Fiche publication


Date publication

janvier 1996

Journal

Journal of clinical epidemiology

Auteurs

Membres identifiés du Cancéropôle Est :
Pr BRIANCON Serge , Pr GUILLEMIN Francis


Tous les auteurs :
Bouchet C, Guillemin F, Briançon S

Résumé

A variety of factors may influence outcome measures in longitudinal studies, including placebo, Hawthorne, or natural history effects. Quality of life (QoL) measures are particularly subject to these phenomena. This 2-month postal survey was set up to examine the extent of nonspecific effects in a treatment (vitamin supplementation) group (n = 180), a placebo group (n = 180), as part of a stratified, randomized controlled trial, and two control groups (n = 768 each). Quality of life was measured using the SF36. The placebo effect had a significant impact on improving physical, mental, and pain dimensions (p = 0.02 to 0.04), and the Hawthorne effect was significant (p = 0.03 to 0.009) for psychological dimensions. Although the impact of natural history was not significant, it tended to worsen all QoL dimensions. Vitamin supplementation had no effect on QoL. These results demonstrate the importance of placebo and Hawthorne effects and suggest that they may be responsible for misleading results or reductions in the power of controlled trials.

Mots clés

Adult, Analysis of Variance, Bias (Epidemiology), Effect Modifier, Epidemiologic, Female, Humans, Longitudinal Studies, Male, Mental Health, Middle Aged, Pain, Physical Fitness, Placebo Effect, Quality of Life, Surveys and Questionnaires, Vitamins, administration & dosage

Référence

J Clin Epidemiol. 1996 Jan;49(1):15-20