- Home
- Find a Therapist
- Topic Streams
- Get Help
Mental Health
Addiction
ADHD
Anxiety
Asperger's
Autism
Bipolar Disorder
Depression
Eating Disorders
Insomnia
OCDPersonality
Passive Aggression
Personality
ShynessPersonal Growth
Happiness
Goal Setting
Positive PsychologyRelationships
Low Sexual Desire
Relationships
SexEmotion Management
Anger
Procrastination
StressFamily Life
Adolescents
Child Development
Elder Care
Parenting
SiblingsRecently Diagnosed?
Diagnosis Dictionary
- Magazine
- Tests
- Psych Basics
- Experts
There is an active debate underway in the popular literature about whether antidepressant medications actually do anything chemically helpful for depressed patients. No one doubts that many patients report feeling better, and that most evidence less depression on standardized rating scales, following treatment. Read More
Oh yes they do!
I went off Lexapro for 9 months after which I had the most debilitating nervous breakdown. Lexapro is an SSRI, but is also used for Generalized Anxiety Disorder. Lexapro acts like a "buffer" for me, without it, it's like I'm an exposed raw nerve...an albino picking cotton in the intense sun. I think I'd be dead without it. I was totally dysfunctional.
Oh yes they do too!
After finally finding a resource that told me I suffered from depression and probably had since I was a teenager, the effect of taking prozac made me stand up and say "so, this is what normal people must feel like?" Then suffering through depressive episodes and being afraid that the prozac was failing after taking it for several years, my doctor switched me to a combination of wellbutrin and lexapro. I got relief. I have had depressive episodes since the combination of drugs, but get through them. Without the antidepressants, I think I would be physchologically crippled!
Greetings Dr. Reidbord
Based on my clinical experience over the past thirty-five years, it seems clear to me that antidepressant medications can be very helpful to patients suffering from depressive symptoms. While I don't believe this significant therapeutic effect on symptoms such as sleep and appetite disturbance, anxiety, anhedonia, low motivation and chronic fatigue are purely placebo based, I think there is always some measure of placebo effect involved in taking most medications: The patient is prescribed something that his or her doctor believes will be helpful, which in turn can bolster the patient's hope and provide some much needed encouragement. It is also difficult to tease out how much medications help in psychotherapy patients as opposed to the psychotherapy itself. Yet, in my experience, when psychotherapy alone fails to sufficiently mitigate depressive symptoms, the addition of antidepressant medication can make a significant difference in treatment. I would also point out that, as you know, since different patients respond differently to the same medications, it is often necessary to try out several medications before finding one that the patient can both tolerate re: side effects and benefit from re: symptom relief. Part of the problem regarding these medications is, for me, really about not knowing exactly how they work to relieve depression, not fully appreciating the psychological nature and meaning of depression, and a need to recognize their clinical limitations: Just as psychotherapy (depending in part upon who provides it) is sometimes unable to ameliorate severe depressive symptoms by itself, antidepressants alone only do so much, and cannot provide what psychotherapy can. This is the main reason why combining pharmacological treatment with psychotherapy and psychotherapy with pharmacological treatment is typically more effective in the treatment of more debilitating depressive conditions than either alone.
Disagreeing with the blind
Only the emotionally dense think antidepressants don't work. I think they always work to one degree or another.
Studies are well known to be superficial. For example: Do spouses get to add their thoughts to the assessments?
Bipolars are mixed in with chronic Depressives. Bipolars need different medicines. Bipolars are natural actors,so responding to a sugar pill favorably, is no surprise.
Those who are bottomed out ''in psychosis'' could be mixed in, which is very unfair, considering it usually takes longer for recovery, and more medicines of different kinds to get them out. Sometimes even needing ECT.
Psychiatrists don't need testing,only the(insensitive) Bipolar critics do. For the same reasons Bipolar Republicans need the constitution, to make a judgement call. Psychiatrists need ''only eyes''. Bipolars need to call home and check with mom first.
Nothing has changed. This is a disagreement involving the differences of emotions, and the blind spots that come with faulty ones. ''Mostly normal'' people against the'' insensitive Bipolars''.
Oh yes they do too!
I should add that I have had seen therapist throughout the years and I agree - there was relief by going through therapy in conjunction with the medication. For me, the two go hand in hand. For someone that is depressed and they think medication is going to snap them out of it! I personally don't see that happening.
I would also think the therapist would have gained a lot of experience in working with patients to see if the medicine is helping! Of course I guess that goes back to is the therapy helping or is it the medicine or is it both. Too complicated from a simple layperson like me, who has experienced being uncomfortably depressed.
Thanks for your comments
There are a great many people who report improvement from antidepressants — enough that I have to believe they do something clinically significant above and beyond the placebo effect. But the results of randomized controlled trials (RCTs) aren't terribly impressive. Interestingly, the same is true of psychotherapy: Lots of satisfied "customers", modest proof of efficacy in RCTs.
Maybe the studies aren't looking in the right place. In clinical practice, treatment is individualized. Medications are changed if ineffective, and differently nuanced psychotherapy happens with each client. As Dr Diamond mentions, combining medication and psychotherapy may help more than either kind of treatment alone. It's hard to study treatments that are so individualized, but that's a shortcoming of research methodology, not necessarily the treatments themselves.
We're left unsure about what has helped a specific patient. As I wrote in the original post, it doesn't really matter for that one patient — improvement is improvement. But it matters a lot for social policy, insurance coverage, risk-benefit ratios, health care dollars, and so forth. We don't have the data we need, and it isn't even clear what such data would look like since everyone's mental health issues are different.
Missing the point
I think this post and respondents are missing the point of what was argued in Marcia Angell's piece and the books she reviewed. None of them argue that taking antidepressants doesn't "work." They just say that the large majority of their beneficial effects are placebo. Placebos can be very helpful. I think people are taking these reports as negating their experience as users and doctors seeing benefits when taking drugs, and that is not at all what they are doing.
However, the problem with using user and clinician accounts is that you do not test yourself or your patients with placebos. So a psychiatrist who prescribes medication and sees that the drug is "working" cannot necessarily attribute symptom reduction to the drug itself, because they have no way of knowing if the drug or placebo effect made folks better. For all we know (and the books reviewed discuss), the beneficial effects of antidepressants could be largely placebo.
The meta-analyses discussed showed that the improvements that could be directly attributed to antidepressants people experienced was very small in comparison to what could be contributed to placebo, small enough that it would be hard for a clinician or user to really notice the effect when interviewing/diagnosing. However, the drug plus placebo effect is large enough for people to feel better. Again, taking antidepressants make a lot people feel better and taking meds may have helped a lot of readers feel better, but the active ingredients in the drugs are not really responsible for that benefit.
