"A study in 2005, published in the Journal of the American Medical Association, found that one-third of all medical studies turn out to be wrong."
http://www.livescience.com/health/091026-medical-research-fraud-errors.html
In addition, many remedies of non-scientific origin may be vindicated despite earlier claims from doctors.
"In 2007, scientists showed that honey works better than cough medicines in soothing children's coughs.
"A new review of studies finds 29 percent of cancer research published in high-profile journals had disclosed a conflict of interest.
While it's a good thing that the conflicts were disclosed, the review also found conflicts affect the research outcomes."
"The findings add to a mountain of evidence suggesting you should be skeptical of health and medical advice."
http://www.livescience.com/health/090511-medical-studies-conflict.html
While many may argue that these disclosures are important since they were revealed. In addition one can hardly argue that all scientific research is flawed in this fashion. Similarly, this doesn't represent a loop hole where every quack remedy suddenly acquires legitimacy.
In particular, the problem is more insidious because it undermines the very credibility that scientists and doctors rely on when attempting to deal with public health issues. It is little wonder that counter claims regarding medical practice gain such traction when so much research can be questioned based on conflicts of interest.
One can certainly understand that sometimes errors can occur and peer review can also ensure errors or faulty research can be identified and addressed. However, that process only works within the scientific community. The far greater problem is how such events are viewed by the public which is expected to take advice and recommendations seriously.
Until this issue is addressed by the scientific peer community, one shouldn't be surprised when the general public views scientific pronouncements as debatable and subject to alternate interpretations.
Actually, the JAMA article doesn't say that. What it does say is that "controversy and uncertainty ensue when the results of clinical research on the effectiveness of interventions are subsequently contradicted. Controversies are most prominent when high-impact research is involved." This is no surprise. The authors also note that "controversies are most common with highly cited non-randomized studies, but even the most highly cited randomized trials may be challenged and refuted over time, especially small ones." Again, no surprise. That's why you do multiple trials - to make sure the effect you saw is correct. And you should read non-randomized trials, and other certain types of trials, with a grain of salt anyway. I feel like a statement like the one above hurts science more than it helps - people think, well, since scientists have a 33% chance of being wrong, why should I trust what I read? That's not true - we're constantly refining, and it's not black and white. Besides, the numbers can always be massaged. If you take wrong to mean contradicted, then in actuality only 15.5% of studies were contradicted by subsequent studies (7 of 45). Another 15.5% found that the effects were actually weaker than the first study suggested, which isn't wrong, just clarified. But even then, only 14 of 45 studies had different outcomes, great or small, from original studies, and we don't even know whether the subsequent studies were just more rigorous and therefore found more true effects, or were better designed, or what. That's 31% of studies, which means 69% of studies are not "wrong," no matter how you define it.