SEEK Study - Study on Emergency Health Care Workers' Job Strain Evaluated by Karasek Questionnaire (SEEK)

Chronic stress at work is a public health problem increasing morbidity and mortality. Emergency physicians is particularly an at risk population. Indeed, their work is a complex interaction between stress due to life-threatening emergencies, the lack of sleep and accumulated fatigue.

A way of measuring stress at work is the use of self-reported psychological questionnaire. Among them, the Job control demand model from Karasek is widely used to evaluate psychosocial factors at work. The Karasek job strain model has been assessed in 25 000 workers in the Sumer study and allowed to classify the main types of occupations. However, no data regarding emergency physicians scores at the Karasek questionnaire were available, even though burnout exposure is a well-known problematic among health care workers. This absence is particularly unfortunate as emergency physicians often needs to deals with life threatening emergency in a short amount of time, following regularly long and craving shifts. A preliminary study with a low sample size reported in concrete terms the atypical "out of limit" situations of emergency physicians, with a decision latitude among the lowest and the highest psychological demand ever reported. Moreover, variation of stress at work has been linked with professional experiences. Karasek questionnaire is supposed to assess a long-term stable perception of work, despite the absence of study comparing Karasek scores within different conditions in the same participant. It is also important to explore protective factors that might help alleviate perceived stress and that are quite easy to integrate into medical organizations or in daily lifestyle.

Finally, several biomarkers of stress have been proposed. Among them, our team previously highlighted the potential of DHEAS to assess stress, due to its stability (steroid) and its long half-life of 16 hours. To our knowledge, the relationships between biomarkers of stress and the Karasek job strain model have never been investigated. Comparisons between emergency health care workers are also lacking.

Therefore, the general aim of this study is to enhance psychological well-being in emergency physicians. The primary objective is to verify our preliminary findings at Karasek scores on a large sample size of emergency physicians. The secondary objectives are to evaluate whether there is a change in perception of work both at short-term (on leave and after a 24-hour shift) and long-term (with professional experience). We will further explore its relationships with some biomarkers of stress and protective factors. Moreover, we aim to compare emergency health care workers specificity.

Questionnaires and saliva collection will be performed twice: at the beginning of a day work and at the end of a night work. Both measures will be recorded at the same time within each participant, between 6.30 am and 8.30 am depending of the health care workers. The time required by participant is evaluated to about 15min each time.

The same measures will be repeated every 5 year to assess long term changes in subjective and objective measures of stress.

The population description will be achieved with descriptive statistics. Continuous variables will be presented as mean and standard deviation, under the normality of distribution condition (Shapiro-Wilk needed).

The primary analysis will be mainly descriptive and will express (as a quantitative and categorical variables) the perception scores of emergency physicians work according to the Karasek questionnaire, for each dimensions composing the score (decision latitude, psychological demands and social support).

The evolution of the Karasek questionnaire score, obtained for each dimension, will be tested with a paired two-sample emergency physician's test (or nonparametric Wilcoxon test if necessary). Linear regression models or aleatory effects models (as described by Klar and Darlington, 2004, in Statistics in Medicine) will eventualy complete these analyzes. This will permit, with multivariate analyzes and the introduction of covariates, to seek for possible explanatory factors of the Karasek score and its evolution. The same process will be used for quantitative parameters that will be collected.

All statistical analyses will be performed as bilateral tests with a p <0.05 considered as significant. Statistical analyzes will be performed using STATA 16 software (Stata Corporation, College Station).