Sleep and health in shift working nurses

Abstract

In this thesis, the prevalence of Shift Work Disorder (SWD) in a sample of nurses, and the association between SWD and work characteristics, health outcome and personality (paper 1) have been investigated. Next, the psychometric properties of the “Bergen Shift Work Sleep Questionnaire” (BSWSQ) were evaluated (paper 2). The BSWSQ evaluates sleep and sleepiness problems in relation to day shifts, evening shifts, night shifts and rest-days. The validated BSWSQ was used to investigate how different work shifts and work schedules affected problems with sleep and sleepiness (paper 3). In paper 1, three symptom-based questions were used to assess the prevalence of symptoms indicative of SWD. To evaluate the accuracy of these questions, two additional measurement procedures were used for comparison. All three procedures were based on the International Classification of Sleep Disorders, second edition (ICSD-2) diagnostic criteria that entail the presence of a sleep or sleepiness problem associated with a work schedule that imposes wakefulness during the usual time for sleep (American Academy of Sleep Medicine (AASM), 2005). Logistic regression analyses were performed with symptoms indicative of SWD as outcome variables. Nurses (n = 1968) reported their age, gender, type of work schedule, commute duration, fraction of a full-time position, presence of children in household (yes/no), number of nights and number of shifts separated by less than 11 hours worked during the previous year, use of sleep medication, melatonin and bright light therapy. All of the nurses completed the Bergen Insomnia Scale, the Epworth Sleepiness Scale, the Global Sleep Assessment Questionnaire, the Diurnal Scale, the Revised Circadian Type Inventory, the Dispositional Resilience (Hardiness) Scale – Revised, the Fatigue Questionnaire, questions about alcohol and caffeine consumption, as well as the Hospital Anxiety and Depression Scale. The SWD prevalence varied among nurses depending on assessment procedure (32.4– 37.6%) and on type of work schedule (4.8–44.3%). SWD symptoms were reported among over one third of the nurses in the study sample with greater prevalence among those working night shifts and/or rotating shifts. Risk of SWD significantly increased with older age, higher languidity scores (circadian type), night work, more shifts worked in the previous year separated by less than 11 hours, more nights worked during the previous year, and with higher insomnia and anxiety scores. High scores of the circadian type of flexibility and being a female were associated with reduced risk of SWD. The three different SWD-assessment procedures provided similar results. The three symptom-based questions were considered adequate at assessing symptoms indicative of SWD for the purposes of epidemiological investigation. Paper 2 investigated the psychometric properties of the BSWSQ among 760 nurses working three-shift rotation schedules. The BSWSQ assesses seven symptoms including: >30 min sleep onset latency, >30 min wake after sleep onset, >30 min premature awakenings, non-restorative sleep, being sleepy either at work, during free periods on work days, or on rest-days. The factor structure was investigated using structural equation modeling (SEM) and the test-retest reliability by calculating intraclass correlation coefficients. Pearson-product moment correlation coefficients were calculated to test the convergent/divergent validity between the BSWSQ and the Epworth Sleepiness Scale, the Fatigue Questionnaire, and the Hospital Anxiety Depression Scale. Differences in insomnia symptom mean scores with respect to different shifts and rest-days were also investigated. BSWSQ demonstrated an adequate model fit using SEM, good test-retest reliability and good convergent/discriminant validity. Post hoc comparisons showed that working night shifts predicted higher overall/composite scores on the BSWSQ compared to day and evening shifts and rest-days. The BSWSQ was concluded to meet necessary psychometric standards and to successfully assess variations in sleep and sleepiness problems by type of shifts and rest-days. In paper 3, shift-related sleep and sleepiness problems among nurses working different work schedules were studied using the BSWSQ. To assess the severity of the sleep and sleepiness problems, shift-related insomnia was defined as often/always experiencing sleep and sleepiness problems (separately for rest days and day, evening and night shifts). Logistic regression analyses were used to assess the associations between shift-related insomnia (during day, evening and night shifts) and type of work schedule (i.e., permanent as opposed to rotational schedules). There were no differences between the work schedules in prevalence of day shift insomnia. Evening shift insomnia was more prevalent in two-shift (29.8%) compared to threeshift rotations (19.8%). Night shift insomnia was more common among three-shift rotation workers (67.7%) than among permanent night workers (41.7%). Rest-day insomnia was more prevalent among permanent night workers (11.4%) compared to workers with two- (4.2%) and three-shift rotations (3.6%). In paper 3, it was concluded that shift-related insomnia is most prevalent among nurses working three-shift rotations and night shifts. Nevertheless, sleep and sleepiness problems were present among nurses during all types of shifts and with all types of work schedules.