Therapy doesn't work how we think it works

[FreudianSlippers]

Let me preface this by saying I am 100% about empiricism. I was trained hardcore in evidence-based practices, the importance of research informed therapy, and to be a good consumer/critic of research. I currently use CBT, ACT, and exposure and response prevention as my main modalities. So...don't get too angry. I want you all to think about something though. The Contextual model of psychotherapy proposed in The Great Psychotherapy Debate by Wampold and Imel (a freaking MUST read) provides a **** ton of evidence that change doesn't happen in psychotherapy like it does in medical treatments. In medicine, you have high blood pressure, you take propanolol, blood pressure reduces, you're good to go. However, with therapy change seems to happen when 4 criteria are met: 1. There is a good therapeutic alliance. 2. Empathetic and caring therapist who believes the treatment will work 3. Some sort of psychologically valid explanation of the client's problem is given, which is culturally acceptable and the client believes (creating hope and expectation). 4. The client engages in SOME sort of healthy behavior change (it doesn't matter exactly what they do). He shows that it doesn't matter what you do or how you do it...as long as it meets these criteria. He provides a rational for why this works (e.g. placebo effect and evolutionary history of humans healing within a social context), but you can read for more details on this.

When you browse through the literature and see how every bonified (i.e. established, widely used, evidence based intervention) therapy is basically the same effectiveness for almost every disorder. I HATE this....I hate how a recent study showed EMDR was as effective as ERP in treating OCD (like WTF...I cited this at the end*), Let me get this straight...it's not saying you can just chat with a client and expect change, but as long as those 4 criteria are met people can't help but get better. Whether its ERP, EMDR, brainspotting, emotion-focused, ISTDP, etc. People will argue that the so called "dodo bird hypothesis" is wrong for various reasons. Do your due diligence and read up on both sides, but Wampold's arguments are solid and grounded in so many studies.

Cognitive therapy seems much more legit than EMDR...but how do we know that cognitive restructuring is affecting change by re-shaping information processing systems, when behavioral activation with no cognitive re-structuring works, increased exercise works, interpersonal therapy works, and psychodynamic therapy works, etc. It's annoying that EMDR is so glaringly in your face about how the eye movements don't make sense, but it has helped thousands of people. Maybe the eye part helps people buy into trauma treatment that wouldn't otherwise receive treatment? I believe we NEED evidence-based practices as vehicles to deliver the components of treatment that work (and to know what doesn't work), but i'm not sure it matters exactly what we do as long as we follow those guidelines.

Feel free to rip apart any of what I said. I'm a nerd about this stuff and enjoy seeing things from new perspectives. I LOVE evidence-based practices, but the evidence for the contextual model/common factors seem undeniable. Please remember Wampold is NOT saying "common factors" like a good relationship and positive regard will cure any psychological ailment. There has to be a valid therapeutic ritual, and positive behavior change. Also, check out the article below for a great review of the Contextual Model and current state of the literature. TLDR: We've been doing psychological research for 50 years and still don't know how therapy works.

Cuijpers, P., Reijnders, M., & Huibers, M. J. (2019). The role of common factors in psychotherapy outcomes. Annual Review of Clinical Psychology, 15, 207-231.

Bonus points if you can rip the below study apart and tell me how EMDR matched CBT for OCD
* Marsden, Z., Lovell, K., Blore, D., Ali, S., & Delgadillo, J. (2018). A randomized controlled trial comparing EMDR and CBT for obsessive–compulsive disorder. Clinical psychology & psychotherapy, 25(1), e10-e18.

---

Members don't see this ad.

 

[WisNeuro]

I'd strongly urge you to read some of the criticisms of Wampold, there are significant issue with methodology within the "Dodo Bird" effect.