The larger dilemmas here I think a primarily two-fold. While the total experience of being prescribed and taking an antidepressant can reduce depressive symptoms, the experience also come with potential serious risks. If the actual drug effect is small, it becomes more of a problem to prescribe medications with potential for harm when perhaps other interventions/medications/therapies show greater effect with fewer risks. Additionally, psychiatrists, insurance agencies, and pharmaceutical companies make a lot of money in prescribing, covering, manufacturing, and selling these drugs that in themselves don't do a whole lot. If there are cheaper, more effective options available, it may be problematic to continue to make/prescribe such medications.
It would be great if we could more systematically identify people who may be more responsive to drug effects to make their use valid, but attempts to do so have come up short. We really don't know beforehand who will benefit from largely from drug effects. What we do know is that when depressed people take antidepressants, the "active" ingredient is mostly placebo.
Did you read the post?
In my opinion, you repeated what Dr. Reidbord said. I don't see any points he missed that you made.
I suppose was responding to
I suppose was responding to the majority of people making comments, as well as responding to a few particular points. I will go through them below:
"This is my own experience, by the way - it's nearly inconceivable to me that antidepressants are no more than placebos. I've seen too many patients improve before my very eyes."
Though Dr. Reidbord acknowledges the findings of the meta-analyses, he is in near disbelief that the experience of taking antidepressants is largely placebo and doesn't seem to think that a placebo effect(though shown to be large) cannot help the many patients he has seen get better before his eyes. As I said earlier, the studies do not say that taking antidepressants are not effective in treating depression, but that placebo is the main component of their benefit. Dr. Reidbord seems to think that just because they are primarily placebo medications, they can't be as helpful as "the real thing", which is not the case and devalues the power of placebo.
"Some patients get dramatically better on antidepressants (in entirely believable ways, as opposed to reactive "flight into health" and the like), some only a little, and others appear not to change at all. Widely varying responses can easily "average out" in the usual randomized controlled trials used to assess efficacy, and could account for lackluster findings in group studies."
It would be great if heterogeneity of depressions could help explain who would receive more "drug" benefit from antidepressants. Added to the placebo effect, these folks would probably do best with antidepressants. Though I have never seen a study that compares different symptom clusters people may exhibit when diagnosed with depression with the effect of antidepressants, the studies mentioned did try to see if heterogeneity of severity of depression (as measured by the HAM-D), accounted for any difference. What was found was that the drug effect remained pretty constant (and clinically insignificant) whether people were mild, moderately, severely, or very severely depressed. What changed was the effect of placebo and hence the ratio of placebo to drug effect size. The drug to placebo ratio effect is higher for very severely depressed people, not because the drug effect is stronger, just because the placebo effect is smaller. It has been theorized that more depressed people are less hopeful that any intervention will work, so they are less likely to see placebo benefit.
"If a patient feels better, I don't worry too much about who or what gets the credit. Maybe it's the citalopram or sertraline in the pill. Maybe it's the patient's belief in the pill and in the medical science behind it. Maybe it's the fact that I gave the patient something that our culture imbues with symbolic healing powers. Maybe my words were healing and the prescription was a mere distraction. Or maybe I had no effect at all, and the patient healed himself or herself. Usually it's impossible to know."
I think this is an important and unsettling sentiment, though I appreciate Dr. Reidbord's humility in the matter. Why would a seeker of mental health treatment pick a treatment plan that has an unpredictable outcome? If I went to a surgeon and said, we are going to this surgery, but we have no idea the likelihood of success or why it would work if it did, I would not consent to the surgery, given that I couldn't really weigh the the unknown benefits against the very real risks.
Additionally, why pay a great deal of money for a psychiatrists time or for medication if they cannot be shown to be effective for treatment (as the post suggests). There is strong evidence for other treatments (ex. CBT) that shows high likelihood of success in treating depression and no real risks.
I think it is deeply problematic for anyone to profit off symbolic healing powers of doctor/patient relationships or the belief that the science behind antidepressants is strong (it isn't) without first saying to patients that benefit people get from these drugs is primarily symbolic/placebo, especially when ingesting antidepressants come with risks directly attributed to the drug (not placebo). So unlike Dr. Reidbord who doesn't care how a person benefits from treatment (even though he gets paid no matter what the treatment could be attributed to), I think it is better for everyone to be upfront and direct about how these medications work (that they do not have much of an effect in themselves and we don't have a good understanding of their mechanism of action to begin with), that their needs to a large, honest public discussion about the economics of antidepressant, use, a greater focus on actually using strong research evidence to inform clinical practice.
Thanks for writing.
I confess that I too thought you had restated the points I made in the original post. But with your further clarification, I'd like to respond. My line about it being "nearly inconceivable to me" was certainly not an argument aimed to refute the empirical evidence. Even if clinical experience with patients makes it hard to believe, I nonetheless recognize that randomized controlled trials attribute most (or all) of the benefit of antidepressants to placebo effect. The point of my post was to highlight the strange situation this creates for clinicians and patients.
In the eyes of many clinicians, including me, there's an odd disjunction between these research findings and what we see in the office. The most obvious explanation is that we're misled by the placebo effect — our patients don't come with control groups. We like to imagine that our treatments help people; cognitive dissonance leads us to rationalize efficacy since we (and pharmacies, drug companies, etc) are paid well for these treatments. And so forth. We could just leave it at that, i.e., those who think antidepressants work better than sugar pills are just plain mistaken.
I was positing another possibility or two to account for research evidence that flies in the face of the subjective experience of so many people. After all, it's a little hard to explain, say, two failed antidepressant trials followed by a third successful one using only placebo mechanisms. But these were only my ideas.
I fully agree with you that psychotherapy has a high likelihood of success in treating depression and no real risks. I'm a strong advocate for psychotherapy. However, psychotherapy involves significant time and expense which dissuade many patients, and in any case many depressed people "vote with their feet" and seek antidepressants from primary care doctors etc. Given the strong and helpful placebo component, wouldn't we do patients a disservice by denigrating and downplaying these treatments? They wouldn't work as well. I do agree that as a society we need a "large, honest public discussion." I alluded to that in my comment just above yours. But one-on-one, with an individual, suffering patient? I'd rather see them feeling better.
Thanks for Responding
Dr. Reidbord,
Thanks for your response and clarification. I appreciate your willingness to entertain the notion that clinicians are often misled by placebo effects and can experience a dissonance when faced with evidence that antidepressants in themselves are not particularly effective after they have prescribed them to many people.
To your "voting with their feet" comment, I think people choose antidepressants due to continued false assertions that 1) the active components of them work, 2) they work by a specific and agreed upon mode of action (neurotransmitter "balancing"), 3) that the medication is easy to take and very effective for most people, and 4)that doctors and scientist support their effectiveness.
These false assertions influence client behavior, the kinds of treatments doctors offer, and the kind of treatments insurance companies cover. This is a deeply systemic problem I think due to all sorts of factors (false marketing, massive drug company funding of psychiatry research and influence over regulatory agencies, scientist/clinicians pushing ideological claims despite strong scientific counter-evidence, a lot of providers, insurers, and manufacturers make a lot of money from these false assertions and want to keep on doing so, etc.)
