from Hacker News

Comparative efficacy and acceptability of 21 antidepressant drugs

by rb2e on 2/22/18, 1:59 AM with 108 comments

  • by amputect on 2/22/18, 4:27 AM

    I think that the United States is getting better about dealing with mental health, but we still have a long way to go. This is a little soap-boxy and anecdotal, and I apologize if it's inappropriate here.

    If I hadn't gotten a prescription for Prozac I probably would have killed myself by now. And I definitely would have destroyed my marriage and most of my important friendships. That's not an over-dramatization, it's the honest truth just based on the direction my life was heading without them.

    If you can't make your own neurotransmitters, store bought are fine. I'm not writing this to say "RARGH YOU MUST USE THESE DRUGS", but I absolutely am writing it to say "hey, this worked for me and got me out of a really dark and bad place". If you are reading this from a dark and bad place, please know that you're not alone. You have a lot of options, and I promise that if you take that first step, things can get better.

  • by harbie on 2/22/18, 3:28 AM

    >"We excluded quasi-randomised trials and trials that were incomplete or included 20% or more of participants with bipolar disorder, psychotic depression, or treatment-resistant depression."

    Wikipedia defines treatment resistant depression as "cases of major depressive disorder that do not respond adequately to appropriate courses of at least two antidepressants."

    Maybe I'm unfamiliar with study methodology, but doesn't this undermine the study's conclusion? It's essentially stating that forms of deppresesion that respond well to antidepressants respond well to antidepressants.

  • by ggm on 2/22/18, 4:19 AM

    A professional I know in the field (who repudiates drugs for CBT, and so is probably biassed, but still...) says that the lack of clarity over how an antagonist and a suppressant can both be claimed to operate on the same underlying problem and have adherents, points to bad understanding of what actually causes the problem.
  • by comex on 2/22/18, 5:41 AM

    The problem with studies of antidepressants is that they're almost all based on short-term, fixed-length treatment courses. After all, you can't do a double blind trial that lasts years on end; it would be unethical to give someone a placebo for that long, when antidepressants are believed to work better. Yet (AFAIK) people who go on antidepressants typically stay on them for years, so the studies are reviewing a fairly artificial use case. It's like the handful of controlled trials that attempt to compare programming languages, which by necessity ask participants to solve short, fixed programming exercises, even though that bears little resemblance to the process of real software development.

    In the case of antidepressants - on one hand, there are some reasons to expect short-term interventions to be the best-case scenario in terms of evaluating benefit, such as the greater risk of side effects with long-term use, and drug tolerance effects. On the other hand, I suspect (but don't have data to prove) that placebos are much less effective in the long term. People think the placebo effect is in part a reaction to the social experience of interacting with a doctor, getting personal attention and concern for whatever condition is supposedly being treated. To the extent this is true, the novelty of the experience is probably a large factor, and over the long term you'd expect a reversion to the mean. So even if antidepressants are less effective in the long term than the short term, they might be more compelling as a treatment option, because the alternative (placebo) loses even more of its effectiveness.

    Edit: Another factor is that the effects of reduced depression may take a long time to be fully apparent. Depression tends to work in feedback loops: as an oversimplified example, you feel bad about yourself, so you lose motivation to take care of your life, so you start neglecting essential tasks, the consequences of which make you feel even worse about yourself. And lifting yourself out of depression is the same thing in reverse. So if an antidepressant has the effect of reducing your susceptibility to depression - i.e. under the same life circumstances, you wouldn't lose quite as much motivation, or see things quite as darkly - then even a small change might tip the balance and let you stay at equilibrium in a more functional state of mind. But before you can reach that equilibrium, you have to go through a long process of getting your life back in order and regaining self-confidence.

  • by gehwartzen on 2/22/18, 4:59 AM

    Reading through the paper it seems like all the statistics are presented as Odds Raios vs placebo. But I couldn't find the actual efficacy of placebo. Did I miss it or does someone know?

    Something being twice as effective as a placebo is great when the placebo effect helps 30% of patients, not so much when it's 3%

  • by random_throw on 2/22/18, 4:02 AM

    Key bit:

    "In head-to-head studies, agomelatine, amitriptyline, escitalopram, mirtazapine, paroxetine, venlafaxine, and vortioxetine were more effective than other antidepressants (range of ORs 1·19–1·96), whereas fluoxetine, fluvoxamine, reboxetine, and trazodone were the least efficacious drugs (0·51–0·84)."

  • by askvictor on 2/22/18, 7:23 AM

    This analysis look at effects over 8 weeks. The first four weeks on antidepressants you're feeling terrible due to onset side effects (on top of your depression). Once that settles, you feel better. But there is no way to tell if you feel objectively better than before the medication started.
  • by jkmcf on 2/22/18, 5:29 AM

    PSA: The first thing to realize is that psychiatrists, and regular practitioners, have no clue what will work for you.

    In my case, it took 8+ years of trying, giving up, and trying again. E.g., Prozac didn't affect me at all, but the doctor never suggested increasing the initial dose. Paxil didn't work, Lexapro kinda worked, and now on the maximum dose of Effexor.

