Initial Severity and Antidepressant Benefits: A Meta-Analysis of Data Submitted to the Food and Drug Administration

Last week a new article analyzing ALL the data from ALL the short term clinical trials submitted to the FDA for the licensing of fluoxetine, venlafaxine, nefazodone, and paroxetine was published in PLOS Medicine, [Public Library of Science, a peer reviewed open-access journal.]

Click Here to Read: “Initial Severity and Antidepressant Benefits: A Meta-Analysis of Data Submitted to the Food and Drug Administration” Irving Kirsch, Brett J. Deacon, Tania B. Huedo-Medina, Alan Scoboria, Thomas J. Moore, Blair T. Johnson, February 26, 2008.

The Editors’ Summary includes the following:


These findings suggest that, compared with placebo, the new-generation antidepressants do not produce clinically significant improvements in depression in patients who initially have moderate or even very severe depression, but show significant effects only in the most severely depressed patients. The findings also show that the effect for these patients seems to be due to decreased responsiveness to placebo, rather than increased responsiveness to medication. Given these results, the researchers conclude that there is little reason to prescribe new-generation antidepressant medications to any but the most the severely depressed patients unless alternative treatments have been ineffective. In addition, the finding that extremely depressed patients are less responsive to placebo than less severely depressed patients but have similar responses to antidepressants is a potentially important insight into how patients with depression respond to antidepressants and placebos that should be investigated further.” 

So was the marketing of these drugs not justified based on the submitted data, or are the authors biased as some advocates of SSRIs as first line treatment for depression claim?
Some interesting comments on the article are at:

Paul Mosher

Explore posts in the same categories: Controversial

34 Comments on “Initial Severity and Antidepressant Benefits: A Meta-Analysis of Data Submitted to the Food and Drug Administration”

  1. Tamar Schwartz Says:

    Comment from Arthur Meyerson:

    The efficacy of psychotropic medications such as the SSRIs is invariably variable as to efficacy across a wide range of studies which may vary in dose, phase of illness, duration of treatment, combination with other drugs and psychotherapy. While this particular meta-analyis is of studies done for FDA approval, and shows little efficacy in a small percent of all depressed patients, there are now hundreds of post approval studies that do demonstrate efficacy in the range of 50-60% with adequate dose, duration and appropriately selected patients. However, since analysis has never undergone testing for FDA approval (generally randomized trials vs. placebo) nor post FDA trials of its efficacy, one would have to favor SSRIs if one is in the practice of providing the treatment with the best supporting evidence. One would also have to prescribe CBT or Interpersonal Treatment before analysis since there are also good studies supporting the efficacy of these treatments in depression (against placebo and/or medication) but there aren’t any for analysis.

    I consider Paul Mosher an extraordinarily intelligent practitioner and contributor to our lists but in this case the article he cited is being misused as if it were a definitive analysis of all good research on the clinical efficacy of SSRIs. It isn’t and Paul’s summary may have been misleading to some readers of this list who aren’t familiar with studies done for FDA approval and those done afterwards.

    Arthur Meyerson

  2. Tamar Schwartz Says:

    Comment from Greg Roeben:

    As a psychopharmacologic researcher in a former life, let me say that I’m extremely skeptical of the efficacy claims made for anti-depressants. Paul and Paul have offered cogent and valid observations, and I have little to add.

    I would direct interested parties to The link is to the AP(sychological)A (now-defunct) on-line peer reviewed
    journal called Prevention and Treatment, which devoted an entire issue to the fact that, if one pools all of the available placebo-controlled trials of SSRIs, the anti-depressants perform no better than placebo. One respondent to the primary article is Michael Thase, a prolific psychopharmacologic researcher. In his response, he essentially resorts to anecdotal support for the efficacy of anti-depressants. The title of the primary article is “The Emperor’s New Drugs: An Analysis of Antidepressant Medication Data Submitted to the U.S. Food and Drug Administration.” Dr. Thase’s response–“Antidepressant Effects: The Suit May Be Small, but the Fabric is Real”—is essentially an admission that claims for the efficacy of anti-depressants are exaggerated.

    Like Dr. Thase, I have seen patients improve on anti-depressants; some I have no doubt would not have improved without anti-depressants. Some patients get better with analytic treatment. Some with ECT. Some with CBT. Would I treat all patients with all of these
    modalities? Of course not. Would I treat all patients with only one of these modalities? Of course not. Unfortunately, anti-depressants are promoted and marketed as THE cure-all.

    Most of you are probably not familiar with the CGI (Clinical Global Improvement) scale. The CGI measures clinical improvement of subjects in efficacy studies, including double-blinded, placebo controlled trials of anti-depressants. The CGI is often divided into two
    subscales: CGI-S (severity) and CGI-I (improvement). The CGI-I is sometimes a primary endpoint of a study, which means that it is the main comparator between the control (placebo) and experimental (anti-depressant groups). When it is not the primary endpoint, it usually is a secondary comparator, which means that it is used to support the findings of the primary comparator (in depression studies, the Hamilton Depression scale (HAM-D) of the Montgomery-Asberg Depression Rating Scale (MADRS)).

    The CGI-I is a seven point rating scale that asks the clinician: Has the patient improved a lot, a little, not at all, or worsened a little or a lot?

    Here’s a thought experiment: How would you rate the CGI-I of patients that you’ve treated with psychoanalysis? Wouldn’t you love to publish that data and be able to promote it as “objective data?”

  3. Tamar Schwartz Says:

    Comment from Henry J. Friedman:

    In a recent exchange of notes Paul Mosher and Mark Leffert clashed over the meaning of a large scale study of multiple studies of SSRI’s and placebos that concluded that there is good reason to question the claims of therapeutic effectiveness of this category of anti
    depressants. Mark later explained that he wasn’t calling Paul “an idiot”, a claim that on rereading his note appears to be true. Instead he was being harsh about psychoanalysts who seem to revel in the idea that anti-depressants are ineffective in order to cover
    their ignorance regarding current psychopharmacology. While I wouldn’t be inclined to put it in quite the way that Mark does I have been aware of many psychoanalysts who simply relegate medication and its prescription to a psychiatric colleague who does keep current with medication. It is still claimed that giving medication will interfere with the transference and its development.

    This position was quite strongly presented at the panel on Medication and Psychoanalysis at the winter APsaA meetings in New York. On that panel a psychoanalyst presented a case that in his view illustrated the way in which the patient’s use of prozac prevented him from becoming fully involved in the transference i.e. he felt that the patient avoided full dependency upon him the analyst because of the use of this anti depressant. The patient had been on prozac for several years before beginning analysis and simply wouldn’t give it
    up but continued to have it prescribed by his primary care physician. I was struck by the counter-transference of the analyst who was clearly angry that the patient wouldn’t give up the medication. More importantly the patient had a form of sexual behavior that I would have doubted as being amenable to change through psychoanalysis.

