In CAM research this pattern is often not followed. Once systematic reviews and meta-analyses start to show that there is no evidence that the CAM treatment works, more small trials of poor methodology are conducted, many of which inevitably have (spurious) positive results. This allows CAM advocates to claim that there is lots of evidence in favour of their intervention, because they don't bother to account for study size and quality.
There is a fine example of this CAM tradition in the latest issue of Homeopathy. A systematic review, published in 2006, of homeopathy for treatment of allergic rhinitis concluded that "Some positive results were described in rhinitis with homeopathy in good-quality trials, but an equal number of negative studies counterbalance the positive ones. Therefore it is not possible to provide evidence-based recommendations for the use of homeopathy to treat allergic rhinitis, and further randomized controlled trials are needed". Well, perhaps: some would argue that the prior probability (close to nil) and currently existing evidence are enough to conclude that homeopathy does not work for allergic rhinitis (or, indeed, anything else). Be that as it may, it should be clear that the only useful new evidence would come from large and well-conducted RCTs. So what do Maria Goossens and a football team of colleagues do in the latest issue of Homeopathy? Why, publish a "prospective, open, non-comparative study" on homeopathy and allergic rhinitis, of course.
The methodology of the study consists of "treating" some patients suffering symptoms of allergic rhinitis with homeopathy, and getting them to fill in a quality of life questionaire at the start of the study, and after three weeks and four weeks. The physicians involved also assessed the severity of symptoms at baseline, three weeks, and four weeks. Unsurprisingly, the study found that people felt better with homeopathic treatment. But the methodological problems with this study are straightforward to point out. There is no control group. As a result, there can be no randomisation or blinding. Don't take my word for it; here's what the authors say in the discussion of their paper:
We did not distinguish between intermittent and persistent allergic rhinitis. All patients with intermittent allergic rhinitis (symptoms present less than four consecutive weeks a year) will be better after four weeks without any treatment. Patients who consult a homeopathic physician for allergic rhinitis usually have been suffering for a long time and from severe symptoms as the high level of the RQLQ score at baseline indicates. This study cannot be conclusive because there is no control group. Neither the physician, nor the patient was blinded. We cannot conclude that the degree of certainty of the physician about the appropriateness of the homeopathic prescription of a homeopathic remedy and the physician’s impression whether he had sufficient information about the patient’s condition influenced the outcome...it is not possible to draw a conclusion on the effect of the homeopathic treatment. This would require an RCT. To evaluate the effect of homeopathic treatment for allergic rhinitis an RCT should be performed.
So there you have it. The study cannot come to any useful conclusions. And, in the introduction to the paper, the authors write "This study was originally considered as a preliminary to a Randomized Clinical Trial (RCT) comparing standard conventional therapy with homeopathy (non-inferiority study). The RCT was never performed because sponsorship was withdrawn".
OK, that's life. Sometimes planned research funding fails to come off. These things happen. But why then publish the pilot study? Methodologically, it is useless, and it could never have added anything to the previously existing evidence from RCTs and systematic reviews. This study would never have been published anywhere other than a CAM journal, where scientific usefulness can take a back seat to an ideological desire to publish any evidence that looks as though it is in favour of homeopathy, no matter how methodologically weak it is, and in defiance of the higher level evidence that already exists.
Well, no-one is going to die from allergic rhinitis, so how much does it matter? The problem is that homeopaths don't stop at self-limiting conditions like hayfever. Some insist that homeopathy is a complete system of medicine and it can cure anything, including AIDS and malaria. Ben Goldacre's miniblog points to Jeremy Sherr's blog, for example, where Sherr is preparing to begin an unethical experiment on AIDS sufferers. This is a long road of madness, to be sure, but it begins where people believe they can cure hayfever through the use of magic sugar pills.