 |
 |
(let me begin with an apology to anybody who gets this note twice, and a grovel to anyone who gets it 3 times!) as you know, we are doing some large, simple studies of the precision and accuracy of the clinical examination (with wonderful collaboration and exciting results in our initial pilot). for one of our next studies, we'll want our clinical collaborators to test two alternative approaches to determining a specific clinical sign, and one obvious approach to achieve comparability would be to allocate patients to the alternative diagnostic manoeuvres in a fashion that achieves the balancing properties of randomisation. the challenges we've identified are three: 1. to achieve concealment from the examining clinician, we can't use any prior information that the examiner might know (eg, odd or even birth date), since that might influence the decision to enter the pateint into the study (eg, if the clinician preferred one method of examination over the other, or if one way of determining the sign took a lot longer than the other). 2. and to prevent unconscious or conscious bias in determining a "random" patient feature (eg, odd or even diastolic pressure or pulse), we can't use any finding that is subject to end-digit preference or other systematic mis-reading. 3. finally, allocation has to be accomplished in the blink of an eye, so it can't require asking a third party or calling some hot-line. although tossing a coin or rolling a die would meet our objective, we would be uncomfortable about doing this in front of our patients (and reckon that their confidence in our subsequent therapeutic advice might be shaken!). can any of you, from prior personal experience or reading, offer advice to us? what allocation strategies could you suggest that would be: 1. capable of generating comparable groups of roughly equivalent size (a split not more extreme than 60-40) 2. concealed from the examiner and patient up to point of allocation 3. unobtrusive 4. instantaneous 5. capable of application in any clinical setting in any country i'll summarise the advice and report back to you. many many thanks dls for sharon straus and finlay mcalister ............................................................................ Prof David L. Sackett Director, NHS R&D Centre for Evidence-Based Medicine Consultant in Medicine Editor, Evidence-Based Medicine Nuffield Department of Medicine, University of Oxford Level 5, John Radcliffe Hospital, Oxford OX3 9DU, England Phone: +44-(0)1865-221320 Fax: +44-(0)1865 222901 Email: [log in to unmask] WWW: http://cebm.jr2.ox.ac.uk ............................................................................ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%