You need of 95% confidence intervals to estimate uncertainty around the point estimate (in your example NNT 25) Any NNT without 95% CIs is useless Regards Nino Cartabellotta Director Centro Studi GIMBEŽ Via Amendola 2 - 40121 Bologna Tel. (+39) 051 5883920 Fax (+39) 051 3372195 www.gimbe.org _____ Da: Evidence based health (EBH) [mailto:[log in to unmask]] Per conto di Jon Brassey Inviato: lunedě 3 agosto 2009 13.59 A: [log in to unmask] Oggetto: Uncertainty Hello All, I may be missing the point, but do others get frustrated with the term uncertainty in medicine? I've just received another pile of promotional material from a publisher who's highlighting uncertainty in medicine and insinuating their evidence is the answer. Surely, 'all' evidence does is quantify the uncertainty? For instance a drug may have a NNT of 25 - where's the certainty in that? Here's hoping for some serious and lighthearted responses. Best wishes jon Jon Brassey TRIP Database www.tripdatabase.com <http://www.tripdatabase.com/>