As with other criticisms of EMDR, it works because it has a fairly extensive exposure component. In the EMDR described in this study, they also added a fairly significant CBT aspect with restructuring. So in essence, it works because you did PE and CBT, the eye movements likely added nothing to the therapy. So, you're rerally just comparing PE/CBT to CBTR in this study. If they really wanted to study EMDR, they would have had a dismantled condition that did the EMDR protocol without the eye movements. When this has been done in the past, the movements had no effect. Other than costing a practitioner several thousands of dollars for "training" and an overpriced LED strip/

 

[Sanman]

The question, boils down to how much of what we do is related to the placebo effect vs actual intervention. Issue becomes it is simply much harder to test that in psychotherapy than in medicine because we cannot just use a sugar pill. That said, it is not that different from medicine, How much of feeling more awake while I am drinking my morning coffee is from caffeine vs expectation of the drug vs a conditioned Pavlovian response? It certainly takes effect faster than the drug gets into my system. The real question then becomes, do I care? Does the mechanism matter if the result works? Can you really remove decades of marketing from CBT and normalizing it to the public to find out? How can you measure internalized expectations of a client to truly get your answer?

 

[cara susanna]

Here's a review of EMDR dismantling studies:

Does EMDR work? And if so, why?: a critical review of controlled outcome and dismantling research - PubMed

Research on Eye Movement Desensitization and Reprocessing therapy (EMDR) was reviewed to answer the questions "Does EMDR work?" and "If so, Why?" This first question was further subdivided on the basis of the control group: (a) no-treatment (or wait list control), (b) nonvalidated treatments...

pubmed.ncbi.nlm.nih.gov

 

[WisNeuro]

The question, boils down to how much of what we do is related to the placebo effect vs actual intervention. Issue becomes it is simply much harder to test that in psychotherapy than in medicine because we cannot just use a sugar pill. That said, it is not that different from medicine, How much of feeling more awake while I am drinking my morning coffee is from caffeine vs expectation of the drug vs a conditioned Pavlovian response? It certainly takes effect faster than the drug gets into my system. The real question then becomes, do I care? Does the mechanism matter if the result works? Can you really remove decades of marketing from CBT and normalizing it to the public to find out? How can you measure internalized expectations of a client to truly get your answer?

There is the placebo effect, sure, but we even have to take a step back to the premise posed by the OP. The premise has never been established all that well due to issues with design. OP, read some of Crits-Christop and Cuipers work in this area. The common factors stuff is all too commonly taught in grad school without presenting critiques of the work. Fairly negligent didactic in my opinion.

 

[Sanman]

There is the placebo effect, sure, but we even have to take a step back to the premise posed by the OP. The premise has never been established all that well due to issues with design. OP, read some of Crits-Christop and Cuipers work in this area. The common factors stuff is all too commonly taught in grad school without presenting critiques of the work. Fairly negligent didactic in my opinion.

To say it more succinctly. I think most therapy is effective due to a combination of placebo effect, common factors, and EST. That does not mean the EST has no effect. Remove any of those factors and therapy is less effective in the real world.

 

[WisNeuro]

To say it more succinctly. I think most therapy is effective due to a combination of placebo effect, common factors, and EST. That does not mean the EST has no effect. Remove any of those factors and therapy is less effective in the real world.

Agreed, and we need to do more work at looking at how these three affect outcomes. Unfortunately, the common factors stuff gets preached like gospel a little too much, with no mention of its severe limitations.

 

[deleted343839]

How can you measure internalized expectations of a client to truly get your answer?

Expectancies are measured mostly in research, it seems, but they are clearly tied to outcomes.

The Contextual model of psychotherapy proposed in The Great Psychotherapy Debate by Wampold and Imel (a freaking MUST read) provides a **** ton of evidence that change doesn't happen in psychotherapy like it does in medical treatments.

Placebo response (including but not limited to the role of expectancies) influences just about any subjective/patient-reported outcome you can measure, and this clearly affects the outcomes of medical treatments, especially for conditions that are not curable.

A couple of books I've enjoyed lately are Hayes and Hoffman's "Process-based CBT" and the more recent "Principles of Change: How Psychotherapists Implement Research in Practice" by Castonguay, Constantino, and Beutler. Among other things, they blur the distinction between "common factors" and empirically supported change techniques.

