirty years earlier, or something like that.
And some therapy books go beyond historicism into purposefulism: symptoms serve some quasi-logical purpose relating to the life history. I recently read a therapy book that included a case like this. Bob had a history of failing at work. He would go from job to job, making various mistakes and doing crappy work until he got fired. He went to a therapist for help. During the therapy, it came out that Bob’s abusive father had always pushed him really hard to succeed. The therapist suggested that maybe Bob failed at work to send a message to his father; ie to prove that his father’s abusive parenting had been a bad idea and would not make Bob successful. The therapist asked Bob to imagine confronting his father about this. After he worked through his anger at his father, Bob was able to succeed at work. In this story, the apparently dysfunctional symptom (failing at work) ended up having a legible purpose within Bob’s life history (it helped him send a message to his father). Only by teasing out the purpose and finding some other way to achieve it could the dysfunctional behavior be prevented.
A non-historical, non-purposeful account might argue that Bob failed at work because he was bad at work. Maybe he was bad at the specific jobs he was holding (in which case he should get more training). Maybe he was bad at social skills (in which case he should learn to communicate better). Maybe he had ADHD and kept getting distracted (in which he should get treatment for ADHD). In any case, him being bad at work isn’t related to any past traumas or serving any hidden purposes. It’s just an unfortunate fact.
I am constantly worried by the history of how many things we historically applied historical-purposeful reasoning to, totally confident at the time that our explanations made sense – which we now know are not historical-purposeful at all. Psychologists "knew" that autism was caused by distant mothers, and schizophrenia by overbearing mothers, right up until we discovered both conditions are about 80% genetic. And when they "knew" these things, they were able to come up with long lists of how exactly each individual patient fit the mold, and reported great progress by helping patients overcome their maternal attachment issues. Back when homosexuality was considered a disorder, historical-purposeful therapists would tell gay people patients they must be so angry at their mother that they had sworn off all female companionship and switched to men instead as a way of sending her a giant "F–K YOU" message; while homosexuality is mostly not genetic, few people today think this is a plausible explanation.
I sometimes see if I can come up with these kinds of historical-purposeful accounts of my patients’ symptoms. These always fit into place freakishly well – so well that either the historical-purposeful perspective is completely true, or there is some very strong bias that makes it extra-convincing despite its falsehood. But we already know there’s some very strong bias that makes it extra-convincing despite its falsehood! That bias must have been at work in all the therapists who applied historical-purposeful narratives to autistics, schizophrenics, and gays! At some point I notice the road I’m on is littered with skulls and start wondering if I should reconsider.
All therapy books propose an answer: the proof is that the patients get better. But my patients do not get better. When I tell them the historical-purposeful accounts I have devised for their symptoms, they usually shrug and say it sounds plausible and they’ve thought along those lines before, but what are they going to do? When I try all the exciting new therapies on them, they just sort of nod, say that this sounds like an interesting perspective, and then go off and keep having symptoms. It’s very rude!
I’ve told this story before: when I was a teenager, I got really into pseudohistory for a while. What snapped me out of it wasn’t the sober historians, who totally went AWOL on their job of explaining why they were right and the whackos were wrong. It was that a bunch of mutually exclusive pseudohistories all sounded equally plausible: the Pyramids couldn’t have been built by Atlanteans and Lemurians and mole-people! At that point I was able to halt, melt, catch fire, and realize there was something really wrong with my reasoning processes, which I continue to worry about and work on twenty years later.
I bring this up because I’m going to be reviewing some specific psychotherapy books. Each of them on their own can be convincing. But they should be taken in the context of All Therapy Books, which as a category are pretty worrying.
This entry was posted in Uncategorized and tagged book review, psychiatry. Bookmark the permalink.