[Validation of the ICE model to assess social class in the adult population]

Social class has commonly been defined by the type of employment and it is assessed as a categorical variable. However, this approach has a number of drawbacks. The objective of this article is to develop and validate a readily standardizable quantitative indicator of social class and to show its ability to measure the impact of social class as a health determinant.

In 6729 individuals we measured income, crowding index, education, occupation and employment status. Two models were adjusted to study the neighborhood, dietary pattern and health problems.

The model that included only income, crowding index and education (ICE) yielded an indicator that correlated with age (r = -0.28; p < 0.001) and consumption of potatoes (r = -0.17; p < 0.001) and salads (r = 0.10; p < 0.001). This indicator estimated that poor social classes were at significant risk for unemployment (OR = 5,), blue collar jobs (OR = 40,9), residing in poor neighborhoods (OR = 30.2), low salad consumption (OR = 2.2) and high consumption of potatoes (OR = 4.5). They also had, especially in women, a higher risk of sedentarism (OR = 1.8), obesity (OR = 4.4), metabolic syndrome (OR = 3.4) and diabetes mellitus (OR = 2.0).

The ICE index was valid, not based on occupation or employment status, readily standardizable, and suitable for measuring social class and its impact of on health.