    What was important about the panel and this presentation was the view of this analyst and apparently many others that medication or prescribing medication was incompatible with psychoanalysis because it interfered with the transference. In this regard I agree with
    Mark Leffert that this is an unfortunate position that derives from the legacy of classical psychoanalysis and its unwarranted belief in the primacy of intra-psychic conflict, the unconscious and the transference in all patients. While I agree that there are instances
    in which patients resist insight and insist on relying upon medication when medication is doing little or nothing it is quite important not to introduce the idea of medication as resistance to analyzing. Many patients, as Mark notes, can’t progress in analysis
    without the use of an anti-depressant and in some instances an anti psychotic.

    In my experience patients in analysis who benefit from psychopharmacological medications often are eager to stop their medication for long periods of analytic work once they have enough freedom from uncontrollable depression or anxiety. There are many
    patients who resist going deeper into a dependent relationship without being on medication. Analysis requires acceptance on the analyst’s part of the patient’s limited ability to use the analyst as a helpful object. In general, there appear to be more individuals who as patients resist using medication as a helpful object than there are those who use it to resist a dependent transference.

    Henry J. Friedman

  4. Tamar Schwartz Says:

    Comment from Sidney Blatt:

    In the NIMH sponsored Treatment of Depression Collaborative Research Program (TDCRP) comparing 16 weeks of active medication, placebo, and two forms of psychotherapy (CBT and IPT),no significant differences were found in degree of symptom reduction among all four treatment groips at a flollow-up evaluation conducted 18 months after the termination of treatment (see Blatt & Zuroff, 2005). The NIMH TDCRP is probably, to date, the most extensive, comprehensive and well conducted randomized
    clinical trial comparing psychotherapy and medication See also a recent (2007) paper by Lance Hawley and colleagues (Stress reactivity following brief treatment for depression: Differential effects of psychotherapy and medication) in Journal of Consulting and Clinical Psychology, 75, 244-256. Sid Blatt

  5. Tamar Schwartz Says:

    Comment from Mike Kowitt:

    It was randomized clinical trials that brought us the SSRI’s and convinced us that not using them was malpractice. Now it is randomized clinical trials that tell us they don’t work any better than placebo. Some of our members are convinced that they really do work, in spite of these findings. What does all this say about the randomized clinical trial itself? The method seems to provide interesting findings. It raises interesting questions, but is it really the gold standard that we seem to envy? It seems to be vulnerable to all the same subjectivities and biases that we generally attribute to the clinical method. Perhaps we should rethink our inferiority complex around the alleged lack of scientific basis for

    Mike Kowitt

  6. Tamar Schwartz Says:

    Comment from Paul Mosher:

    Art Meyerson wrote:

    “While this particular meta-analysis is of studies done for FDA approval, and shows little efficacy in a small percent of all depressed patients, there are now hundreds of post approval studies that do demonstrate efficacy in the range of 50-60% with adequate
    dose, duration and appropriately selected patients.” “… in this case the article [Paul] cited is being misused as if it were a definitive analysis of all good research on the clinical
    efficacy of SSRIs. It isn’t and Paul’s summary may have been misleading to some readers of this list who aren’t familiar with studies done for FDA approval and those done afterwards.


    Thank you for your thoughtful response. I am amazed at some of the other replies on this list (and some on the members list where half the subscribers are unable to follow the conversation) which seem to be attacks of a somewhat personal nature, implying that skeptics about SSRIs are either ignoramuses regarding psychopharmacology or are some kind of drug Luddites. If that’s the best the SSRI-adherents can do, its a pretty weak form of debate.

    This issue of the possible lack of efficacy of SSRIs is not a debate of “crackpots vs. scientists.” It is a truly serious matter involving millions of patients and billions of dollars, with respectable and scientifically knowledgeable people on both sides of the discussion.
    The research community is struggling to try to figure out how it can deal with the confounding effect of the extremely high placebo response rate in studies of drugs for depression which tend to bury any possible specific beneficial effects — if there are any — in the “noise.” New kinds of research strategies are being considered.
    See, for example:

    Klein, D.F, (2008) “The Loss of Serendipity in Psychopharmacology” JAMA,

    Citing “personal experience” and offering anecdotes is an equally weak strategy in such a discussion. All of us have seen patients who appear to respond to SSRIs — some in ways that even seem dramatic — BUT THAT PROVES NOTHING, and in the light of past false paths in the field of medicine, some lasting for decades, and especially in the tricky area of
    mental disorders, we should by now be extremely “wised up” — skeptical about any so-called “remedies” until they are clearly demonstrated to be helpful via some SPECIFIC EFFECT (as opposed to a placebo effect.) Even single controlled studies with positive results, no matter how well designed, are no longer taken as proving anything unless ALL COMPARABLE NEGATIVE STUDIES are also taken into account.

    You now base your argument on a claim that published studies of SSRIs following FDA approval show benefits which cannot be detected in the overall corpus of initial SSRI data submitted to the FDA.

    While it is true that the article by Kirsch et. al. which I cited dealt with studies submitted to the FDA, that’s not the whole story. The Turner et. al. article published in the January NEJM MATCHED studies submitted to the FDA with comparable studies published in the psychiatric literature.

    The point of that article was to show that the studies published in the literature appear BY COMPARISON to be SELECTED so as to exclude negative studies. It is those PUBLISHED studies, not the entire body of FDA submissions, which appear to overstate any benefit that SSRIs provide, a benefit which seems to be marginally detectable and difficult to discern within the total body of studies submitted to the FDA (which include all negative results along with the positive studies.)


    As to some of the complaints about meta-analytic reviews of published studies, I think the entire story is yet to be told. I’d like once again to call attention to the massive study of psychotherapy efficacy published by Wampold, essentially a meta-analysis of 3,000 published controlled studies of psychotherapy which demonstrated a VERY LARGE effect size (0.81,) for psychotherapy much larger than has been reported, as far as I
    know for the class of drugs under discussion, even excluding all the negative studies. Wampold’s book contains a very detailed discussion of the strengths, pitfalls and mathematics of a meta-analytic approach, and a very sophisticated discussion of what a non-specific “placebo response” actually means. I can’t recall, however, how Wampold dealt with the problem of possible unpublished negative studies of psychotherapy.

    Paul Mosher

  7. Tamar Schwartz Says:

    Comment from Herbert Gross:

    If we assume that the mechanism of action of the SSRIS is to enhance
    “considered” frontal lobe influence by routing more neuronal activity
    from posterior to anterior and that psychoanalytic therapy is designed
    to have has a similar influence couldn’t we also assume that the same is
    true of the influence of a placebo? We might also assume that
    “considered” frontal lobe influence in the more entrenched depressions
    is overwhelmed by fixed patterns and the effect of placebo induced hope
    is not strong enough. Many of us also know how tough it is to treat
    severe entrenched depression.

    I do believe that neuroscience is clinically relevant and that we won’t
    successfully market our “product” in the 21st century until we include
    neuroscientific thinking in our daily clinical discourse. Restricting
    our discourse to 20th Century psychoanalysis continues the isolation
    noted in the “marketing report.”

