Do anti-depressants actually work?

[hebel]

This may be kind of a dumb question for a PGY-3, but what are we actually doing with anti-depressants?

A family debate recently had me looking more into this again, and I'm pretty surprised to see how little of an effect over placebo they have...and that even the measured effect is a rather small drop in HAM D scores.


Here's an example of such articles, by someone many of you are probably aware of:
Antidepressants and the Placebo Effect

If we consider giving a placebo unethical, what makes giving an active placebo with pontential long term side effects ok?

I'd really appreciate some insight on this!

---

Members don't see this ad.

 

[hamstergang]

I just searched this forum for Kirsch and found these 2 threads. I'm about to sleep and haven't reviewed them, but they seem like they'd be of interest until someone else responds.

How do psychiatrists feel about the SSRI vs Placebo issue?
0.3 Effect Size of Antidepressants, Implications?

 

[aim-agm]

Antidepressants have an effect size of 0.3. Proponents and opponents argue whether that counts as "working."

 

[clausewitz2]

Antidepressants have an effect size of 0.3. Proponents and opponents argue whether that counts as "working."

This argument exposes a weakness of Cohen's d, namely that because it is dimensionless, it is quite difficult to interpret except as a way of comparing studies within the same domain. It is great for comparing two antidepressant studies, say, but is almost useless for saying from an outside perspective what the practical implications ought to be. There is a convention about saying 0.3 is such and such an effect size.and 0.7 is such and such an effect size but this was literally and explicitly an arbitrary scale Cohen outlined on the basis of his intuition and not at all principled; even he said that it was likely to vary dramatically based on domain and the kind of decision making the evidence was going to be used for.

Mean(1) - mean(2) has drawbacks but is more interpretable for sure. Statistics are an auxiliary measure for making decisions based on evidence and not magical numbers that can do your thinking or arguing for you.

 

[ridethecliche]

I feel that number will have more clinical relevance coupled with an MCID style analysis. Haven't opened the study yet, but effect size needs to be rooted in an analysis that puts a clinical relevance to the metric using functional survey data.

One of the links discusses 1-2 point changes in HAMD17 but again... I'm not sure if this is a meaningful change in effect. The effect size could be high or low, but the survey has to be sensitive enough to find the difference... but even that doesn't mean that the difference is clinically significant even if it's statistically significant.

Here's a little primer on it. I've done some research in this and it's pretty interesting.

Clinimetrics corner: a closer look at the minimal clinically important difference (MCID)

https://www.ahrq.gov/sites/default/...h/findings/ta/drafts-for-review/trd-draft.pdf

According to this AHRQ doc on page 48, MCID for the HDRS is HAM-D17 of ≥27.1% ± 25.7% change HAM-D21 of ≥27% ± 25.1% change HAM-D24 of 28% ± 25.2% change, from this citation:
Rush AJ; Trivedi MH; Ibrahim HM; Carmody TJ; Arnow B; Klein DN; Markowitz JC; Ninan PT; Kornstein S; Manber R; Thase ME; Kocsis JH; Keller MB. The 16-Item Quick Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in patients with chronic major depression. Biol Psychiatry. 2003 Sep 01;54(5):573-83. PMID: 12946886.

The 16-Item Quick Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in... - PubMed - NCBI

 

[clausewitz2]

I feel that number will have more clinical relevance coupled with an MCID style analysis. Haven't opened the study yet, but effect size needs to be rooted in an analysis that puts a clinical relevance to the metric using functional survey data.

One of the links discusses 1-2 point changes in HAMD17 but again... I'm not sure if this is a meaningful change in effect. The effect size could be high or low, but the survey has to be sensitive enough to find the difference... but even that doesn't mean that the difference is clinically significant even if it's statistically significant.

Here's a little primer on it. I've done some research in this and it's pretty interesting.

Clinimetrics corner: a closer look at the minimal clinically important difference (MCID)

https://www.ahrq.gov/sites/default/...h/findings/ta/drafts-for-review/trd-draft.pdf

According to this AHRQ doc on page 48, MCID for the HDRS is HAM-D17 of ≥27.1% ± 25.7% change HAM-D21 of ≥27% ± 25.1% change HAM-D24 of 28% ± 25.2% change, from this citation:
Rush AJ; Trivedi MH; Ibrahim HM; Carmody TJ; Arnow B; Klein DN; Markowitz JC; Ninan PT; Kornstein S; Manber R; Thase ME; Kocsis JH; Keller MB. The 16-Item Quick Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in patients with chronic major depression. Biol Psychiatry. 2003 Sep 01;54(5):573-83. PMID: 12946886.

