Emmanuelle Cambois1, Sébastien Grobon2, Herman Van Oyen3, Jean-Marie Robine4. 1. Institut national d'études démographiques (INED), Paris, France cambois@ined.fr. 2. Direction de la recherche de l'évaluation, des études et des statistiques, Ministry of health (DREES), France. 3. Scientific Institute of Public Health (WIV-ISP), Brussels, Belgium. 4. Institut National de la Santé et de la Recherche Médicale (INSERM), Ecole Pratique des Hautes Etudes (EPHE) and Institut National d'Etudes Démographiques (INED), France.
Abstract
OBJECTIVE: The Global Activity Limitation Indicator (GALI), a single question measuring disability, had been introduced in various European surveys since 2004. The complexity of its wording has been questioned. Our study compares alternative variants aiming to simplify the wording. METHOD: We used the Health-Related Opinion Survey run in 2014 in France (N = 3,009). Its split sample design allows testing four variants of the questions. We analyzed the prevalence of activity limitation (AL) resulting from the four different constructs of the GALI using multinomial logistic regressions, adjusted for background variables and functional limitations (FLs). RESULTS: The alternative GALI variants result in significantly different prevalences compared with the original question, in particular for people with FL. The current variant is more inclusive than the routed variants. DISCUSSION: Our study suggests limited benefits of changing the GALI construct which do not outweigh the costs of breaking the established chronological series of the current variant.
OBJECTIVE: The Global Activity Limitation Indicator (GALI), a single question measuring disability, had been introduced in various European surveys since 2004. The complexity of its wording has been questioned. Our study compares alternative variants aiming to simplify the wording. METHOD: We used the Health-Related Opinion Survey run in 2014 in France (N = 3,009). Its split sample design allows testing four variants of the questions. We analyzed the prevalence of activity limitation (AL) resulting from the four different constructs of the GALI using multinomial logistic regressions, adjusted for background variables and functional limitations (FLs). RESULTS: The alternative GALI variants result in significantly different prevalences compared with the original question, in particular for people with FL. The current variant is more inclusive than the routed variants. DISCUSSION: Our study suggests limited benefits of changing the GALI construct which do not outweigh the costs of breaking the established chronological series of the current variant.
Authors: Adája E Baars; Jose R Rubio-Valverde; Yannan Hu; Matthias Bopp; Henrik Brønnum-Hansen; Ramune Kalediene; Mall Leinsalu; Pekka Martikainen; Enrique Regidor; Chris White; Bogdan Wojtyniak; Johan P Mackenbach; Wilma J Nusselder Journal: Int J Public Health Date: 2019-06-11 Impact factor: 3.380