    Herbert S. Gross, M.D.
    Clinical Professor of Psychiatry
    University of Maryland School of Medicine
    Washington Center for Psychoanalysis

  8. Tamar Schwartz Says:

    Comment from Henry Kaminer

    The very informative exchange between Art Myerson and Paul Mosher leads me to expand Paul’s point- he reminds us that case reports, even a large number of them, are really only anecdotal evidence and not scientific proof. I want to remind our colleagues of the humbling reality that clinical reports of analytic patients are also only anecdotal evidence- often startling, often provocative, often giving us leads for hypotheses, but not scientific proof of the effect of psychoanalysis to any generalized degree. The difficulty of creating scientific comparisons and evaluations are well known to us all, but that should not be a reason to despair of attaining such information.
    henry kaminer

  9. Tamar Schwartz Says:

    Comment from Henry Kaminer:

    Red or Black placebo pills work better than white or beige ones. A pill dispensed by a doctor with a firm voice or with an empathic and soothing manner works better than one handed over by a clerk thru a barred window. (my speculations)- and there is an interesting story about Sigmund and Anna- about the use of suggestion as a placebo- the story goes that Anna was with her father when he spoke to a new patient, and and he answered the patient’s question by saying, Yes indeed, this treatment will help you. Later Anna reproached him for being dishonest- she reminded him there was no guarantee that this treatment would help the patient. Sigmund replied that this was part of clinical medicine and the patient needed this reassurance for the treatment to work. Despite all the doubt about their efficacy, placebos are here to stay.
    henry kaminer

  10. Tamar Schwartz Says:

    Comment from Sidney Blatt:

    Mike Kowitt correctly points out that randomized clinical trials (RCTs) have limitations, though I disagree that they have the “same subjectivities and biases we generally attribute to the clinical method.” In fact, David Zuroff and I discuss in detail some of the limitations of RCTs in the discussion section of our 2005 paper in Clinical Psychology Review. (I would be glad to transmit a PDF file of this paper). We point out, for example, that patients with different types of personality organization may react differentially to the
    procedures of being in and RCT–not deciding which treatment one will receive and when treatment will end. We speculate that introjective patients (e.g., Blatt, 2006, 2008) who are concerned about maintaining control and autonomy, may have a negative therapeutic response to the procedures of a RCT.
    But let me expand a bit on the findings from the TDCRP because they
    may resolve to some degree the differences in impression about the
    contributions of anti-depressive medication. The original TDCRP
    investigators (Elkin et al ) reported that medication resulted in a significantly more rapid reduction of symptoms than psychotherapy. (CBT and IPT)–at mid treatment (after 8 weeks of treatment) the group receiving Imipramine had significantly less symptoms than the double blind placebo group or either of the two psychotherapy groups. At termination (after 16 weeks of treatment), the Imipramine group had significantly fewer symptoms than the placebo group, but there were no significant differences in level of symptoms between the Impramine group and the two psychotherapy groups after 16 weeks of treatment. The data set from the TDCRP became available to the scientific community in 1994
    and our research team began exploring this data set. Among a number of our interesting findings (summarized in Blatt and Zuroff 2005), we found major differences between the group receiving medication and the two psychotherapy groups–the two psychotherapy group had significantly less stress reactivity–that is, they were significantly less vulnerability to stressful life events in the 18 months following the termination of treatment. To summarize: medication results in a more rapid reduction in symptoms but brief psychotherapy results in the development of greater adaptive capacity and significantly greater resilience to stressful life events. These findings from the TDCRP data set may explain some of the differences in opinion about the role of medication in the treatment of depression. Sid.

  11. Tamar Schwartz Says:

    Comment from Harvey Schwartz:

    I have long felt troubled and confused by the discrepancy between the research reports which minimize the efficacy of the SSRIs and my clinical experience with them. My findings are in accord with those who report significant and predictable effects that are reproducible and dose related. I would go further in that I often, not always, find that the medication improves impulse control, increases recognition of and tolerance for affects, and supports tolerance for an analyzably optimistic vs masochistic character orientation. An example is a patient who was stuck in a prolonged sado-masochistic transference stalemate who responded to an SSRI by then being able to recognize and engage in a libidinal whole object transference relatedness. This led to his then being able to analyze in an affectively alive manner the many layers of conflict that this relationship represented
    to him. That is, he became analytically accessible.

    I have seen my share of patients who have come from either failed or limited treatments where the use of medication was to me peculiarly omitted and when its use was instituted resulted in a deepening of the treatment. (I have also of course seen cases where medication was used unnecessarily). In a number of cases the prior analyst was from the oldest cohort of analysts still practicing, ie before the widespread training in psychopharmacology was commonplace. This at times has led me to feel sad about and dismissive of the thinking of these analysts who often deride the use of medications. On the other hand, I have also wondered about their cases who do not seek additional treatment and who in fact may have benefitted from their analyst’s commitment to struggling longer and perhaps more skillfully with their patients’ regressive states to the ultimate betterment of all concerned.

    I’m loath to generalize from my meaningless N though I of course remain influenced by it. I remain confused about where the truth lies. To state the obvious, however inexact and idiosyncratic are the findings from our individual clinical data, it far surpasses in usefulness that which is learned from prejudice.

    Harvey Schwartz

  12. Tamar Schwartz Says:

    Comment from Michael Feinberg:

    Clinical trials have become big business, and the typical clinical trial is run to make money, not to advance knowledge. Consider a trial of an antidepressant. To obtain good (statistically significant) data, the patients must be ill enough to be significantly
    “better” after successful treatment but not so ill as to require expensive hospitalization nor suicidal enough to raise ethical questions about possibly giving them placebo. Such patients are hard to find, and investigators have been known to inflate admission rating
    scale scores to meet entrance criteria. One result is that placebo response rates are disastrously high, and such trials are not infrequently classed as “failed” because they cannot differentiate a standard treatment from placebo.

    Antidepressants do work well, and can be separated from placebo with good effect size (references on request) IF patients are carefully screened on admission and placebo responders are weeded out. Such rigorous trials are now out of fashion because the patient population differs from the “real world” patients who walk into a doctor’s office.

    Severely ill patients, and those with abnormal biology, are likely to respond to antidepressants but not to psychotherapy or to placebo (references on request).

    There’s the rub. Scientific rigor is all but unknown in the real clinical world, and mildly ill patients still deserve treatment.

    In short, some depressive illness requires somatic therapy, some will respond better to psychotherapy, and some will improve with any or no treatment. Treatment response, or lack of it, is not generalizable across clinical populations.

  13. Tamar Schwartz Says:

    Comment from Robert Galatzer-Levy

    The summaries of the several important studies showing that SSRIs are less effective than commonly thought have been extremely interesting and illuminating if slightly marred by echoes of the long standing battle between psychological and pharmacological interventions.

    I think that we as psychoanalysts have something to contribute to the issue of drug efficacy and effects beyond general hypotheses about the psychological action of the drugs. All of the studies discussed so far in this thread have the common elements that they are based on large samples of patients studied in a relatively superficial fashion usually over short time periods. All of these factors are necessary to achieve the kind of statistical significance for the studies that are currently the norm for medication studies.

