Those 1.8 points are 1.8 points over placebo when it is well known that depression responds particularly well to placebo. This study also only looked at first line (least side effects, relatively low effect) antidepressants, it averaged in Serzone which is a particularly weak one that was discontinued 4 years before the study, and it calculated the difference between average drug responses and average placebo responses rather than the average of differences between drug responses and placebo responses, which reduces the perceived effect.
Do you think that in the past depressives didn’t get any treatment that’s as good as placebo because they were told that their depression doesn’t matter?
But of course the framing wasn’t “your depression doesn’t matter”—the concept of “depression” wouldn’t even come up. Even doctors would frequently not think to check for it. If it manifested psychosomatically, you’d get painkillers. If it manifested as difficulty sleeping, you’d get sleeping pills. And of course there was always self-medication with alcohol. All of these are worse than placebo in the long term. You could get a depression diagnosis, especially if you made a suicide attempt, but it was much rarer, and not a desirable diagnosis because you could get institutionalized into a psychiatric ward and stay there for months or years (especially if being there depressed you).
Before modern times of course, it was basically nothing. If you were lucky enough to see a doctor, and so lucky the doctor was at least equipped with the concept of “melancholia” (basically a catch-all category for mental illnesses), you might be prescribed hot baths and opium, which would have temporarily helped a little. But other than that, your prescription would be prayer and other placebos.
But what’s the problem with that if the difference between placebo’s and the drugs is only so small?
On the treatment side, the difference between placebos and SSRIs is small. But again, SSRIs are only a small part of “the drugs”. They’re what we try first because they require very little oversight by the doctor and they work often enough, whether by large placebo response or by smaller genuine effect. But if they don’t, there are second and third and fourth and more things we can try, ending with electroconvulsive therapy that works really well but is super difficult/expensive to administer.
But the bigger difference isn’t in the treatment, it is in the diagnosis. Today we distinguish between about a dozen different mood disorders with different treatment plans (although most of these plans do involve SSRIs at some point). And most of all, we’ve gotten a lot better (though still not perfect) at distinguishing what is pathological and what is, say, a normal reaction to bereavement or an adaptive response to an abusive partner.
Without this knowledge, depression would often go unnoticed or one form of it would be mistaken for another. You have depression caused by a testosterone deficiency? You seem to be troubled, go pray. You have depression caused by lead poisoning? You seem to be troubled, go pray. You’re decompensating and probably going to have a psychotic break within the month? You seem to be troubled, go pray.
It’s not even clear that praying to a god who supposedly loves you has no positive effects and is worse than the the drugs.
Oh, prayer does have some positive effects. We call it religious coping and there are studies that prove it can help with (for example) fear of death in end stage cancer patients. But you have to be fairly intensely religious to get a measurable benefit, and it doesn’t help a lot.
There are a few studies showing religiosity to be weakly negatively correlated with depression (example), and a few others showing it to be weakly positively correlated (example). If prayer helped a lot with depression the evidence would be much more clear cut.
There are also a variety of traditional remedies that should at least have placebo effects.
Yes. And terrible side effects, too. St John’s Wort and Kava are among the worst examples. Or, again, Opium.
And if you count religious practice as a remedy, you should count the time and sacrifice it requires, as well as the wrong beliefs it entails, as side effects as well.
If prayer helped a lot with depression the evidence would be much more clear cut.
It doesn’t have to be “a lot” to be better than the 1.80 points for SSRI’s.
You have depression caused by a testosterone deficiency?
Do you have an estimate for the detection rates? How many people who have depression due to testosterone deficiency do you think do get accurately identified by our system?
Let alone people who have depression due to lead poisoning. I would estimate most of those not being identified by our present system.
And if you count religious practice as a remedy, you should count the time and sacrifice it requires, as well as the wrong beliefs it entails, as side effects as well.
Last week I spoke to a person who went to a Freudian psychologists for four years. Partly multiple times per week. They thought it didn’t bring them much forward.
