Should we banish the term ‘negative trials’?
24th September 2013
Sir Iain Chalmers of The James Lind Alliance writes:
I wrote about this in 1985 and I think it is still relevant. Too often people refer to trials as either ‘negative’ or ‘positive’ based on whether or not statistically significant differences between treatments have been found. This dichotomisation of clinical trials is misconceived and unscientific. Statistically significant differences can suggest that treatments have differential effects when in fact they don’t; and lack of statistically significant differences can suggest that treatments don’t differ when actually they do. The risks of being misled in these ways can be reduced by basing judgements about the effects of treatments on systematic reviews and meta-analyses and by using confidence intervals showing the range of possible values of true treatment differences. The terms ‘positive trials’ and ‘negative trials’ should be outlawed.