I agree that people should have a great deal of responsibility in deciding their best own treatment, but people in distress (people seeking antidepressants to treat depression), reasonably put a lot of trust in an expert psychiatrist who has gone through medical school, a psychiatry residency program, and I think would be safe to assume are supposed to have a lot more knowledge about treating mental illnesses than patients do. If not, why see a psychiatrist in the first place? I think the responsibility is now on the psychiatric community to reestablish trust in their treatment of depression and to be very forthcoming about the science and theory behind antidepressants and how/why they prescribe them in light of it.
In terms of cost differentials between psychotherapy and medication use, medication may be cheaper for some people with good prescription coverage. However, if the effects of the medication are not clinically relevant, psychotherapy is definitely the better value no matter how cheap the medication.
I also want to push a little on the use of psychotherapy to treat depression. The term psychotherapy here I think is being used as a broad category, but specific kinds of psychotherapies have been shown to be more effective for more people with depressive symptoms than others, and I would say those most effective treatments should be used as first line therapeutic treatments. Broadly, these are cognitive behavioral therapy (CBT), interpersonal therapy (IPT) or derivative therapies of those methods/theories (http://www.psychology.sunysb.edu/eklonsky-/division12/disorders/depressi...). I think it important to emphasize that not all psychotherapies are the same, and many have been shown to lead to different outcomes when treating particular mental illnesses. As you have said in another post, psychodynamic therapies are the hallmark of therapeutic interventions of psychiatrists.
Psychodynamic therapies are often very expensive because they are often open-ended (though not always), sometimes call for meeting several times a week (though again not always), and when provided by a psychiatrist, people have to pay for psychiatrists' time (which is more expensive than the time of other mental health professionals offering therapy). While in extensive us by psychiatrists who offer therapy, the major consideration beyond expense and time is that there is less research support showing their effectiveness and the research support is far less robust for the effects of psychodynamic therapies for treating depression than there is for CBT. I find it strange and upsetting that psychiatrists often use psychodynamic therapy as first line therapy treatment for depression when there is less evidence of its efficacy than CBT, CBT usually beats it in direct comparative studies, takes more time, and is more expensive. Additionally, psychodynamic therapies have higher rates of negative therapeutic reactions (worsening of symptoms) than CBT, making using psychodynamic therapies technically more risky for people who are already experiencing distress than the more scientifically supported therapies.
This being said, I have nothing particularly against psychodynamic therapies when they have been shown to work well for particular mental illnesses, just that they have less scientific support for treating depression than other therapies that already exist and I think then should not be considered a good first therapy option for people with depression. It is good that the option exists, because not everyone will respond to CBT or IPT, and they might to psychodynamic therapies. Again, we still don't know how to determine good fit for therapies before treatment starts, so we still have to rely on evaluations of outcomes from studies that have already occurred to determine how to best intervene to likely support the most amount of people the most of the time.
Relating back to cost then,3 months of time-bound CBT for instance can be comparatively cheap and very valuable when compared to paying for antidepressants for several months or years (as withdrawal can also be difficult), along with bi/tri-weekly consultation with a doctor. It can be much cheaper than taking antidepressants and seeing a psychiatrist for psychodynamic therapy.
I don't want this comment to come across as just psychiatry bashing, but I feel there are some major ethical considerations when a doctor can charge a large fee to prescribe often an expensive medication that has no clinical relevancy. I find this exploitive of people who are often experiencing considerable distress. It is great to want to help people when they come to you in need, but there are ways to do so that are based on good science and won't damage the bank accounts of people already in distress.
Psychiatrists have a lot to offer their patients and the field of medicine more broadly, I just think it is important that good science inform clinical practice and policy, not just ideology, historical norms, or market forces. Psychiatrists can bring a range of knowledge to the table when theorizing, developing, and assessing mental health treatments that optimize effectiveness, cost, and length of time in treatment.
another reply...
I appreciate your concern and close reading of these issues, but frankly I don't have time to fully address each of your points. As a general comment, you are conflating the very thing you said I conflated: the demonstrated efficacy of antidepressants (by whatever mechanism), and the findings of randomized controlled trials that fail to separate, or only slightly separate, active drug from placebo.
People get a lot of benefit from antidepressants, there's just no denying it. The question is why. I bristle at "chemical imbalance" rationales, which are false and misleading, but I do tell patients with major depressive disorder that they are apt to feel better if they take an antidepressant. Because that happens to be true. As I mentioned in my prior reply, "macro" level questions such as economics, the integrity of psychiatry as an institution, and so forth are out of place in one to one consultations with distressed patients. I save that for blogs and similar forums, and in those settings I would venture that you and I agree more than we disagree.
Your lengthy discussion of CBT vs dynamic therapy is off topic, so I'll only say this. CBT and other manualized therapies for specific conditions are much easier to study. (I was a research fellow for 3 years doing psychodynamic therapy process research at UCSF.) Thus, it's really no surprise there ARE more such studies... it's like looking for your keys under the lamppost because the light is better there. There's a curious lack of recognition of the considerable empirical evidence for the efficacy of dynamic psychotherapy, see last year's review by Shedler _American Psychologist_ 65(2): 98-109. Aside from this, dynamic therapy has different goals than CBT. If someone is seeking relief of depressive symptoms, pure and simple, I am happy to refer them to a CBT therapist (and have on several occasions).
Cost comparisons of medications vs psychotherapy involve many variables, and it can go either way, depending. But the perception, and often the reality, is that a generic antidepressant costs less and works faster than psychotherapy "on the average". That of course doesn't mean it's better, or lasts as long, or anything else. But it does explain why so much depression is treated by primary care MDs (not psychiatrists). I'll let it go at that for now.
Thanks for engaging
I acknowledge the distinctions between effectiveness of psychotherapies was a bit off topic from your original post. I raised it in order illustrate the importance of using strong evidence to inform clinical decision-making. Just like the ethical consideration of using antidepressants or not based on lack of evidence for the drug component of their effectiveness, so too are their serious considerations to be had when choosing therapeutic techniques (let alone considerations of cost, time, risk, etc.).
Quickly, as for Shedler's review, there have been very large and reasonable criticisms to his methodology (choice of studies to include in his meta-analysis and how they were weighted), the validity of many of the studies and other meta-analyses that were analyzed (many showing computational errors that led to impossible effect sizes as well as weak design), and some logic flaws in some of his conclusions. I would point to this sarcastic but useful comment in another post by Dr. Stefan Hoffman with associated reference and critiques by Dr. Mike Anestis on his website.
http://www.psychologytoday.com/blog/psychoanalytic-excavation/201002/gro...
http://www.psychotherapybrownbag.com/psychotherapy_brown_bag_a/2010/01/w...
