By Matthew Oyer

Andrew Gerber, MD, PhD, is an Assistant Professor of Clinical Psychiatry in the Division of Child and Adolescent Psychiatry at New York Presbyterian Hospital, Columbia University.  He completed his PhD in psychology at the Anna Freud Center and University College London, where he studied under Peter Fonagy and Joseph Sandler.  He did his medical and psychiatric training at Harvard University, Cambridge Hospital, and Weill-Cornell Medical College – Payne Whitney Clinic.  He also trained as a research fellow with Bradley Peterson at the New York State Psychiatric Institute in brain imaging and child psychiatry.  Dr. Gerber’s primary research interests are in functional imaging and psychophysiology of developmental processes, particularly social cognition, as well as designing and evaluating trials of psychotherapeutic process and outcome.

Matthew Oyer: I thought I might start out by asking you a little about the Psychodynamic Psychoanalytic Research Society (PPRS), which you’re a founding member of.  What do you think CUNY students should know about PPRS?

Andrew Gerber: Sure.  So, PPRS has a long and somewhat complicated history.  Probably the first thing I should say is there have been many, many attempts in the history of psychoanalysis, really over a hundred years, of people who are interested in empirical research, or quantitative, empirical research in psychoanalysis and psychodynamic ideas to get together in an organized way.  And the fights about how to do that go back to the early days of the American Psychoanalytic Association (APsaA), to fights within the American Psychoanalytic, then in Division 39 in APA, and in a variety of other forums.  So, the last thing in the world that I think, or anyone in PPRS thinks, is, naively, this is just some simple thing where you put up a flag and say psychoanalytic researchers get together.  We know how complicated this is.  But that said, there’s a group of us that felt that really none of the major psychoanalytic organizations, whether it’s the American Psychoanalytic Association, Division 39, the International Psychoanalytic, really had the structure or the political will to make empirical research a big focus.  It’s not to say there weren’t many people within each of those organizations who were interested in empirical research, who were friendly to it, who at various times have supported it in many ways.  It’s just to say there were other big, big things they were struggling with, and this wasn’t top on their agenda.  And that for any of us who were empirical researchers, who were interested in empirical research in psychoanalysis, to invest the kind of time it would need in one of those organizations, to build it from within, which in theory was possible –  there’s Section Six within Division 39, there are certainly numerous committees within the American Psychoanalytic, the IPA has numerous committees devoted to research – but the amount of time and energy it would take within that organization to fight for it – which some people had done, which Stuart Hauser was a big champion of, which Peter Fonagy was a champion of in the International, and which other more contemporaries of ours had done, guys like Bill Gottdiener and Ken Levy in APA, and Pam Foelsch had really fought very hard for – we just thought we didn’t have that resource, we didn’t have the time and energy to do that.  But if we could do it a little separately, maybe we could do it without offending anybody, without trying to replace anybody, but just sort of making a smaller, more efficient organization that would have less of the historical baggage of each of the rest of these organizations.  That was really the idea.

Practical terms, it arose out of a bunch of discussions that took place at the American Psychoanalytic that were led by Stuart Hauser, who was one of my mentors and who was a wonderful guy who really embodied this idea of getting people together in a non-turf battle kind of way, and Linda Mayes, who was another one of my mentors at Yale and who, again, is somebody who just doesn’t have a jealous or territorial bone in her body but wants to do good work and does good work, and wants to do it with colleagues she enjoys.  And, I think, particularly when Stuart died, we all felt, Linda but all the rest of us, a real responsibility to see his vision through.  And that was the motivation for trying to bring this group together.  Obviously there are bumps along the way, and we can certainly talk about them if you want, but what I would say I want, and this is all sort of a preface to saying what I would want a City student to know is that despite all the history of rivalries in our field and the political complications, there is a core group of people, I believe, internationally, not just in the U.S., but in Belgium, England, South America, who are MDs and PhDs, who are men and women, who are more Freudian and more Kleinian, who are imagers, and geneticists, and psychotherapy researchers, who really span the spectrum, who want to have a place where they can go and talk about psychoanalytic research in a friendly way, and where they can bring their students and where they can meet other students, a place that kind of feels home for them for that sort of thing.  And certainly one of the hopes would be that grad students at City, who are so well trained in psychoanalytic thinking and, I think, most of whom are trained and most of whom are interested in some form of empirical research, even if it’s not the career path, the primary career path they intend to choose, there’s a lot of people at City who would be interested in that.  And I guess we always hoped that if there were barriers or political reasons why they weren’t, that we would learn about that and use that to make our organization more inviting.

MO: And what are some of the projects, ongoing ways you guys get together, that type of thing?

