ELIZA and the efficacy of "professional" therapists versus the rest of us

April 20, 2009 2:48 PM

Meta-Meta-RelationshipFilter: Have there been any scientific trials to study the efficacy of talk therapy that involved a placebo as control?

I'm thinking, is there any definitive proof that a professional talk therapist is substantially more effective than just talking about shit with a friend who doesn't charge a fee (which, granted, RelationshipFilter isn't necessarily like), or even substantially more effective than simply sitting someone down, telling them "we're going to have conversations that will explore and fix your psychological problems" but then having a discussion about any old thing that will run the clock out?

My own experience with years of visiting a variety of talk therapists would appear to indicate that at least with the average therapist I've seen the answer would be "No, there is no substantial difference" but I'm curious to see what if anything science has said. I'm definitely not saying that RelationshipFilter is any replacement for hours of ongoing, focused personal discussion, I'm just curious as to whether "professionalism" in this area, even of practitioners who are licensed or certified, has any demonstrated and quantifiable value.

posted by XMLicious to science & nature (21 comments total)

Placebo would have to be no therapy, as talking about any old thing could actually make you feel better.
posted by Ironmouth at 2:55 PM on April 20

I really don't see how you could do a double-blind study like this, because the people who weren't getting therapy would know they weren't getting therapy--there isn't a large enough or statistically diverse enough population of experimentally naive humans who could mistake a 50-minute discussion of the Mets' pitching staff for psychotherapy to make the study work.
posted by Sidhedevil at 3:07 PM on April 20 [1 favorite has favorites]

Here's a review article that touches on some of the difficulties you see and some of the attempts to get around them.
posted by grouse at 3:15 PM on April 20

Short answer: yes, and the correct psychotherapy is more helpful than non-clinical chatting given the correct circumstances.

Longer answer: There have been many, many studies in this area. The most interesting (and practical) ones compare different types of talk therapy for efficacy in treating particular psychological disorders. Here's an article that compares different treatments for post-traumatic stress disorder.

There are even meta-meta-meta studies looking at the researcher's allegiances to different therapy techniques influencing comparative studies.
posted by emyd at 3:22 PM on April 20 [1 favorite has favorites]

About fifty years ago there was a study done where patients in an asylum were divided into three groups, at random. One group was assigned to psychotherapists who did classic "laying down on the couch" talking therapy. The second group was assigned to psychologists who tried doing touchy-feely social work kinds of things.

The third group was assigned to attendants at the asylum who were told to do what they thought would help. They were the control.

After a period of time, outside doctors were brought in to evaluate the patients. The group assigned to the psychologists did better than the one assigned to the psychotherapists, but the group assigned to the attendants did the best overall.

That result was so surprising that the study was repeated, with a different group of patients, and the second time the same thing happened: the attendants did the best of all.

I wish I could give you a link or reference but I don't have one. I read about this in my college psychology text book, which I no longer have. (Perhaps someone else here will recognize my description.)
posted by Chocolate Pickle at 3:29 PM on April 20 [4 favorites has favorites]

I really don't see how you could do a double-blind study like this, because the people who weren't getting therapy would know they weren't getting therapy--there isn't a large enough or statistically diverse enough population of experimentally naive humans who could mistake a 50-minute discussion of the Mets' pitching staff for psychotherapy to make the study work


....I think you're missing the point. Thats pretty much what some therapists do (or at least that what it seems like from the other side of the couch, so to speak). You come in and talk about whatever until your time is up but it costs you a fortune.
posted by missmagenta at 3:50 PM on April 20 [1 favorite has favorites]

The most interesting (and practical) ones compare different types of talk therapy for efficacy in treating particular psychological disorders.

I'm not asking "have there been any studies on talk therapy done at all", I'm certain that there have been; I'm basically asking about studies comparing "professional" talk therapy to what I suppose we might call "amateur" talk therapy to evaluate whether or not the difference in efficacy is a substantial one. I'm going to tick Chocolate Pickle as a best answer there because that's exactly the sort of thing I'm looking for but it certainly would be great if anyone can find a citation on that.

And Sidhedevil, missmagenta has it.
posted by XMLicious at 4:59 PM on April 20

...and the general discussion of control design is interesting too, thanks grouse.
posted by XMLicious at 5:06 PM on April 20

I don't have a citation, but my intro psych course at MIT a few years ago really pushed the idea that it didn't matter what qualification a therapist had (PhD/MD/MSW/None). I was - and am - somewhat skeptical of this, but the professor didn't go into detail about what specifically they were measuring or how they dealt with control groups and subject selection, and the TA didn't know any of these specifics.

