Comparison of Biomarkers of Stress in Emergency Physicians Working a 24-hour Shift or a 14-hour Night Shift - the JOBSTRESS Randomized Trial

A stressful state can lead to symptoms of mental exhaustion and physical fatigue, detachment from work, and feelings of diminished competence. One of the health consequences of chronic stress at work is an increased risk of coronary heart disease. Emergency physicians subjectively complain of stress related to changes in work shifts. In occupational medicine, no marker of stress or fatigue has achieved consensus. Several potential biomarkers of stress have been described, but never investigated in emergency physician, who may represent a good model of stress due to the complex interplay between stress (life-and-death emergencies, which is the defining characteristic of their job), lack of sleep and fatigue due to repeated changes in shifts.The aim of this study was to compare HRV in emergency physicians working a 24-hour shift (24hS) or a 14-hour night shift (14hS), and in those working a control day (clerical work on return from leave). We will also follow these markers three days following each shift (D3/24hS and D3/14hS).

The psychological consequences of shifts will be assessed in terms of perceived stress and fatigue using visual analog scales. Potential biomarkers will be assessed through urine and saliva collections. Heart rate variability will be measured using 5-lead electrocardiogram. Psychological questionnaires will be completed only once during the control day. The workload during each shift will be estimated by: the total number of entries, the number of admissions, the number of outpatients (collected by computer), and the number of life-and- death emergencies (given by the emergency physician). Sleep duration, including naps, will be assessed by questionnaire (bed time - wake time) on the three-day tracking of each shift and on the control day.

Shift randomization: Latin squares were used to randomize the pattern of shifts and control day (24hS then 14hS then control day or any other combination).

Gaussian distribution of the data wil be tested by the Kolmogorov-Smirnov test. Data will be presented as mean ± standard deviation (SD). Comparisons between shifts will be made with ANOVA. Relationships between data will be assessed by Pearson correlation. Significance will be accepted at the p<0.05 level. Statistical procedures will be performed using SPSS Advanced Statistics software (SPSS Inc., Chicago, IL).