 

[FreudianSlippers]

I love the responses so far, you all have given me a lot to think about.

Placebo response (including but not limited to the role of expectancies) influences just about any subjective/patient-reported outcome you can measure, and this clearly affects the outcomes of medical treatments, especially for conditions that are not curable.

A couple of books I've enjoyed lately are Hayes and Hoffman's "Process-based CBT" and the more recent "Principles of Change: How Psychotherapists Implement Research in Practice" by Castonguay, Constantino, and Beutler. Among other things, they blur the distinction between "common factors" and empirically supported change techniques.

Funny that you mention this MamaPhD...i'm literally reading Process-Based CBT right now! I've found it a great way to bridge the gap of common factors and specific factors. Regardless of "how" it works, its highly logical that tailoring psychologically sound interventions to the unique contextual variables of each client is going to be helpful.

I think Routine Outcome Monitoring and Feedback Informed Treatment are also ways to bridge this gap.

I like the EMDR example because like many of you, I despise the treatment. Why not just use exposure with no weird eye movements? I get it! BUT...if i'm being real with the research...I can say the same about behavioral activation. If it's AS effective as cognitive techniques, why just stop doing cognitive techniques. Well...because some people respond better to one than the other. I think the difference is EMDR is literally grounded in nothing..it's like me saying "lets juggle while talking about your trauma, but pay me 5000 dollars to be trained"....we have to draw the line somewhere I guess lol.

ALSO (my last point I swear)...I think therapy is just too dynamic and complex to really ever know what works for whom under what circumstances. Everybody has such unique learning histories and biological predispositions, a cognitive technique for person A might work by giving them a new meaning, for person B might work by disproving negative thoughts, for person C might work by leading to behavioral activation which improves their mood, for person D might work by increasing psychological flexibility....multiply that by the thousands of interventions nested under hundreds of evidence based practices.

 

[cara susanna]

I love the responses so far, you all have given me a lot to think about.



Funny that you mention this MamaPhD...i'm literally reading Process-Based CBT right now! I've found it a great way to bridge the gap of common factors and specific factors. Regardless of "how" it works, its highly logical that tailoring psychologically sound interventions to the unique contextual variables of each client is going to be helpful.

I think Routine Outcome Monitoring and Feedback Informed Treatment are also ways to bridge this gap.

I like the EMDR example because like many of you, I despise the treatment. Why not just use exposure with no weird eye movements? I get it! BUT...if i'm being real with the research...I can say the same about behavioral activation. If it's AS effective as cognitive techniques, why just stop doing cognitive techniques. Well...because some people respond better to one than the other. I think the difference is EMDR is literally grounded in nothing..it's like me saying "lets juggle while talking about your trauma, but pay me 5000 dollars to be trained"....we have to draw the line somewhere I guess lol.

ALSO (my last point I swear)...I think therapy is just too dynamic and complex to really ever know what works for whom under what circumstances. Everybody has such unique learning histories and biological predispositions, a cognitive technique for person A might work by giving them a new meaning, for person B might work by disproving negative thoughts, for person C might work by leading to behavioral activation which improves their mood, for person D might work by increasing psychological flexibility....multiply that by the thousands of interventions nested under hundreds of evidence based practices.

IIRC the research on behavioral activation being as effective as cognitive therapy is mixed. Let me know if I'm wrong though!

 

[FreudianSlippers]

IIRC the research on behavioral activation being as effective as cognitive therapy is mixed. Let me know if I'm wrong though!

All the research i've seen has seen has shown them to be equivalent (see below)
Study 1
Study 2

 

[DynamicDidactic]

.I think therapy is just too dynamic and complex to really ever know what works for whom under what circumstances.

I have a much more optimistic view of research and the future. Just think about the state of the science related to mental health even 50 years ago, I imagine we will be light years ahead in another 50 years. I see no reason why there would not be new discoveries if the scientific community keeps grinding. If anything, we can at least get the harmful stuff out of the way (e.g., trigger warnings). Someone will come along and revolutionize something. At one point, psychoanalysis was the most common option, we are doing much better nowadays.

