@giordanopg recently tweeted a link about a paper in pre-publication from The Journal Of Pediatrics. It’s an analysis of a bunch of Cochrane reviews that had both child and adult RCTs included in a meta-analysis and asked the question “Do children and adults really differ?”
Their technique was to compare the key outcomes by comparing the odds ratios … and they found that for the majority of cases THERE WAS NO OBSERVABLE DIFFERENCE (overall relative odds ratio 0.96; 95% confidence intervals, 0.86 to 1.08)
This does hide the fact that in only 1 / 127 could it be said there was very unlikely to be a difference (ie no more than 20% better or worse), and that in 4 / 127 there was evidence of a real, clinically important difference.
How we read this, I think, depends on where we sit. Do we want to lump … and draw down adult evidence … or do we want to particularise, and require truly direct evidence? (Or do we go for a Balanced Parliament and have a bit of both with electoral reform thrown in?)
Acknowledgement: Picture from the Online Independent