Background The aim of this study is to investigate the impact of job strain, as measured by the Karasek demand/control model (DCM), on smoking cessation and relapse in a representative general population sample.

Methods A secondary analysis of data from the Canadian National Population Health Survey (NPHS) was undertaken. Daily smokers and former daily smokers (n=1287 and 1184, respectively) at cycle 1 (1994/1995) of the NPHS were followed up at cycle 2 (1996/1997). Measures of job strain (the independent variables) were based on data from cycle 1, predicting smoking status at cycle 2. Logistic regression analysis was employed in two ways. Individuals were stratified into job strain quartiles while continuous measures were also employed in separate analyses for job strain and its component dimensions.

Results In the quartile analysis, no effect of job strain was observed on the likelihood of cessation, while a non-linear effect was observed on the likelihood of relapse, although this relationship lost significance (p>0.05 and <0.10) after controlling for personal characteristics. No effect was observed using the continuous measure of job strain or the continuous measure of job demand on either cessation or relapse. For job control, no effect was observed on the likelihood of cessation, but increased control was found to decrease the likelihood of relapse in the unadjusted model only.

Conclusions Psychosocial work environments may be too diverse for uniform trends in the relationship between job stress and smoking behaviour to emerge in a population sample. Future research should avoid use of the scaled-down DCM instrument where possible.