The use of homeopathy to treat depression in peri- and postmenopausal women seems widespread, but there is a lack of clinical trials testing its efficacy. The aim of this new study was therefore to assess efficacy and safety of individualized homeopathic treatment versus placebo and fluoxetine versus placebo in peri- and postmenopausal women with moderate to severe depression.
A randomized, placebo-controlled, double-blind, double-dummy, superiority, three-arm trial with a 6 week follow-up study was conducted. The study was performed in a Mexican outpatient service of homeopathy. One hundred thirty-three peri- and postmenopausal women diagnosed with major depression according to DSM-IV (moderate to severe intensity) were included. The outcomes were:
- the change in the mean total score among groups on the 17-item Hamilton Rating Scale for Depression,
- the Beck Depression Inventory;
- the Greene Scale, after 6 weeks of treatment,
- response rates,
- remission rates,
- safety.
Efficacy data were analyzed in the intention-to-treat population (ANOVA with Bonferroni post-hoc test).
After a 6-week treatment, the results of homeopathic group showed more effectiveness than placebo in the Hamilton Scale. Response rate was 54.5% and remission rate was 15.9%. There was a significant difference between groups in response rate, but not in remission rate. The fluoxetine-placebo difference was 3.2 points. No differences were observed between groups in the Beck Depression Inventory. The results of the homeopathic group were superior to placebo regarding Greene Climacteric Scale (8.6 points). Fluoxetine was not different from placebo in the Greene Climacteric Scale.
The authors concluded that homeopathy and fluoxetine are effective and safe antidepressants for climacteric women. Homeopathy and fluoxetine were significantly different from placebo in response definition only. Homeopathy, but not fluoxetine, improves menopausal symptoms scored by Greene Climacteric Scale.
The article is interesting but highly confusing and poorly reported. The trial is small and short-term only. The way I see it, the finding that individualised homeopathy is better than a standard anti-depressant might be due to a range of phenomena:
- residual bias; (for instance, it is conceivable that some patients were ‘de-blinded’ due to the well-known side-effects of the conventional anti-depressant);
- inappropriate statistical analysis if the data;
- chance;
- fraud;
- or the effectiveness of individualised homeopathy.
Even if the findings of this study turned out to be real, it would most certainly be premature to advise patients to opt for homeopathy. At the very minimum, we would need an independent replication of this study – and somehow I doubt that it would confirm the results of this Mexican trial.
21 Responses to A new study seems to show that homeopathy works
Leave a Reply to Acleron Cancel reply
This site uses Akismet to reduce spam. Learn how your comment data is processed.
While I don’t have a reference right at hand, I am quite sure that recent research has cast doubt on the effectiveness of anti-depressants for mild depression. The current thinking is that talk therapy (especially Cognitive Behavioural Therapy or CBT) is the most effective treatment. While not formally talk therapy, some benefit can come from simply spending time with a patient and offering a some modicum of empathy. Thus mild depression is a condition that can have show a relatively strong placebo response.
It is important to note that more serious forms of depression do require medication.
yes, but this does not explain the difference between the homeopathy and the anti-depressant groups.
If there were any de-blinding if the patients (either deliberately or accidentally) it could lead to the difference.
Knowing the side effects of antidepressants I think we can take it as a given that at least some patients were accidentally de-blinded. The effects are noticeable.
“It is important to note that more serious forms of depression do require medication.”
Require? In order to make such an assertion, there would have to be a known biological basis for depression. Are you saying there is?
I would counter that there is not a single condition that requires psych drugs. Not one. Maybe in some cases a person in extreme crisis could benefit from being drugged, just as one would use street drugs to alter mood or behavior.
It would need far, far more than this study or even a more powerful version of it. There is no reason at all to think that homeopathy could work except as a form of comforting care and even there, cheaper alternatives abound.
It would need many trials, replicated by respected clinicians and uncontaminated by homeopaths. And before any recommendation for treatment it will require much improvement in standards of manufacture and quality control.
Perform enough studies and you will have some positive ones. Significant result only means that the null hypothesis seems to fail. But it does not tell anything about the probability of the alternative hypothesys. If the alternative hypothesys is so implausible as homeopathy, the following explanations of the study results are far more probable than that homeopathy works:
– just by chance we got this
– something happened and we have no clue what.
That is why it is totally contraproductive to cite individual positive homeopathy studies. If multiple evidences of different kinds will show consistency with the proposed theoretical background then we will have something to discuss.
The selection criteria, based on DSM-IV, has no solid evidence that I can see in the report of having been adequately screened against experimenter bias. One clue to this is the statement “The Diagnostic and Statistical Manual of Mental Disorders, abbreviated as DSM, is the diagnostic tool that serves as a universal authority for psychiatric diagnosis.” This is a huge red warning flag because: it is an appeal to authority; it is inferring that all other evidence-based diagnostic tools are invalid/useless, which is patently false; even worse, it it inferring that the selected participants had a mental disorder that was efficaciously treated via homeopathy.