 

[DynamicDidactic]

@FreudianSlippers
All that Wampold/common factors stuff is based on one fundamental flaw; correlation does not equal causation.

Cuijpers, P., Reijnders, M., & Huibers, M. J. (2019). The role of common factors in psychotherapy outcomes. Annual Review of Clinical Psychology, 15, 207-231.

Your own citation, which I love to cite to make my point.

To date, research on the working mechanisms and mediators of therapies has always been correlational, and in order to establish that a mediator is indeed a causal factor in the recovery process of a patient, studies must show a temporal relationship between the mediator and an outcome, a dose-response association, evidence that no third variable causes changes in the mediator and the outcome, supportive experimental research, and have a strong theoretical framework. Currently, no common or specific factor meets these criteria and can be considered an empirically validated working mechanism. Therefore, it is still unknown whether therapies work through common or specific factors, or both.

For a guy like Wampold, with all the pull he (or his supporters) have, to never attempt a single controlled trial of a common factor is very telling. We can easily do a trial of, lets say, BA. We can randomize cold/uncaring therapists and warm/empathic therapists. Then lets see if there is a difference. Not a single study (that I know of) has ever tried this yet Wampold keeps beating his correlational drum.

Or, as I have said a few times on here, no matter how warm and caring you are that isn't going to help the person with borderline personality disorder and a recent suicide attempt if you are practicing psychodynamic therapy as opposed to DBT (or at least you are likely to expect half as many suicide attempts in DBT).

Two-Year Randomized Controlled Trial and Follow-up of Dialectical Behavior Therapy vs Therapy by Experts for Suicidal Behaviors and Borderline Personality Disorder

Context Dialectical behavior therapy (DBT) is a treatment for suicidal behavior and borderline personality disorder with well-documented efficacy.Objective To evaluate the hypothesis that unique aspects of DBT are more efficacious compared with treatment offered by non–behavioral...

jamanetwork.com

If you are worried about allegiance effects, how about this study on psychoanalysis for bulimia (no surprise, CBT works better)

Psychiatry Online

ajp.psychiatryonline.org

Our lack of understanding etiologies and clumping heterogeneous disorders together is likely interfering with our ability to better understand which treatments work (and how they work). Therefore, it makes it look like many treatments work similarly well. When the empirical evidence is murky we have to go to the theory. Which theory do you stand by:

The only thing that can help people is warmth, empathy, an idea of why you have a problem, and doing something/anything.
Or
There are behavioral/social/biological interventions that can, on top of common factors, provide incremental therapeutic validity.

And even if we don't have those treatments now only means we have to keep working on it. Not sit back and throw our hands up in the air and exclaim that its all the relationship.

 

[representwitpixels]

@FreudianSlippers

Our lack of understanding etiologies and clumping heterogeneous disorders together is likely interfering with our ability to better understand which treatments work (and how they work). Therefore, it makes it look like many treatments work similarly well. When the empirical evidence is murky we have to go to the theory. Which theory do you stand by

Very well said. Not sure if I quoted that correctly, I'm a noob, but I think what you wrote should be amplified.

 

[clausewitz2]

bonified

Fascinating and super important discussion, glad you kicked this off. However.

Bona fide bona fide bona fide bona fide.

Ahem. That is all, hopefully my eye will stop twitching.

 

[FreudianSlippers]

For a guy like Wampold, with all the pull he (or his supporters) have, to never attempt a single controlled trial of a common factor is very telling. We can easily do a trial of, lets say, BA. We can randomize cold/uncaring therapists and warm/empathic therapists. Then lets see if there is a difference. Not a single study (that I know of) has ever tried this yet Wampold keeps beating his correlational drum.

I agree 100%. I have had dreams about doing a study like this. Or alternatively, doing ERP for OCD and in one group explaining the rationale and in another group giving less positive rationale (to manipulation the CF of high expectations/hope).

