Is depression really an illness? Or is it a normal reaction to a crazy world? If antidepressants can fix the problem in six weeks, then why spend six years on the couch soliloquizing about your bad potty training? Or, if they can’t, then what the hell is wrong with psychiatry?
On and on it goes. If you want to get up to speed fast on a subject vital to your life, then I highly recommend the article.
The piece zoomed in on the same set of studies that Newsweek made the basis for its outrageous cover story, Why Antidepressants Don’t Work (see my highly-critical post). Consider the New Yorker piece chess to the Newsweek’s checkers, with a much more nuanced look at the issue.
As you will recall, Newsweek cited two extremely convincing meta-analyses by Irving Kirsch in support of the proposition that antidepressants are basically placebos with side effects. Over eight or so years, no one has been able to shoot holes in these studies. Trust me, the second Kirsch meta-analysis is bullet-proof, but the results are open to interpretation.
Namely: In the real world, patients are likely to try a second antidepressant if the first one fails. Various small studies at the time indicated that the odds of success go way up when patients adopt this approach. A later large-scale series of trials underwritten by the NIMH, STAR*D, confirmed this.
That was how I reported the issue eight years ago and in various follow-ups and this was the approach taken by the New Yorker. (Both the New Yorker and I also shot to pieces a bogus meta-analysis recently published in JAMA that Newsweek took at face value).
Not so fast, says Robert Whitaker, author of "Mad in America." In his blog on Psychology Today, Whitaker accurately points out that STAR*D used statistical hocus-pocus to come up with an otherwise unsupportable claim that two-thirds of the patients in the study recovered on antidepressants.
I too, found this conclusion difficult to believe, and didn’t feature it in my STAR*D reporting. What I did feature were two key results: 1) It is worth trying a second antidepressant after the first one fails. 2) Trying a third after the second one fails is problematic.
The second result is the real STAR*D story, one that features in many of the pieces I write here (and one the New Yorker hinted at) but that Newsweek and Whitaker and just about everyone else missed. Here’s the deal:
The DSM depression diagnosis is an emperor with no clothes. It doesn’t tell us anything we don’t already know. One of it’s nine symptoms is “depressed mood.” Huh? So, if you have “depressed mood” and aren’t sleeping right or eating right and have low energy and seem to be moving in slow motion, what state of mind are you in?
DSM depression is a plain vanilla diagnosis that disguises the fact that depression comes in many flavors with many different ingredients. The plain vanilla approach encourages clinicians and researchers to treat all depressions as if they were the same.
This is the major reason clinical drug trials - and for that matter talking therapy trials - tell us so little. In any given trial, we can predict in advance that 50 percent of patients are going to get 50 percent better. The catch is which 50 percent? No wonder the results for the treatment group and the placebo group are about the same.
This is plain vanilla diagnostics at work. In all likelihood, there is a subgroup that is getting 80 percent better 80 percent of the time, as well as a large group of those who should never be taking antidepressants. But who are these people?
Of all things, STAR*D serves up a hint. As Frederick Goodwin, former head of the NIMH, pointed out to me, about two-thirds of the STAR*D study subjects had recurrent depression.
It seems likely that those in the STAR*D study with recurrent depression would not have fared so well on antidepressants, as recurrent depression is a close cousin of bipolar, but we’ll never know. STAR*D made no attempt to separate out this population. To the investigators, depression was depression.
The other lesson to be gleaned from STAR*D is this: After your second antidepressant fails, you need revisit your diagnosis. Maybe you don’t have depression. Maybe you have bipolar or a depression that behaves like bipolar. Maybe you have borderline personality disorder. Maybe depression is part of your baseline temperament.
But plain vanilla DSM depression offers no guidance. Chances are a large population of individuals in the STAR*D study did not even have depression. Same with clinical drug trials. Treatment works only if the diagnosis is correct.
Now we’re getting to the moral of this story:
Yesterday, I premiered “The People’s DSM.” My first installment featured Part I to My Alternative Depression Diagnosis. In the intro to the piece, I joked that, “if you want anything done right, you have to do it yourself.” But I’m not fooling around.
The DSM-IV depression diagnosis is based on an antiquated and totally arbitrary symptom list from 1980. The people working on the DSM-5, if they are serious, need to rip up that list and start over. Maybe then, psychiatry will get serious about what is really going on in our brains and come up with answers. Instead, they plan to leave the list intact.
As I said, sometimes you have to do it yourself. I’m not joking.
Posts
8 comments:
Looks really interesting, John, and I think you are on to something. I'm going to need to do some thinking about this.
Many thanks, Fern. I'm looking forward to hearing back from you.
From what I read of the New Yorker article, it seems that the people who criticize pharma for selling drugs have something to sell themselves. They want people to buy their books and use there "treatment". They are as biased as pharma-funded doctors. Their attack on medications reeks of getting rid of the competition.
Hi, Tony. I would give a qualified "yes" to that. Talking therapists tend to jump on every study critical of meds, with no critical thinking evident. Various antipsychiatry advocates make a living on government largesse and as speakers. Vitamins/supplements is an industry far more corrupt than the drug industry.
By the same token, a number of courageous individuals have risked their careers and chances of advancement by speaking out against outrageous abuses in the drug industry and psychiatry. The industry is notorious for handing out lucrative bribes disguised as speaking engagements and consulting fees and authorship credit for those willing to play the game.