It doesn’t have to be “a lot” to be better than the 1.80 points for SSRI’s.
For a single individual, no. But to beat 1.8 points on average across multiple studies with hundreds of subjects, yes that would have to be “a lot”. And it simply isn’t.
Testosterone levels are a standard test, lead poisoning has thankfully become so rare it isn’t usually tested. But those are object level distractions from the point that mental health has advanced enormously, and a big part of that is the diagnostic side.
Freudian psychoanalysis doesn’t (usually) help, of course. That’s why I didn’t include it in the lists of things that can.
I thought we were arguing about actual clinical practice. Testosterone tests do exist but from what I read they are seldomly done in actual clinical practice.
Freudian psychoanalysis is still a large part of actual clinical practice.
Obviously. What’s your point?
I’m not that certain that St John’s Wort really has much worse side effects than many of the regular drugs. It might have more drug-drug interactions than various drugs because it has more active components.
There’s much money invested into proving that existing drugs do better than something like St John’s Wort and we know that this money skrews study results.
Testosterone tests are common in the group that tends to need them (men over 40).
Freudian psychoanalysis continues to be paid for by health insurers in Germany for historical reasons and there’s an aging cohort of psychoanalysts making their living with it in private practice, but clinics overwhelmingly do CBT instead, even in Germany.
What would convince you that St John’s Wort is inferior to modern antidepressants?
For moral reasons I would also want a clinic who reserves it’s right to take people’s freedom away to predict outcomes of it’s decisions.
There might be other reasons to see changes in the system as success. If suicide rates go down, that might be a sign that depression get’s treated better.
US numbers suggest 15–24 years olds are more likely to commit suicide while people over 55 are less likely to commit suicide.
That doesn’t suggest a much better system.
If you can point to other things besides suicide that caused by depression and the prevalence went down a lot, that might be a sign that our system is more effective.
I might also be convinced by an inside view account but it would have to be quite conclusive to overrule the biases inside the system for finding that patented drugs are more effective.
Those 1.8 points are 1.8 points over placebo when it is well known that depression responds particularly well to placebo. This study also only looked at first line (least side effects, relatively low effect) antidepressants, it averaged in Serzone which is a particularly weak one that was discontinued 4 years before the study, and it calculated the difference between average drug responses and average placebo responses rather than the average of differences between drug responses and placebo responses, which reduces the perceived effect.
For much more, check out http://slatestarcodex.com/2014/07/07/ssris-much-more-than-you-wanted-to-know/
Do you think that in the past depressives didn’t get any treatment that’s as good as placebo because they were told that their depression doesn’t matter?
Yes.
But of course the framing wasn’t “your depression doesn’t matter”—the concept of “depression” wouldn’t even come up. Even doctors would frequently not think to check for it. If it manifested psychosomatically, you’d get painkillers. If it manifested as difficulty sleeping, you’d get sleeping pills. And of course there was always self-medication with alcohol. All of these are worse than placebo in the long term. You could get a depression diagnosis, especially if you made a suicide attempt, but it was much rarer, and not a desirable diagnosis because you could get institutionalized into a psychiatric ward and stay there for months or years (especially if being there depressed you).
Before modern times of course, it was basically nothing. If you were lucky enough to see a doctor, and so lucky the doctor was at least equipped with the concept of “melancholia” (basically a catch-all category for mental illnesses), you might be prescribed hot baths and opium, which would have temporarily helped a little. But other than that, your prescription would be prayer and other placebos.
But what’s the problem with that if the difference between placebo’s and the drugs is only so small?
It’s not even clear that praying to a god who supposedly loves you has no positive effects and is worse than the the drugs.
There are also a variety of traditional remedies that should at least have placebo effects.
On the treatment side, the difference between placebos and SSRIs is small. But again, SSRIs are only a small part of “the drugs”. They’re what we try first because they require very little oversight by the doctor and they work often enough, whether by large placebo response or by smaller genuine effect. But if they don’t, there are second and third and fourth and more things we can try, ending with electroconvulsive therapy that works really well but is super difficult/expensive to administer.