    Given Norman Doidge’s finding that a large proportion of analytic patients receive medication either from their analysts or another prescriber we are in an extraordinary position to observe the impact of these medications on patients including individual differences in response, the psychological significance of the medication, and longer term effects. For example, many of us observed that both the SSRIs and Lithium taken of extended periods seem to produce feelings of general indifference sometimes amounting to an “amotivational syndrome” but often more subtly taking something of the desirable edge off people’s personalities. Long before it came to public notice we observed the agitated states of varying intensity associated with starting and stopping SSRIs. We also learned clinically a great deal more about the sexual side effects of these medications then tend to be learned in large studies which either avoid the question or raise it in a manner
    that elicits little information.

    Even more important we are well aware of the complex psychological significances of medication which extend far beyond the placebo effect. For example, response to medication is commonly used defensively (“Since my depression responded to medication it has not psychological meaning.” Etc) or in terms of concerns about self efficacy (“Since I responded to medication I have no power to control my life and psychology.”)

    As a community we have a huge amount of experience about how medications effect real people (in the jargon we have access to “ecologically valid” information)and we have this information in a depth that is largely untouched in most of the published research. It would be of great value if such information were collected and analyzed systematically as it would provide a truer and more complete picture of how these drugs actually affect
    people’s lives.

    Robert M. Galatzer-Levy, M.D.

    P.S. After writing this I found Harvey Schwartz’ latest post which is exactly the sort of thing I mean.

  14. Tamar Schwartz Says:

    Comment from Peter Badgio:

    Harvey Schwartz offers some compelling clinical experience on the effects of SSRIs from his work with patients. Others also have spoken of their conviction form clinical experience that SSRIs are effective in a variety of ways. These testimonials, however, miss the point of the research cited by Paul Mosher and others. The point is not that patients do not have such responses to SSRIs. They do. The point is that they also have such responses to placebos. Thus, much of the scientific posturing of the biopsychiatry/pharmaceutical complex is called into questions. Clearly SSRIs do something. It remains to be learned what it is that they do different from placebos. Moreover, the claimed greater scientific basis to medication treatment compared to analytic treatment is largely a fiction.
    Peter Badgio

  15. Tamar Schwartz Says:

    Comment from Ron Abramson:

    In a message dated 3/6/2008 10:21:21 PM Eastern Standard Time, writes:

    < < There's the rub. Scientific rigor is all but unknown in the real clinical world, and mildly ill patients still deserve treatment. >>

    In my clinical practice, which includes a lot of psychoanalytic psychotherapy and one case approaching psychoanalysis (I don’t practice in an area where most patients who see me have ever heard of psychoanalysis , and most can’t come close to affording it anyway.), I have practiced according to the principle that I have an affirmative obligation to meet the needs of a patient to feel better, however I can think to do it. I am not in love with
    psychopharmacology but I believe that I have an affirmative obligation to use it when it
    makes sense to do so. My, admittedly unscientific, clinical experience with SSRI’s is that they are not very good for many people but they have been very good for some. For all I know, this may be placebo effect, but my hunch is that it is not.

    Some patients don’t want to talk that much and some of those respond very well to an antidepressant. Others want to talk. Yesterday, a 14 y.o. girl I saw for the first time told me that she had stopped her previously prescribed antidepressant because she did not want to think that she has to be dependent on a medication. This 14 y.o. wants to talk.

    Psychoanalysis is not the indicated treatment for everybody, and exploring unconscious conflicts is not for everybody, and if somebody just wants a pill and that pill works for them, I see no problem with that. For myself, I am not interested in treating somebody who does not want to talk and I make that clear at the outset. But that has to do with my interests, not necessarily the interests of the patient. (I refer patients who don’t want to talk to psycho pharm guys I know.)

    I see no conflict between somebody taking pharmacological agents and doing exploratory work in psychotherapy. I have a number of patients like that. The pharmacology does not create lifelessness (That would be bad psychopharmacology.) and generally, by reducing overwhelming affect, helps ego strength and the ability to integrate understanding of unconscious conflicts. The idea that psychopharmacology strengthens defenses and therefore blocks meaningful psychoanalytic psychotherapy does not at all square with my clinical experience.

    So, in general, I view a conflict between psychopharmacology and psychotherapy as a “pseudo” conflict. The way I think of it is that psychopharmacology is a “bottom up” (bloodstream to brain) way to relieve painful affects and psychotherapy (and psychoanalysis) is a “top down” (mind to brain) way to do the same thing. They don’t oppose each other.

    Ron Abramson

  16. Tamar Schwartz Says:

    Comment from Michael Kowitt:

    Since we’re talking about the narrowing scope of SSRI’s, I wonder what others have thought of their use in anorexia and bulimia. I’ve worked with a number of these patients over the years and have yet to see any discernible treatment response. Sometimes the patient seems a bit less volatile, but the eating symptoms seem largely unchanged. As I recall, the Amer Psychiatric Assoc’s practice standards include them as a first line of
    treatment. Have others had better experience in this area?

    Michael P. Kowitt, PhD
    Director of Psychology, Pennsylvania Hospital
    University of Pennsylvania Health System

  17. Tamar Schwartz Says:

    Comment from Elio Frattaroli:

    Mike’s point cannot be overemphasized. Psychoanalysts as a group have allowed ourselves to be intimidated and demoralized by those who brandish double blind statistical research findings and use them to discredit anything that isn’t validated by such studies. In other words, we have simply accepted the claims that analytic evidence isn’t real evidence on which real knowledge can be based and that statistical evidence from double-blind studies is in fact real evidence on which real knowledge is based.
    I hereby declare that NOBODY GETS TO HAVE IT BOTH WAYS. If you are going to base your belief in the value of antidepressants on statistical evidence funded by drug companies, then you have to admit now that your belief was unfounded. Either the numbers don’t lie or the numbers lie. If you want to place your trust in numbers then live with your choice.
    If, on the other hand, you base your belief in the value of antidepressants on your experience working with patients analytically while they are taking antidepressants, then you can feel free to discount the drug-company funded research. Your evidence-based knowledge isn’t threatened or compromised by new statistical findings or newly discovered suppressed findings or new interpretations of old findings.