One problem here is that there is a lot of noise in the measurements for depression. We are at the same time dealing with a treatment with a very small effect. We also have a medication that does not show an immediate effect. Six weeks to show an effect may be enough for most patients but not all. I would be surprised if this result could be replicated on a regular basis. This is more a demonstration of the limitations of our ability to do a clean study than it is for supporting a particular treatment.
A good point, David. I’m not entirely sure how valid any trial on the effects of antidepressants can be if it ends after just six weeks. By this time there will be a number of participants who do not yet feel the full effects of the medication, considering the average time for antidepressants to kick in is between three to for four weeks… One could probably think up settings for a study where the effects of placebos for mild depressions could be tested within that relative short trial time. But that’s another story as this wasn’t the aim of the study analyzed here at all.
Isn’t the clue in the synopsis?
“assess efficacy and safety of individualized homeopathic treatment versus placebo and fluoxetine versus placebo in peri- and postmenopausal women with moderate to severe depression”
~
Any strong placebo effect might be due to the individualized “treatment” (I don’t like using that word in regard to practices which defy Avogadro’s Number). It wouldn’t surprise me if the practitioners knew which was which and the more individualised (I’m Australian, hence the s) interaction was with the sugar water.
~
I agree with Pete about the use of the DSM. I knew (very past tense) a sociopathic psychiatrist who thought the DSM was worthless; after all, it once listed his homosexuality. 🙂
Each type of measuring instrument (including psychometrics and statistics) has a very limited and well defined area of operation (scope). Anyone and everyone can use these instruments, but only experts can obtain both meaningful and useful data from them.
Experts in pseudoscience and anti-science love misusing these instruments at the margins of their intended scope, usually well beyond their scope, in order to produce a study that they know will convince the majority of the general public simply because the study managed to obtain a p-value less than 0.05, or some other statistical metric, that will be just enough to qualify having the study published.
The pills/water used in homeopathy contain no active ingredient therefore I could easily engineer very similar results as this ‘study’ has produced by using my Pink Unicorn Therapy in place of homeopathy. However, I would never conduct such a diabolical ‘study’ because I actually care about the suffering and the well-being of others.
As always, pseudoscience and anti-science starts with a conclusion then it works backwards, always devising yet more meagre scraps of evidence to support its agenda and its literally mind-numbing levels of bullshit used in its promotional rhetoric.
Critical thinking skills alone are seriously underpowered for enabling people to develop an immunity to 21st Century quackery. The quacks are fully exploiting even more than the many techniques described by Robert B. Cialdini in his book Influence: The Psychology of Persuasion. After reading and fully understanding this book, it becomes much easier to spot (and to avoid falling foul of) most of the techniques used in advertising, telemarketing, door-to-door canvassing, and quackery.
“Isn’t the clue in the synopsis? . . . Any strong placebo effect might be due to the individualized “treatment”
I think that’s where the double-blind, double-dummy bit comes in, you know, in the synopsis. Everyone goes through the individualised prescribing bit, but neither patient nor the practitioner knows who got what. So no, that isn’t a very reasonable explanation based on what we know of the study design.
I think there is an oversight in reading completely the study. The study did not conclude individualized homeopathic treatment (IHT) is better than a standard anti-depressant. If you carefully read the conclusion you can find that IHT was better than placebo. No significant difference was found between IHT and fluoxetine. Regarding study duration, many conventional studies evaluating antidepressant response have a 6-week duration as you can read in the literature [Nemeroff CB, Thase M (2007). A double-blind, placebo controlled comparison of venlafaxine and fluoxetine treatment in depressed outpatients. J Psychiatr Res 41: 351-359; Gibbons RD, Hur K, Brown H, Davis JM, Mann JJ (2012). Who benefits from antidepressants? Synthesis of 6-week patient-level outcomes from double-blind placebo controlled randomized trials of fluoxetine and venlafaxine. Arch Gen Psychiatry 69 (6): 572-579. doi:10.1001/archgenpsychiatry.2011.2044].
As you mentioned, the positive results might be due to a range of phenomena, but this applies for any study, conventional or homeopathic. Methodological flaws are not exclusive of homeopathy studies. I agree with you that more studies should be conducted, in other settings, but the results of this study are supported by a RCT, which is the design suitable for evaluating efficacy.
It is true that a single study with positive results is insufficient for fully recommending a treatment for a specific condition, but the same thing is if the result is negative, it can’t be claimed that the treatment does not work for that condition either.
No argument will be sufficient for those skeptics that claim, based on their beliefs, that homeopathy does not work.
I am the author of the article mentioned above
EmmaMC wrote: “No argument will be sufficient for those skeptics that claim, based on their beliefs, that homeopathy does not work.”
