Comments on: Confidence Intervals http://blogs.bmj.com/ebn-confessions/2009/11/06/confidence-intervals/ Evidence-based confessions of a newly-qualified nurse Sat, 07 Nov 2009 07:51:34 +0000 hourly 1 http://wordpress.org/?v=3.5.1 By: Ohad Oren http://blogs.bmj.com/ebn-confessions/2009/11/06/confidence-intervals/#comment-10 Ohad Oren Sat, 07 Nov 2009 07:51:34 +0000 http://blogs.bmj.com/ebn-confessions/?p=19#comment-10 Thanks for the enriching piece, John. I think that your clarification of the indeed-confusing concept of confidence intervals is an excellent starting point for any person who truly wishes to get to the bottom of this issue. Why the so-called arbitrary parameter triumphed above all others, being stressed in every research article we read nowadays is interesting. What makes 5% (not slightly lower, not slightly higher) the acceptable compromise we permit for a statistically significant conclusion to be made? I always ask myself what would have happen if we developed a more individualistic approach to the patient, rather than reflecting from a population’s results to the specific patient. Those at the extremes of results (that typically create the false positives and false negatives) would have been analyzed in a more objective manner, while less epidemiologic assumptions would have diverted our thinking to some diagnoses. The question is what tools do we have to come up with a patient-directed clinical approach that would still benefit from research studies but that would, at the same time, prioritize each case as the core of our attention, from the presenting symptoms to the definitive treatment.

Ohad Oren

ohadoren@gmail.com

]]>