As a disgruntled grad student and former evaluator of evidence-based HIV prevention programs and those seeking to show evidence, I am really sensitive to the difference between doing science to show effectiveness of particular interventions(controlling for variables that are not being tested), and simple pre-post-follow up evaluations (that can show improvement, but is lacking sufficient evidence to show that an intervention is what led to outcomes). the methodology of many of the studies and meta-analyses included in Shedler's review would be considered of poor scientific quality. I recognize that the world is complex, people are complex, doing good studies on human thoughts, feelings and behavior takes lots of time and money, and that they will always be questions that need to be answered that just haven't been yet. I think it is important however to weight good science over good evaluation, use strong scientific evidence when making decisions as much as you can, and use other evidence (naturalistic studies, evaluation, clinical experience) when the science isn't there and with a healthy decrease in confidence and increase in humility when you have to use them.
I suppose my final point of disagreement is that ultimately I don't think prescribing placebo treatment without informing someone that it is in fact a placebo treatment is ethical. I don't think it is found acceptable in other branches of medicine, and it shouldn't be in psychiatry or the other mental health professions. Again, I think it leads to poorer science being conducted, unjust healthcare systems, and most importantly, poorer overall health outcomes.
Shedler & the psychodynamic evidence base
Now we're far afield of "Do antidepressants work?" and into complex debates about psychotherapy research methodology. Hoffman chides Shedler's selection of meta-analyses, but meta-analyses of CBT and even purely medical issues are routinely criticized and debated in the same way. He claims psychotherapy studies need a control group, although what constitutes a control is problematic in psychotherapy research (because the active treatment itself operates according to placebo principles by design). And he doesn't like vagueness in the presenting problem, even though that's how most patients present to psychodynamic clinicians. Anestis offers a more fine-grained critique, again of the meta-analyses mainly. I'm not in a position to refute him, so maybe we shouldn't grant Shedler's data our utmost respect. I found the comments interesting, including yours; as in any such forum they went both ways.
Since you've felt free to go on at considerable length, I too will take one last shot at presenting my viewpoint on this. If you care to reply, I'll let you have the last word.
Western medicine's great strides are largely thanks to nosology (ie, defined disease categories), and to treatments aimed at the category, not the individual patient. Take 100 healthy volunteers, swab their throats with Streptococcus, and 88 (let's say) will soon evidence the disease we call strep throat. Do this experiment enough times and with proper controls and we have good evidence that Streptococcus causes the disease. This type of knowledge, of disease categories, is amenable to randomized controlled trials (RCTs). The equivalent in psychology, called nomothetic research, aims to find general laws of behavior. Intro psych texts are full of such laws, and the proud evidence base for manualized therapies such as CBT likewise depends on putting subjects into categories, and finding statistically significant *group* mean differences.
However, there's another kind of knowledge that is important too. Why didn't the other 12 subjects get strep throat? Is it the same reason, or do each of the 12 have a different reason? Looking at what makes people unique, as opposed to a member of a category, is called idiographic research in psychology. This is the nature of psychodynamic theory and treatment, and why it resists the usual RCT approach to research. Most patients who present for such treatment don't fit neatly into a category like "depressed". They vaguely say their lives aren't working well for them, or that their relationships are unsatisfying in a particular way. Or they get a "funny feeling" when dealing with competition. Or their boss rekindles authority issues for them. Or they can't trust their spouse's fidelity. And on and on. There *are* ways to study such therapy — pre and post measures in single-case designs, self-reports, analog markers like medical utilization and interpersonal connectedness — but you won't like these, because there may not be control groups and they're "weak science." Meanwhile, countless patients report high satisfaction with the approach (see the Consumer Report's survey from some years back), and it is based on a rich theoretical foundation that has been scrutinized and refined for the past 120 years.
I'm not saying the nomothetic approach isn't important. I'd be a witch-doctor (or worse) without it. But it's not the only route to knowledge. I'll close with this commentary by Dr. Jay Einhorn. Again, you can have the last word if you want it.
http://psychatlarge.blogspot.com/2010/05/truth-and-turf-in-psychotherapy...
Last word? More like last lots of words and an expression of gratitude for the experience.
In keeping with the original blog post and to my intended extend of deviation, I want to make explicit that my goal has always been to discuss effective treatments for depressions. Your original post focused on the lack of evidence that antidepressants have any drug effect on depression symptoms, and I expanded the debate to examine other interventions’ effects (various therapies) on depression symptoms. Antidepressants may be effective for other mental illnesses (some are indicated for general anxiety disorder), and psychodynamic therapies may also be effective for other mental illnesses or non-pathological life issues (though I ask to be shown quality evidence), but I am trying to keep my focus on depression.
SHEDLER AND META-ANALYSES
Both Hoffman and Anestis don’t have problems with meta-analyses per se. They have problems with misuse and misunderstanding of them. Meta-analyses are as useful as the studies that they analyze. (Hoffman referenced one of his own analyses of a meta-analysis of effectiveness of CBT in treating various anxiety disorders). I think they would both agree (but I won’t presume to actually know what they think) that meta-analyses are strong (their results are convincing and meaningful) if they:
1. Include only strongly controlled studies.
2. Weight studies that are less controlled (by using perhaps the Jadad scale in determining inclusion or weighting of other studies)
3. Use studies with similar narrow focuses.
4. Explicitly state any potentially known biases that could be occurring, and address how those biases are being mitigated.
So I would agree that meta-analyses that don’t follow these points, no matter what they are studying, would need to be severely questioned and critiqued. However, I think meta-analyses are useful in systematically making sense of a lots of different studies related to a particular subject. I don’t think Shedler’s review of meta-analyses overall did not really follow these points, and is an example of a review that should be aggressively questioned.
As for control groups, I agree that controlling for human variables in psychotherapy research/experiments is tremendously difficult and often requires a lot of thought and creativity to do so. It is part of the reason why I love and have so much respect for such research and the people who try to do psychological research -- who says human sciences are soft and physical sciences are hard!? There are however, ways and study designs that better control for variables than others, and I would advocate that those methods that do so are the ones for which their results can be more useful, generalizable, and help establish evidence for particular treatments.
Furthermore, Shedler’s review did not specifically address how effective psychodynamic therapy is in treating depression, and as Dr. Anestis shows, many of the quality studies included that assess that effectiveness show that other treatments are more efficacious. My original assertion was that because some treatments (medications, therapies, or anything else) can be shown to be more effective at treating depression than others, those treatments should be more highly considered when deciding first-line treatment of depression. I know that cost, time, access, etc. all play a factor in determining treatment as well, but ultimately, unless it is shown to be at least somewhat effective, the treatment is not really of value to be considered in the first place. I don’t think Shedler’s review adds much to the evidence pool of psychodynamic therapy’s treatment of depression, and that other EBTs should likely be preferential first treatments.
MULTIPLE ANTIDEPRESSANT TRIALS AND PLACEBO
Irving Kirsch’s The Emperor’s New Drugs, one of the books Marcia Angell reviewed in her NYT review, addresses I believe two possible ways to account for why the third or nth trial might “work” after several failures. He is quick to acknowledge that the evidence for both of these theories needs some bolstering, but I think they are promising places to look. I am also quick to note that I do not have the book handy and am relying on my own memory for describing these theories (so please check my book for my accuracy, though I will do my best to be accurate).