    My own analytic experience (including long-term psychodynamic psychotherapy) is that antidepressants work really well less than half the time I prescribe them and that in most of the cases in which they work really well they eventually (sometimes after a period of years) stop working at all.
    I am quite confident that there is a medication antidepressant effect that is different from the placebo effect. Because across the board I have seen that the biological effect of SSRI antidepressants is to diminish the intensity of all affects. That’s why so many people talk about their numbing effect. They have the same effect on affects generally that they have on sexual desire. That’s also why they are just as effective for anxiety disorders as they are for depression — it follows from basic psychoanalytic principles that if you diminish the intensity of all affect, including unconscious affect, then naturally you will diminish the intensity of inner conflict, hence relieve anxiety and reduce the psychological need for symptoms.
    Note that this opinion is based on extensive analytic evidence acquired thru analytic experience with patients before, during and after they have taken SSRIs. There isn’t a statistical study or PET scan in the world that could ever convince me this evidence or my interpretation of it is invalid. Nor do I need such studies to confirm or validate what I already know. (I wouldn’t object to validation however. I think neurological research would be much more interesting and valuable if it were guided by analytic evidence.)
    For many patients who are overwhelmed by traumatic levels of affect, to the point that it interferes with their self-reflective capability, medications can facilitate analytic work by making it easier for the paients to get in touch with and understand their feelings. For others whose self-reflective capacity is intact despite their disturbing emotions, medication can potentially interfere with analytic work by making it more difficult to get in touch with the painful feelings that are at the heart of the problem. The other problem with using antidepressants is that the decision to intervene with medication can easily be made out of the analyst’s inability to tolerate what the patient is feeling or is on the verge of feeling — which may suggest to the patient that it is better not to feel such feelings. It’s a possibility I always try to keep in mind whenever I suggest medication — am I treating a discomfort that the patient can’t tolerate and work with or a discomfort that I myself can’t tolerate and work with?

    Some recent posters have attacked a straw man psychoanalyst who knows nothing about medication, never uses it, yet believes (without evidence) that medication interferes with analysis. I’m sure that such benighted attitudes exist among us (though not among Paul Mosher) but they are far less common and far less dangerous than the widespread attitude that priveleges the use of medication to reduce symptoms — because it is “scientific” — and buys into the standard medical model philosophy that biological treatments should or must be used for supposedly biological disorders. The misguided idea behind this attitude is that medications have been proven by research to be effective treatment and that it is therefore malpractice or unethical or at least ignorant for an analyst not to use them when the patient suffers from the appropriate DSM diagnosis. THAT is a scary idea.
    In 1993 I attended a session of the discussion group on “use of antidepressants in psychoanalysis” (I think that was the title) at the Winter meeting. Out of 50 analysts in the group I was the only one who insisted that all cases of depression are caused by inner conflict. Everyone else had accepted some version of the idea that some/many/all cases of depression are purely biological diseases — not in any way the product of psychological conflict — and that any conflictual material they entail is so-to-speak tacked on after the fact from the stress of having to cope with such a painful biological illness. This was in 1993. I was shocked. I am quite sure that no one in 1983 or even in 1987 would have disagreed with me.
    And what was the basis for this sea-change in psychoanalytic opinion? Nothing other than double-blind placebo controlled drug-company sponsored outcome research.
    But wait, you will say, even if we accept the validity of the outcome research, it says nothing about a biological or psychological cause of the illness. Correct, it says nothing about causation at all and in fact there is no biological OR statistical evidence that has ever had anything valid to say about the cause (etiology) of depression or of any other DSM diagnosis. But such research has nevertheless been almost universally misinterpreted — including by 49 out of 50 analysts in that 1993 meeting — as showing that, since a biological treatment cures the disease, then the disease must have been biological to begin with. (The logic of which is so patently specious that it is astonishing how many intelligent people have accepted it) And this misinterpretation has then been used to discredit psychoanalysis because we continue to believe (at least a few of us do) that analysis is a useful and effective treatment for depression and for many other “diseases” even though these “diseases” are “known” to be biological and it has been “proven” that their symptoms can be effectively reduced (some of the time, for at least 6 weeks or maybe 3 months if it’s a really long study) with medication.
    With all this in mind, I could really do without the pious disparagement from people who believe that it is naive, unscientific, ignorant or unethical for people like me to be skeptical of or even outraged by practice patterns and patterns of thinking that are based on the tendentiously systematically misinterpreted findings of statistical placebo controlled double-blind drug-company sponsored research.
    Having said that, I believe that there really is some good statistical research out there and thank heavens we have people like sidney blatt who know how to think clearly about such research.
    End of rant.


    Elio Frattaroli, M.D.

  18. Tamar Schwartz Says:

    Comment from Elio Frattaroli:

    Michael Feinberg writes, “I spent a research career studying this, and believe that there are biological data supporting the “two type” hypothesis of depression.” Mich ael, I take it you mean the hypothesis that some depressions are based on inner conflict while other depressions are purely biological? If so, I would say that there can be no biological data that support any conclusion at all about this hypothesis.

    First, biological data can never reveal causes of psychological distress, only correlations, and there will always be biological orrelations. Second, the only way to know if a depression is based on inner conflict is through analytic data, which has always been there in abundance in every case of depression I have ever treated. The problem is that if we go into a situation with the ASSUMPTION that some cases of depression are purely biological, it is almost a sure thing that we will not notice the analytic data that would disconfirm the assumption because we will have already medicated the patient and won’t be looking for evidence of inner conflict.


    Elio Frattaroli, M.D.

  19. Tamar Schwartz Says:

    Comment from Sidney Blatt:

    Elio: An extensive research literature (e.g., Blatt, 2004) indicates the validity of differentiating between a depression focused on interpersonal issues of loss, abandonment and loneliness from a depression focused on issues of self-regard (self-worth,
    self-criticisim, failure and guilt). Relatively little research, however, has addressed possible neuro-biological correlates of these two sources of depressive affect and, if they exist, whether they are causal or consequential. Also little research has explored the relative
    response of these two types of depression to different types of medication, though some research has addressed their differential response to different types of psychotherapeutic intervention. Sid

  20. Tamar Schwartz Says:

    Comment from Michael Good:

    Regarding the ongoing exchanges about placebos and antidepressants, I have not
    done placebo-controlled, double-blind pharmacologic studies with children (or adults). However, certain child patients, I believe, shed additional light on the question of whether SSRIs have more than a placebo effect.

    In my experience seeing children over time, there are some who have benefited greatly from the careful use of SSRIs, usually in quite low doses. Most commonly, these have been girls of about 8 to 10 years of age who have tremendous anxiety or irritability, often with sleep disturbance. Their adjustment in school and at home is significantly compromised, and they have psychic pain. Their parents suffer along with them.

    While ongoing external factors may be contributory, the early history suggests more
    complex developmental or temperamental factors. In taking the history from
    parents, often I hear an interest in using some kind of medication, or, after an
    extended evaluation or trial of psychotherapy, I raise the possibility of medication with them. We discuss potential risks and benefits in detail. With enough time to build a solid-enough working relationship, a medication trial is a often a natural consideration, along with psychotherapy.

    Although the relatively low effective dose could suggest a placebo effect (e.g., 12.5 or 25 mg of sertraline), it continues to be effective until the child has grown more, whereupon there appears to be less pharmacologic benefit. The child may then require a higher dose, which again is effective.

    Another reason I suspect it is not a placebo effect is the lag time to efficacy. The
    child usually does not know that it initially takes a while to work (though parents
    could tell them this), yet after 10 to 14 days children often show remarkable
    benefit. Affect tolerance, family relationships, and school functioning improve.
    Development is thereby furthered, and the child generally becomes better able to
    make use of psychotherapy.

