RESUMO
This work was originated from the concern for the availability of a short, acceptable and reliable instrument for self-assessment of perceived stress. The questionnaire was designed to be routinely applied within a visit at an Occupational Medical Center. A new questionnaire was developed with this aim, consisting of 9 subscales presented in a Lickert form, with 4 modes of answer, scored from 0 to 3. The content of the 9 items covers the multiple facets of perceived stress and its consequences: "feeling of being under pressure", "impatience", "irritability", "intrusive thoughts about work", "inability to entertain", "discouragement", "morning fatigue", "food compensation", "compensation by smoking". Stress global score is defined as the sum of the 9 elementary scores. The first 7 items are similar in their construct to the 7 factor solutions of the principal component factor analysis performed on Levenstein Perceived Stress Questionnaire (1993). On the other hand, in comparison with Cohen Perceived Stress Scale (1983), our instrument keeps an important place for affective, physiologic and behavioral impact of stressing situations. A study on the homogeneity of the scale, its factorial structure, and its time reproducibility after 6 weeks of interval, was carried out on a first population of 91 subjects seen during an occupational medical visit in several companies of Paris district (PCV-Metra group). The coefficient of internal consistency is very high (Cronbach's alpha = 0.82). Principal component analysis extracted two factors, which were unchanged after a Varimax rotation and respectively represented 42% and 13% of the total variance: they can be interpretated as a general perceived stress component (being overwhelmed, loss of control) and a behavioral bipolar component opposing food compensation to smoking, whilst facing stressing situations. Test-retest correlation coefficient is 0.88 (Pearson r as well as intraclass correlation coefficient), without any significant gap between the first and the second assessment. A second study was carried out on 761 working individuals seen in the same conditions during an occupational medical visit in the same companies (596 males and 65 females, aged 40.1 +/- 8.8 years). Socio-demographic data analysis showed higher stress scores in females (10.2 +/- 4.0), than in males (8.5 +/- 3.7) (p < 0.0001). Detailed analysis showed differences related to gender in the same direction for the items "irritability", "discouragement", "morning fatigue" and "food compensation". Highest stress scores, for both males and females, were found for the items "intrusive thoughts about work" and "morning fatigue". Stress global score was correlated with socioprofessional status (SPS): unskilled and skilled workers (n = 39) as well as technicians (n = 346) exhibited lower scores than clerical workers (n = 108) or engineers (n = 162) (ANOVA, p < 0.0001). The comparison between mean scores performed separately by gender, given the different sex-ratio according to SPS (employees were mostly females) confirmed the association between stress score and SPS only in males, with blue collars and technicians looking less under stress than engineers. Metrological properties of this Perceived Stress Questionnaire incite to perform studies focused on the associations between such a stress index measured in an occupational setting, and several other clinical or biological variables, especially those supposed to constitute classical cardiovascular risk factors.