The first way would use some of the results of antidepressant research trials that compare antidepressants to an active placebo, one that mimics a common side effect of an antidepressant (like nausea or dry mouth.) While in his meta-analyses of effectiveness of antidepressants found that 82% of the effect of antidepressants could confidently attributed to placebo, in 7 of 9 comparisons of antidepressants to active placebos, the drug effect disappeared altogether. This implies that when compared to an active placebo, nearly all of antidepressant effects may be attributed to placebo. Kirsch argues that perhaps it is the side effects people feel from antidepressants that make them feel that it is working that leads them to get better, not the drug itself. Having some tolerable side effects, enough to know that the drug is affecting them in some way, may be enough to convince people that the drug will work for them. Applied to your situation of Mr. X who has a successful 3rd drug trial, perhaps the third drug had the right kind of side effects at the right severity for him to think that the drug was going to work and hence he started to feel better.
Kirsch also references a study of people who are given an agent that makes them vomit some for a period of a day, and then try subsequent rounds of “anti-vomiting medications” to counter the effects of the first agent. People would stay in the research trial until 100% of them found an “anti-vomiting medication” that kept them from vomiting after ingesting the first agent. Having done a preliminary placebo controlled test showing that the anti-vomiting agent in fact was the cause of people’s vomiting (not placebo), researchers began to how people would respond to the “anti-vomiting medications” they were providing. All of the participants in the study eventually found a medication that kept them from vomiting after taking the initial agent. I believe it took some people up to 6 different trials with different “medications” to find one that worked, but ultimately, everyone was able to find a substance that kept them from vomiting, or so they think. The trick of this study is that every “anti-vomiting medication” provided was a non-active placebo. People just thought they were getting medications. Despite that their treatment was 100% placebo, 100% of participants found that the treatment worked in treating an observable, physical experience (vomiting). The mechanisms involved to explain why this happened is difficult to parse just from this study, but it is clear that expectancy effects are very powerful in shaping experience of taking medications. Kirsch compares this study’s methodology to the one used for the STAR*D trials of antidepressants, which used a similar design (but with actual antidepressants and not placebos) and showed similar result, implying the same expectancy effects were at work to account for why people could benefit in their 3rd or 4th drug trial and not their first.
ANALOGIES
I acknowledge that I try to stay away from using analogies as argument or even as evidence for an argument. It just takes so much explanation to be precise with it that I find it is just easier to make a logical argument than relying on analogy. More so, analogies can be interpreted to fit any argument’s needs (as we saw in our exchange), rendering them particularly useful. That being said, it seems like in your own analogy that you think that as a key “psychodynamic treatments…open doors and unlock potential,” and also that EBT treatments for depression like CBT are also keys. I don’t see that we disagree that EBTs also can “open doors and unlock potential.” I don’t want to go so far into your analogy because I don’t want again to be accused of overanalyzing, but I think you are trying to say that somehow because EBTs are more studied, they do not assess anything deep, but somehow psychodynamic studies, because they are less studied, actually address deep issues in “the dark.” You might be implying that EBTs are easier studied because of “agreed upon dependent variables for measure,” but I can’t imagine that it is any more difficult to design a study of how psychodynamic approaches affect those same variables. I do not imply then that psychodynamic therapies cannot be effective for a variety of things or that they aren’t used with people who come to treatment with vague/non-pathological symptoms, but it isn’t hard to test their effects on particular (vague or otherwise) symptoms/variables, researchers just have to take the time to do it.
NOMOTHETIC RESEARCH AND IDIOGRAPHIC RESEARCH
Reading your comments on these topics, I don’t think we disagree much on definitions or that there various and ways of meaningfully knowing, but I do believe that our conclusions based on our understanding of these definitions differ.
I will agree that idiographic knowledge is useful. My undergraduate background focused on ethnography and qualitative methods in social sciences (particularly cultural anthropology). Learning about individual and unique experiences can help show how history, communities, important events, etc. come together in a person’s experience in meaningful ways. In terms of health sciences, I find idiographic knowledge tremendously useful in helping determine health care needs of different people/communities and what might be useful in addressing those needs. Particularly, and I think this is where you were headed with your line of questions about the people who did not end up getting sick with strep after exposure, idiographic knowledge helps to create new research questions and hypotheses that no one would have had a reason to think about beforehand.
However, I think there is a difference between using idiographic knowledge as support for developing research questions and hypotheses and using idiographic knowledge as answers/evidence for questions or hypotheses. It is one thing to notice that (to use and vary your analogy a bit) perhaps 10 of 12 people who did not come down with strep after exposure to Streptococcus were routinely taking antibiotics for a more benign condition (perhaps acne), and then doing RCTs that assess whether antibiotics are an effective treatment for people who have developed strep throat. In non-analogy terms, idiographic knowledge can shape research hypotheses, but it takes nomothetic research to determine whether those hypotheses are actually valid. You cannot generalize anything from exclusively idiographic knowledge, nor make any predictions about future events/phenomena from them. In designing treatments for particular conditions like depression (or even vague symptoms like feeling mistrustful), it takes nomothetic research/knowledge to actually determine things like how effective a treatment might be to those who receive it, to what extent that treatment can be generalized (ex. Particular population groups like ones based on gender or people with single or comorbid mental illnesses), and to determine how confidently it can be predicted that people receiving a treatment will experience benefit from it.
Yes, these kinds of knowledge are both important and they work together to make important strides in human health. Yes research needs to be done on both fronts. My argument, however, has always been about applying EBTs to depression, and I maintain that it takes nomothetic research to build convincing evidence.
PSYCHODYNAMIC THEORY/THERAPY AND WAYS OF KNOWING
Unlike ethnographic methods in cultural anthropology that do not claim to be based on scientific methods, nor confidently generalizable or predictive, psychodynamic theory and therapy somehow claim to still be so. Unlike ethnographies of particular people’s in particular places, times, and cultural contexts that cannot be expected to be replicated in any naturally occurring or lab context, effectiveness of psychodynamic therapies on particular symptomatic or diagnostic variables can be assessed by RCT study design.
As I’ve noted previously, evaluation studies of pre/post measures of single-case designs, exclusive self-reports, and analog markers, do not establish whether the therapy experience can be attributed any of the change seen in any of those measures. You are right in saying that I think these methods are “weak science,” and now that we are using more specific terms, poor ways of generating nomothetic knowledge which I have described above as I think needed to show effectiveness of treatments. Pre/post single-case designs are vulnerable to all sorts of biases - dissonance related to having invested a lot of time into a treatment that not having results would be painful, no control/comparison to account for any other component of therapy is accounting for the effects being seen (the benefits of just having a space where someone is actively listening to a client, where clients don’t feel judged, etc.), no assessment of other things that may be happening in someone’s life that isn’t treatment that could account for change, not accounting for drop-outs as potential treatment failures, etc.
