Editor’s Note: Over the next several months, Mad in America is publishing a serialized version of Les Ruthven’s book, Much of U.S. Healthcare is Broken: How to Fix It. In this blog, he addresses the scientific literature on antidepressant efficacy and FDA approval. Each Monday, a new section of the book is published, and all chapters are archived here.
A brief review of the scientific literature on depression drug treatment outcome studies.
Some of the following studies appeared in my book Antidepressants: Science, Magic or Marketing,1 and in the next section, I will report on some of the more recent depression treatment outcome studies.
In some published FDA antidepressant placebo-controlled clinical trials, the response rate in terms of symptom reduction is 50% for the drug and 32% for placebo.1 If my arithmetic serves me, this response rate difference (18%) means that one would have to give approximately five depressed patients ADMs to be sure that one of the five patients would have more than a placebo response if all five had been on placebo.
I must report that the difference observed between drug and placebo was between the drug and an inert placebo and not the more scientifically appropriate active placebo, a non-antidepressant drug with side effects which helps to improve the blind of trial subjects and prescriber-raters of improvement.
When we see results—such as these—on how minimally effective the ADM is in treating depression, one must question if there is indeed a drug effect in addition to a placebo effect? For example, if a physician had to give five patients with infections antibiotic drugs in order to be sure one of the five had more than a placebo response, antibiotics would simply not be utilized in medical care today! However, antibiotics are used because they are substantially more powerful than placebo in treating infections as opposed to antidepressants in treating depression.
At one time I thought that the FDA efficacy standard for psychiatric drug treatment of mental health problems has a lower standard than for general medicine, but later in my review of health research I found that in many cases this was not true (as I will discuss in Chapter 6).
What is the power of cognitive behavioral therapy (CBT) in treating depression versus drug treatment efficacy? Investigators compared CBT and waitlist controls on improvement of depression.2 The similarly depressed control subjects received no treatment in the interim. Psychotherapy groups had mean effect sizes of .28 to 1.00 with regard to reduction of depressive symptoms. An effect size is a statistical term reflecting (in healthcare) the power of any treatment versus placebo, and the larger the effect size, the more effective the treatment. In one of the several waitlist groups under discussion, the majority of subjects worsened (effect size of .28) even though some improved.
Just as with the FDA depressed placebo control subjects, most subjects in the waitlist improved, but the level of improvement was far less than the psychotherapy treated subjects. The psychotherapy groups had a mean, on average, effect size of 1.60 versus 0.37 for the waitlist or no treatment controls. The CBT superiority over the waitlist improvement (1.6 vs. 0.37) was quite impressive. CBT treatment of depression was over four times better than waitlist improvement!
The larger the positive effect size of any therapy, the greater the degree of improvement. Keep the above in mind when later observing the effect size of ADMs in treating depression. It is interesting that the wait list treatment effect of 0.37 is typical of the drug treatment effect in many FDA depression clinical trials!
The above 19 studies in the meta-analysis showed a substantially greater improvement for CBT in treating depression than no treatment. It is not surprising that some on the waiting list improved with no treatment (just as some on placebo improved in the FDA drug studies); when a depressed person has hope (e.g., even getting on a waiting list for treatment), the severity of their depression lessons in most cases. However, how does the effectiveness of drug therapy in depression treatment compare with the effects of CBT alone in treating depression?
Steinbrueck conducted a meta-analysis of 56 depression outcome studies and found that psychotherapy had a substantially greater impact (mean effect size of 1.22) than drug therapy (mean effect size of 0.61) in unipolar depression.3 The effect size difference means that psychotherapy had about twice the power of reducing depression as did the drug.
Both CBT and interpersonal psychotherapy have proven their effectiveness in research studies in treating depression, but in the clinical or real world both are modified by the trained therapist to meet the depressed patient’s individual needs, their current life difficulties, their personality and other factors. As I have reiterated, if you have seen one case of depression you have seen one case of depression. In treating depression with psychotherapy, the therapist’s clinical understanding of the patient’s personality and life situation will significantly modify the individual treatment approach.
In antidepressant drug treatment of depression, one size fits all. The drug therapy proponent will object to my last statement by saying “We have about five to six types of ADMs from which to choose” and I would counter with: “Yes, but those six types are only marginally more effective than placebo”! The above is not to say that depressed patients as a group do not have many characteristics in common—because they do—but there are individual differences between depressed patients and these differences must be addressed and treated by the competent therapist. However, in the world of research, scientific controls must be rigorously considered and the treatment (e.g., drugs or type of psychotherapy) must be uniform across all patients in the study.
