Despite the fundamental importance of measurement in medicine, studies which compare methods of clinical measurement are often analysed inappropriately and may thus come to incorrect conclusions.
We explain the problems with the main methods used, notably correlation and regression, and describe the limits of agreement approach which we have proposed.
We consider why incorrect methods are so widely used and describe our own experience in bringing this issue to the attention of medical researchers.
Mots-clés Pascal : Méthode mesure, Etude comparative, Méthode statistique, Médecine, Evaluation, Validation test
Mots-clés Pascal anglais : Measurement method, Comparative study, Statistical method, Medicine, Evaluation, Test validation
Notice produite par :
Inist-CNRS - Institut de l'Information Scientifique et Technique
Cote : 95-0448352
Code Inist : 002B28F. Création : 01/03/1996.