    Could there have been a placebo effect? Yes, but if so, it may enhance adherence and ultimately benefit the pharmacologic effect. I don’t think it’s an either-or proposition.

    I would be interested in hearing of others’ experiences on this subject.

    Michael Good

  21. Tamar Schwartz Says:

    Comment from Jerome Blackwell:

    Some contributions to the discussion:

    1. Affects are like fever — they indicate something is conflicted, and, if exaggerated, that there is something else wrong. One can give medication to try to relieve the painful sensation of affects (cf. Brenner, 1982) — it’s sort of like giving ASA for fever.

    2. Affects (like anxiety and depressive affect) have varying intensities. When affect-tolerance (generated in the first 3 years of life and again in the first 6 years of adolescence or so by optimal soothing and optimal frustration by parental figures) is poor in an
    adult, breakdown in ego functions may ensue — e.g., in severe disruption in integration, abstraction, self-preservation, reality testing, sleep-wake cycle, memory, and speech — which can be melted down by the intense affects. When this happens, the patient can not
    make use of ego functions to understand conflict, and supportive measures are indicated. These include hospitalization, offering empathic understanding of the overwhelming state, and medicines. If the overwhelming affect is depressive, an antidepressant may be
    helpful. If anxiety is breaking down functions, an anxiolytic may be ameliorative. When medications are withheld in such cases, trouble may follow (cf. the infamous Chestnut Lodge case).

    3. Painful affects, such as depression and anxiety, may not be accompanied by any breakdown in autonomous ego functioning (either because affect tolerance is good, or because the intensity of the affect is not too great). In such cases, where integration,
    abstraction, and reality testing are intact, that analyst may use interpretive interventions to aid in understanding the nature of the depressive feelings. A la Blatt, e.g., introjections can cause self-hatred as a defense against anger, or a la Volkan (“Linking Objects
    and Linking Phenomena”) there may be “established pathological mourning” comprising defenses against grieving (anaclitic type) by keeping too many symbolic objects of the lost object and through identification with the lost object (Freud, 1917). In such cases,
    interpretive interventions seem indicated, and medication contraindicated.

    4. Intactness of object relatedness bears on the type of treatment instituted, as well, and often weakness go hand and hand with poor affect tolerance. If self and object constancy have been achieved, clinically the patient shows no deficits in “Warm – ETHICS”: warmth,
    empathy, trust, holding environment, identity, closeness, and stability in relationships (cf. “101 Defenses”). In such cases, the analyst can develop a trusting, warm relationship with the patient fairly easily, and interpretive work can be helpful. If not, where there are deficits in trust or empathy, and the patient suffers with either anaclitic or introjective depression, the analytic relationship will not provide enough constancy for support, and many supportive measures may again be indicated, including something to relieve the
    UNPLEASANT SENSATION of the depressive affect (C. Brenner, 1982, 2006). Even if medicine makes someone feel a bit better, it’s just the sensation – no effect on the thought content.

    5. Many depressions are what I call “secondary depressions.” They occur because the patients have severe character problems that have caused havoc in their lives: e.g., obnoxiousness, passivity, infidelity, procrastination. The depression occurs because they feel sad, frustrated, and helpless to change anything, and their lives are a mess. It is a mistake to medicate such people for depression usually, I think. As long as they have some object relatedness and reasonable ego functions (and affect tolerance), it seems to me it’s usually better to try to begin to understand the nature of the character pathology, so that their character problems get under control and they can realistically get a bit more optimistic about their situations.

    6. From a unique experience of my own regarding the conflicts about prescribing, I published “Dynamic Supervision Concerning a Patient’s Request for Medication” in the Psychoanalytic Quarterly in 2003.

    I have enjoyed reading the comments on the openline about this important subject, look forward to further discussion, and thank Paul Mosher for getting this started.


  22. Tamar Schwartz Says:

    Comment from Paul Mosher

    In regard to some of the reports on this list of individuals’ clinical observations, anecdotes, and testimonials about the effects of SSRIs, Peter Badgio wrote:

    “These testimonials, however, miss the point of the research cited by Paul Mosher and others. The point is not that patients do not have such responses to SSRIs. They do. The point is that they also have such responses to placebos.”


    Thanks. It seems to me that some of those offering recent comments on this haven’t taken the time to read and understand the articles I cited and thus don’t grasp what the current controversy is about. They therefore have responded as if they imagine I were saying that SSRIs do NOTHING, or that depression can not, at least in some patients, have a major “biological” basis. Of course, I said neither.

    1. SSRIs do appear to have an effect which exceeds that provided by placebos. The incremental efficacy, however, seems to be very small.

    The Kirsch et. al. study showed that SSRIs DO appear to have an effect in depressed patients. The issue is that, compared with the effect of a placebo, the SSRIs’ advantage over placebos seems to be VERY SMALL. In fact, the increment is so small that for SSRIs it falls well below the widely accepted threshold for “clinical effectiveness.” You
    cannot understand this point unless you can get your mind around the concept that even if a procedure or drug produces a great positive change in a clinical condition, but a placebo produces a similar great improvement, the active procedure or drug is (by definition) NOT
    “clinically effective.”

    In comparing a drug with a placebo, or one drug with another, the difference in response can be minuscule and yet can reach STATISTICAL SIGNIFICANCE, even a high level of significance, if the sample is large enough. Such a result does NOT DEMONSTRATE that the drug being tested is “clinically effective.”

    2. In many cases Depression may have a “biologic basis” which takes the form of a PREDISPOSITION to develop depression as a response to stresses such as major losses. About 20% of the population has this vulnerability. This does NOT mean that psychological factors don’t matter, and in fact the psychodynamic factors with which we work may in many patients be necessary but not sufficient conditions to cause a depressive illness. Similarly, psychotherapy may be sufficient to relieve the depression
    without any “biological” treatment.

    In some cases it is plausible to assume that there needn’t even be a noticeable psychological stressor (in particularly vulnerable individuals.) It has long been known, for example, that in SOME PEOPLE administration of the antihypertensive drug, reserpine, can trigger a serious depression and even lead to suicide. Reserpine affects (depletes) the presynaptic storage of transmitter amines (e.g., serotonin.)


