A reliability and validity study of the Bergen Shift Work Sleep Questionnaire in nurses working three-shift rotations.Chronobiol Int. 2012 Aug; 29(7):937-46.CI
The "Bergen Shift Work Sleep Questionnaire" (BSWSQ) was developed to systematically assess discrete sleep problems related to different work shifts (day, evening, night shifts) and rest days. In this study, we assessed the psychometric properties of the BSWSQ using a sample of 760 nurses, all working in a three-shift rotation schedule: day, evening, and night shifts. BSWSQ measures insomnia symptoms using seven questions: >30-min sleep onset latency, >30-min wake after sleep onset, >30-min premature awakenings, nonrestorative sleep, being tired/sleepy at work, during free time on work days, and when not working/on vacation. Symptoms are assessed separately for each work shift and rest days, as "never," "rarely," "sometimes," "often," "always," or "not applicable." We investigated the BSWSQ model fit, reliability (test-retest of a subsample, n = 234), and convergent and discriminant validity between the BSWSQ and Epworth Sleepiness Scale, Fatigue Questionnaire, and Hospital Anxiety Depression Scale. We also investigated differences in mean scores between the different insomnia symptoms with respect to different work shifts and rest days. BSWSQ demonstrated an adequate model fit using structural equation modeling: root mean square error of approximation = .071 (90% confidence interval [CI] = .066-.076), comparative fit index = .91, and chi-square/degrees of freedom = 4.41. The BSWSQ demonstrated good reliability (test-retest coefficients p < .001). We found good convergent and discriminant validity between BSWSQ and the other scales (all coefficients p < .001). There were significant differences between the overall/composite scores of the various work shifts. Night shift showed the highest score compared to day and evening shifts as well as to rest days (all post hoc comparisons p < .001). Mean scores of different symptoms also varied significantly within the individual work shifts. We conclude that the BSWSQ meets the necessary psychometric standards, enabling systematic study of discrete insomnia symptoms in different work shifts.