Actually, probably the most valuable thing I learned in that course was to be skeptical of people saying "studies show X"; in reducing a study to a soundbite, people often ignore what variables were measured and how, who the test population was, and other very relevant details. I also saw a few instances of accepted "common wisdom" parroted that had been debunked long ago by people actually in the relevant field, but still used as examples in a sort of allegorical fashion. (Koko the gorilla was the most cringeworthy example of this.) This is not to say that it's all bullshit; just that being an expert in one thing is enough to teach an intro course in the broader field that contains your subject, and to repeat things that seem like they should be true, but that don't hold up under closer examination.

Is my cynicism showing?
posted by spaceman_spiff at 8:36 PM on April 20

I am weeks away from graduating from a counseling program. Needless to say, I am biased. Here is my opinion. What you are talking about is the difference between release and change. The release of getting something off of your chest is useful and feels good temporarily. Admittedly, this can be accomplished with a therapist or a friend. However, true change comes from unconditional acceptance and therapeutic techniques/interventions (both counselor driven) which the average person couldn't emulate. These skills take years of practice to develop. Believe it or not, the therapist who appears to just sit there and listen, very carefully crafts their questions, answers and responses to help the client gain insight, perspective and understanding.
posted by rglass at 8:54 PM on April 20 [1 favorite has favorites]

Well, there are several studies on the efficacy of different types of psychotherapy, and there is a large movement toward Evidence-Based Treatment, which means 'stuff we know works because we have studied it at length over time against many other treatment modalities'.

The reason that therapists CAN "work better" than just chatting with friends, for people with serious psychological difficulties, is that we are trained specifically to work with those difficulties and help an individual to initiate changes that they have not been able to do alone and without the support and specific, focused collaborative work that a therapist brings. It's also helpful to most people to know that they won't be judged or gossiped about, and can control the level of involvement that a therapist has in the knowledge about their lives and experiences. Sometimes, therapy is sitting and talking. Sometimes, it is a lot more than that, and no therapist worth his/her salt is going to make anyone pay for "just sitting and talking" if it's not directed toward a specific goal for that client. And if you are just a person who had a crappy day at work, by all means, do not go to a therapist for that, go to a friend. If you are a person who can no longer function on a daily basis because you are so debilitated by depression, anxiety, hallucinations, whatever it is for you, therapy can be the catalyst for bringing you back to a healthy, happy life.
posted by so_gracefully at 11:04 PM on April 20 [2 favorites has favorites]

I suspect it would depend on the kind of talk therapy being done.

If it's just come-in-and-talk-about-your-problems "therapy", then I suspect the evidence is pretty much conclusive that it doesn't work any better than talking to a close friend.

On the other hand, specific therapeutic techniques like Cognitive Therapy for depression or therapies for phobias do definitely work better than just talking with a friend.
posted by rhartong at 5:09 AM on April 21

There are a couple of semi-contradictory answers to your larger question, which I take to be, is therapy more effective than talking with your friends. (The placebo question is problematic for a bunch of different reasons, not least because as someone pointed out upthread, one factor which influences the efficacy of therapy is therapist allegiance.)

There are two places I'd look for information. The first is Bruce Wampold's book The Great Psychotherapy Debate which reviews the research on psychotherapy outcomes and then uses them for meta-analysis. He covers the problem with placebo studies of psychotherapy reasonably in-depth. His results, consonant with basically all meta-analyses and most stand-alone studies is that psychotherapy is very effective (an Effect Size of 0.80) for addressing mental distress. 79% of people who want therapy and get it do better than people who want therapy but do not get it. He does not specifically address the issue of whether or not those who want therapy but do not get it are talking to friends in the meantime, but you're welcome to draw your own inferences from your knowledge of human behavior. His results also show that this effect size is independent of what type of therapy is being practiced. (In other words, CTR does as well as interpersonal does as well as behavior modification.) The most important constituents of therapeutic effectiveness are a positive relationship with the therapist, hope for change, and a plan for how the change will proceed. (Michael Lambert has also done work on this.) Wampold's work puts the lie to the notion of Evidence Based Therapies.

The other thing I'd look at is The Heroic Client by Duncan, Miller and Sparks (Or their book The Heart and Soul of Change). In general, they show that while therapy is really important, new therapists, and even students, tend to just as well as (or maybe slightly better than) established therapists. Therapists who have been therapists for longer tend to be less responsive (as a group, this obviously varies by therapist) to the needs of patients, and therefore have slightly worse outcomes. They also show that there is no difference between type of training and outcome, and indeed, even medications work substantially better for some psychiatrists than for others.