I was an evaluator of HIV prevention programs that were being conducted where HIV prevalence rates and transmission rates were high. I get that sometimes when the science isn’t available and that the risk is too great to take the time to do all of the science beforehand, evaluation techniques like the ones you describe are sometimes the best evidence available to make a case for choosing and maintaining the use of a particular intervention. However, even after 120 years of psychodynamic theories and therapies, research needed to show strong, empirical evidence of effectiveness are not available. I’m not saying that it can’t be done; I am saying that the studies should be done. It can only improve overall mental health and improve treatment options/quality if psychodynamic therapies are rigorously studied. I would be very excited if psychodynamic therapies for depression could be added to the list of EBTs. In meantime, other theoretical models and treatments have already gathered a great deal of evidence of their effectiveness for doing so, and I maintain again that those should be considered first-line treatments for now.
As for satisfaction ratings, they do not in themselves say anything about the effectiveness of a treatment. I will agree that no matter how effective an EBT is, it would probably be difficult for people to agree to do one if they have low satisfaction of the experience. This is not the case for the current EBTs for depression. I tend to think of outcomes as only as beneficial changes people experience because of their experience in a treatment/intervention. For treating depression, these outcomes might include reduced symptoms of depression (increased mood, more sleep, better concentration, etc.) and the ability to do things again that depression was preventing (hold a job, feel hopeful about planning for the future, etc.) Satisfaction with an experience could not be considered an outcome, because is not a meaningful benefit of a treatment. This is a little off-track, but I just want to note that no matter how many people are satisfied or not, it has no bearing on determining if a treatment is effective.
EINHORN
This is probably more on Einhorn’s piece than you care to read, but I wanted to do some close reading of his words.
“Psychodynamic psychotherapy is better explained as a method, with the minimum of theorizing, and better understood, as much as possible, through models of neuroactivity, as Allan Schore is developing. This brings us closer to the truth about psychotherapy.”
Theoretical validity or not, the method can still be tested by RCTs. Neuroimaging studies are great too, but they are still often in the idiographic state of knowledge and unable to determine causalities. I think neuroscience will have a lot to teach all of us about mechanisms of emotion, thought, and behavior, and changing them, but it will still take scientifically rigorous studies to build a strong case for any conclusion that can be drawn from them.
“Its authors, implicitly defining randomized clinical trials (RCT) as science in psychology, explicitly define evidenced-based treatments as those which have been validated by RCT studies, and consign all other methods of therapy to the garbage pail of superstition and uninformed personal preference.”
I think I have been clear in my previous writing that I am more nuanced in my belief in how to establish rankings of strength of evidence for psychological research than Einhorn describes of the beliefs of the researchers he cites. RCTs provide really strong evidence for therapeutic interventions. It doesn’t mean that other studies cannot provide evidence, just that they are a lot weaker and we cannot be as confident about their conclusions. Fortunately, interventions can be tested through RCTs and there has been an abundance of research done that is able to show that well-controlled studies RCTs are possible. RCTs actually allow the opportunity for “superstition and uninformed personal preference” to be shown to be not superstition and informed. They just have to be tested.
“There is nothing scientific about [psychoanalytic theory], and the claim of scientific validity for theoretical pronouncements given without a shred of evidence (even liberally defined) to support them is justly characterized by radical empiricists as ludicrous. On the other hand, reducing human nature in order to fit it into the scientific method available at the time has ever been the problem of behaviorism’s search for scientific respectability in psychology. While the radical empiricists are right in asserting that merely claiming that what one is doing is scientific doesn’t make it so, this applies to their own position of defining science as equal to RCT as well; that’s a philosophy of science, not science itself.”
Even if there is nothing scientific about the psychoanalytic theory, like Einhorn, I believe the methods can be tested against other treatments to show effectiveness. Like discussed earlier, idiographic knowledge may be used to inform research questions, hypotheses and theories, but then systematically collected data can through nomothetic research can provide evidence for or counter those theories, and likely those theories will change through a continuing process of testing assumptions of them. This is a good thing because it is how science happens. Psychodynamic therapy research can be included in this endeavor, but for some reason many psychodynamic practitioners and theorists do not participate or even care to. It is only good for mental health research and studies of mind and brain if they do.
The other aspect of this quote is the strange tension that Einhorn demonstrates by first highlighting Shedler’s scientific evidence of psychodynamic therapy’s efficacy in one section of the piece and then denies that psychodynamic theories are not scientific and denounces a philosophy of science that is widely used to plan, conduct, and interpret research. Somehow Shedler’s review can at once be seen as efficacious using the “radical empiricist” non-normative understanding of science, but also because it has to deal with psychodynamic theories, it also undercuts the relevancy of that approach to science.
“In fact, most of what we know about human nature, and particularly brain structure and function, has come about through autopsy studies of people with brain injuries, by neuropsychological and imaging studies of brain-injured people, by animal studies, and by imaging studies of normal people; not by RCT studies, although of course they have made a contribution.”
I highly disagree that the most about what we know about human nature has been made through autopsy and neuro-psychological imaging. There are too many places I can take this disagreement, but they are off topic. I just want to highlight that a close reading of this assertion of “fact,” needs be done.
“...therapy is often more like education--in which there are lots of different schools and methods of teaching, and students may have to find the ones that work best for them--than medical treatment, which at least aims for an expert consensus of recognized best practice for any disorder.”
I admit that I do not like the therapy as education analogy. If it is, I feel like people with mental health disorders or other issues in their life would be more effectively and efficiently served by well-trained educators on human living (religious leaders, counselors with background in education, coaches, etc.) than by mental health professionals (clinical social workers, psychologists, and medical doctors etc.). I also find the sentiment disheartening and disagree with it that people in distress have to search blindly for methods of treatment that work best for them. I don’t believe that, as Einhorn implies, there is no way to make good predictions about what kind of treatment would work best for a particular condition/issue. EBTs, whether therapy or medication, allow for just that.
“The truth is that different methods and treatment relationships may work better for different people, or for the same person at different times. Twenty sessions of cognitive-behavioral therapy will work better for some people, five years of analytic therapy for others; or maybe both will work better for the same person, at different times in his or her life.”
I certainly don’t disagree, and I don’t think anyone who supports the widespread use of EBTs, disagrees with the first sentence of this quote. I just question how Einhorn would set out to determine which methods/treatments would be better for different people at different times. My belief would be that rigorous studies of treatment effectiveness that take into account personal characteristics, traits, diagnostic history, other resources etc., while very far from perfect, would provide more insight into making treatment decisions that just guessing based on potentially irrelevant criteria (which is what I think Einhorn implicitly indicates in his piece).
Even if there was no way of determining what treatment would work best for someone at a particular time, that would mean you have to assume that any treatment would be as good as any when making a decision. If that is the case (for argument’s sake I’ll say that both treatments are risk free), then there are still probabilistic ways of making decisions that can be useful. Using the example Einhorn gives between twenty sessions (let’s say 5 months) of CBT and 5 years of psychoanalysis, given his supposition of not being able to know what might work best, CBT I think would still be considered the best first choice. I would consider the best choice available given the condition the treatment that works in the least amount of time with the least amount of wasted time possible.
