Re: Effort-Reward Imbalance at Work and Incident Coronary Heart Disease.

Abstract

To the Editor: Recently, the Individual Patient Data (IPD)-Work consortium published results on the association between coronary heart disease (CHD) and work-related stress exposures defined as effort–reward imbalance (ERI) and job strain.1 The study found that effort–reward imbalance was a risk factor for CHD, independent of job strain, and that the risk associated with a joint exposure to effort–reward imbalance and job strain was even higher. However, the results are ambiguous and may be due to only one of the components of (efforts and rewards, or to only one of the components of job strain (demands and control), or possibly to only confounding from socioeconomic status (SES). The study found an association between rewards and CHD (hazard ratio [HR] = 1.18, 95% confidence interval [CI] = 1.04, 1.33) but no association between efforts and CHD (HR = 0.99, 95% CI = 0.87, 1.13), adjusted for age and sex. These results indicate that the effect of the effort–reward imbalance ratio (HR = 1.16, 95% CI = 1.00, 1.35) was attributable to the effect of only rewards and not to a combined effect of efforts and rewards. The separate effects of efforts and rewards were not adjusted for SES. The effect of rewards could very well be attributable to confounding from SES. In a previous study, the IPD-Work consortium found that job strain, adjusted for age and sex, was associated with CHD (HR = 1.23, 95% CI = 1.08, 1.39), so was control (HR = 1.16, 95% CI = 1.05, 1.27), but demands were not.2 Based on age- and sex-adjusted results, it was later shown that there was no multiplicative interaction between demands and control and that the effect of job strain could be explained by only control.3 Moreover, the effect of control was reduced to near unity (HR = 1.04, 95% CI = 0.94, 1.15) after further adjustment for SES. Thus, there seemed to be no combined or separate effects of demands and control when analyzed by a standard regression model specifying the main effects of demands and control, with or without their interaction term, adjusted for age, sex, and SES, contrary to the published result. Considered together, these results indicate that there may be no combined effects of efforts and rewards or of demands and control if analyzed in a standard regression model with adjustment for the main effects of the components of the effort–reward imbalance ratio and the job strain variable. The use of multiplicative interaction terms or ratios in regression analyses may lead to incorrect or misleading inferences if the main effects of the components are ignored.4 Sigurd MikkelsenDepartment of Occupational and Environmental MedicineBispebjerg University HospitalCopenhagen, Denmark[email protected] Johan Hviid AndersenDepartment of Occupational Medicine, Danish Ramazzini CentreRegional Hospital West Jutland–University Research ClinicHerning, Denmark Michael IngreIndependent ResearcherStockholm, Sweden