Both sets of research and review, which converge at many points, support a point that would seem to support your hypothesis: the placebo effect (expectation) is a huge constituent of psychological and psychiatric care. I think it takes a willful misreading of the research to suggest that this is not true. However, it would be a serious misunderstanding of the placebo effect to suggest that this means that talking to your friends constitutes "amateur psychotherapy." The placebo effect works because of our expectation that the intervention will work. There would have to be a substantial change in how we view psychotherapy on the one hand and talking with our friends on the other for them to be equivalent in this way. This is not a limitation of the friendships, but of our way of framing these issues.

In addition, I do think that the specificity of the therapeutic relationship has an effect on how people approach the encounter. I've never known a therapist who was not changed by it, nor a patient. (This is different than saying that the relationship works in all cases.) In general, therapists are trained to excite and amplify those parts of the relationship and the clients life that might lead to change. Friends are not necessarily cognizant enough about those things to do that, and since they are major constituents of change, that might work out to the detriment of the people involved. Finally, but by no means beside the point, many people don't want to talk to their friends or loved ones about problems, especially problems they might attribute to those friends and loved ones. To go back to Wampold's 79% helped figure: if those not in therapy were talking to their friends, then something obviously wasn't working for them in that set-up; if they weren't why weren't they? I haven't seen any anti-talking-to-your-friends messages spouted about.

All that said, I do think what we know about what works in therapy suggests that there's plenty of room for a very effective model of "amateur" or "peer" psychotherapy, I just don't think that this is the same as talking to your friends. (Neither, by the way, is Chocolate Pickles anecdote.)

As a coda to my comment, I'd like to point out that you've basically obviated the usefulness of your own question, XMLicious, by marking Chocolate Pickles' answer as best. It's clear from your question what you want the answer to it to be, but picking a non-cited story that reads like an anecdote as your best answer makes it look like you've asked your question in bad faith. Because that question appropriates the trappings of scientific language, your choice makes it that much worse. (I'm not, by the way, suggesting that CP's anecdote isn't true, I can well believe that it is. One of my favorite anecdotes of this sort is about a psych hospital in Italy where the level of acute symptomology fell appreciably after the medical director asked the patients what would help them most, and they said "Better lunches in the cafeteria." He changed them, and people got better." I believe that anecdote as well, but it isn't science, at least as quoted here.)
posted by OmieWise at 6:48 AM on April 21 [3 favorites has favorites]

OmieWise: Thank you for the links and your response.

If anyone had come up with the same sort of thing cited, I wouldn't have marked Chocolate Pickles' anecdote as a best answer; I did so because people were presenting links to inaccessible papers which would take me quite an amount of effort to get ahold of (to get to a library and go through an inter-library loan process, though I'd be willing to do that) and read, which from their abstracts appear to simply be comparing different types of talk therapy. Even if his anecdote had a different outcome I would have marked it thus in the absence of cited examples of what I'm talking about. I also marked grouse's paper on control design in experiments related to talk therapy as a best answer.

As I realize that RelationshipFilter isn't the same thing as a one-on-one conversation, I also do realize that a concerted conversation exploring someone's problems and informed by research is still not the same thing as a casual chat. But I'm still skeptical that "professionals" are really all that good at delivering the former. And my question is still specifically about the use of these types of placebos in scientific experiments.

To give a little more context: the most effective therapist I've visited myself in the past, the only one whose meandering conversations appeared to actually have a point more than once in a blue moon and the only one whose mannerisms and approach over the course of several months appeared to follow some coherent plan, didn't have the PhD or the MD or the MSW that spaceman_spiff mentioned. Instead he turned out to be a guy who was an insurance company administrator for most of his life who had gone back to graduate school to fulfill the minimum requirements for licensure of mental health counselors in my U.S. state. This inclines me to wonder how much of his acumen actually comes from the profession; and this particular AskMe question, about trying to compare the efficacy of therapy to a well-designed placebo as a baseline, is the way I chose to start inquiring about that.
posted by XMLicious at 3:11 PM on April 21

Instead he turned out to be a guy who was an insurance company administrator for most of his life who had gone back to graduate school to fulfill the minimum requirements for licensure of mental health counselors in my U.S. state.

This is like saying that the best mechanic you ever had, who just happened to fix cars for a living, was really only good at that because she pursued painting as an avocation.

I'm glad the guy was helpful for you, and it may well have been his previous experiences that made him helpful, but he was a professional counselor. He wasn't an amateur in any sense of that term. If he was licensed, he probably had a Masters. At the very least he had a Bachelors and was supervised by someone who was a Masters level therapist. (I don't know any states where the minimum reqs for licensure don't include one of those two things.) Again, you seem to be confusing your desire for him to be something other than what he was for him actually being other than he was.

