Talk:Clinical trials with surprising outcomes

From Wikipedia, the free encyclopedia

[edit] misleading title

Andrew there is a problem with this page's title. In most contexts, a "negative study" is one that shows no effect, no difference between the treatment groups, not necessarily one that shows a harmful effect. When I clicked on the link in EBM that is what I expected this article would cover but it doesnt. I agree the term has some ambiguity but EBM has been paying increasing attention to the need to publicize negative studies in the first sense, so it is especially misleading to invoke EBM and then use the term in a different sense. I was trying to think of a way to characterize the three studies you describe here: each compared a treatment outcome to a control group and demonstrated no net benefit. Perhaps what they had in common is that most physicians might have bet on the opposite outcome, so it might actually be more accurate to describe them as "treatment trials with surprising outcomes" or "treatment trials that did not support a promising or widely used treatment". Do you understand the two different senses of "negative study"? How should we handle this? alteripse 05:13, 27 October 2005 (UTC)

Thanks for your comments. I kind of anticipated that the title was going to draw some ire as I completely agree that the title doesn't do a good job of conveying the topic, but at the time same time, it was hard to come up with something that was succint. "Treatment trials with a surprising outcome" is an improvement. Andrew73 12:00, 27 October 2005 (UTC)

[edit] Criteria

What are criteria for inclusion on this page, given that we're documenting a trend? JFW | T@lk 07:26, 27 October 2005 (UTC)

JFW what do you think about the meaning of the term "negative study" as described above? alteripse 10:32, 27 October 2005 (UTC)

I wasn't anticipating any strict inclusion criteria, I was going to leave that open ended. I wasn't sure though if this is necessarily a trend but perhaps reflects the evolution of evidence-based medicine and perhaps an increased acceptance of publishing studies that had negative outcomes. Some other studies that come to mind include the emerging data about the lack of effectivenss of vitamin E, conservative instead of liberal transfusion thresholds, albumin. I'm sure that some of these topics could generate controversy as well. Andrew73 12:04, 27 October 2005 (UTC)

The phenomenon is well recognised, but the question is what constitutes a "negative trial" (I would have found "disappointing trial" more appropriate but equally difficult to maintain). Some recent trials on hepatic encephalopathy show that lactulose is rubbish and protein restriction nonsense. That is very significant - so-called "time tested" treatments suddenly blown away because there is finally statistical material available.

This is why I support the premise of this article but wonder how disappointing a study needs to be to merit inclusion here. And, to Alteripse's point: perhaps we should modify the title accordingly. But I have no idea what title would better cover such an effort. JFW | T@lk 00:35, 28 October 2005 (UTC)

Andrew, JFW, I am not trying to be argumentative, but your answers suggest I wasn't clear. There has been criticism, especially from an EBM perspective, that medical journals have had a longstanding publication bias in favor of positive studies over similarly powered, similarly significant negative studies (i.e., those that found no outcome difference between the groups). Here are two brief editorials that discuss this issue:[1], and [2]. In this context a positive result doesn't mean an experimental treatment is better and negative result that it is worse, but rather a positive result shows a difference between the treatments and a negative study found no outcome difference. Please look at those two brief editorials and perhaps you will understand that what they are referring to as a negative study is different from the examples Andrew gave in this article. In fact, from the descriptions of the three studies, the first is a positive study, the second might be a negative study but you didnt give enough info to exclude that it might be positive in the same way the arrhythmia study was positive, and the description of the third study also leaves it unclear whether there was a benefit advantage to no treatment (which would be a positive result). Do you understand now or have I thoroughly confused you? For which sense of negative study should we use the term negative study: one that shows a popular treatment is not as much better than an unpopular treatment as many people think, or the one that reports that there was no outcome difference? As described in the editorials I cited, EBM usage is this second sense. I think we should probably stick to their usage, but we should certainly make the difference clear to the reader in the first paragraph if you want to use this term differently. alteripse 01:33, 28 October 2005 (UTC)

The purpose of the article was to draw attention to the power of randomized clinical trials to debunk practices that initially found favor based on observational studies or theoretical principles. In this respect, I favored the first sense that Alteripse refers to rather than the lack of outcome difference...otherwise, I'm sure the article could be full of me-too drug A is not significantly better than me-also drug B!

I think the bottom line, the article needs a better title! Andrew73 02:35, 28 October 2005 (UTC)

[edit] New title suggestion

After further discussion with InvictaHOG, how about "Importance of randomized clinical trials in medicine"? Andrew73 15:23, 30 October 2005 (UTC)

I just renamed it to "Clinical trials with surprising outcomes." Perhaps this will be more accurate and appropriate! Andrew73 03:26, 6 November 2005 (UTC)