Article Text
Abstract
Background/objectives Socio-economic status (SES) has long been recognized as determinant of cardiovascular risk factors and disease. Recent studies suggest an association of neighbourhood SES with risk factors independent of the individual's own SES, but the mechanisms have not fully been understood. Our aim was to assess the association of neighbourhood and individual SES with cardiovascular risk factors in an Eastern German population with exceptionally high cardiovascular mortality and unemployment rates.
Methods We used cross-sectional data of 1779 inhabitants of the city of Halle (Saale), aged 45–83 years, who participated in the population-based CARLA study. We calculated linear mixed models to assess the age-adjusted influence of neighbourhood SES (defined as neighbourhood-specific unemployment rates for 39 administrative districts of the city) and individual SES (defined as number of education years) on smoking (defined as number of currently smoked cigarettes/day), systolic blood pressure (SBP), and body mass index (BMI). Spatial dependencies within and between neighbourhoods were adjusted for by using ICAR models.
Results The unemployment rate ranged from 6.3 to 35.3% between neighbourhoods. For smoking, there was a statistically significant increase of 0.11 cigarettes smoked/day per 1% increase in the neighbourhood's unemployment rate in men (95% CI 0.09 to 0.12), and a decrease of 0.59 per increase in education years (CI −0.62 to −0.56), but a weaker association in women (regression coefficients (β) for unemployment rate and education years 0.054 (CI 0.039 to 0.067), and −0.21 (CI −0.24 to −0.19)). There was no statistically significant association of SBP with SES in men (β=−0.07 (CI −0.22 to 0.08) for unemployment rate, and −0.15 (CI −0.69 to 0.38) for education years), while in women, there was a statistically significant decrease in SBP of 0.79 mmHg per increase in education years (CI −0.82 to −0.76), and an increase with unemployment rate (β=0.04, CI 0.03 to 0.06). BMI was statistically significantly associated with education in men and women (0.11 decrease in BMI per increase in education years in men (CI −0.14 to −0.08), and 0.35 in women (CI −0.38 to −0.33)), but only for women with unemployment (increase in BMI per 1% increase in unemployment rate 0.008 (CI −0.008 to 0.02) in men, and 0.036 (CI −0.38 to −0.33) in women. Spatial correlations within and between neighbourhoods were small for all of the assessed outcomes.
Conclusions Our findings confirm the previously described association of neighbourhood SES with smoking independent of individual SES, while we found inconsistent associations with SBP and BMI. The neighbourhood environment may be more relevant for behavioural than for biomedical risk factors.