Therapy: The Gold Standard?
People often think therapy is the gold standard for mental health treatment. It’s not entirely unwarranted. There is indeed a decent body of evidence that supports the effectiveness of therapy. But I think this is overrated. Here, I argue that therapists are not uniquely effective in helping people with their mental health problems. There are several suggestive lines of evidence that support this:
It’s unclear why therapy works at all.
Talking with a non-therapist is as effective as talking to a therapist.
Guided self-help is as effective as face-to-face therapy.
Amateur therapists are as effective as experienced therapists.
Therapists overestimate their own effectiveness.
I. No Consensus
One might think that if therapists were especially effective in helping with mental health problems, this would have something to do with their specialized training. The problem, though, is that different therapists are trained in different ways. Some are taught to focus on clients’ childhood upbringings, some are taught to focus on getting clients to challenge irrational thoughts, some are taught to regard their clients with unconditional positive regard. But as it turns out, none of these strategies is uniquely effective! This finding—that most therapies produce roughly equivalent outcomes—is called the Dodo Bird Verdict, named after the Dodo in Alice in Wonderland, who declares, “Everybody has won, and all must have prizes.”
The Dodo Bird Verdict has led some researchers to form new hypotheses about why therapy works. Specifically, they believe the success of the different therapy modalities can be explained by a few common factors. E.g. the structured environment where people can express their concerns, the supportive relationship between therapist and client, and the client’s expectation that their life will get better. But if mental health improvements are driven primarily by these common factors, then this undermines the view that therapists would be uniquely effective. Anyone else providing these benefits would help their clients just as much. And indeed some evidence suggests exactly that.
II. Professionals vs. Non-professionals
In one classic (but small) study, college students experiencing loneliness, anxiety, and depression were assigned to meet twice-weekly over 3-4 months with either trained therapists (with an average 23 years of experience) or college professors selected for their warmth and trustworthiness (but with no mental health training). Researchers compared how much each group improved their mental health over time. What did they conclude?
Patients undergoing psychotherapy with college professors showed, on the average, quantitatively as much improvement as patients treated by experienced professional psychotherapists.
Additionally, in a classic meta-analysis from 1979, researchers examined 42 studies comparing the effectiveness of mental health professionals with paraprofessionals (who had relatively little mental health training). For example, one study assigned schizophrenic patients to group therapy sessions led by either 1) college undergraduates with no training or 2) mental health professionals with 5-17 years of training. The authors of this study concluded:
By comparison to an untreated control group the lay therapists achieved slightly better results than psychiatrists and psychiatric social workers doing group therapy with similar patients.
And what about the results of the meta-analysis?
Paraprofessionals achieve clinical outcomes equal to or significantly better than those obtained by professionals. In terms of measurable outcome, professionals may not possess demonstrably superior clinical skills when compared with paraprofessionals.
Follow-up work appears to have confirmed these findings. So it seems the training that the professionals received was not enough to make them any more effective than mere amateurs. Simply having someone—anyone—to talk to may be sufficient to improve a person’s mental health.
III. Professionals vs. Self-help
It might actually be possible to skip the talk sessions entirely. For example, one meta-analysis has found that “guided self-help” programs produce outcomes comparable to standard face-to-face therapy for depression. These guided self-help programs are a heterogenous bunch, but they typically involve a structured set of steps for the client to follow at home to treat their mental health problems. Sometimes a therapist will be available via email to answer clients’ questions. Sometimes there are brief weekly phone calls to check on the clients’ progress. What’s common among them is that there is relatively limited contact with the therapist. Yet these self-help programs work nonetheless!
Granted, the meta-analysis did find that unguided self-help was less effective. So maybe it’s helpful to have someone in the loop to support a client and hold them accountable. But evidently, it’s not necessary for therapists to have substantial involvement in the process.
IV. Experts vs amateurs
One might also think that if the training therapists received was uniquely helpful, then therapists with more experience would be substantially more effective than those with limited experience. This is not borne out by the data. In one study, researchers compared the effectiveness of therapists who varied in experience, ranging from half a year to 17 years of practice. Shockingly, the therapists with more experience had worse client outcomes than those with less experience. Evidently, all that practice hadn’t helped the more experienced therapists treat their clients any more effectively.
Granted, the overall literature has mixed results. Other studies have found null effects or small positive effects of therapist experience on client outcomes. This is not very encouraging though. Experienced therapists should be beating the pants off of amateurs! Compare this to other domains. For example in sports competitions the professionals reliably outperform the beginners, and we can clearly see it in the metrics. Usain Bolt objectively sprints faster than anyone else. Michael Phelps swims faster. And it’s not just physical feats either. No amateur chess player would be able to reliably beat Magnus Carlsen. And yet, when it comes to helping people improve their mental health, it’s a toss up whether the therapist you receive does any better than someone who’s almost a complete novice. This should raise your skepticism about whether there’s some special sauce that comes from training as a therapist.
V. Overconfidence
Of course, therapists will say that they have special insight into treating people with mental health problems. But there’s an old saying: “Don't ask the barber whether you need a haircut.” Therapists have an obvious motive for inflating their own importance, and indeed, there’s some evidence that that is in fact what they do. They’re overconfident in how effective they are with their clients. In one study, for example, when comparing themselves to other metal health professionals, 100% believed that they were above average and 25% believed they were at the 90th percentile in effectiveness. And close to half of therapists said that none of their clients had ever deteriorated throughout the course of their therapy.
This overconfidence isn’t unusual. It’s common across professions and across humans in general. But what it suggests is that one should practice a fair bit of skepticism when “asking the barber.” When a therapist recommends therapy, it’s worth considering the possibility that other treatments may be just as effective.
Conclusion.
A couple caveats might be necessary to head off some misinterpretations. First, to reiterate the point at the top: Therapy is useful for many people! The purpose of this post is to argue that it isn’t uniquely useful.
And second, the quality of research in general is quite poor, so it would be premature to be extremely confident one way or another. For example, some studies are decades old, use questionable methods, and have not yet been replicated. This post is not intended as a knock-down case against therapists. Instead, as I indicated from the start, these are suggestive lines of evidence.
But taken together, 1) the uncertainty about therapeutic mechanisms, 2) the comparable outcomes with non-therapists, 3) the comparable outcomes with minimal-contact self-help, 4) the comparable outcomes with inexperienced therapists, and 5) the overconfidence among therapists all suggest for the lay person that seeking help outside of therapy may be just as effective in improving their mental health. And for therapists, it suggests they should be leveling up their skills to help their clients better.