    3. My impression is that many of our colleagues tend to underestimate or dismiss the enormity of the placebo effect in the treatment of depression. That effect seems to be significantly larger than is seen in other mental disorders. See:

    With the placebo response in depression being as large as it is, all of us, even the “true believers,” should try to maintain a certain degree of skepticism before attributing improvement of depressive symptoms in one or a few patients to a particular drug or to what we like to imagine are our unique skills as therapists. In addition, half of patients suffering from a major depressive episode recover (without any special treatment) in three

    4. For those who have doubts that the issue of publication bias, which leads to overestimation of drug efficacy, is one that is being taken seriously in the medical community I would call your attention to a recent account of the Turner et. al. article in AMA Medical News at:

    Paul Mosher

  23. Tamar Schwartz Says:

    Comment from Elaine Zickler:

    It seems to me that the placebo effect in medicine would translate in psychoanalysis to a transference effect, one that would have to be assessed individually with each patient-regardless of diagnosis, age or sex. We can only speculate about some of the obvious symbolic gratifications of medication, of the prescribing doctor, etc. and whether medication comes to symbolize caring or not caring, wanted or unwanted incorporation, or intrusion. Paranoid patients can become markedly worse, for example, if the medication serves to bolster up delusions of being controlled or poisoned. If patients who are very depressed and possibly suicidal are given large quantities of a medication
    instead of small, they may feel abandoned, or pushed toward the suicidal gesture, perhaps. A prescribing doctor may in some way mediate a family and/or intrapsychic crisis for a young child that is then represented by the pill, which may involve the parents in the child’s care in ways that are gratifying or perhaps even more developmentally necessary. My point is that it all has potential meanings that the research studies, no matter how well-designed, do not ferret out and, in fact, may strive to filter out as so much confounding static. Psychoanalytic treatment is attuned to finding and interpreting this kind of data, unique to each case.
    Elaine Zickler

  24. Tamar Schwartz Says:

    Comment from Zvi Lothane:

    Drugs are our friends, “drugs” in Russian (pronounced droogs), as once expressed by Anthony Burgess in Clockwork Orange. That work and the later film were prophetic about the rise of violence in post WW II societies.

    We need drugs to blunt the edge of overpowering feelings, as stated by Jerry Blackman.

    That does not mean that drugs are a causal cure for depression. Pursuing the simile of fever, drugs are like aspirin that reduces the fever but does not cure the bacterial infection that caused the fever in the first place, for which penicillin is needed.

    The causal treatment of depression is psychotherapy, as argued persuasively, from a scientific point of view, by Blatt and Luyten. However, in situtations where the feelings become overhelming, psychotherapy may not always (never say never) be able to provide either the necessary relief or the necesseary control.

    Control is an essential issue in mitigating the onslaught of feelings, and one genus in particular, not mentioned by Jerry: anger, rage, fury. In controlling these violent emotions drugs act like anesthetics, a legacy harking back to the original tranquilizer revolution that started around 1950 with Largactil or Thorazine and irrevocably changed the scene in asylums, psychiatric units in general hospital, and the population at large.

    Zvi Lothane

  25. Tamar Schwartz Says:

    Comment from Joseph Schachter:

    Tolstoy (1869) made essentially the same point:

    The human mind cannot grasp the causes of phenomena in the aggregate. But the need to find these causes is inherent in man’s soul. And the human intellect, without investigating the multiplicity and complexity of the conditions of phenomena, any one of which taken
    separately may seem to be the cause, snatches at the first, the most intelligible approximation to a cause and says : ‘This is the cause!'” (War and Peace, Trans. A. Dunnigan, New York, N.Y., p. 1178).

    Joe Schachter

  26. Tamar Schwartz Says:

    Comment from Paul Mosher:

    Maybe it would be more reasonable to think of an infinite number of types of depression (i.e. a spectrum) rather than trying to divide the category artificially into two (or even three or more) types.

    On the issue of the causation of emotional disorders, and please excuse my repeating myself, here’s my paraphrase of the Strachey rendition of the way Freud put it over 90 years ago. I first posted this passage on the members list in 1998:

    >>I would like to take this opportunity to defend myself against the mistaken charge that, because I have stressed the role of early childhood experience, I have denied the importance of genetic/constitutional factors [in personality formation]. This sort of charge arises from the simplistic way people think about causation: in contrast to what actually goes on in reality, PEOPLE PREFER TO GRAB ONTO A SINGLE CAUSE WHEN TRYING TO EXPLAIN SOMETHING. Psychoanalysis has talked a lot about the role of childhood experience as a cause of personality development, and little about genetic makeup and other physical factors; but that is only because we had something new to say about the role of childhood experience, while, to start out with, we knew no more about the physical factors involved in mental development than anyone else. But that doesn’t mean that we accept any idea that these factors of early experience and genetics should be considered separately; on the contrary, we assume that early experience and genetic endowment act together in bringing about development.

    [Genetic Endowment and Experience] determine a man’s fate — RARELY OR NEVER ONE OF THESE ALONE. The extent to which each factor is causative appears to vary from one person to another. We assume that people are arrayed along a spectrum according to the extent of the role played by genetics and experience, and no doubt there are some individuals near each end of that spectrum. As our knowledge increases we may change our estimate of the relative importance of genetics and experience in specific cases…(1) [emphases added]<< Paul Mosher ------------------------------------------------- (1)The Dynamics of Transference footnote 2, paraphrased Standard Edition, volume XII, p. 99

  27. Tamar Schwartz Says:

    Comment from Tom Bartlett:

    I once did intermittent diagnostic interviews for patients randomly assigned in a study of an SSRI versus a placebo, along with manualized cognitive therapy, in the treatment of depressed alcoholics. I don’t remember all the details of the study, but remember two interesting kinds of patients.

    A small number of patients had to be withdrawn from the study because they couldn’t tolerate the “side effects”. Once the blind was broken, it turned out some of these had been on the placebo. I don’t know the numbers, but I think a study of this phenomenon would be very interesting, placebo induced side effects.

    At the other end of the spectrum, there were patients who were convinced that they WERE on the placebo, and were not. “This isn’t doing anything”, “I don’t feel anything from it” etc. In some instances *I* was fairly certain that the patients WERE on the SSRI, from the very evident changes in the patients. They would be people who came in saying, “I joined a gym this week”, and looking upbeat in a way many SSRI patients tend to feel in what is nicknamed the “honeymoon period” in weeks 3 to 6 or so. When we broke the blind, we were usually right in our guess that they’d been on the medicine.

    It is a phenomenon many therapists have seen when patients they’ve treated for a long while start on an antidepressant. Suddenly (and I can’t rule out placebo aspects in the patients’ AND MY expectations playing in here), but relatively suddenly in the overall course of things you start to see changes you just know are the medication, because you know how therapy has gone for the preceding months and months, and now see changes you feel you just cant attribute to any major change in the therapeutic work you’re doing. Like the patient who was constantly feeling victimized and overlooked in a sales job, who suddenly begins asserting herself with her boss, making sales, starting sessions with what a good week it had been, etc. And then she says, ” I think I might as well stop taking this medicine, I don’t feel it doing anything.” It is really curious, and very common, to see this, ie, changes very likely being brought about by a medication in people who (consciously) believe the medicine has no effect on them. Who knows, maybe it is the placebo effect at play despite their conscious belief.

    As someone who came into psychotherapy with a strong bias against medications, changes like those seen with antidepressants were always humbling for me, and if nothing else, helped cure me of a certain hubris about psychotherapy as the panacea I’d previously taken it to be.

    I am not trying to use this anecdotal evidence as a counter to well conducted studies, for what they can show. I just find these would be ‘false negatives’ and ‘false positives’ intriguing.