There is a place for partnerships between academia and industry, but the research that has emerged has been nowhere near as productive as its defenders claim, but has been every bit as corrupting as its critics charge.
The lesson is not to take the claims from anyone at face value. Everyone claiming to act in our interest has an agenda and is trying to sell us something. Most are honorable people, but opportunists abound. Also, honorable people are prone to appalling lapses in judgment.
No one but us is acting in our own interest. But you already know that. :)
I have enormous problems with the STAR*D study. Let's start with the absence of a placebo group. In its absence, we really can't tell whether 2/3rds get better with one or two antidepressants. Cycling or not, depression goes away (at least temporarily) on its own. Given the high level of clinical follow-up all the way through, I expect that the placebo rate would have been higher than usual, if it had been measured.
And given that, by the time they got to the fifth trial, (what was that in real time, if you include the first antidepressant they tried before they entered the study -- two years?) a 7-10% success rate certainly raises the question, was the treatment making people worse?
I don't think STAR*D supports your recommendation that people revisit their diagnosis after two trials. I think it's an important recommendation -- but STAR*D doesn't support it. The third trial included a switch to either a thyroid medicine or lithium -- a tacit admission that maybe the initial diagnosis was wrong. But these meds had the same declining efficacy rate.
And again, since there was no placebo, nor follow-up of those who quit before the chemistry experiment was over to examine whether they might have gotten better on their own, there was no way to measure just how ineffective lithium was. Given its real world efficacy among those who are already taking it for bipolar, these lousy results do not suggest that they had some unidentified bipolars in their pool.
Hey, Willa. I agree. The non-inclusion of a placebo arm makes any STARD*D finding suspect. But I nevertheless submit that the findings do support the wisdom of taking a second antidepressant after the first one failed.
I also found the finding that trying a third offered scant chance of recovery to be a real eye-opener.
The results from STAR*D were not quite what the investigators expected or hoped for. (For instance, they were expecting better results from the second antidepressant trial and seemed shocked by the failure of the third.)
Unfortunately, the investigators tried to spin the results instead of looking to the real lessons: Namely, if two successive antidepressants fail it is time to revisit the diagnosis. It may turn out you still have depression, but it doesn't hurt to double-check.
I would submit that it is best practice for clinicians to do a complete diagnostic revaluation after two failed antidepressants, especially with the strong likelihood a third is not going to work or may only make the patient worse - and with the extremely likely possibility that one's untreatable unipolar depression may be treatable bipolar.
You're right: The study did not "prove" this, but this is the lesson we need to draw from this very imperfect study.
Re lithium: I'm going from memory here, but my impression is that both the lithium and thyroid were used as augmenters to antidepressant medication, to "boost" the antidepressant. It's not surprising lithium failed in this context.
What would have been interesting is if lithium or Depakote or Lamictal had been used in the third trial as a mood stabilizer. But that would have required a change in treatment objectives (from eliminating the depression to stabilizing the cycle).
Anyway, I love it when you disagree with me. As all of us appreciate, depression is way more complex than our clinicians and Pharma make it out to be. There are no easy answers, and there is no one explanation that fits all depressions. If there were, we would all get better real fast and I would have no reason to continue with this blog. Heaven help if I had to find a real job. :)
There are a couple of things to consider.
First, understanding statistical mumbo jumbo is essential, along with other design like length of study, sample size, cutoff level for degree of significance, lack of a placebo and even "placebos" that aren't really inert. I look at studies every day citing percent changes between percentage rates, which will exaggerate the effects significantly. And there is so much innumeracy even among doctors that the true results are rarely reported.
Secondly,it is absolutely true that western medicine does not make differential diagnoses between types (or etiologies) of bipolar or any other mental disease and the new DSM doesn't help at all. If they followed the practice of Chinese medicine where there are several patterns of bipolar, they might prescribe more accurately.
Third, while it may help to look for a second or combination antidepressant, you have to also look longitudinally. I have any seen SSRIs make a major difference initially, then stop being as effective as the body accommodates. "Prozac poop out" is what one psychiatrist of my acquaintance calls it. Unless you treat the underlying imbalances, nutritional deficits and lifestyle issues I doubt that the pharmaceuticals will offer sufficient long term effects.
Conversely, medicines could be seen as ways for the brain to learn new patterns, then be stopped. I have a friend with OCD who learned new mental states from Strattera, but doesn't use it most of the time. If she drops back into old patterns she may repeat a short course, but she knows she doesn't have a chronic Strattera imbalance.
Forth, our bodies don't recognize lots of the pharmaceutical medicines at a cellular level. Due to testing standards it is difficult to patent a drug with many different actions as one normally finds with minerals and botanical medicines. An herb like St. John's wort has 25 different antidepressant actions while Prozac has one, and some of the other SSRIs have two.
I have a client who combines St. John's wort with Zoloft- since it has an additive effect he keeps the individual doses low to avoid Serotonin syndrome. The herb rounds out the missing actions of the drug which helped him rapidly when he was feeling suicidal. In the meantime he is building up his magnesium, trace minerals, Vitamin D, fish oil and is sleeping enough and volunteering in an area he finds meaningful. I am betting that chronic depression will abate to the point his medication can be reduced or eliminated with the changes he has made in lifestyle. While the propensity may not change, the results may be manageable with occasional forays into pharmaceuticals.
Hey, Karen. This is very very interesting. I'd love to have you do some guest blogs.
Post a Comment