AG: So, it’s a good question.  The whole idea of a community has to get down at some level to what do we actually do.  And that’s been somewhat slow.  The first thing we did, really, was, well two things.  We started a website that was going to be just a 21st century kind of bulletin board forum where people could know about us, and we could list people who were members and they could get in touch with each other if they wanted to, and some of the things we’re interested in doing.  So that’s one thing.  And the second was really to kind of consolidate a bunch of the activities that go on at the American Psychoanalytic [conference] that were very separate and, in a way, even though individually all very cool, not so efficient because they weren’t organized and the American didn’t have a structure to organize them.  But really to make a structure that could organize that.  So, for example, to bring together seminars on Saturday, which used to be primarily a function of RAAPA, to bring together the poster session, to bring together the big lecture on Friday, which used to be run by Stuart Hauser and Bob Waldinger.  And to some extent to build that in a unified way, to then find more time, and more space, and to argue for more time in the program for these things, not in an antagonistic way but in a collaborative way.  And I’ve been on the program committee at the American for the last five years or something.  Amazing bunch of people, but not researchers; that’s just not who’s been attracted to that sort of thing.  And to basically say to the program committee, “Hey guys, we love all the stuff you do, and we know most of you are interested in the stuff we do too, but this isn’t your field so much, why don’t you let us help you organize it?”  And, in general, that message was extremely well received, in saying, “This is great; we love the idea; we don’t know how to do it; it’s not our thing, and we’d love help organizing it.”  So, we’ve done that now.  We’re on our third year.  And the goal every year, larger and larger, to make the process by which we select things more open and democratic, in the sense that it’s not a popularity contest, as much as you can prevent those kinds of things, but actually one with very specific criteria that then goes to a kind of public selection committee, really on the model of other scientific organizations that do that.  Not to suggest that that’s any sort of perfect way, but I do think it’s better than what historically has been the process in the American, which was a lot of turf.  In other words, somebody had a seminar, and it was up to them to choose someone, and they could be great and choose exactly the right person, but it’s also an incentive to choose someone who agreed with them or a friend.  And so much of the program, the general program but also the scientific program, that we thought it was discouraging to people who were used to the other kind of model in the scientific world, and also discouraging to students because, if you’re a new person and nobody knows who you are yet, how are you supposed to get on the program unless you buddy up with someone who is in charge, as opposed to you send in a proposal and people read it without even looking at names and think, that’s cool stuff, and they don’t know if you’re 75 or you’re 25, if you’re at Adelphi or Berkeley.  So, that was a big part of it.  Those are really the two things.  We had ideas for other things, which we’ve been mulling over, but we’re trying to be careful not to spread our resources, which are not huge, too thin, because then we could end up doing a lot of things poorly instead of doing a couple things well.

What else would we like to do?  We’ve always been interested in having a listserv, but the psychodynamic research listserv, which everyone confuses for the PPRS listserv, and I don’t know any way to, other than just explaining it, they’re different, they’re separate.  Of course, heavily overlapping groups of people, and I was heavily involved in both, but they are unrelated to each other.  We don’t need a new listserv to some extent because that listserv has been so successful.  Now, we could have a new listserv in one respect.  The psychodynamic research listserv, not the PPRS listserv, was always intended from the beginning, and this has sometimes made people not like it, was always intended to be for people who were doing and publishing research; it was not intended to be a place where people who were not doing research came to learn about research or to kind of kibbutz about the issues around research and dynamic thinking.  And it was because when we founded the listserv, a bunch of people said to us those listservs have existed, and we need a place where we can do our work, not where we can get into old and never ending debates about whether research is valuable or not.  That’s not where we want to be.  We want to talk about measures, and, if you’re not a researcher, that’s boring.  We want to talk about references, and so that’s why we designed it that way.  And so there is theoretically room for a new listserv that would not have that kind of exclusivity to it, that would just be a more general place where people who are interested in research, no matter what their backgrounds, can talk.  The problem is that I think all listservs need, in some sense, they need to be driven by a core group that writes a lot.  And we’re all on listservs; we all know when those can get annoying, but my sense, being on a couple listservs, is that, without those people, things kind of tail off.  And, obviously there are good and bad versions of it, and I’m not trying to say everything is perfect.  But we don’t have that yet for a new listserv.  We have the structure in place; there is a listserv.  It’s not really used.  We think maybe someday it will be nice for it to serve that purpose, but we haven’t gotten there.

We’ve always wanted to have a blog.  That is some kind of place where people post kind of news updates, like the latest article in Science, or the New England Journal, or the New York Review of Books that’s relevant to this discussion could get posted, and, therefore, people could follow research in psychoanalytic areas without making it a full time job to read all of the journals.  We’ve talked about that.  The task is finding someone who’s got the time to do it, which isn’t easy.  A lot of people are qualified to do it, but the time is hard.

And then, down the line, we, of course, would love to have more scholarship opportunities.  So, one idea would be, as soon as there’s enough money, to give prizes for papers, to give prizes for research projects like pilot grants, maybe travel money to come to meetings.  All that, obviously, would require a lot more fundraising, which we haven’t yet embarked on.  But I think the fundamental idea that if you get like-minded people together, who really want to make this thing work, they find fun, cool things to do together, and we sort of have faith in that process.

MO: And you guys did something that I think was somewhat unusual and kind of great, in that you, at least at the last American Psychoanalytic conference, you recorded everything and they’re posted on the website.

AG:  Thank you for reminding me of that.  I forgot to mention that.  That’s right.  So, for the last meeting, exactly what you said.  We recorded everything and posted the slides and MP3s of almost everything on the website.  Now we did make it open only to members.  And, listen, in some theoretical sense, I wish we didn’t do that; I wish we could make it public to everybody because that’s what we want.  But we did that because otherwise you get asked the question, and it’s a very reasonable question, “Why should I pay the dues if I get everything without paying them?”  And altruism goes only so far when you’re a poor student or a poor faulty member, and its $100 or $50 even.  So, we did make that, and it did help us actually a lot to raise money we really needed.  And so far, our largest expenditure of the money has been paying for foreign speakers at the PPRS meeting because there are so many wonderful people around the world doing psychodynamic research, and it’s one thing to go to those meetings when you live in New York, which we do, but it’s another if you live in Belgium, or London, or Germany.  And we wanted to help those people.  It’s not like we paid their whole tickets because we didn’t have that kind of money, but we helped them, and that’s, to me, that justifies whatever trickery we have to use to get a little more money because it enriches the group.

