Please tell me what you think of this study.

[jdc324]

Meta analysis is never a good methodology, it depends upon the methodology of the studies under it, often they have different controls and protocols. Good meta analysis has large sample groups and demographic controls and rule outs.

Now it is good way of looking at what studies might show.

I'm not sure I agree that "meta analysis is never a good methodology".

I do take your point regarding inclusion/exclusion criteria for meta analyses though. I would point to Shang et al for an example of a meta analysis that included large studies with good methodologies: http://www.ncbi.nlm.nih.gov/pubmed/16125589

In both groups, smaller trials and those of lower quality showed more beneficial treatment effects than larger and higher-quality trials. When the analysis was restricted to large trials of higher quality, the odds ratio was 0.88 (95% CI 0.65-1.19) for homoeopathy (eight trials) and 0.58 (0.39-0.85) for conventional medicine (six trials).

Biases are present in placebo-controlled trials of both homoeopathy and conventional medicine. When account was taken for these biases in the analysis, there was weak evidence for a specific effect of homoeopathic remedies, but strong evidence for specific effects of conventional interventions. This finding is compatible with the notion that the clinical effects of homoeopathy are placebo effects.

 

[jdc324]

@Este

Thanks for the reply to my earlier post.

Now that you point it out, the similarity to the thimerosal argument seems obvious. I'm a bit surprised it hadn't occurred to me when I was writing that post.

 

[Dancing David]

I'm not sure I agree that "meta analysis is never a good methodology".

I do take your point regarding inclusion/exclusion criteria for meta analyses though. I would point to Shang et al for an example of a meta analysis that included large studies with good methodologies: http://www.ncbi.nlm.nih.gov/pubmed/16125589

The more common uses of meta analysis over emphasize the findings, and don't make enough rules out and exclusions. Almost all studies have different recruitment methods which makes for bad meta analysis from the start. A good paper using meta analysis will spend almost half the paper talking about the rule outs and exclusions.

Good ones do exist but they are in the minority.

Shang's paper may be a good one, I can't tell from the abstract.

http://www.sld.cu/galerias/pdf/siti...al-effects-of-homoeopathy_placebo-effects.pdf

I would say that the study is more of an inclusive study, which more likely reflects the nature of publication in homeopathy. So the caveats I would normally want to roll out about the nature of the study given that there is a large mix of protocols. But my guess is this reflects the nature homeopathy publications.

The results are interesting and not too surprising.

The slightly lower heterogeneity of the homeopathy results is very interesting as is the wash out in larger trials of the alleged effects.