logo BDSP

Base documentaire


Votre avis nous intéresse

Le réseau BDSP met en oeuvre un projet d'innovation et d'amélioration de ses services, dans le souci constant de proposer des contenus de qualité adaptés aux besoins des utilisateurs.

Identifier de nouvelles sources de financements est la condition nécessaire pour assurer la poursuite et la pérennité de cet outil unique qu'est la BDSP, tout en le faisant évoluer.

Pour définir un nouveau modèle économique, nous avons besoin de votre avis : merci de répondre à notre enquête (temps estimé : 5 minutes).

Participer maintenant
Participer plus tard J'ai déjà participé

  1. Fulltext. Mismeasurement and the resonance of strong confounders : Uncorrelated errors.

    Article - En anglais

    Fulltext.

    Greenland first documented (Am J Epidemiol 1980 ; 112 : 564-9) that error in the measurement of a confounder could resonate-that it could bias estimates of other study variables, and that the bias could persist even with statistical adjustment for the confounder as measured.

    An important question is raised by this finding : can such bias be more than trivial within the bounds of realistic data configurations ?

    The authors examine several situations involving dichotomous and continuous data in which a confounder and a null variable are measured with error, and they assess the extent of resultant bias in estimates of the effect of the null variable.

    They show that, with continuous variables, measurement error amounting to 40% of observed variance in the confounder could cause the observed impact of the null study variable to appear to alter risk by as much as 30%. Similarly, they show, with dichotomous independent variables, that 15% measurement error in the form of misclassification could lead the null study variable to appear to alter risk by as much as 50%. Such bias would result only from strong confounding.

    Measurement error would obscure the evidence that strong confounding is a likely problem.

    These results support the need for every epidemiologic inquiry to include evaluations of measurement error in each variable considered.

    Mots-clés Pascal : Epidémiologie, Méthodologie, Erreur mesure, Classification, Facteur confusion

    Mots-clés Pascal anglais : Epidemiology, Methodology, Measurement error, Classification

    Logo du centre Notice produite par :
    Inist-CNRS - Institut de l'Information Scientifique et Technique

    Cote : 96-0254159

    Code Inist : 002B30A01A1. Création : 199608.