MO: And your dues are not terribly much for students as well.

AG: $50 for students.  I actually campaigned for $25, but I lost that.

MO: Ok.  Next I wanted to ask about your meta-analysis of RCTs for psychodynamic therapies that appeared in the American Journal of Psychiatry in January.  I guess, just to get the ball rolling, can you highlight your findings?

AG: I mean there are a few things I would say to summarize it.  I would start by saying, when we reviewed the literature, and we really made a big effort to look high and low for any published paper that described a Randomized Controlled Trial (RCT) of a psychoanalytic treatment, self-defined psychoanalytic treatment, it’s not like we were going and verifying each one, compared to something else, whether it be another psychotherapy, a medication, or even another psychodynamic therapy, we came up with quite a decent number, I think 94 trials… which is surprising to pretty much everybody who hears that.  People think of psychodynamic and psychoanalytic therapies as pretty much not having anything in the way of randomized controlled data, and then get into all sorts of discussions about whether that’s good or bad, and that’s a separate issue.  But to at least know that these 94 trials exist was an interesting thing.  And, of those 94, you can pull out 103 comparisons between a dynamic treatment and a non-dynamic treatment, and the reason those numbers are different is because some of those trials you have to throw out because they’re comparisons of two dynamic treatments, that’s just a few, and some of those have more than one comparison of a dynamic and a non-dynamic treatment: it could be CBT and IPT, or CBT and meds, or something like that.  So, that’s where you get the 103 number.  So, that was the first kind of interesting thing.

The second kind of interesting thing is that if you go and rigorously, as rigorously as you can, look at the quality of those trials using well accepted, standard measures of how one rates quality, or thinks of quality in Randomized Controlled Trials, that the trials are not that bad.  That is, they have been gradually improving over the years, that the quality of those trials meet kind of basic, minimal standards for being well done, some of them are extremely well done, and that our sense from that is that, in that, they are probably not different from trials of other psychotherapies.  Now, parenthetically, and it’s not in that article, we’ve now submitted a second article with data to show that it’s really not different than, for example, CBT for depression, that those lines that show the quality improving over time of dynamic RCTs and CBT for depression are completely indistinguishable and overlapping.  In other words, they’re both increasing and neither is higher or lower, neither has a larger variance or smaller variance; they’re identical… which is really remarkable because we think of CBT as being the gold standard, and it is extremely well demonstrated.  Now, the number of CBT trials is much, much greater, but the quality of them is probably not that much different, and that really addresses a critique you’ll often hear, even when somebody says, well, maybe you have 94 trials of dynamic therapy, but they’re crappy.  Well, it turns out they’re no more nor less crappy than anybody else’s RCTs of psychotherapy.  So, that’s number two.

The third thing, and this is, I think, the one that raises the most controversy, is how do we summarize those RCTs, what do we do when we put them together?  And different people do that in different ways.  And, like with any kind of piece of data, that’s what data is for, so people can come up with different models and then test them.  One could look at those trials and say a significant number of them show dynamic treatment was effective.  And that’s, I think, from my reading of the literature, mostly what’s been done up to this point… which is important, and I don’t think wrong, but, I think insufficient.  And we tried to take that a step further, and we did it in the following way: rather than focus on an effect size of every individual trial and then bringing them together in a meta-analysis, which is obviously a time tested and very rigorous method for bringing together multiple trials, we had a slightly different idea of how to do this.  The first thing we did was divide the trials up into those that compared a dynamic treatment against what we called an active comparator.  And, by active, what we meant was, and it’s a little hard to operationalize this, but by having gone through all these trials we came to agreement about this, and I don’t think it’s a particularly, when you get right down to the nitty-gritty, it’s not that political a question, that is, you can agree on these things even if you don’t have the same theoretical perspective: an active comparator is a type of treatment that the person who designed it, usually the primary investigator of the study or the primary investigator in that cell, believed in that treatment, used their best abilities to make that treatment work, and was motivated to make that treatment the best treatment it could be for these people.  That’s broadly what I would call an active treatment.  An inactive treatment is anything that the primary investigator or the person leading that cell did not have that motivation to do, and very understandably, and I’m not criticizing this motivation; in fact, I think in some ways these people did something very smart that we should all be doing more of, they planned a comparator that was okay, but was missing fundamental ingredients that they thought would be necessary for a good treatment.  That is, they designed a treatment that was meant to be mediocre.

MO: And this is what Wampold is getting at when he talks about allegiance effects.

