Interesting question. Psychoanalysis and psychodynamic therapies seem to be quite healthy, despite the strength of the em[pirically-grounded treatment approach. The dodo hasn't died, has it? Is it an issue of peaceful coexistence? I'll be curious to see what others think, but I'm quite sure psychotherapy is not dead.

Every year 10s of millions of dollars are spent on testing the efficacy of CBT alone. That is a conservative estimate that doesn't take into account eCBT. The money that is still spent on the already tested 300+ evidence based therapies would be more, and this has been going on since the mid 90s. Yet the results are always the same, and the results do not translate into better clinical practice. The studies that match person to treatment or compare treatment with treatment also do not enlighten us. In fact back in 1986 when there was only 200 EBTs it was estimated that it would take 47 million studies to do all the necessary comparisons to find out which is best. My colleagues who currently command grant money to study EBTs fiercely object that their research is redundant, and are vocal detractors of common factors research. They take particular objection to meta-analysis of 50 years of research which suggests that specific interventions account for so little of the clinical outcome. Neither the Dodo or psychotherapy die if the debate is over. But if indeed it is over, then should we not be directing the millions of dollars in annual research funding, and man hours in scholarly efforts that improve clinical outcomes?

Such an interesting question. Among some theorists and researchers, the dodo debate is alive and well. As you suggest, tests of efficacy are often criticized for having too little "real world" value - subjects are often too homogenous, protocols are inflexible, and outcomes are only published if significant results are found.

I would love to see more research on mechanisms of change. Much to many empiricists' dismay, the mechanisms generally considered crucial to treatment success often do little for outcome predictions (e.g., cognitive change in cognitive therapies). Does this mean that common factors are driving effects of all psychotherapies? What if individual differences are taken into account (e.g., psychoanalysis is generally considered detrimental to people woth severe alexithymic traits)? It's possible, perhaps even likely, that different mechanisms are necessary for different psychopathologies, personalities, etc. Yet, too few people are concerned with whether a treatment works that they fail to investigate why it works. I believe this kind of mechanistic research is needed, not only to resolve the age-old common v. specific factors debate, but also to improve upon the treatments we already have.

Well, I was asked by my psychiatrist to do psychotherapy and CBT by clinical psychologists who also practice in his mental health clinic. It costs money and is not cheap ... These are private sector specialists and not from government hospitals ... so I think if there are people willing to pay money for these services, I guess at least for some clients, there is some benefit, or according to the economic law of supply and demand, these psychologists will not be able to exist and make a living?

Hi Emily & Bob,

Thank you for some great answers. I like how you emphasise the need for knowing why treatment works for people or better still why it doesn't work. We have asked so many more questions that focus on what the clinician does that what is happening for the patient remains undiscovered territory. Perhaps the risks of not finding anything are unattractive for the researcher are prohibitive. However there has to be a point where repitition is redundant or even ethically questionable. We have technology now without the constraints on paper and ink technology, so we should be able to disseminate more resesearch regardless of any "discovery" and perhaps provide the data sets too. Can you imagine having the datasets at your disposal for almost every study you cite in your work? We don't even have a name for the "type" of research design!

Help seeking for psychological distress will continue to drive clinical professions. I don't see an eradication of psychopathology. However in the other clinical sciences, money spent in research usually results in a reduction of cost to treat. But in psychotherapy cost to treat rises. 

Perhaps it is helpful to conclude that in the light of not having better questions to ask, the great psychotherapy debate is over. Especially if this drives better research that translates in better clinical outcomes for more people. 

Mechanism of change is an interesting area, but I think we can safely say that if it had anything to do with the model, we would see vastly different outcomes between the different models. And we don't.

For me, I believe healing happens in the human, and we know very little about what speeds up that healing and what slows it down...

The fact that some therapists are vastly better than other therapists is way more worthy of investigation.

Allegiance is another area, where the impact on outcome is far greater than the contribution of the model. That is something therapists can actively work on while academics find out more about it.

It is hard for me to understand why so many resources are spent repeating what we already know about models, that got us no where, and so little spent on areas that have been clearly shown to contribute more to outcome than the model used.

Very interesting question and discussion.  The many millions of dollars that have been spent on psychotherapy research have been useful to the extent that it is now unequivocally clear that psychotherapy is effective. However, in their systematic review of psychotherapy research  the American Psychological Association concluded that there is no evidence for the superiority of one treatment over another  (http://apa.org/about/policy/resolution-psychotherapy.aspx ).  The  common factors answer to the debate is not very satisfying in that it fails to provide sufficient clarity clinically or empirically.  For instance, emphasizing the importance of the therapeutic relationship is fine as far as it goes but does not say anything about how to create an optimal relationship with a specific patient.  I addressed some of these issues in my presidential address to the Society for Psychotherapy Research and proposed strategies for identifying and empirically investigating psychotherapy change mechanisms.

