Hacker News Reader

Comment by cpncrunch | Hacker News Reader

No it doesnt. Placebo controlled trials show it has no effect. Im surprised at the nhs for this shoddy review.

Checking the meta-analysis paper on sci-hub, Fig. 2 has 7 RCTs, all of which found large effect sizes. I think you're simply mistaken.

http://sci-hub.bz/http://onlinelibrary.wiley.com/doi/10.1111...

I think you didnt read it. It says there were no placebo controlled trials.

"The benefits, risks, burdens, and effectiveness of the new methods should be tested against those of the best current prophylactic, diagnostic, and therapeutic methods. This does not exclude the use of placebo or no treatment in studies where no proven prophylactic, or therapeutic methods exist." World Medical Association (WMA) Declaration of Helsinki.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2647647/ for more.

Are you familiar with c diff? Placebo's don't work. It's dangerous to have placebo controlled trials with this disease. The current standard of care is a course of very strong antibiotics which still leave > 25% chance of recurrence. These antibiotics are most often used as a control group in c diff infections.

In general, placebo controlled trials are usually conducted for diseases which don't have an effective standard of care, or for conditions which aren't serious enough to be considered life threatening. For more serious diseases where a standard of care exists (standard of care is usually scientifically discovered but validated by placebo controlled trials), you conduct trials and compare the performance of the new drug/therapy with the existing standard of care as the control group.

Yup. It's like wanting placebo controlled trials for cancer. It would be unethical and ungodly harmful to condemn people to suffering and death just so people could gather data that wouldn't even compare the results to the current best practices.

Also the result you care about is often not "does it work?" but "does it work better than the current approach?".

>Are you familiar with c diff? Placebo's don't work

More familiar than you apparently, because placebos have been proven to work:

https://www.medscape.com/viewarticle/853175

I think the problem is the subjective measure, so it's impossible to know which symptoms are caused by c.diff and which from IBS.

Placebo controlled trials for C Diff would be unethical thus would likely not be legal in any country that conduct their trials with any ethical standards. Just like you wouldn't have a placebo group in a trial for an new AIDS or cancer treatment. In this case we're looking for a treatment that works just as well or better than current treatments.

Well clearly that's not true because placebo controlled trials into FMT for c.diff have been done in the USA. AIDS and cancer have objective measures, whereas with c.diff the trials just look at subjective measures.

Also, given that the placebo improvement rates are 40-90% in all c.diff trials, it would be unethical to not use a placebo arm.

Here's a randomized trial with placebos that did work. I wasn't able to find one where it didn't work.

http://annals.org/aim/article-abstract/2545886/effect-fecal-...

I submitted on a while back. Will check later if people reverse the downvotes and assume good faith.

You'll provide evidence for your claims if people down downvote you for making claims without evidence?

I'm not terribly impressed by that approach.

In my world people ask politely for refs or search for it themselves, rather than downvoting first (which is saying "your comment isn't worth reading").

I got up out of bed, posted a comment and was going to look for the link later if people asked. (I did a quick search and didn't find it). However, from the comments here, it's pretty clear people aren't interested in the science. Carry on...

Is this the post you're referring to?

"There have been a total of 2 placebo-controlled trials into fecal transplants for C.Diff AFAIK. One showed some benefit over placebo, and the other showed no difference."

https://news.ycombinator.com/item?id=10679823#10680173

In case people can't read the article as it's behind a paywall, it's not the primary source. However, to summarize the differences, this one study should not be representative of fact, as it very well could be that in the trial that did not show any benefit, patients could have had very aggressive treatment prior to beginning the study, and thus may have already been clinically cured prior to the study actually beginning. I would definitely trust the OPs meta analysis and systematic review of multiple different trials before believing this placebo controlled trial, for a number of reasons:

1. What I already mentioned above regarding the conflicting results of these two placebo controlled trials

2. The studies used in the nhs article are all done for patients that we're currently undergoing treatment for c diff, not after a long treatment time.

3. Meta analyses and systematic reviews are always going to be higher powered statistically speaking. As long as the papers looked at don't stink of publication bias it's a very valid way of looking at all the research done on fecal transplants for c diff.

The primary source is the one posted above by elsherbini.

I would definitely trust a small negative placebo controlled trial over a large meta-analysis that didn't include any placebo controlled trials, especially for a treatment that has a 90% placebo responder rate (as the trial posted by elsherbini shows).

Why not publish your own meta-analysis, then? Or if not publish, at least write an amateur meta-analysis?

Because I'm not a scientist. And anyway, the reason they didn't include this is because back in July 2016 when this review was written, there were no published placebo controlled trials. The Kelly trial was only published in Nov 2016. Presumably newer meta analyses will take this into account.

But how can you draw a conclusion if you have conflicting results from both placebo controlled studies mentioned? One with a significant difference, and the other without.

That's the study I was referring to. You'll see that there was no difference between donor and placebo in the NY group.

But there was a difference in the other group, right? Isn't that still relevant? The study's authors certainly seem to think so. I've only read the abstract.

Yes, that group had a 40% placebo rate vs the 90% placebo response of the NY group. However it is a very small group, and only marginally significant. It needs a larger, better designed study (preferably one which objectively measures c.diff infection status before and after), but if one study fails to show a response that's a pretty good indication that the treatment is suspect.

Are you thinking of IBS?

No, Clostridium difficile (Kelly et al, 2016, as discussed in the comments above).

You mean shitty review?

Would you please stop posting unsubstantive comments to HN? We're trying for better than that here, and I'm sorry to see that you've done it before.

Reply on news.ycombinator.com