|
|
 |
 |
| |
 Printable version |
Effect size assessment in clinical and epidemiological research |
Sang Thrombose Vaisseaux. Volume 23, Number 10, 552-60, Décembre 2011, Note méthodologique
|
 Résumé
Texte intégral
|
Author(s) : Julien Labreuche |
Summary : The finality of medical research is to improve disease management by identifying the risk factors (epidemiology) and treating the disease or the underlying causes (clinical research). The assessment of the relationship between exposure to a risk factor or a treatment and the disease is a crucial step. As statistical tests provide information only on whether the association is statistically significant, it is necessary to quantify the strength of the effect using measurements called “effect sizes” in statistics. Several effect sizes (risk difference, relative risk, odds ratio) are commonly used depending on the study design. This statistical note describes how to use and interpret effect sizes when both exposure and disease are dichotomous variables, i.e. the most common situation in practice. |
Keywords : effect sizes, risk difference, relative risk, odds ratio, risk factor |
|
