Table 3– Multiple linear regression coefficients for the effects of different socioeconomic factors on the years of exposure with symptoms before removal from exposure
VariableStrataA priori model 1A posteriori model 2A posteriori model 3
βp-valueβp-valueβp-value
Age>40 yrs versus ≤40 yrs (R)0.499±0.2310.0360.539±0.2080.2080.489±0.2000.018
Revenue<30,000 CAD versus ≥30,000 CAD (R)-0.585±0.2450.020-0.578±0.2030.203-0.433±0.2030.038
Dependent children≥1 versus none (R)0.357±0.2340.1330.371±0.2070.0780.376±0.1980.063
AgentHigh molecular weight versus low molecular weight (R)NI0.475±0.1960.0190.409±0.1890.036
Asthma severity#≥18% versus <18% (R)NINI0.515±0.2110.018
Adjusted r20.2010.3050.364
  • Data are presented as mean±sem, unless otherwise stated. The dependent variable was log (years in the workplace with symptoms prior to claim). We initially performed a multivariate linear regression by including socioeconomic variables, as defined a priori in our research hypotheses (a priori model 1). We then performed a multivariate linear regression using economic variables (a posteriori model 2) and asthma related variables (a posteriori model 3) with a significant relation to years of exposure without symptoms before removal from exposure in the univariate analysis (table 2). The covariates sex, immigration status, marital status and education status were included in a priori model 1 but are not displayed in the table as the p-values of the multiple linear regression coefficient β were >0.2. R: referent; NI: not included in the model. #: asthma severity was defined according to the Quebec Workers' Compensation Board's definition (0%: low severity; 100%: maximum severity).