AG: Correct.  The term Wampold uses is a “bona fide” treatment.  Same general idea.  We didn’t end up using exactly Wampold’s classification system just because of small differences here and there, which I don’t want to get bogged down into, but it’s essentially the same idea.  I don’t know his work well enough to know exactly his use of the word allegiance.  Allegiance has been used by other people as well, but it captures the same idea.  Now it’s often hard to get people to tell you exactly what their allegiance is, but if you have someone running the CBT arm of a trial who is a sworn dynamic therapist, it’s a little hard to convince me that they want the CBT to be just as good as the dynamic, as opposed to you partner up with your friend down the hall who is a CBT researcher tried and true, and, in a friendly but adversarial way, wants their treatment to beat yours.  Those are different types of comparisons.  Now, I’m not saying that one is right and the other is wrong because I don’t think that, actually.  But I think you have to make the distinction.  Because here’s the thing, when you make that distinction, and we did that, the numbers of trials that come out in different ways are drastically different.  I’m not going to remember all these numbers, but the quick summary is when you compare dynamic treatment against what we would call an inactive treatment, something that was designed to be mediocre, dynamic treatment wins almost all the time, never loses.  Because dynamic treatment is good treatment, right.  It’s better than not good treatment.  However, when you compare dynamic treatment against another good treatment, CBT, IPT, medication, something that’s been really thought about and done well, it almost always ties.  And sometimes dynamic comes out a little ahead, sometimes the other one comes out a little ahead, but it’s about equal.  The vast majority of the time it’s a tie.  Now, do I think that means there is no difference between dynamic and other treatments?  No, of course I believe there are differences, but is the design of the studies we have to date,  the design where we throw pretty heterogeneous groups, maybe they’re homogeneous on diagnosis, depression, or maybe they’ve got panic disorder, or borderline personality disorder, but within those, we all know, there is extraordinary heterogeneity, and that heterogeneity has not been studied in this way, or at least not very well.  We have these incredibly heterogeneous groups; we have small sample sizes because these studies are hard to do – so the largest sample size in the whole thing was this group in Scandinavia that had several hundred, but most of these trials we’re talking 30-40, if you’re lucky, in each cell – and then you had treatments that were clearly not uniform either, and I’m not even saying that treatments have to be uniform.  Maybe it’s ridiculous to call treatments uniform.  We can get into that debate separately.  But the point being, that with all those things, the idea that we’re going to get really clear differences between two active treatments, I think, is something that theoretically we should probably give up on.  That there is one right treatment for depression, there is one right treatment for anxiety disorders, there is one right treatment for borderline personality disorder, that seems to me an old idea and one that has been written a lot about in terms of the Dodo bird hypothesis and so on.

Now, that being said, I don’t think, and I’m sort of moving now from the results of the study to the conclusions as I see them, what does this mean?  A, I’m saying it doesn’t mean that we should conclude all treatments are the same and stop worrying about them.  So, that’s one.  B, I don’t think it means we should never do RCTs anymore because RCTs, to me, actually have a value.  And this is also a somewhat controversial idea, particularly in the dynamic world.  RCTs, in my opinion, are the best way we have of protecting ourselves from only seeing the results we want to see when we look at treatment outcome.  Now, they’re not foolproof ways, as we just talked about; allegiance effects go beyond that even when you have an RCT.  And there are other scientific controls people put into studies that try to get at the same thing.  So, I’m not saying they’re the only way or a perfect way, but they are a very good way.  In fact, I think the best way in this particular area.  And I do think it’s a reasonable claim, and this gets me into trouble sometimes, I do think it’s a reasonable claim when the general public, when insurance companies, when reporters ask the question, have you demonstrated that this treatment is better than either no treatment or some type of not very good alternative in a Randomized Controlled Trial, I think that’s a reasonable question.  The history of medicine and psychotherapy is filled with proposed treatments that theoretically, practically sounded wonderful, and, when you put them to that test, failed.  And, in retrospect, in my opinion, and I could give examples from other areas of medicine, should have failed, and it is good we found that out because all the theory, and hope, and faith in the world is not going to protect you from sometimes we just get these things wrong.  And, again, it’s not perfect, it’s not the only way to do these things, but it’s a pretty good way, and, therefore, I don’t regard the evidence based movement as a uniformly bad thing.  I think the notion that we apply some kinds of general standards and think very critically about our treatments is not only good, it’s necessary.  Now, how that gets done and the politics of how that gets done is another matter.  So, I’m not trying to say I endorse everything called evidence based, which can often sound, or even be, very anti-dynamic.  That’s not what I mean.  But I do think that if we’re going to help shape those standards from the dynamic community, we’ve got to participate, and the excuse they’re all against us, they wrote these standards in such a way that we can’t win, to me, wears thin after a little while because it means you gave up the opportunity to get into the fray and actually influence these things yourself.  So, anyway, that’s kind of my long answer to the second major thing that I would conclude.

The third is a little glib, but I think it follows from what I just said.  I don’t think there is anything wrong with designing a study where you put your good, active treatment up against an inactive comparator.  As long as you’re not hurting anybody, you’re not being unethical – that is you’re finding a way to get those people treatment when the study period is done, and it doesn’t last years, it last twelve weeks or even six months – that’s okay.  And that’s scientifically and ethically a worthwhile thing to do because it lets us know that, yes, this treatment is worth something.  And CBT, IPT, medications have been doing that sort of thing for decades, and dynamic researchers have done much, much less of that, not none, but less, in part because less research is done, but also in part because they think that, if we’re going to go up against somebody, we’re going to take on the toughest thing.  And I think that’s a mistake because when you then tie, nobody knows what to think.  Did you tie because your sample size was too small, because they’re equivalent?  It’s very hard to know that because you would need enormous sample sizes to be able to tell that.  So, I do think the dynamic treatment world should be doing more trials against inactive comparators, replicate those trials at multiple sites, and, in a matter of a couple years, dynamic treatments could be represented in every single category of the evidence based treatments, that is for depression, anxiety, personality disorders, where they’re only sparsely represented right now.  They could be all over that with a relatively small amount of work.  And we don’t have them to blame for that.  We have ourselves to blame because we haven’t done that.

