"primary prevention for patients without clinical evidence of CHD, statins significantly reduced all-cause mortality, fatal MI and non-fatal MI in the meta-analysis carried out for the NICE appraisal however, no statistically significant differences were seen for cardiovascular mortality or stroke" Yes, it seems a confusing stance. As far as I can gather, statistical significance concerns itself with academic "proof" (ie more than 95% sure of certainty-or more technically, less than 5% percent of innacuracy), whereas the bottom line revolves around "clinical significance," ie the ability to to do what it says on the tin. It works the other way round as well. I once looked at some statistics around smoking cessation-the intervention showed a statistically significant difference in the number of cig's smoked-when we looked at this in more detail, the amount of cig's smoked was reduced from 20 a day, to 18. Statistically significant-yes, clinically significant-no. (in other words, the cig's will still kill you.) I think I've got the correct end of the stick, there's a really good BMJ book that can probably explain things better than me, or perhaps an Academic would care to step into the breach :-) It's important to note that the statement concerns itself with chd/cva MORTALITY only. I'm wondering if it means "no proof of effect" or "proof of no effect" (Very different things.) Notably, statins reduce "all cause" mortality, and I'm assuming, chd/cva MORBIDITY, which I guess is the ultimate bottom line.
Thanks for the questionning, it's good to shake around the grey matter.
Last edited by gingernut; 10th Sep 2012 at 20:09.