In the September editorial,1 surgical researchers were reminded to submit manuscripts to the ARCHIVES only if they use adequate statistical methods and "quote the value of the test statistics as well as the P value; and, whenever possible, report confidence intervals."2 The November issue then contained an article in which all of these laudable principles seem to have been ignored.
Sauerland S. Too Much Confidence (Intervals) in Advanced Trauma Life Support?. Arch Surg. 2002;137(7):867. doi: