Study of the validity of a job-exposure matrix for psychosocial work factors: results from the national French SUMER survey

Objective To construct and evaluate the validity of a job-exposure matrix (JEM) for psychosocial work factors defined by Karasek's model using national representative data of the French working population. Methods National sample of 24,486 men and women who filled in the Job Content Questionnaire (JCQ) by Karasek measuring the scores of psychological demands, decision latitude, and social support (individual scores) in 2003 (response rate 96.5%). Median values of the three scores in the total sample of men and women were used to define high demands, low latitude, and low support (individual binary exposures). Job title was defined by both occupation and economic activity that were coded using detailed national classifications (PCS and NAF/NACE). Two JEM measures were calculated from the individual scores of demands, latitude and support for each job title: JEM scores (mean of the individual score) and JEM binary exposures (JEM score dichotomized at the median). Results The analysis of the variance of the individual scores of demands, latitude, and support explained by occupations and economic activities, of the correlation and agreement between individual measures and JEM measures, and of the sensitivity and specificity of JEM exposures, as well as the study of the associations with self-reported health showed a low validity of JEM measures for psychological demands and social support, and a relatively higher validity for decision latitude compared with individual measures. Conclusion Job-exposure matrix measure for decision latitude might be used as a complementary exposure assessment. Further research is needed to evaluate the validity of JEM for psychosocial work factors.