If you are worried about allegiance effects, how about this study on psychoanalysis for bulimia (no surprise, CBT works better)

Great study...I love seeing studies like this because it reinforces what, at my core, feels so true! Thanks for citing it here. However, I have two take aways. 1. I don't put AS MUCH value into results of one study compared to all the meta-analyses i've seen. A quick Google Scholar search came up with several studies showing that psychodynamic was AS effective as CBT in treating ED:

Link 1
Link 2

Furthermore, in the discussion of the article you cited about psychoanalysis and bulimia, they authors note "the fact that CBT has been tested and revised extensively, whereas this is the first trial of the present version of longer-term psychoanalytic psychotherapy for bulimia nervosa, may have also contributed to the difference in outcome". In the common factors theory, this would explain A HUGE disadvantage for psychoanalysis, as the clients wouldn't feel like the treatment is as helpful as something that was more targeted and applicable. I think psychoanalysis inherently is an inferior treatment since it tends to be SOOOO unstructured...according to the CF model this would give less opportunity for clients to make changes/feel hopeful, therefore be less effective.

Our lack of understanding etiologies and clumping heterogeneous disorders together is likely interfering with our ability to better understand which treatments work (and how they work). Therefore, it makes it look like many treatments work similarly well. When the empirical evidence is murky we have to go to the theory. Which theory do you stand by:

The only thing that can help people is warmth, empathy, an idea of why you have a problem, and doing something/anything.
Or
There are behavioral/social/biological interventions that can, on top of common factors, provide incremental therapeutic validity.

And even if we don't have those treatments now only means we have to keep working on it. Not sit back and throw our hands up in the air and exclaim that its all the relationship.

It's not "do something/anything"....it's "do something/anything that is a healthy behavior change that the client and therapist both believe will be helpful." I believe specific ingredients DO provide incremental validity....just not for the reasons we THINK they do.

I think the fact that you wrote "throw our hands up in the air and exclaim its all in the relationship" makes me think you might be misunderstanding the whole contextual model. If the contextual model is accurate, it will be more important than ever to continue doing research into various aspects of mental health pathology and mechanisms...there's no throwing up our hands. It will be incredibly important to give the most up-to date and rational explanations of how the human mind and body work. All it will mean is shifting the focus to ways we can best tailor evidence based treatments to meet the unique background and needs of clients, in a flexible way. I think this will mean more emphasis on feedback informed therapy and routine outcome monitoring than "oh, this person has depression, we must do cognitive re-structuring". That is doing our clients a huge injustice. On the other end, saying something like "oh, all we need is a good relationship and to talk about your childhood" is another huge injustice (i'd argue MUCH worse than the former).

 

[tr]

Great study...I love seeing studies like this because it reinforces what, at my core, feels so true! Thanks for citing it here. However, I have two take aways. 1. I don't put AS MUCH value into results of one study compared to all the meta-analyses i've seen. A quick Google Scholar search came up with several studies showing that psychodynamic was AS effective as CBT in treating ED:

Link 1
Link 2

OK but I kind of think the apparent statistical equivalence here is mostly about the huge CIs. But if you look across all the studies from the Steinert meta graphed out, the trend really does not look so good for psychodynamic.

 

[bcliff]

Another Cuijpers paper (link) that I often cite when discussing EST/EBPs:

Trials comparing different types of psychotherapy for adult depression do not have sufficient power to detect clinically relevant effect sizes. In order to demonstrate a clinically significant effect size of d=0.24, a trial would need to include 548 patients, but the largest comparative trial we found in three major meta-analyses included only 221 patients. This largest trial had only enough power to detect an effect size of d=0.34, and even this trial did not have enough statistical power to detect the mean difference between antidepressant medication and placebo. The implication is that individual trials are heavily underpowered and do not even come close to having sufficient power for detecting clinically relevant effect sizes—let alone smaller effect sizes that may not be clinically significant—but are nevertheless interesting from a scientific point of view.

I agree with others that common factors likely represent the bare minimum of what's necessary for behavior change, but I hypothesize that ESTs likely yield larger effects on behavior change than common factors alone. Failure to detect a difference is not the same as demonstrating equivalence.