I'll admit, I take it a bit personally when someone dismisses the thing I do with so little actual curiosity about what that thing is. Despite that, I presented a load of (secondhand) data that goes part of the way to getting you where you claim to want to be. It's not clear from your response that you considered any of that at all, and certainly none of the disconfirming bits. Instead you provided a justification for 1) why the unscientific response with which you agreed was the best one, and 2) why a professional counselor who was helpful to you didn't disconfirm your hypothesis. Forgive me if I don't buy your purported interest in science.
posted by OmieWise at 4:05 PM on April 21 [1 favorite has favorites]

This is like saying that the best mechanic you ever had, who just happened to fix cars for a living, was really only good at that because she pursued painting as an avocation.

It would be like that, if it was the only experience I've ever had that lead me to inquire about this. It's not, but I really am looking for the specific information on these types of placebos in experiments rather than a qualitative discussion of why professional psychotherapy is better. I welcome that too but it's the stuff about placebos that is guiding my use of the best answer marker.

Re-reading your initial answer, I realize that I missed the sentence where you said that Wampold's book covers the use of placebos in-depth. So I'm marking yours as a best answer too but I'm not going to retract the one on Chocolate Pickle's comment.
posted by XMLicious at 4:29 PM on April 21

For reference to all, the table of contents and some parts of Wampold's book are available on Google Books.
posted by XMLicious at 4:32 PM on April 21

One other thing: I did not mean to say that the therapist I mention was especially helpful, I meant to say that what he was doing in exchange for my time, effort in getting to the appointments, and money seemed much more appropriate, a difference dramatic enough to in some cases wonder what the heck the other licensed therapists I'd seen before had been doing, for years in some cases, as well as the ones I've seen since.

Props to your profession and all OmieWise, but given the large amounts of time, effort, and money that professional therapy can require over the course of years and years, I'm given to think that at least part of that might be better spent in other endeavors. For just one example, I seriously have to wonder how far I might get towards getting a degree in psychology myself during, say, the next ten years. If I could do that without spending the equivalent amount of time in traffic, and it was a useful experience, it would be nice.

Maybe the current arrangement, which seems to be sort of styled after a physician treating a patient with a chronic illness, just isn't such a great model.
posted by XMLicious at 5:13 PM on April 21

Maybe the current arrangement, which seems to be sort of styled after a physician treating a patient with a chronic illness, just isn't such a great model.

I actually couldn't agree more. The research shows that therapy is not only HIGHLY efficacious, but also works very quickly. Most people who enter therapy seeking change find relief very quickly. I much prefer a dental model of therapy, where if acute treatment is needed it happens relatively quickly and the door is open for more acute treatment (or simply maintenance treatment) in the future if it is needed.
posted by OmieWise at 6:45 PM on April 21

Hmmm... are you saying that there are lots of people who just go to a therapist a few times, briefly, and never go back and never see another therapist because their problems have been resolved? That was the opposite of my impression; lots of people have told me that they've seen a therapist repeatedly in the course of years and I've never heard an anecdote of a talk therapy treatment working very quickly; but perhaps it's just that people who go to a therapist repeatedly are more likely to talk about it.

Is there anything like insurance company statistics out there that show the ratio of claims for individual therapy visits to the total number of individuals receiving this sort of treatment? If that ratio was fairly low in the data of a given year, it would bear out what you seem to be saying.
posted by XMLicious at 8:19 PM on April 21

Continuing to think about it, I realized a way to better articulate part of what I'm getting at by inquiring into this. It was pondering rglass's answer that prompted me:

What you are talking about is the difference between release and change. The release of getting something off of your chest is useful and feels good temporarily. Admittedly, this can be accomplished with a therapist or a friend. However, true change comes from unconditional acceptance and therapeutic techniques/interventions (both counselor driven) which the average person couldn't emulate. These skills take years of practice to develop.

Implicit in these statements is the idea that the therapy recipient cannot achieve the "change" therapy outcome on their own through self-reflection. If this is true, then a placebo of the type I described in the original question - saying "we're going to talk about your problem X" but then having a conversation about anything whatsoever - should genuinely be completely ineffective.

But if the average person does have some capability to achieve for themselves the "change" result as would be pursued by a professional therapist I would think it might be made to show up in long run statistics in a scientific experiment with some strategy like this. And perhaps this inherent capability, if it exists, varies quantifiably in different populations, which might be of interest; or perhaps there are ways to enhance that capability. Or maybe there are other ways to scientifically examine the capability that such a study would lead to.

Possibly this sort of thing is discussed in Wampold's book but Google Books cut me off before I got very far, unfortunately.

(And though I'm certainly critical of some aspects of the profession as I'm familiar with it and not reserved in making those criticisms I didn't pose this AskMe simply for the purpose of criticizing so at least I should have worded that last sentence in the OP differently.)
posted by XMLicious at 5:43 AM on April 26