But the bigger difference isn’t in the treatment, it is in the diagnosis. Today we distinguish between about a dozen different mood disorders with different treatment plans (although most of these plans do involve SSRIs at some point). And most of all, we’ve gotten a lot better (though still not perfect) at distinguishing what is pathological and what is, say, a normal reaction to bereavement or an adaptive response to an abusive partner.
Without this knowledge, depression would often go unnoticed or one form of it would be mistaken for another. You have depression caused by a testosterone deficiency? You seem to be troubled, go pray. You have depression caused by lead poisoning? You seem to be troubled, go pray. You’re decompensating and probably going to have a psychotic break within the month? You seem to be troubled, go pray.
Oh, prayer does have some positive effects. We call it religious coping and there are studies that prove it can help with (for example) fear of death in end stage cancer patients. But you have to be fairly intensely religious to get a measurable benefit, and it doesn’t help a lot.
There are a few studies showing religiosity to be weakly negatively correlated with depression (example), and a few others showing it to be weakly positively correlated (example). If prayer helped a lot with depression the evidence would be much more clear cut.
Yes. And terrible side effects, too. St John’s Wort and Kava are among the worst examples. Or, again, Opium.
And if you count religious practice as a remedy, you should count the time and sacrifice it requires, as well as the wrong beliefs it entails, as side effects as well.
It doesn’t have to be “a lot” to be better than the 1.80 points for SSRI’s.
Do you have an estimate for the detection rates? How many people who have depression due to testosterone deficiency do you think do get accurately identified by our system?
Let alone people who have depression due to lead poisoning. I would estimate most of those not being identified by our present system.
Last week I spoke to a person who went to a Freudian psychologists for four years. Partly multiple times per week. They thought it didn’t bring them much forward.
Depression drugs also have their side effects.
For a single individual, no. But to beat 1.8 points on average across multiple studies with hundreds of subjects, yes that would have to be “a lot”. And it simply isn’t.
Testosterone levels are a standard test, lead poisoning has thankfully become so rare it isn’t usually tested. But those are object level distractions from the point that mental health has advanced enormously, and a big part of that is the diagnostic side.
Freudian psychoanalysis doesn’t (usually) help, of course. That’s why I didn’t include it in the lists of things that can.
Obviously. What’s your point?
I thought we were arguing about actual clinical practice. Testosterone tests do exist but from what I read they are seldomly done in actual clinical practice.
Freudian psychoanalysis is still a large part of actual clinical practice.
I’m not that certain that St John’s Wort really has much worse side effects than many of the regular drugs. It might have more drug-drug interactions than various drugs because it has more active components.
There’s much money invested into proving that existing drugs do better than something like St John’s Wort and we know that this money skrews study results.
Testosterone tests are common in the group that tends to need them (men over 40).
Freudian psychoanalysis continues to be paid for by health insurers in Germany for historical reasons and there’s an aging cohort of psychoanalysts making their living with it in private practice, but clinics overwhelmingly do CBT instead, even in Germany.
What would convince you that St John’s Wort is inferior to modern antidepressants?
I wrote http://lesswrong.com/r/discussion/lw/oe0/predictionbased_medicine_pbm/ for specifying a formal way of how I want to be convinced that something is clinically more effective and better than an alternative.
For moral reasons I would also want a clinic who reserves it’s right to take people’s freedom away to predict outcomes of it’s decisions.
There might be other reasons to see changes in the system as success. If suicide rates go down, that might be a sign that depression get’s treated better.
US numbers suggest 15–24 years olds are more likely to commit suicide while people over 55 are less likely to commit suicide. That doesn’t suggest a much better system.
If you can point to other things besides suicide that caused by depression and the prevalence went down a lot, that might be a sign that our system is more effective.
I might also be convinced by an inside view account but it would have to be quite conclusive to overrule the biases inside the system for finding that patented drugs are more effective.