MO: I wanted to ask, you brought up heterogeneity of the groups and the idea of packaged therapies and whether that’s really what’s going on, and it makes me think of Sid Blatt and his two primary critiques of the philosophy of RCTs.  If I understand them correctly, one he says the type of treatment is not the primary factor, and two that symptom reduction is not the most effective expression of therapeutic gain.

AG: Listen, Sid Blatt was my first dynamic teacher ever when I was a sophomore in college and essentially the reason I got interested in psychodynamics period, and it was only later I found out how important and influential he was in psychoanalytic research, so I love Sid Blatt.  I think he’s, not just a wonderful guy, but an incredibly important, and generative, and productive researcher.  So, it’s hard for me to disagree with anything Sid Blatt says.  So, both of the things you said before you got to the RCT part, I completely agree with.  Is the type of therapy the most important thing?  No, I agree with him.  It’s something much finer in terms of the technique that we’ve got to understand, and that’s absolutely our goal.  I can’t remember if you said this, but my first introduction to his work was the heterogeneity of depression and the distinction between the introjective and the anaclitic types.  Not only did I write my paper in his class, but I loved the idea so much that year all my final papers in all my classes used that model because I loved it so much.  I’m not joking, in architecture, in psychology, and, oh I took a history class, history of Freud, but I used it in there too.  So, I love that idea.  And that I think is also right, depression is a heterogeneous group; that’s what I said before, and, if you don’t recognize that, you’re never going to tease out the nuances between different outcomes.  All that is true.  But where I part ways with him, and reluctantly, is I don’t think that means that RCTs are evil.  I don’t think that means you never do an RCT.  You still do an RCT in my opinion, not because it’s going to unlock the answers to any of these more interesting and fundamental questions, and it may be that I would never want to do these RCTs.  I’m not trying to sell them; they don’t sound that interesting to me, and I know Sid wouldn’t want to do one.  But I want somebody to do that RCT because, without it, I’m a little nervous that we’re going to do what various groups have done for centuries, which is take for granted an idea that may not be right.  May it turn out to be that there are certain ways in which dynamic treatment is not the right treatment?  Yes.  We act like it’s obvious now, but we know that dynamic treatments don’t cure psychosis.  Now, it drives me crazy when I say this to a group of analysts, somebody inevitably raises their hand and says, “Analytic treatments are very useful for psychosis.”  Sure, but if you think it’s ethical to give up all the antipsychotics in the world and replace them all with psychoanalysts, I don’t think there are too many people in the world who would take that extreme a position.  Are dynamic treatments interesting and relevant?  Sure, even in psychoses.  But are they the primary and fundamental treatment?  No, and we’ve known that for quite some time.  However, if you go back to a survey that was done by the American Psychoanalytic Association in the early fifties – very large effort, expensive effort in the era before computers, where they surveyed about a thousand analysts on self report on about 10,000 patients – they overwhelmingly got the response back that patients with psychotic symptoms were not improving in analysis.  And rather than publish that – and this wasn’t even an RCT, this was before RCTs – but rather than publish that survey, they decided not to because they felt it would weaken the position of psychoanalysis.  It wasn’t published until decades later.  And to me, that’s a sign of a mistake.  Analysis has gotten into it on various occasions of claiming it could treat things it can’t.  We know that now, and so we feel better about it.  We know now, I hope everybody knows, that psychoanalysis is not a treatment for homosexuality, that makes people who are gay quote unquote “better”.  But people didn’t know that in the lifetime of many living analysts.  And I think it would be very shortsighted for us to pretend that that sort of thing will never happen again just because we’re smarter now.  Those analysts were pretty damn smart, and they made a couple big mistakes, and it would not surprise me if we’re making some big mistakes.  I don’t know what they are, but I’d like to think there are some mechanisms that we’ve put in place that didn’t exist back then that make those mistakes less likely, and, despite all their shortcomings, I believe RCTs are such a mechanism.

MO: So you see them as a system of checks and balances?

AG: Correct.

MO: And what about looking at the idea that symptom reduction is not the best measure of outcome.  Are there other ways to measures that could be added to RCTs?

