Queen Square MS Centre Journal Club – Why clinical trials produce positive results for treatments that don't work

Speaker: Professor David Colquhoun

It seems to be alarmingly common for clinical trials to give "statistically significant" results for treatments that don't work. And even if the treatment does work, the effect size is often overestimated. To a large extent these problems are a result misunderstanding of statistics. It is scarcely an exaggeration to say that, as it is commonly used, statistics is a "mathematical machine for turning baloney into breakthroughs, and flukes into funding". This is convenient for bean-counting managers, but it's bad for patients.