If someone chooses CBT first and it is helpful, then treatment only takes 5 months.
If someone picks psychoanalysis first and it is helpful, treatment would take 5 years
If someone picks CBT and it isn’t helpful, and then tries psychoanalysis, and that shows to be helpful, then treatment will take 5 years and 5 months.
If someone picks psychoanalysis first and it isn’t helpful, then tries CBT which shows to be helpful, treatment would take 5 years and 5 months.
If someone tries both treatments and both fail, that person is still in need of treatment after 5 years and 5 months
So just based on time in treatment, starting with CBT and it working would be the ideal situation. But as Einhorn implies, we have no way of really knowing whether CBT or psychoanalysis would be more likely to work, we have to compare possible outcomes of choosing either treatment.
If someone starts with CBT, then after failure of treatment success switches to psychoanalysis, only 7.7% of their time would be in a failed treatment.
If someone starts with psychoanalysis, and then after it fails switches to a successful treatment of CBT, they would have spent the same amount of time in treatment as the above scenario but 92.3% of their time in treatment would have been in a failed treatment.
The risk of starting with CBT and switching to psychoanalysis is so much less than starting with psychoanalysis and switching to CBT, especially because either switch condition takes 5 years and 5 months and if they both don’t work, you still end up at the same place. Because I would want someone to be in a treatment that takes the least amount of time with as little wasted time as possible, starting the significantly shorter treatment would be preferential. Of course in this scenario, the only meaningful difference between CBT and analysis is the time it takes to do them, but if as Einhorn suggests that you can’t know how to systematically determine if a treatment will work for someone, length of time is as good a criteria as any when making a decision to choose a treatment.
My point is that two-fold: 1) even if we don’t know exact effectiveness of a treatment for a particular person at a particular time (as Einhorn suggests), there are still systematic ways of making decisions that don’t rely on ideology or what a particular treatment provider is capable of providing. 2) The value of EBTs is that they are able to show likelihood that someone may respond to a treatment and to what extent that treatment could be predicted to help. They don’t guarantee success, but they say that a lot of people with a similar condition (depression), found a particular treatment more beneficial than another treatment or no treatment in treating their condition. Unlike Einhorn, EBTs indicate that there are distinctions of effectiveness of treatments and they can be used to inform treatment decision-making. This goes for therapy, medications, or any other intervention.
CONCLUSIONS
Again, I want to assert that treatment decisions for depression are based on strong evidence so that people can be treated more effectively (and quickly and cheaply, etc.) when they are in need. Doing research to establish EBTs and using them allows for the best known treatments to be used, for better ones to be researched, and that resources in our mental health care system are optimized.
Dr. Reidbord, I truly thank you for engaging me so critically and quickly about a blog post that is several months old and that went off topic. I really learned a lot from this exchange, and valued your willingness to have this discussion. You obviously care a great deal about your patients, your work, and your profession, and I hope to model your care and passion in the future.
I'll break my own promise... briefly
I can't resist one final comment. I believe you still miss my point about idiographic methods of treatment and research. Idiographic methods are not merely for hypothesis generation as a precursor to nomothetic research. Some approaches to knowledge are idiographic by nature.
If a patient presents with major depression, nomothetic research can and should guide treatment. If the problem is defined this way — as a disease category — then psychodynamic therapy must stand or fall on the same RCT basis as other treatments. Given the original topic regarding shaky evidence for antidepressant efficacy, manualized psychotherapies such as CBT, IPT, and a few others have the best evidence base to treat this disease category. Shedler (and Einhorn, and I) may be misguided in citing nomothetic support for open-ended dynamic therapy if this evidence is weak or contradictory.
However, psychodynamic therapists and psychoanalysts say this whole nomothetic approach is barking up the wrong tree — or looking for keys where the light is brightest, instead of where you actually lost them. Most patients present with unique, idiosyncratic complaints that don't fall into a diagnostic category; I listed a small sample above. Therapists like us often find DSM-IV and similar nosology nearly useless, as it sheds no light on the particular patient in the office, with his unique history, dreams, fears, hopes, etc. The psychoanalytic/dynamic perspective is to understand the uniqueness of that specific patient, and to help induce unique changes that may not help any other person seen in the practice. Idiographic research methods have been used to study such therapy, both whether it works and how (this was my own research focus for a few years, about 20 years ago). From the nomothetic viewpoint, it is inevitably "weak science." From the psychodynamic viewpoint, it's the best that this domain of inquiry allows.
And now I will really, truly sign off from this thread. You've been a very thoughtful and earnest interlocutor, and I really appreciate your civil tone and careful thinking. Take care.
Clearing things up
I hope my comments did not come across as implying that I think idiographic methods are merely for hypothesis generation. I think there are many fields where approaches to knowledge are ideographic by nature (ethnography, history, qualitative sociology, basically all of the humanities). I think we may be talking past each other on our understanding of scientific methodology and philosophy of science, as Einhorn pointed out is a debatable point.
I will push further on why I think it is strange that exclusively psychodynamic therapists (who are not willing/able to learn/use other methodologies when indicated by nomothetic research) still can’t put their knowledge gained from idiographic research to the test. I do not believe as you say, “idiographic research methods…[are] the best that this domain of inquiry [allows].”
I can envision a study design where folks coming to receive therapy for issues that do not have a clear diagnosis and who are relatively functional/healthy are randomized into psychodynamic therapy, another kind of therapy, and an intent-to-treat group. Psychodynamic therapists will use their theory and knowledge to work with clients individually to and find unique ways to address their unique issues, another therapy will do what it does (though I really think it is giving manualized treatments an unfair rap to imply that they are not sensitive to individuals’ unique needs and circumstances, I think they very much are), and the folks in the intent-to-treat group will wait for treatment. I think a generous time cap can be put on treatment (maybe two years), but if treatment can end before that time than it certainly should. At the beginning, a few times during treatment, and at the end of treatment, predetermined outcomes across groups can be assessed. These outcomes may include to what extent clients primary goals for treatment were achieved and how long it took to achieve them, positive changes in quality of life measures, positive changes in satisfaction with life measures, maybe some depression and anxiety scales for good measure, etc. Perhaps follow-up data can be collected a year after treatment ends as well to see if the outcomes remain/change over time. If these outcomes are more likely to occur and occur to greater extents in psychodynamic therapies than the intent to treat group and shows outcomes similar or better than the other therapy, I think those would be convincing results that psychodynamic therapy can be useful for people who come to treatment without descriptive, DSM disorders hoping to live better.