AG: Yeah, and I again, I completely agree with Sid, and I wouldn’t be a dynamic therapist and an analyst if I didn’t think that.  And, listen, any good CBT therapist that I know, and I know lots of them, they don’t think that either.  They don’t think symptom reduction is the be all, end all.  Everybody thinks that one has to get at the underlying roots to these things.  The question isn’t, “Is symptom reduction the only game in town?”  The question is, “How do you do it?”  And in some cases focusing on symptoms really helps with the underlying things.  It doesn’t mean it’s your only goal, and it doesn’t mean it’s the only thing you should measure.  It’s the easiest thing to measure.  But I don’t think there’s anybody who would disagree with the idea that if you could develop techniques to get at more underlying constructs, you should be doing that, whether it’s the SWAP, or whether it’s measures of social functioning, or measures of cognitive functioning, that’s not controversial, it seems to me.  And criticizing RCTs because they use symptom measures as primary outcomes, it’s setting up a straw man.  They’re not claiming that symptom measures are the whole story.  Now could it be that it gives CBT an unfair advantage over a dynamic treatment because CBT is more focused, and there’s a good name for this but I’ve forgotten the term, sort of a more specific, measurable goal?  Yes.  So, if we saw a rash of studies showing that CBT was better than good dynamic therapy, I would bring that up.  We don’t see such a rash of studies; they don’t exist because they usually tie.  So, I’m not that worried about that problem.  I like more interesting measures, and I think every study, CBT or otherwise, should include them, but I also recognize they’re hard to do, and we should fund guys like Shedler and Westen to do the SWAP everywhere because it’s a wonderful measure.

One last thing, and this is where I think analysis gets itself into what I consider to be slightly crazy territory.  A person, who will remain nameless, who is an analyst, once said to me, “If you can measure it, by definition, it’s not psychoanalytic.”  Now, that definition is foolproof, right.  I can’t prove to him that I can measure something psychoanalytic because he has defined psychoanalytic as that which I can’t measure.  As soon as I measure it, he’s not interested in it.  Now, as a debate tactic, it’s airtight.  As a way to think of our field, and as a way to think of our relation to research, I think it’s crazy.  Are there things I think we’ll never be able to measure?  Absolutely.  I’m not so naïve that I think we can put a number on everything.  That would be crazy.  But I want to try to measure as much as I can.  And I want to try to measure things that are really important, not just symptoms.  And I think we can do an okay job.  And what does okay mean?  Okay means we can do it well enough that we can learn something to make our treatments better, to make sure we’re not doing anybody harm, that we constantly refine our ideas, that we test hypotheses when there are disagreements, and that we move the field intellectually forward.  That’s good enough for me.  It doesn’t have to be perfect or that we have to expect to measure everything, just the way in physics, or in chemistry, or in cell biology you can’t measure everything, and nobody gets worried about that.  But if we decide that the only thing we’re interested in in psychoanalysis are the things we can’t measure, we have so narrowed our world that I fear that we make ourselves so much less relevant to patients and to helping people.

MO: I wanted to broaden a little based on that idea.  And, looking a little more globally, I wanted to get your thoughts for the general frameworks for the future of psychodynamic research.  In the late nineties Eric Kandel wrote a series of papers urging a closer relationship between psychoanalysis and cognitive neuroscience, and we’ve seen a lot of that in the last ten years.  In the same sort of broad vein, what paradigms do you see as being especially important in the future of analytic research?

AG: I think where I would start, it’s a good question, is that what seems to me the most important thing is a cultural shift.  There are many, many ways in which, almost an unlimited number of ways in which the dynamic model, which is such a powerful, interesting, comprehensive model – there’s just nothing like it in psychology – can interface with other areas of empirical research, whether it be trials and therapy outcomes, or cognitive neuroscience, or sociology, or genetics.  There’s virtually an unlimited number, and I could talk about my favorites, but the problem is the culture because, and this sort of gets back to what I was saying before, if the analytic world stays separate from that, my fear, and I think what is already happening, is the world, the scientific, intellectual world will move on without them.  And that is, I think that, ultimately, many, if not all, the ideas that are embedded in psychoanalysis are going to emerge in other fields as well.  And we’ve seen this; there’s an explosion of this.  Cognitive neuroscience, an enormous percentage of it, whole journals are devoted to social and affective neuroscience, which twenty years ago was not something they talked about in cognitive science but which we always talked about in analysis.  You could say it was our idea except I don’t think it started being talked about in cognitive science because analysts suggested it.  They had to get there on their own.  Now, maybe they could have gotten there earlier if we had been more involved with them.  There is an awful lot of discussion, I don’t think there’s a whole journal on it yet, but I bet there will be relatively soon, on unconscious processing.  Now, you may say it’s not the analytic unconscious, and it’s not totally.  They talk about the cognitive unconscious, although, to be honest, I think the distinction is a little bit arbitrary, and, when you actually analyze it, it sort of breaks down.  But no cognitive scientist questions in any way what we would call descriptively unconscious, or preconscious, or unconscious with a small “u” processing.  How much of that is dynamically unconscious, how much of that is actively repressed, well that’s a separate question, but I actually don’t think too many cognitive neuroscientists would argue with the idea that some things are, there is motivation for keeping them out of awareness; there’s quite a bit of evidence for that now too.  The question is how much, and that’s a different question.  So, again, an idea that analysts have had for 100 years and cognitive neuroscientists have only been exploring for the last, what, ten, since Kihlstrom’s paper[g1]  in Science.  So, again, I think they could have gotten there faster if we’d been more involved.  And I think that’s going to continue to happen.  Concepts like intersubjectivity, narrative coherence, projective identification, transference, the role of development, the role of sexuality, all of those are extremely powerful ideas that have not totally entered the mainstream of cognitive science.  Little bits here and there.  But engagement with that world is helpful to everybody.  But you can’t engage with a bunch of smart intellectual scientists by saying, A) you guys have nothing to teach me; I only have things to teach you because it’s arrogant and it’s just wrong; and B) you can’t expect them not to be skeptical and to then want to devise ways to test your hypotheses.  Now, it may be that their tests are not good tests, and you tell them that, but you can’t be offended by that, you can’t be afraid of it, you can’t run screaming in the opposite direction and say, “You’re all out to get me!”  Because, once you do that, they’re going to retreat too, and then you’re going to have these two very separate worlds.  Will analysis continue in some sense?  Of course, but I think it will lose its scientific credibility, it won’t progress as much, and I think, ultimately, most of what, if not all of what it’s accomplished over the last 100 years will have to be re-accomplished by the world of cognitive science, which is a pity, and it shouldn’t be that way.  So, to me, that’s the most important thing that we could shift that.  Now, obviously, that’s very hard to shift because there are many, many years of animosity between two groups that, like all civil wars, it’s very hard to say who started it.  Are there many examples of analysts being mistreated in psychology departments by anti-analytic colleagues, or by funding organization, or by journal reviews?  Of course.  Are there lots of examples from a somewhat earlier era of non-dynamic clinicians and researchers being incredibly mistreated by analytically-inclined departments?  Yeah.  We’re forgetting those because that’s older now; that’s certainly not happening much anymore, but it happened too.  And, when you really get down to it, I think there’s an enormous amount of tit for tat where the people who are excluding another group are doing it because of exclusion they experienced earlier on.  It doesn’t take an analyst to know that that’s kind of human nature.  But that’s a cycle we have to break, and we’re the ones to do it because we have to say the fights our parents, and our grandparents, and our great grandparents had are never ending and, ultimately, counterproductive.  And if there are ways to get along with the cognitive scientists, and to learn from them, and to have them learn from us, and maybe even, at some point, to blur the distinctions – because I’m not sure the distinctions even make that much sense, to intermarry you might say – I think that’s better for all of us.  It’s hard, and I think it takes a conscious paradigm shift, but it’s happening in certain places.  Certainly, when I think of my mentors, whether it was Fonagy, or Hauser, or Mayes, or my senior colleagues for whom I have so much respect, whether it’s Ken Levy, or Mark Hilsenroth, or Elliot Jurist, or Arietta Slade, or Diana Diamond, I believe that this is the spirit that they’re bringing to it, not one of division and of combat, but of let’s figure out how to do this together.