Despite this, CBT shows up very well against drugs in research studies on treating depression and from such research data, CBT should be the first-line treatment for depression rather than second-line treatment when ADMs continually fail. Moreover, in drugs vs. psychotherapy in treating depression, the most important measure of efficacy is not efficacy in the short term but rather, which treatment is best in keeping the depression from returning?
Many of us psychologists believe that depression is the final common pathway that can be reached as the result of a variety of pre-depression causes and unfortunate life events that emotionally impact the individual in a negative way. For example, in Chapter 4 (the section on diagnostic errors) I will address how I diagnosed a male adolescent’s depression as arising from his anxiety about having any close contact with female peers. Once he was desensitized to this fear of relating to female peers in the real word, his depression lifted.
The best cure in most cases for a depression following a job loss is not drugs or CBT but another good paying job, which requires astute assessment of the patient by a well-trained therapist! This is not to say that the depressed out of work person cannot and does not benefit from behavioral treatment, treatment that counters the behavioral consequences of depression, namely the diminished hope, behavioral inactivity, reduced self-esteem, and the like.
The official position of the American Psychiatric Association is that combined therapy (drugs and psychotherapy) is more effective than either alone. However, the APA’s own committee review of 12 studies concluded there was no demonstrable relationship between endogenous depressions (a depression without an “apparent” cause in the patient’s life) and treatment outcome—that is, drug therapy was no more effective than psychotherapy for severe depression as well as for the milder depressions.
For years, psychiatrists would tell me that psychotherapy works for mild depression, but drugs are required to treat severe depression. As a psychiatrist you may say that, but scientific studies do not. The propaganda of combined therapy (drugs and psychotherapy) has even infiltrated the profession of psychology to some extent; at a recent dinner party a fellow psychologist was quite convinced of the superiority of drugs combined with psychotherapy in treating depression. In my experience it seems that psychologists who are in practice with psychiatrists are more prone to believe the drug propaganda than psychologists who do not office with or are employed by psychiatrists.
I apologize for using the word propaganda, but the theory that depression is a brain disease is simply unproven!
As mentioned previously, some researchers have a problem using inert placebos in drug studies to measure the true effectiveness of the active drug because of the drug’s side effects, which compromises the “blind.” The inert placebo has no side effects of course, which reduces the strength of the blind since psychiatrist researchers in many of these studies rate the improvement of drug and placebo subjects. The subjects themselves and the psychiatrist raters of improvement may be aware of the condition (placebo or drug) which may falsely increase the “power” of the drug over placebo.
To strengthen the blind and level the playing field some researchers use active placebos, non-ADM drugs that have dry mouth and other side effects similar to those of ADMs. Use of such active “placebos”, non-antidepressant drugs that have side effects as do ADM drugs, reduces the bias arising over the use of inert placebos in such antidepressant effectiveness studies. Subjects on inert placebos in these clinical trials are likely to have a less powerful placebo effect because of the absence of side effects, which may enhance the power of the drug under study.
Joanna Moncrieff and colleagues conducted a meta-analysis of several active placebo-depression studies and found the active placebos had a stronger blind (as expected) than the inert placebos, and the ADM drug effects were only slightly superior to the effects of the active placebos.4
Also, where do these findings as the Moncrieff study leave the “theory” that depression is caused by and due to a reduction of serotonin in the brain? Within such a context, the belief is that the SSRI, by enhancing serotonin at brain synapses, the patient’s depression is effectively treated. Any such “theory,” including the serotonin theory of depression, also has a problem, however, when it comes up against new facts.
An antidepressant drug licensed in France,5 tianepentine (brand name Stabolon), has the opposite effect on the brain as in the SSRI drugs; rather than enhancing serotonin in the brain tianeptine reduces serotonin at brain synapses! From past experience, I am sure proponents of the serotonin theory will somehow find a way to get around such an annoying fact. It is of interest that the above French-licensed antidepressant could not or did not even try to gain FDA approval for their drug in the U.S.
I do not want to imply that there are no biological or brain factors in depression or indeed in many of the mental disorders, because there are. First, psychology is a bio-psychosocial science and I have previously argued we cannot legitimately separate mind from body as many physicians in their daily practice seem to do. With regard to depression, there are certainly genetic and personality factors involved in people who become depressed or not depressed from similar adverse environmental life circumstance. As a group, people who tend in their personality organization to be overly conscientious and perfectionistic (including morally) compared to many of us are more prone to bouts of depression, and their depression will take more than exercise or a placebo to effectively treat their depression. These individuals with perfectionistic traits, when depressed, are likely to have a more serious depression than less perfectionistic people (as well as a higher risk for suicide) and are likely to require more intensive psychological therapy than others.