The 16-Item Quick Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in... - PubMed - NCBI

It doesn't help that the HAMD17 is structured in such a way that you can easily achieve 3 point changes by making sure your active drug makes people hungrier and sleepier than placebo. This may have something to do with the "efficacy" of SGAs for "treatment-resistant depression"...

 

[calvnandhobbs68]

It's interesting, but these findings aren't limited to anti-depressants. The most dramatic one you can bring up are the sham knee arthroscopies and back surgeries, which in half the studies were actually equivalent to real surgeries (Surgery Is One Hell Of A Placebo). There was an article in Scientific American a few years ago about how drug companies are having a hard time getting pain medications past clinical trials because it seems like the placebo effect is actually getting STRONGER in america (Placebo Effect Grows in U.S., Thwarting Development of Painkillers).

There's this underlying theory that when you're primed to expect benefit from the drugs/procedure, no matter what you actually take you'll tend to have improved symptoms because your body expects it. Interestingly, some of the articles speculate that part of this lies in the direct to consumer marketing of drugs/procedures in the United States, which is why you see a stronger placebo effect here. You should look into the "nocebo" effect too...the fact that patients will suffer "side effects" from the drugs even though there's nothing there.

I think all this proves is that when we educate patients on the mind-body connection when talking about conversion disorder/pseudoseizures/etc., we should really be buying into it cause it's actually pretty powerful.

 

[clausewitz2]

I think all this proves is that when we educate patients on the mind-body connection when talking about conversion disorder/pseudoseizures/etc., we should really be buying into it cause it's actually pretty powerful.

Yes, and even better, actually believing in what you are saying when having that particular conversation rather than just telling a pretty story that we hope might stigmatize less helps the functional neurological/NEBS diagnosis encounter go a lot smoother in my experience.

 

[Armadillos]

It doesn't help that the HAMD17 is structured in such a way that you can easily achieve 3 point changes by making sure your active drug makes people hungrier and sleepier than placebo. This may have something to do with the "efficacy" of SGAs for "treatment-resistant depression"...

Although to be fair for most folks with treatment resistant depression getting a good nights sleep and being hungry enough to eat is a great outcome.

 

[Sushirolls]

Although to be fair for most folks with treatment resistant depression getting a good nights sleep and being hungry enough to eat is a great outcome.

Treatment resistant depression lacks a uniform definition, whether it is 1-4 failed interventions, be it medication or therapy. After 2, or even 3 trials, all patients should be evaluated by a TMS/ECT Psychiatrist. I had accrued a remission rate of 64%, and 18% response rate, and 18% non-response with ECT.

The paradigm is, and needs to continue to shift towards early intervention with Neurostimulation. TRD often times isn't TRD...

 

[birchswing]

What are SSRIs most effective at treating (if it's not depression)? They're used for anxiety, OCD, hot flashes (Brisdelle), premature ejaculation, and probably other things I don't know. Maybe they should be renamed for whatever they're most effective at.

Although giving an elderly depressed woman anti-premature-ejaculation tablets might not go over well.

 

[clausewitz2]

Although to be fair for most folks with treatment resistant depression getting a good nights sleep and being hungry enough to eat is a great outcome.

I'll grant you the hunger part but I am not so sure Ambien is a worse choice if you are mainly aiming to make people sleep and certainly has fewer negative long-term effects than Seroquel.

 

[Shufflin]

In the Peruvian Amazon basin, spiritual healers known ayauasqueros have used ayahuasca brews for thousands of years to cure depression.

Ayahuasca brews are made of Banisteriopsis caapi, a plant containing the following harmala alkaloids:
- Harmine, 0.31–8.43%
- Harmaline, 0.03–0.83%
- Tetrahydroharmine, 0.05–2.94%

Guess what. These alkaloids of the beta-carboline class act as monoamine oxidase inhibitors (MAOIs).

Probably more than a placebo effect. MAOIs strike depression like a tiger.

 

[Armadillos]

Treatment resistant depression lacks a uniform definition, whether it is 1-4 failed interventions, be it medication or therapy. After 2, or even 3 trials, all patients should be evaluated by a TMS/ECT Psychiatrist. I had accrued a remission rate of 64%, and 18% response rate, and 18% non-response with ECT.

The paradigm is, and needs to continue to shift towards early intervention with Neurostimulation. TRD often times isn't TRD...