It’s hard for me to say then where that should happen.  I mean I absolutely think it should happen in cognitive science, and I was giving those mostly as my examples, but I think it should happen in psychotherapy research too.  There’s no reason why CBT, IPT, DBT, and dynamic researcher couldn’t work together and not be secluded into different departments or divisions, and that’s happening.  I think it could happen in all sorts of areas, and that’s the exciting part.

MO: I wonder if I could sneak in two somewhat related last questions.  The first is if you could give advice for grad students who are interested in both psychoanalysis and empirical research, what sort of paths, or what are some of the necessary experiences they should seek out?  And related by the idea of necessary experiences, but also very different, what do you see as the necessary experiences in becoming a psychoanalyst?

AG: They’re both very important questions.  They’re two totally different questions.  The first question, the first thing to say, and this is a very practical issue is to figure out, given the constraints of the world we work in, what one wants to devote one’s time to.  And, obviously, the theoretical ideal for me is that we could all find time to do both, but the world just isn’t set up that way.  The demands of clinical practice, and of being a good clinician, and of paying the bills through that means are such that to have a second job at the same time is usually not possible.  And at the same time, the demands of being an academic researcher, and to get grants, and to write papers, and do all the stuff that goes along with that, means that to be a full fledged clinician is not so practical.  So, I think one has to figure out where one wants to fit in.  And I hate this; I wish it didn’t have to be this way, and believe me, when this was said to me earlier in my career and is still said to me, I get angry because I don’t want to be told I can only be one and not the other.  Do I do clinical work and research?  Yes.  Would I do better research if I wasn’t spending time doing clinical work?  I’d be more successful at research; I don’t think my research would be better because it would lack the clinical element, but I’d certainly have written more papers and have gotten more grants.  No question.  Would I be a better clinician if I were not doing research?  It’s harder for me to say this, but I think I probably would be because I would spend more time, knowing more patients, learning more techniques, discussing my patients with more colleagues, doing supervision, all the things that are essential for good clinical work.  Now, I don’t think I’m a terrible clinician.  I wouldn’t do clinical work if I thought that meant I couldn’t be good at all.  But am I as good as I could possibly be?  No, things just take time, and I have to be honest with myself about that.  So, I think people have to make a decision.  I think it’s possible to do what I do, and there’s a bunch of people who do, and accept the sacrifices on both sides, or one could decide to be somewhere else along the spectrum.  One extreme would be clinically you could be interested in research, you could read about it, but that’s just not what you do or have time for.  Or you could be a full time researcher, be interested in clinical work and be respectful of it, but still never do any of it.  And there’s an infinite variation in the middle.  It gets harder when those niches aren’t as carved out.  Maybe more of those niches should be carved out.  For example, it seems to me one thing that would make so much sense and there are people doing, is to be a full time clinician, that is to be the best clinician you can possibly be, but to spend some of your time, and sometimes its reimbursed and sometimes its not, meeting with and doing clinical work as part of research trials.  Then you have an opportunity to interact with your research colleagues, enrich them with your clinical knowledge, and you get enriched with their research knowledge.  And I think there are a bunch of really wonderful clinicians, some people at Columbia like Larry Sandberg, and Fred Busch, and a couple others who, for example, work in Barbara Milrod’s trials.  They’re not full time researchers; their names go on papers, but they’re not the one’s sitting in the office getting all the grants; that’s really more what Barbara does, who does less clinical work than they do but still does some.  It strikes me that they really enrich everybody by doing that, and I’d like to think that more and more, as things go forward, so many of those things will be available that you can come out of City or Adelphi or wherever, and really just make a very individual choice about where you want to be on that spectrum based on what you like to do, and what kind of environment you want to be in, and what kind of money you need to make, and so on.  So, that’s sort of my idea about that, that no one gets tracked or pigeon-holed, but that you learn a lot, and then you can find your way.