Some personalities, unlike perfectionistic people, tend to externalize blame for their life difficulties and tend to avoid depression as a result, and others such as those susceptible to depression err on the side of being self-blaming in response to life difficulties. Depressed patients who are self-blaming and overly conscientious are likely to require more intensive psychotherapy as mentioned than depressed patients whose personality make-up is less-self blaming. Effective psychotherapy with overly conscientious depressed patients will require modification of the patient’s over-conscientiousness and perfectionism, which are both self-defeating attitudes and must be changed to prevent further episodes. These depressed patients will need to undergo an attitude change and CBT would be the treatment of choice. With some non-perfectionistic depressed people, brief psychotherapy or in some cases an exercise program or other self- or other-directed behavior change will be sufficient to remit their depression, but not so for a self-blaming depressed person.
The above conceptualization of depression does not preclude brain factors being involved (particularly the brain effects caused by the patient’s depression itself) and certainly with such a serious disorder that affects almost all behavior, the organization of the brain must show those behavioral effects of the depression as they do in brain scans of all depressed and non-depressed persons as well.
Under the Freedom of Information Act, Irving Kirsch and colleagues secured clinical trial data for the six most widely prescribed SSRIs approved by the FDA between 1987 and 1999.6 There were 47 FDA-submitted drug and placebo-controlled short-term efficacy trials for Prozac (5 trials), Paxil (16), Zoloft (7), Effexor (6), Serzone (8), and Celexa (5), which involved 6,944 drug-treated patients. The data does not include those studies with negative findings that were not submitted (about 50% of all such studies) to the FDA. One must assume that those 50 completed but unsubmitted studies came up with negative findings for drug effectiveness. Later, the FDA added a much-needed requirement that the results of all FDA clinical trials must be submitted to the FDA, not only trials with a positive drug outcome.
In 22 of the 46 clinical trials (48%) submitted to the FDA, the difference in drug and placebo groups on improvement was not significantly different. This statistic—that 48% of all FDA submitted clinical trials found no differential improvement between drug and inert placebo cases, is the most serious indictment in the whole study of ADM effectiveness in treating depression. Several of these drugs that initially failed to gain FDA approval continued to be the subject of further trials and finally gained FDA approval. As I recall, one popular SSRI gained FDA approval on the basis of three submitted clinical trials, one of which was positive for the drug and two of which were negative for the drug!
Irving Kirsch and colleagues in the above-referenced study found that about 80% of improvement of depressive symptoms, measured by the HAMD-Scale, was duplicated by the placebo control groups. The mean difference between drug and placebo was less than two points on the HAM-D-Scale in favor of the ADM versus placebo patients, a statistical “difference” which is within what psychologists refer to as the standard error of measurement. A ruler gives an exact measurement, but in psychology a behavioral measurement (e.g., an IQ of 84 on a test of intelligence) means the IQ of 84 is not exact but the real number (depending on the test) is within perhaps four points, the standard error of measurement on the particular instrument.
In order to generalize the drug findings of the clinical trials to a larger population, FDA reviewers seek a completion rate of 70% or better for these typically six-week trials. Kirsch found, however, that only four of 45 trials reached this FDA objective! With such trial data why did the FDA approve those 41 ineffective drugs? Moderate to severely depressed subjects in these clinical trials are certainly, as a group, highly motivated for treatment. Add to it that when so few trials match or complete the 70% completion rate, this is a very strong indictment against the use of antidepressants to treat depression in the real world. Transferring this very poor trial completion rate to real world use means that of 100 moderate and severely depressed patients going on these six popular antidepressants, at most 70 or fewer than 100 such patients would complete six weeks of such therapy in the real world! Do we need any further evidence than this that we had better find a treatment other than drugs to treat depression?
To see the list of all references cited, click here.
Mad in America hosts blogs by a diverse group of writers. These posts are designed to serve as a public forum for a discussion—broadly speaking—of psychiatry and its treatments. The opinions expressed are the writers’ own.
Mad in America has made some changes to the commenting process. You no longer need to login or create an account on our site to comment. The only information needed is your name, email and comment text. Comments made with an account prior to this change will remain visible on the site.