Abstract
Associations between SES and asthma are not always consistent, partly due to the complexity of the SES assessment at individual- or area-level. A first step to better understand this point is to examine consistency between SES measures.
The aim was to define the most appropriate area-level SES in the French context by 1) measuring agreement between area-level SES 2) testing the well-known hypothesis that French elderly women with low SES have lowest probability to smoke than those with high SES.
Analyses were performed in Asthma-E3N (nested case-control study on asthma; n=19,346; mean age=70 yrs; 72% never smokers, 12% low education level (number of school years (NSY) ≤12)). Two individual (education & occupation) and 3 area-level Deprivation Index (DI) (Townsend, French European Deprivation Index (FEDI), French Deprivation Index (Fdep) defined at the smallest French area-level and classified in quintiles) were used.
According to Townsend, FEDI, Fdep, 23%, 20% and 12% of the women lived in most deprived areas, respectively. Agreement between Townsend and the 2 French DIs was slight (weighted kappa: 0.21 and 0.06, respectively) and moderate between FEDI & Fdep (k=0.54). As expected, using individual measures, elderly women with low SES were significantly less often smokers than those with high SES (NSY ≤12 vs. >17: OR adjusted for age [CI95%]=0.40[0.35-0.46]). At area-level, significant associations were only found for the Fdep (most vs. least deprived area: 0.81[0.72-0.91]).
Fdep seems to be adapted for E3N elderly women, mostly teachers living in urban areas. Analyses will be pursued by studying associations between body mass index, alcohol intake, asthma and SES.
- © 2014 ERS