Believe me I am on the bandwagon as far as that goes, but I don’t think your going to be jumping to give TMS to all my uninsured patients.

 

[Armadillos]

I'll grant you the hunger part but I am not so sure Ambien is a worse choice if you are mainly aiming to make people sleep and certainly has fewer negative long-term effects than Seroquel.

I don’t disagree, I was mostly just pointing out that alleviating two the core symptoms of depression isn’t something to necessarily shrug off.

 

[Trismegistus4]

What are SSRIs most effective at treating (if it's not depression)? They're used for anxiety, OCD, hot flashes (Brisdelle), premature ejaculation, and probably other things I don't know. Maybe they should be renamed for whatever they're most effective at.

Although giving an elderly depressed woman anti-premature-ejaculation tablets might not go over well.

If nothing else, renaming them something other than "antidepressants" would at least foreclose the "why do they keep giving me antidepressants? I'm not depressed, I have anxiety!" objection from the benzo-seekers.

Anyway, of course antidepressants don't work. I have it on good authority that Xanax is "the only thing that works." That, or Adderall, depending on your condition. Either one of those drugs, if not both together, is a panacea for all psychiatric ills.

 

[clausewitz2]

I don’t disagree, I was mostly just pointing out that alleviating two the core symptoms of depression isn’t something to necessarily shrug off.

I feel like "anhedonia" or possibly some manner of hopelessness are closer to the phenomenological core of depression, given sleep and appetite can routinely be off in either direction .

 

[Armadillos]

I feel like "anhedonia" or possibly some manner of hopelessness are closer to the phenomenological core of depression, given sleep and appetite can routinely be off in either direction .

No argument about that, literally all I was saying is helping a severely depressed patient eat and sleep is good and improves their life.

 

[FlowRate]

The paradigm is, and needs to continue to shift towards early intervention with Neurostimulation. TRD often times isn't TRD...

Yes but it seems to me there's much better evidence for ECT than for TMS, unless I missed something in lecture.

 

[clausewitz2]

No argument about that, literally all I was saying is helping a severely depressed patient eat and sleep is good and improves their life.

Agreed. If we were only giving abilify to the truly depressed your point to would be very hard to argue with. Maybe your community prescribing patterns are different but I end up seeing a lot of people who get Seroquel who are not really having a problem with underconsumption of food.

Mostly I just get very annoyed with psychiatrists who reach for neuroleptics for depression 100% of the time after something first line because I guess it is hard to sue them for the patient getting obese ten years down the line but in the vanishingly unlikely case they overdose on their TCAs they think they will be sued.

 

[Armadillos]

Yes but it seems to me there's much better evidence for ECT than for TMS, unless I missed something in lecture.

Acutely definitely. Though I think there is an argument that if you get someone into remission with TMS it is more durable and less likely to relapse than ECT. Also worth remembering that TMS is so young that some studies have what we now know were suboptimal treatment courses such as too few treatments or sub optimal coil placement still make it into the lectures/metaanalysis etc.

That being said there was the recent negative study at the VA, not sure what to make of it. My personal anecdotal experience at VA was 1/3rd to a half of patients were malingering, so wonder if that makes it hard to run studies there.

 

[clausewitz2]

Acutely definitely. Though I think there is an argument that if you get someone into remission with TMS it is more durable and less likely to relapse than ECT. Also worth remembering that TMS is so young that some studies have what we now know were suboptimal treatment courses such as too few treatments or sub optimal coil placement still make it into the lectures/metaanalysis etc.

That being said there was the recent negative study at the VA, not sure what to make of it. My personal anecdotal experience at VA was 1/3rd to a half of patients were malingering, so wonder if that makes it hard to run studies there.

I think as long as your are applying these procedures to a category as amorphous and heterogenous as "major depressive disorder" you are in for a lot of noise in your results. Do you know of anyone trying to analyze characteristics of TMS responders to try and get a better grasp on this?

 

[WisNeuro]

Acutely definitely. Though I think there is an argument that if you get someone into remission with TMS it is more durable and less likely to relapse than ECT. Also worth remembering that TMS is so young that some studies have what we now know were suboptimal treatment courses such as too few treatments or sub optimal coil placement still make it into the lectures/metaanalysis etc.

That being said there was the recent negative study at the VA, not sure what to make of it. My personal anecdotal experience at VA was 1/3rd to a half of patients were malingering, so wonder if that makes it hard to run studies there.

Most definitely. If there are validity measures in place you can clean up the data quite a bit, but your N's generally have to be much higher in the VA to account for the data that you throw out.