Dear All A few weeks ago I sought your advice regarding statins and their statistics. Thank you to the 30+ respondents, whose replies I shall be collaging and distributing FYI in due course. Naturally there was some overlap. The most-repeated comment was "David Spiegelhalter will know the answer to this". And indeed he did! The second-most common reference was to the Framingham Heart Study, which is the source for much of this data - see http://en.wikipedia.org/wiki/Framingham_Heart_Study My question now - dear statisticians - is maybe not truly statistical, but affects the statinistics. It concerns the definition of "cardiovascular event", which is the Framingham output measure. I assume that this includes everything from ministrokes to permanent paraplegia and death - am I right in this? And if so, what is the conditional likelihood of the various points on this spectrum, given that I have such an event? e.g.I am told I have a 30% 10-year risk of a cardiovascular event. This will reduce to 23% if I take statins like a good boy. (Strong steer from doctor to do this.) My decision will or may be different depending upon whether the bulk of these events are ministrokes, or whether the bulk is at the serious end of the spectrum. (I guess it's mainly in the middle - but what is the middle; indeed, what is the metric?) Thanks for any help or advice you can provide. If I do not reply it is more likely due to email overload than to any other adverse event. JOHN BIBBY You may leave the list at any time by sending the command SIGNOFF allstat to [log in to unmask], leaving the subject line blank.