So finally the 'scientific' community is admitting what we knew all along - from Medscape

Study Identifies Bias in Favor of Publishing Positive Antidepressant Trials
Marlene Busko

January 17, 2008
A study of Food and Drug Administration (FDA)–registered clinical trials of 12 antidepressants found a bias toward publication of positive results. Almost all studies viewed by the FDA as positive were published. The clinical trials that the FDA deemed negative or questionable were largely not published or, in some cases, were published as positive outcomes.

For each of the 12 drugs, at least 1 study was not published or was reported in the literature as positive despite a conflicting judgment by the FDA.

The overall effect size of the antidepressants (vs placebo) that was reported in the published literature was nearly one-third larger than the effect size for these agents that was derived from FDA data.

"Selective reporting of clinical-trial results may have adverse consequences for researchers, study participants, healthcare professionals, and patients," they conclude.

These findings are published in the January 17 issue of the New England Journal of Medicine.[snip]

'Evidence Based Medicine' is currently the be-all and end-all of clinical practice. This is problematical anyway in that whilst it is fine to accept that spending money on unproven treatments, it isn't accpetable to throw the baby out with the bathwater by refusing to fund treatments which are poorly or un- researched. Something like physiotherapy intervention is especially difficult to research, because so much of it is down to choosing the right treatment for the right patient, which is something that comes with practice. Evidence based medicine doesn't allow for any status to be given to treatments which are based on clinical experience.

As an example we can look at the application of heat treatment in low back pain. This is something that physiotherapists and others often do, yet there is very little evidence that it works (or doesn't work), according to the Cochrane Collaboration. Looking at the papers they have used to conclude this, we find that they are deemed poor quality because they are not all the 'gold standard' of randomised (patients randomly assigned to the intervention or control group) controlled (treatment group versus placebo) trials. Yet in clinical practice, my experience has been that it is pointless using heat therapy in anyone other than frail elderly ladies with a history of previous back pain and previous successful heat therapy, in which case it is more effective than anything else. But who will fund a randomised controlled trial of heat therapy on such people? How long would it take to find a suitable amount of patients to make a meaningful sample? Who would care about the results?

This is why treatment efficacy is usually decided on mass samples of lots of different people, thus hiding any efficacy for particular groups of people, or indeed for individuals. As physiotherapists, we are always urged to treat people holistically, to consider them as whole people. In practice this doesn't happen in the NHS, where taking a 'social history' usually means asking if the patient has stairs. Could it be that evidence based practice is making professionals even less likely to treat patients as individuals by leading them to disregard their own clinical experience?

I'm all for evidence based practice (strangely in psychiatry the evidence base is ignored when it comes to choosing between drugs or other treatments): but the evidence base has to be unbiased (which the above would seem to suggest isn't the case), and clinical experience shouldn't be discounted as evidence too.