Thanks George

I really appreciate the way your paper lays out the problem, the way that we need to accept that we can never have a homogeneous group of people that will behave like a predictable machine for the purposes of research. I like the description that "patient and treatment are an inseperable pair" and was also reminded that Jay Hayley said that The therapist needs to realise that they are "part of the truth they seek." The idea of making the variability of patients the actual subject of research, rather than treating it as error variance and trying to control for it.... is I think pivotal.

After that... I'm not so sure. I was reminded of one of my heroes Heinz von Foerster who said that tests test the test.... not the person being tested.

I appreciate your contribution to this discussion.

Hi George,

It is encouraging to read your work.  The shear volume of research dedicated to EBTs and common factors makes it difficult to ignore either body of work when planning research. Yet EBTs continue to be funded, and I am convinced this is only partly to blame on funding providers. I tend towards the argument that the responsibility for what research is undertaken is responsibility of the scholar and educator, because government and the public purse rely on the expertise of the scientist to inform municipal planning.

I believe that when a doctoral student is confused over whether to do research that they know will elicit statistically significant results, or do research that matters, the academy is in serious trouble. (See link below)

There seems to be a consensus among researchers that the pusuit for grant dollars through redundant research trumps pursuing the  novel and unknown.  Just this weekend my colleague and I identified several millions of dollars had recently been granted to EBTs in Australia alone. Our calculations did not account for existing research of the same that was in progress, or not clear from grant submission descriptions. Researchers have an incredible impact as mentors on their doctoral students, and by example they instil values that govern what research is important, and what will be more likely to improve clinical outcomes. Or is psychotherapy already good enough?

In medicine a particular body of research can end or change focus when, for example, a treatment is obsolete or the manner in which the treatment is being studied will no longer enliven the researcher or clinician to better knowledge or practice. But that does not seem to happen in psychotherapy research. Is it the professional bodies that lobby government on behalf of clinicians that are hedged in by a paradigm they need to look legitimate compared to primary medicine and psychiatry ?

Furthermore if a researcher considers the meaningful conclusions such as you have stated in your comments and in your article, I struggle to understand how studies testing eCBT (or similar) can be justified. If we have enough evidence to assess  how much the alliance accounts for in clinical outcomes, how can removing the clinician from the equation be moving  research in the best direction ?

Can we be confident that the scientists directly involved in psychotherapy research can bring about the change needed for the better research and better clinical outcomes? Or does the pressure to change need to come from external pressure?

Perhaps if more scientists ask these questions among their peers, and with their students, then we may draw closer to reconsidering psychotherapy scholarship.

http://www.vox.com/2016/7/14/12016710/science-challeges-research-funding-peer-review-process

George,

Thank you for sharing your work. Too little research has ignored the real-world truth that patients rarely look like those in clinical trials - they are complex, have comorbid psychopathology, and require a clear case formulation to guide treatment plans. Your study has truly highlighted this real-world challenge faced by clinicians every day. I look forward to reading more of your work in this area!

The ultimate purpose of conducting research - to inform practice, policy, theory, etc. -  is often forgotten or overlooked in the name of "tenure" or "impact factor." There is pressure to pursue the research that is likely to be funded and to yield significant results and less encouragement of creativity and novel development. As Jason mentioned, part of this may be changed through socialization practices in doctoral programs. I agree that this may begin to alter priorities adopted by the field. I think it is also vital to encourage rising researchers and clinicians to break out of the confines of the lab and academia to research issues faced in real-world contexts, such as community clinics, state hospitals, VA hospitals, etc. Too long has there been a disconnect between research and practice. Our field needs researchers willing to bridge the gap.

No, the debate isn't over, and probably won't be for some time.  And there are probably several reasons for this:

1. Proponents of the technique vs. relationship argument are both right.  That is, there are some people who say that specific techniques matter (e.g., you should use exposure therapy for fearful patients, or behavioral activation for depressed patients), etc.  And that is absolutely correct.  Patients receiving these treatment techniques do better on average than do patients not receiving these techniques.  So clearly what the therapist does, matters.  And, conversely, there are those who argue that much of what is "effective" in psychotherapy can be attributed to factors in the therapeutic relationship (e.g., a therapeutic relationship that is warm, empathic, and supportive will be more effective than one that is not).  And that is also absolutely correct.  So it's not either/or.

