Skip to main content

Table 2 Effect modifications by gender, age and occupation for the prospective associations between occupational factors and the incidence of poor SRH among the study sample of 15,971 employees

From: Prospective associations between psychosocial work factors and self-reported health: study of effect modification by gender, age, and occupation using the national French working conditions survey data

  RR 95% CI p-value for interaction
Significant effect modifications by gender1
Job insecurity      0.019
Men 1.41 1.16 1.73  
Women 1.03 0.87 1.23  
Significant effect modifications by age (years)2
Temporary employment      0.002
< 30 0.67 0.45 1.01  
[30–40] 1.73 1.18 2.55  
[40–50] 1.23 0.82 1.85  
≥50 0.77 0.53 1.12  
Significant effect modifications by occupation3
High changes at work      0.039
Managers - professionals 1.52 1.15 2.01  
Associate professionals - technicians 1.07 0.87 1.33  
Clerks - service workers 1.60 1.33 1.93  
Blue-collar workers 1.26 0.98 1.61  
Shift work      0.006
Managers - professionals 0.12 0.03 0.48  
Associate professionals - technicians 1.07 0.75 1.54  
Clerks - service workers 1.19 0.87 1.64  
Blue-collar workers 0.79 0.60 1.04  
Biomechanical exposure      0.017
Managers - professionals 1.05 0.75 1.46  
Associate professionals - technicians 1.05 0.86 1.28  
Clerks - service workers 1.50 1.23 1.83  
Blue-collar workers 1.81 1.25 2.62  
Toxic/dangerous products      0.018
Managers - professionals 0.95 0.64 1.42  
Associate professionals - technicians 1.03 0.84 1.27  
Clerks - service workers 1.40 1.13 1.72  
Blue-collar workers 0.83 0.64 1.06  
  1. RR incidence rate ratio, CI confidence interval
  2. Each occupational factor was studied separately using discrete time Poisson regression models and weighted data
  3. 1Adjusted for age, marital status, life events, and occupation
  4. 2Adjusted for gender, marital status, life events, and occupation
  5. 3Adjusted for gender, age, marital status, and life events
  6. p > 0.05 after correction for multiple testing (FDR)
  7. 84 tests were done, 4 or 5 would be significant at 5% even if the null hypotheses were true, and 6 were found to be significant (before correction for multiple testing, FDR, and 0 after correction)