The second question, which I understand why it’s related even though I think it’s a different question, is, “what does it take to be a good clinician?”  And, listen, I’m certainly no expert in that.  There are many, many supervisors who know that much better than I do.  I guess there is some research on the subject, but I don’t think there’s very much.  I can tell you from my own experience, I think a personal analysis is essential.  If I took the three components of my psychoanalytic training, which I did at the same time as my psychiatry residency, so they’re kind of intermingled, I think my personal analysis was the most important for me.  I don’t think it’s the only thing, but it’s the most important.  I think seeing patients and having a diverse but good group of supervisors is second.  I just can’t imagine being a clinician without having had a huge range of supervisors.  One of my supervisors used to say to me that it’s when your supervisors disagree that you learn how to do therapy because then you have to decide what you think is right and not do it because someone told you to do it but because it feels or you think it’s right in various ways, and you test it out and see what happens.  And third, and I’m not trying to suggest any one of these is not important, is theoretical training.  The theory of dynamic work, the theory of all psychotherapy is very complicated, and it really does require a lot of study.  Now, I don’t always agree with the way theory is taught.  The first two things, I don’t actually see how one could do them any differently.  I think you have to have a personal analysis, and I think you have to see lots of patients and be supervised by lots of people.  I don’t think there’s any way to change that much.  But I do think theory could be taught differently than it’s taught now.  I think theory is taught in a heavily historical context right now.  That is, you’re taught, in most analytic institutes including Columbia and I think all of them, you’re taught Freud first, and there’s this gradual progression from Freud into other schools.  I don’t think that’s the most effective way of teaching clinical theory.  And my comparison would be it’s not the way we learn math, it’s not the way we learn physics, it’s not the way we learn chemistry.  We don’t learn what Newton thought, and then only gradually more and more recent.  We learn a framework that we have now, that they didn’t have back then because we have the benefit now of all those hundreds of years of change, and the same is true in psychotherapy.  We have frameworks now that are just better, in my opinion, than what Freud had.  It’s not an insult to Freud; he was a genius, he was a pioneer.  But I just don’t see any reason to think of what he wrote as some sort of sacred text that should be read before and is some final authority or resource on every topic.  It’s not.  It was a first attempt, a good first attempt.  And is there stuff in there still that I learn from?  Absolutely.  Should we read it?  Sure, but not in the way that it’s read.  I think we should start with much more contemporary perspectives.  I think we should integrate research and empirical ideas and observations much more from the beginning into the way we think.  We should build the frameworks of different ways of thinking of things, and then, as time goes on, you can learn some more of the details of the individual schools, probably mostly contemporary, going back to early literature when it’s essential.  But, listen, if you look at the curriculum of a lot of analytic institutes, and I don’t know how it is at City – I think it’s probably a lot better than this – but you can have classes where there is no reading more recent than 1980, and in 2011, to me, that’s inexcusable.  There is no area of psychology that stopped in 1980.  The teachers may have stopped reading in 1980, but that’s not a valid reason.  It should be taught by someone who is trying not to include anything before 1980, not because there weren’t smart people who said important things before 1980, but it’s probably been re-said in a better way since.  You want to have history courses, that’s fine.  You want people to have that as their specialty, that’s wonderful; I love history, and I love historians.  But that’s not the work of becoming a therapist, in my opinion, or a good analyst.  It’s more having a contemporary, complex view of the world.  And I know one counter-argument that gets made frequently is, “Well, if you learn too many things, you’ll just confuse them and smoosh them all together, and, therefore, that’s why you have to learn things from the beginning separately.”  Maybe that’s true sometimes, but that is not the way I think learning takes place.  You don’t take your kid and only teach them to see and keep them in a silent room because you don’t want to impair their visual learning.  You don’t even prevent them, in most cases, from being exposed to multiple languages if you can.  In fact it’s good to expose them to multiple languages.  Are there situations where a kid with certain deficits, you have to be a little more conscious and careful about that, and, for example you don’t expose them to a second language because they’re having language problems?  Yes.  So, I’m not trying to make this an airtight, absolute argument, but, in general, I think we learn by getting lots of information – not random information, we need frameworks for it – but by learning many things at the same time and figuring out where to place the different things.  And I think that’s the way theory should be taught more at analytic institutes, and to what extent it’s not, and I’m in no position to comment on this because it may be at City that this is very different now, but I think in grad programs as well.

MO: All right.  Well, thanks, Andrew.