2. Technique and relationship are inseparable, at least with the current data.  As far as I know, the information we have about the effects of the therapeutic relationship are all correlational.  That is, we look at a bunch of therapist-patient dyads, rate them for characteristics such as warmth and empathy, and track the patients' outcomes.  That's a very reasonable thing to do, but we have no way to disentangle the factors.  Often, we find that the therapeutic relationship is going well because the patient is doing better.  Patients who are feeling better are happier, and more cooperative with their therapists, and the therapists are happier too, so the relationship is strong.  But that's not the same thing as being able to say that the strong relationship caused the patient's improvement.

That distinction between technique and relationship gets even fuzzier when we try to define what we mean in terms of therapists' behavior.  If the therapist is supposed to make empathic statements, express support, etc., how is that less of a "technique" than things like approaching scary topics and activities (as in exposure therapy), challenging maladaptive beliefs (as in cognitive restructuring), encouraging self-motivational statements (as in Motivational Interviewing), or reinforcing more healthy in-session behaviors (as in Functional Analytic Psychotherapy)?  Put another way: how do we know when the therapist is doing a "technique" and when he/she is not?  There's no good answer to that.  My personal preference would be that we eventually let this false dichotomy of technique vs. relationship go, recognize that good therapy is a mix of those, and reframe the whole thing as "Stuff a Therapist Should Do."

3. We haven't figured out what the right techniques are, nor have we figured out what the right relationship is. There's clearly no consensus here, and even if there were one, whatever we're doing isn't really all that great.  Let's take, as an example, CBT for anxiety disorders, which many people would suggest is The Thing You Absolutely Must Do.  Response rate, across studies, is only around 50%-- and for other disorders, the percentage is even lower.  I say this as a guy who spends all day thinking about and doing CBT for anxiety disorders.  It's good, but it's not terrific. Same thing can be said for the "relationship" factors.  Yes, they account for a good percentage of the variance (depending on whose data you're looking at).  But we certainly haven't come up with a set of relationship factors (or, especially, therapists' relationship-oriented in-session behaviors) that come anywhere close to guaranteeing a good outcome.  On both sides of the issue, what we don't know still greatly exceeds what we do know.  

4. The available data are a god-awful mess.  Much of this issue gets argued in meta-analyses, in which multiple trials are synthesized and pooled effect sizes are calculated.  The problem is that there is so much heterogeneity in these studies-- diagnosis, personality characteristics, type of treatment, type of therapist, length of treatment, methodological qualities such as blinded assessors and instrument quality-- that you could squint at the data and come up with virtually any answer you like.  This is why the "technique" people and the "relationship" people can look at the same sets of data and reach wildly different conclusions.

So-- the technique researchers have found some techniques that work and are not great.  And the relationship researchers have found some relationship factors that work and are not great.  Neither camp is in a vary good position to claim superiority, and we need a lot more carefully conducted research before we can have a good answer.

Really interesting question and answers, but it seems to me that the client is missing from this discussion, which seems rather inwardly focussed on the therapists's insights and implied capabilities. There seems to be a persistent assumption that there is, somewhere, a magic recipe (ritual?) that can be rolled out for every client. And behind that another assumption that every client is the average person, or pretty much is. I've been through the sausage machine of CBT, and was horrified at how disinterested they were in me. And these were some of the best. I really got the impression that it wasn't me who had trouble thinking clearly, it was them!

But what if people are hugely more individual than we give them credit for, with vastly different natures, preferences, and pathologies? Some clients will say, 'here I am, fix me'. Others will need to work things out for themselves. Many will mainly want to feel that they can actually be understood as an individual. What if choosing a technique involves piling further inferences on the inferences of assessment (and personality), and ignores what is actually known for sure - that the client is always actualizing, seeking to flourish, 'regressing toward the gene'? That the client has been working on this problem for longer than you have, and is quite able to make rational decisions provided their frame of reference starts to make sense again.

What if the actualizing tendency provided the technique, ready-tailored to each individual client, waiting to be nurtured back into the heart of the client's life through the kind of relationship that encourages neural integration? The kind of relationship that is the opposite of the ones behind the psychopathology. Because we can no longer blame random chemical imbalances given what we now know about how our relationships can make or break us.

What, therefore, if therapy is more like love than fixing a car, or medicine? In all but the simplest of cases I suspect that it is. And humans aren't simple. What if therapists were willing to work with their relational attitudes rather hiding behind personas, checklists and the awkward 'active listening' that pretends to be a 'relationship' (and ceased discussing the merits of relationships on that naive basis). What if the nature of the healing relationship could, and should actually be as individual as the client, because it fully incorporates, and emanates from the client?

How about we spend a few million research dollars on all of that?