Slashdot Mirror
Major New Study Confirms Antidepressants Really Do Work (theguardian.com)
According to authors of a groundbreaking study, antidepressants really do work in treating depression, though some are more effective than others. "Millions more people around the world should be prescribed pills or offered talking therapies, which work equally well for moderate to severe depression, say the doctors, noting that just one in six people receive proper treatment in the rich world -- and one in 27 in the developing world," reports The Guardian. From the report: "Antidepressants are an effective tool for depression. Untreated depression is a huge problem because of the burden to society," said Andrea Cipriani of the NIHR Oxford Health Biomedical Research Centre, who led the study. The debate over antidepressants has unfortunately often been ideological, said Cipriani. Some doctors and patients have doubts over whether they work at all and point to the big placebo effect -- in trials, those given dummy pills also improve to some degree. Some people suspect drug companies of fiddling trial results. Some patients simply do not want to take pills for a mental health condition. The study published in the Lancet took six years, Cipriani said, and included all the published and unpublished data that the scientists could find. It was carried out by a team of international experts. They looked at results after eight weeks of more than 500 trials involving either a drug versus placebo or comparing two different medicines. The most famous antidepressant of them all, Prozac -- now out of patent and known by its generic name, fluoxetine -- was one of the least effective but best tolerated, measured by a low drop-out rate in the trials or fewer side-effects reported. The most effective of the drugs was amitriptyline, which was the sixth best tolerated.
Last line in the abstract gives the funding — National Institute for Health Research Oxford Health Biomedical Research Centre and the Japan Society for the Promotion of Science. it’s independent.
The authors are heavy hitters in the evidence synthesis field, with names on key articles. John Ioannidis, author of the famous paper “Why most published research findings are false”, and numerous other articles pointing out the pitfalls and limitations of evidence based medicine. Higgins is a co-author of the Cochrane Collaboration Handbook (the bible for systematic reviews), Cipriani and Salanti are key players in the development of the methods for network meta-analysis.
This is as good as it gets for medical evidence synthesis. A large dataset, expert authors, and findings that will come under scrutiny. Haven’t read the paper yet, but I’d expect the major limitations to be publication bias — the constant concern that negative studies don’t get published — risk of bias in individual studies, since many of those are pharma-funded, and whether the pooled papers fulfil the assumptions needed for network meta-analysis. There are formal methods for assessing the potential impact, and the discussion will get to grips with those questions. So will the commentary around it in other journals. See PLoS Medicine, for instance.