Of course this study is not perfect, but this kind of RCT design I think could put to the test psychodynamic approaches to therapy vs. other therapies or no treatment to the test. The therapists doing psychodynamic therapy will have two years to use their idiographic ways of understanding and treating their randomly assigned clients, and it can be seen if that methodology gets better results than other approaches for the types of non-pathological issues people come to therapy for. It also of course would not indicate psychodynamic therapy for any DSM diagnosis, but it would add other kinds of treatments to the pool of evidence supported treatments for for some of the idiosyncratic symptoms people often show up to treatment for.
Yes it is a study based on changes of group means, but it doesn’t change how dynamic therapists do or conceptualize their work. It also I’m sure will show that many individuals in psychodynamic therapy will get a lot of out of the experience. However, if I were a client seeking therapy or a clinician providing therapy, I would want to know what kind of approach/treatment is most likely to best help whatever the client is there to get help with. If psychodynamic approaches are shown to do it better/more often than another approach great, use those, if not, use the other approach. If there is no strong evidence of anything effective, don’t try anything that might risk harming a client (perhaps antidepressants), and use a treatment, including perhaps psychodynamic treatment, with less evidence but knowing that it might work.
So to bring it more together, my understanding of strong scientific research in developing and evaluating mental health treatments/interventions is really based on nomothetic understandings of coming to knowledge. It is true there are several ways that science could be understood and philosophy of science is not a static field, but this is the perspective that I currently agree with most. I don’t think this understanding is unaccommodating to evaluating any mental treatments, regardless of the way that treatment was developed (idiographic or nomothetic). If a treatment can’t compete with another on how effective it is, how long it takes, etc., then it should be lower on the ranking of which treatment to try to treat what it is intended to treat.
Again, thanks so much for putting the time and thought to engaging with me, especially months after you originally posted this blog entry. This really has been a stimulating discussion and has definitely helped me clarify where I stand and a lot of issues. I wish you all the best!
Lamppost analogy
I really don't see anything problematic with following good evidence for treatments, especially if it has been demonstrated repeatedly. To use your analogy, I think it is more than reasonable to start looking for my lost keys under a lamppost because if they are there, as I can't see in the dark, my only chance of finding the keys there are if I stumble upon them. Extending the analogy, ff I lost my keys on a dark street with only one street lamp working on it, looking for keys there first would be ideal because it would save me time and energy looking around blindly in the dark for my keys. Even if I did not find the keys under the light and had to look in the dark street, there are systematic ways of going about that would increase my chance of finding the keys, like making a grid of the street and searching each part of the darkness one by one. (using less rigorous study findings in a hierarchy of confidence in their results to influence my search)
EBTs are like lampposts on a dark streets. The point of developing them is to increase the number of lammposts you have on a street so you can be more able to better recognize where you should look for lost, light-reflecting keys (solutions to mental health problems. I might posit that EBTs may also be like a GPS service that help pinpoint where your keys are on the street before you start randomly searching to begin with.
It very may well be that the keys are in the dark, but advocating for treatments that do not add any light to the matter does not help find the keys any easier or faster.
I hope that makes sense.
You're over-analyzing, my friend.
Randomized controlled trials shine a bright light in a particular place: where subjects can be categorized, a standardized treatment applied, and agreed-upon dependent variables measured. Manualized treatments such as CBT are keys under this lamp post. Psychodynamic treatments are keys too — they open doors and unlock potential — but they lie a distance away from this light.
It was my analogy, for better or worse. You'll need to make one of your own.
Just another quick note.
Just wanted to add a few more cents to the discussion.
"I was positing another possibility or two to account for research evidence that flies in the face of the subjective experience of so many people. After all, it's a little hard to explain, say, two failed antidepressant trials followed by a third successful one using only placebo mechanisms. But these were only my ideas."
I would say this line of thought again confuses the differences between the drug effect of antidepressants and the total experience of taking them (including the placebo effect). The research evidence does not fly in the face of the subjective experience of clinicians and patients who find benefit from antidepressants. When people take them, many feel better. It's just that the drug effect is not responsible.
To explaining variability of trial outcomes, I would first say that trials have variability. They are not designed to prove or disprove anything, just allow people to make predictions about causality to a certain level of confidence from estimations of samples to broader populations. Sometimes means of participant data are higher or lower than population data, and not 100% representative. Researchers do meta-analyses in some part in order to account for this variability and to counteract some of the confusion mixed results from different trials crate. .
I think more telling is that the significance testing doesn't tell us about the effect size of drugs in these trials, just that a drug effect is detectable and not likely due to chance. I take your meaning of a "successful" drug trial to be that the drug shows statistically significant beneficial effects compared to the placebo control. However, because of the large sample sizes of these trials (many hundred participants), very small statistically significant effect sizes can be detected. The meta-analyses reviewed showed that in the aggregate, even when effect sizes of antidepressants are statistically significant, they are not clinically meaningful. That is, the drugs may have an effect beyond placebo, but people a clinicians/patients are not going to be able to reasonably discern it.
"Given the strong and helpful placebo component, wouldn't we do patients a disservice by denigrating and downplaying these treatments? They wouldn't work as well...But one-on-one, with an individual, suffering patient? I'd rather see them feeling better."
This is the crux of the dilemma with how to deal with these new research findings. It would make sense that doing public education that these drugs are not effective in themselves and that doctors are no longer confident in their effects would reduce the the placebo effect they have in reducing depression symptoms. I don't have studies handy, but there has been interesting research in showing effects of just prescribing placebos as such. These studies show that there is still strong effects of people knowingly taking a non-active drug on depression symptoms. Perhaps the act of doing something to alleviate symptoms everyday is helpful, perhaps just because we have such faith in the medical system that having a pill bottle makes us confident, but I think prescribing placebos as such, instead of prescribing expensive medications that carry substantial risk is far more ethical.
Working at the one-on-one level is not working all that well. By many accounts, very high percentages of people can be diagnosed with Major Depressive Disorder at a given time, a huge amount are not seeking treatment (due to cost, time, lack of access to treatment, etc.), we do not have a mental health care system capable of addressing the needs of everyone at the one-on-one level, and despite a huge surge in antidepressant use over the past twenty years, the rates of people with depression have not gone down (implying that these medications are not actually addressing individual suffering across the board or societal mental health outcomes).
Ultimately, maintaining a false premise of "drugs work" in order to do something to make a patient feel better just leads to continued bad science, bad policy, poor health outcomes, and a denigration of the mental health field's ability to come up with more effective, safer, and ethical treatments.
It's a pity you needed to write all that...
... due to a simple misunderstanding of what I meant. I wrote: "After all, it's a little hard to explain, say, two failed antidepressant trials followed by a third successful one using only placebo mechanisms." I meant drug trials on an individual patient, not research trials. We clinicians talk like that sometimes. E.g., "Mr. X has had 3 antidepressant trials, and only the 3rd was successful." This isn't an uncommon observation in practice, and as I said it's hard (but not impossible) to account for if antidepressants are purely placebos.
This discussion...
... about subjecting dynamic psychotherapy to statistical (nomothetic) testing, continues on my personal blog at
http://blog.stevenreidbordmd.com/?p=506
Post new comment