    The biggest difference makers were the doctors perfectly willing to help me find a solution. Keep trying to find one like this, and don't give up.

  • by djsumdog on 2/22/18, 5:44 AM

    > meta-analysis

    I knew a lot of other grad students, myself included, who would throw anything with meta-analysis in the introduction in the trash. You cannot deal with controls across completely different studies in meaningful ways.

    I'm also hesitant about anything that tries to claim things definitively without question. Science is about continually questioning your axioms. Without doubt[1] there is no progress.

    As someone who has been on various anti-depressants, I will say that some of them "worked" .. but the side effects were quite high. Working only lasted the first few weeks with several different SSRIs. Eventually the side effects ended up being worse than the treatment.

    I found the most effective thing for me was simply a really good therapist. She did try to recommend drugs to me again after I had quit, but she did respect my wishes to not be on them. I feel that having someone who really showed me my options and truly helped examine negative thinking patterns helped a lot more than the drugs ever did.

    That being said, I know people who say they'd be in serious trouble or dead without SSRIs. It's a tough line to talk about. I personally would rather not ever be on them again. Dulling the pain for me also meant dulling life.

    There are trade offs and we need to talk about them and have full discussions on the consequences of mind alternating drugs. When things are written into pure absolutes, it is a means of killing real discussion and dialogue.

    [1]: https://khanism.org/science/doubt/

  • by SpikeDad on 2/22/18, 4:05 PM

    I was worried as to the source of this study and the people who are claiming this "puts to rest the controversy of antidepressant drugs".

    It's all psychologists who naturally benefit greatly from a public perception that medication is effective. If I had seen quotes from psychologists saying "This is better than therapy - I'm stopping therapy and giving them drugs" then I would be on the bandwagon with the rest.

  • by imperio59 on 2/22/18, 7:03 AM

    Horrible side effects, no tests for diagnosis, no actual cures (have you heard of these drugs actually curing depression? No. Once you get on them you are basically on them for life...)...

    I think this "Science" of psychiatry has a long ways to go to actually get repeatable, scientifically proven results.

  • by siliconc0w on 2/22/18, 5:32 AM

    So it should probably be noted that given these are powerful psychotropic drugs so it's difficult to control for them with a placebo and we don't really know their longitudinal effects. Also proper diet, exercise, yoga, or meditation can all help with mood and depression but you can't patent that shit. This isn't to say the drugs have no value but the bar to use them should be high.
  • by dmitripopov on 2/22/18, 3:43 PM

    I don't think that researches like this make sense at all. Each depression is unique in it's pathogenesis, there are no drugs that work for everyone. For example, a depression caused by a chronic desease or the one with deep roots in your childhood - a true research needs to evaluate all 21 drugs in each situation.
  • by alz on 2/22/18, 5:51 PM

    Authors receive consulting fees from big pharma including Merck Sharp & Dohme, LB Pharma, and Pfizer
  • by alz on 2/22/18, 4:27 PM

    96 out of 522 trials identified as low risk of bias (18%)
  • by gliese1337 on 2/22/18, 5:11 AM

    As another person who is on antidepressants, and tried going off them for a while: well, duh. Yeah, I know, anecdotes aren't data, and I'm glad this is being verified by more reliable statistical methods, but this is not news to the many of us who might be dead or worse if someone hadn't said "you are sick--go see a doctor" and if that doctor hadn't said "you need medication--let's see if this works".

    Without antidepressants, my best case scenario is being an unemployed grad-school drop-out. With anti-depressants... well, it's amazing what you can do when you actually have enough neurotransmitters in your brain!

  • by icantdrive55 on 2/22/18, 4:24 AM

    To those struggling with the terminology, this site helped me a bit. I wasn't familiar with Odds Ratios (OR). And I'm still confused, with a headache.

    https://www.students4bestevidence.net/a-beginners-guide-to-i...

    I would love to have someone explain this study in a simple way, with simple percentages.

    For example, what is the efficacy of, say Prozac, compared to placebo?

    "In terms of efficacy, all antidepressants were more effective than placebo, with ORs ranging between 2·13 (95% credible interval [CrI] 1·89–2·41) for amitriptyline and 1·37 (1·16–1·63) for reboxetine."

    This says all antidepressants work between what percentage?

    And in another post, someone mentioned the newer antidepressants were better that the older ones. What percentage better?

    (And yes--I need to brush up on my statistics. I've taken a few of these drugs. I used to know how to read a simple study, and stopped because the results were so depressing. I am throughly confused with this study, and the charts. Then again, I'm not feeling great.)

  • by smt88 on 2/22/18, 4:54 AM

    This is consistent with my understanding, but:

    - how serious are the side effects?

    - how well would harmless placebo pills work in their place?

    I currently believe that there needs to be a safe, placebo-like option for people without severe illnesses. For those who are severe (and can tolerate side effects as serious as suicidality), it still seems to be a good idea to prescribe them.