It isn’t about beliefs, it’s about burden of proof. The default position is that something doesn’t work / isn’t true until there is robust evidence to support the claim. Think of a trial by jury, who owns the burden of proof: is it the accused who must prove their innocence or the prosecution who must prove the guilt? One hopes the latter and that the jury consists of skeptics.
Homeopathy no longer lacks evidence of efficacy, their is a mountain of evidence that it is ineffective beyond being a theatrical placebo. It is also scientifically implausible. Therefore, a trial that shows a positive result has produced a false positive result; a trial that shows a negative result is very likely to be correct.
Continuing with homeopathy trials is as silly as conducting trials that attempt to show the value of pi is, say, 3 rather than 3.14159…
Homeopaths believe (and have a deeply vested interest in maintaining their belief) that homeopathy works; and they mistakenly think that the default, skeptical, and scientific positions are likewise just beliefs. The logically correct wording of EmmaMC’s statement is:
No argument will be sufficient for those homeopaths that claim, based on their beliefs, that homeopathy works.
Here’s a good explanation of “true-believer syndrome”:
http://skepdic.com/truebeliever.html
Using homeopathy for the treatment of any medical condition (including psychological conditions) would be unethical and tantamount to committing health fraud and/or medical fraud. The only thing for which homeopathy is effective is in the teaching of science and critical thinking skills.
EmmaMC said:
I’ll leave Prof Ernst to comment on that, but surely there’s a problem in studying a nominally 28-day cyclical issue for only six weeks? In particular, in perimenopausal women with widely varying time between periods, even six weeks might not be sufficient to capture even one complete menstrual cycle with all the hormonal issues that implies.
Tu quoque. It’s this one that’s being examined here.
But, as has been pointed out, there are reasons to be cautious in interpreting the results.
Red herring.
Ah.
Because Alan has highlighted the study duration of 6 weeks I shall attempt to explain why this duration is so absurd using an analogy from a branch of engineering (applied science)…
Imagine that we record data for roughly one and a half cycles of the 50/60 Hz mains electricity supply voltage at various houses. Such a study, however well designed, will tell us nothing meaningful whatsoever about the quality and quantity of the mains electricity supplied to those houses. If the study data yields a very low p-value, does this mean the data is valid? No, because of the huge fundamental error in the design of the study. The study has failed to take sampling theory into account, either accidentally due to ignorance or deliberately due to vested interests.
Sampling theory clearly shows us these fundamentally essential points: we must sample variable quantities at the rate of at least twice the highest frequency that we wish to measure and observe; we must continue the sampling for many times longer than the period of the lowest frequency we wish to measure and observe in the variable.
In cases were both the prior probability distribution function of the data and the prior plausibility of the scientific tests have been sufficiently well established then it can be appropriate and reasonably valid to conduct simplified tests based on only subsampling the variable quantity being measured. But, as is usually the case in non-engineered systems such as people, whenever the variable contains asymmetry in its amplitude axis and/or its time-domain axis then the statistical analysis of its only subsampled data will yield unanticipated and unpredictably large errors.
Homeopathy, and most other branches of alt-med, are based on very close to zero (homeopathy has, effectively, now well below zero) prior probability and prior plausibility, therefore, they will yield much higher levels of unanticipated and unpredictable statistical errors when using studies, such as RCTs, that are acceptable to the prerequisites of testing in science and medicine. Alt-med refuses to accept the most fundamental prerequisite of the scientific method: Firstly, establish that a phenomenon actually exists before attempting to explain how it might work.
Rather than provide a long list of references to explain sampling theory and the often large anomalies caused by subsampling, all one has to do is sit close to an HDTV and switch back and forth between the same TV programme being broadcast on both standard-definition and high-definition channels. The vastly increased occurrence of anomalies in the heavily subsampled standard-definition audio and video will be completely obvious.
Television systems, of course, are based on well established high levels of prior probability and plausibility. Thus far, the only things a homeopathy-based television could produce is deafening and blinding levels of random noise. If one stared at the screen for long enough, a picture frame resembling the outline of Samuel Hahnemann might once appear, thereby confirming to the true-believers that homeopathy does actually work and that skeptics are wrong!
Since my specialty doesn’t include psychiatry, I cannot comment comprehensively on the validity of the study. But there are two points in the paper that caught my attention. First, as the authors mentioned in the discussion, the Fluoxetine regimen seemed to be used at sub-therapeutic (or not therapeutically-optimal) doses. I am curious as to the reason for choosing this dosage.
Secondly, I was waiting to see a vertical analysis of scores put forth in Table 3; there was none. The authors also didn’t comment on the fact that in this cohort, the placebo (nothing + nothing) seemed to decrease most scores appreciably by 6 weeks. In BDI and GS, this decrease in the placebo group is quite close to that in the IHT group. What does that say about the validity of the study design and/or the interpreted outcome?
It might be worth to note, that the study in question got retracted:
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0232415
thank you for alerting us – most interesting!