    Tom Bartlett

  28. Tamar Schwartz Says:

    Comment from Arthur Meyerson:

    Paul Mosher cites one of my favorite passages of Freud, one that so many analysts seem to ignore or forget. To my mind the best way of thinking about depression (along with other major illnesses) is in terms of a stress/vulnerability model. The stress may be considered some current or ongoing environmental situation, most often interpersonal, which increases the strength of the patients particular central conflict(s). It may also be a biological stress such as those who respond to particular medications like cortisone with depression. The vulnerability may be an inherited genetic matter or a one or more of a set of developmental factors such as parental loss or abuse. These childhood stressors may produce permanent neurochemical changes such as those demonstrated by Nemeroff, et al which render the person more vulnerable to depressive and anxiety disorders.

    Vulnerability is on a curve from being the entire “cause” to being only a contributing cause. Some patients just seem to get depressed with no change in their life and no demonstrable psychological stress even with intense analytic exploration. These rapidly recurrent depressions are difficult to treat in any way but clearly psychological treatments are difficult to bring to bear.

    Psychological stress appears mainly active as a trigger for the underlying vulnerability. After all, there are few stressors that cause depression in everyone subjected to them. Even concentration camp victems, highly traumatized, may not demonstrate clinical depression as a major feature of their reaction.

    In an animal model, learned helplessness can be a universal trigger for intense emotional reaction in rats but rats can be bred to have either depression in some strains or anxiety in other strains and this is an example of the significance of vulnerability.

    Arthur Meyerson

  29. Tamar Schwartz Says:

    Comment from Michael Feinberg:

    First, I will declare that I know nothing about the causes of depressive illness. It reminds me of the wave and particle theories of light, reconciled (superseded) by new physical theories. There are data which are better fit by biological hypotheses and others which
    seem more psychological, and that ignores the interaction of the two.

    Kendell expounded a continuum view of diagnosis, which has some experimental support, as do theories of discontinuous groups. The best mathematical treatment I’ve seen is “grade of membership analysis,” which allowed separation of categories with fuzzy and
    overlapping boundaries.

    Even assuming a continuum, there are biological (sleep and neuro-endocrine) phenomena and clinical features (psychosis) which identify patients who are unlikely to respond to psychotherapy but are likely to respond to somatic therapy.

    This whole discussion started with the issue of whether antidepressants work, and for whom if at all. My point is that they work well for carefully selected patients, but are now more widely used (the widening scope of antidepressant use?). I believe that diagnostic rigor and biological testing where available help to identify patients who are more or less likely to respond to different treatments. Psychological and biological treatments will each be more effective if prescribed to patients selected for probable response. Given sloppy selection, any treatment will seem ineffective. As clinicians, our dilemma is that we are obliged to treat patients whose clinical presentation does not fit neatly into any category, and our best treatments may seem worthless.

  30. Tamar Schwartz Says:

    Comment from Greg Roeben:

    Mar 17 Newsweek column on Placebo effect

    Greg Roeben

  31. Tamar Schwartz Says:

    Comment from Michael Good:

    In my experience seeing children over time, there are some who have benefited greatly from the careful use of SSRIs, usually in quite low doses. Most commonly, these have been girls of about 8 to 10 years of age who have tremendous anxiety or irritability, often with sleep disturbance. Their adjustment in school and at home is significantly compromised, and they have psychic pain. Their parents suffer along with them.

    While ongoing external factors may be contributory, the early history suggests more complex developmental or temperamental factors. In taking the history from parents, often I hear an interest in using some kind of medication, or, after an extended evaluation or trial of psychotherapy, I raise the possibility of medication with them. We discuss potential risks and benefits in detail. With enough time to build a solid-enough working relationship, a medication trial is a often a natural consideration, along with psychotherapy.

    Although the relatively low effective dose could suggest a placebo effect (e.g., 12.5 or 25 mg of sertraline), it continues to be effective until the child has grown more, whereupon there appears to be less pharmacologic benefit. The child may then require a higher dose, which again is effective.

    Another reason I suspect it is not a placebo effect is the lag time to efficacy. The child usually does not know that it initially takes a while to work (though parents could tell them this), yet after 10 to 14 days children often show remarkable benefit. Affect tolerance, family relationships, and school functioning improve. Development is thereby furthered, and the child generally becomes better able to make use of psychotherapy.

    Could there have been a placebo effect? Yes, but if so, it may enhance adherence and ultimately benefit the pharmacologic effect. I don’t think it’s an either-or proposition.

    I would be interested in hearing of others’ experiences on this subject.

    Michael Good

  32. Tamar Schwartz Says:
  33. Tamar Schwartz Says:

    Comment from Elio Frattaroli:

    I especially liked Bob Galatzer-Levy’s post about SSRIs and psychoanalysis and was reminded by it of something I had forgotten to mention: In my experience the most prominent symptom of withdrawal from SSRIs (and from SNRIs as well) is affect rebound, sometimes to the point of affect storm — the patient is suddenly flooded by the very emotions that had caused the problem in the first place, that had been suppressed by the biological action of the drug.
    The good news is that patients are reassured to get this explanation and to be told that it doesn’t mean they are sicker than they thought and have to go back on the medication, only that they need to withdraw from it more gradually.


    Elio Frattaroli, M.D.

  34. Tamar Schwartz Says:

    Comment from Elio Frattaroli:

    Just kidding! I have in fact been misunderstood but I’m not offended or depressed about it. I simply wasn’t clear enough.
    I never said or wrote or meant that biological factors weren’t relevant in the etiology of depression. But there can be no scientific evidence that biological factors “cause” a psychological condition. There can however be scientific evidence that psychological factors “cause” a psychological condition. Perhaps the source of the confusion is that “cause” has a different meaning when we are talking biology than it does when we are talking psychology. In psychology a cause is a motive or set of conflicting motives. “I punched him because I hated him.” “I couldn’t talk to him because I hated him and was afraid that it would show if I opened my mouth.” These are valid statements of psychological cause-and-effect and key word is “be-cause.” Within the psychological/personal domain, the cause-and-effect relationship is transparent and available to be understood by any human being who has ever reflected at all on motivation. In contrast, a statement like “I couldn’t talk to him because I had a biological depression caused by a serotonin imbalance” is ridiculous. It involves a category mistake, mixing the language of motivation (the introspective personal level of observation) with the language of neurochemistry (the extrospective cellular level of observation) as if they were compatible and mutually translatable. Which they aren’t.
    Here’s how I currently think about the symptoms of the so-called biological mental illnesses: Looking at them from the biological level of (extrospective) observation, the biological factors probably determine the type of symptom a person is likely to develop under the stress of inner conflict, and they probably determine how intense the stress has to be in order to produce an inner conflict disturbing enough to produce symptoms (whether by increasing/decreasing the intensity of emotional reactivity — the drives, if you will — or by decreasing/increasing the strength of the ego/personality defenses) But at the introspective psychological level it is inner conflict that triggers (causes) the unmanageable anxiety that then triggers (causes) the development of symptoms, which serve to reduce anxiety and partially solve the psychological/personal problem entailed in the inner conflict.


    Elio Frattaroli, M.D.

Recent Posts