The main aim of the present study was to investigate the effects of a BDLE compared with a STLE in an acute psychiatric unit at St. Olavs Hospital on shift working nurses’ sleep, mood, levels of stress, and caffeine use. To the best of our knowledge, this is the first study to examine how nurses experience working in a BDLE compared with a STLE. Overall, the results showed that most aspects of the nurses’ sleep and functioning were unchanged by exposure to the two light conditions. However, the nurses reported higher levels of sleepiness during evening shifts in the BDLE than in the STLE. In addition, nurses reported consuming a slightly higher number of caffeinated beverages during night shifts in the STLE than in the BDLE.
The fact that subjective sleepiness was higher during evening shifts in BDLE makes sense considering that lower levels of white light exposure during the evening (when it is becoming gradually darker outside), is associated with lower levels of melatonin suppression (45). Melatonin is a hormone that helps the body to know when it’s time to sleep and wake up, and melatonin suppression is associated with shifting of the circadian phase so that sleepiness and sleep occur later in the day (46,47). At least a partial circadian adaptation to night shifts might be desirable, since it could potentially increase job performance and reduce the risk of accidents at work or in commute, as well as improve daytime sleep during days off work when nurses must re-adapt to a daytime schedule (48–50). We did not find support for differing adaptation to shift work between the light conditions, but the lower levels of evening sleepiness is STLE likely reflects that white light has a direct activating effect, e.g., by increasing alertness (3,38,39). As such, increased sleepiness in BDLE may compromise the nurses’ safety, e.g., if they need to react instantly to adverse events. Further, it is comparably surprising that we did not find the same increase in sleepiness during night shifts in the BDLE. One explanation may be that the study does not have statistical power to detect e.g., small within-subject effects of BDLE compared with STLE. Another possible explanation may be the ceiling effect, in which the level of sleepiness during the night shift was high in both light conditions, and that the measurement of sleepiness used in this study is not well enough suited to distinguish between the nuances in the levels of sleepiness.
Our research group previously conducted a pilot study with 12 healthy adults using a randomized cross-over design comparing the effects of a BDLE with a STLE on e.g., sleep, subjective sleepiness, and the experience of side effects (30). In that study, the effects of the BDLE were generally positive. Participants did not report higher levels of sleepiness or negative side effects, the participants’ sleep-wake cycle was phase-advanced (i.e., higher levels of melatonin earlier at night) and they slept marginally longer (8.1 minutes) after residing in the BDLE. One explanation for our contradictory finding of increased sleepiness during evening shifts in BDLE could be the difference between assessments, i.e., hourly ratings of sleepiness from 19:00 to 23:00 versus global assessment of sleepiness during whole shifts retrospectively. Alternatively, differences in demands made on participants just residing in the building compared with nurses performing work-related tasks could further impact to what extent an individual experiences fatigue and sleepiness. Additionally, in the present study, we did not find significant differences on any sleep outcomes after working in BDLE compared with STLE, which might also be explained by the difference between assessments (i.e., polysomnography data versus diary- and actigraphy data), or that artificial light exposure when commuting home from work and at home before bedtime is sufficient to reverse subtle effects of a BDLE at work (51–53).
We also found a small 0.2 increase in number of caffeinated beverages consumed when undertaking night shifts in the STLE compared with the BDLE. While interesting, the available data do not allow us to determine whether this is best explained by differences in energy levels or perceived alertness during night shifts in different light conditions or individual fluctuation in caffeine intake (related to spurious factors), etc. Interestingly, we found no differences between BDLE and STLE on nurses’ reported levels of stress, positive or negative mood during shifts, or sleep patterns after each shift was completed. This was unexpected given that shift work is a known risk factor for e.g., poor sleep (5), medical or mental health problems (4), and impaired attention and alertness during waking hours (54). Some of the negative effects of shift work can generally be attributed to suboptimal shifting of the circadian phase (55). However, as light can be used to shift the circadian phase to better adapt to night work (56), we might expect BDLE, or ‘virtual darkness’ (57,58), to have a different effect on nurses’ sleep and functioning in a naturalistic setting than STLE. In many respects, the limited number of macro-level differences between the light conditions is encouraging as it indicates that a BDLE is not associated with major side effects or harmful effects. However, these findings need to be confirmed in further studies before BDLEs can be assumed to be beneficial to patients but not harmful to clinical staff working in inpatient units.
Limitations and future directions
There are some important limitations of the present study that should be mentioned. Due to the comparatively small sample size, we were limited in the types of analyses we were able to conduct. Further, we did not perform any correction for multiple comparisons (e.g., by adjusting for false discovery rate (59)) as this, with our limited sample size, would have increased the risk of false negatives. Failure to detect e.g., side effects of the light conditions would be potentially harmful to the nurses. A larger sample size is important for e.g., reliable, and meaningful multivariable analysis on the effects of a BDLE compared with a STLE on nurses’ sleep, health, and functioning or effects of switching between light conditions. A larger sample size would also facilitate analysis of whether the BDLE effects the nurses’ circadian rhythms directly and the impact of possible confounding factors (e.g., if patients exposed to BDLE were calmer and as such influenced the nurses). The within-subject design in a naturalistic setting is a strength of the present study but due to the low participation rate (29.1%), we cannot ascertain that our findings are representative of all nurses working at the acute psychiatric unit at St. Olavs Hospital. Although other single-site studies in small workplaces (less than ~100 employees) also necessarily will be bound by an upper limit of available participants, ensuring high participation rates would make sure that any drawn conclusions will be representative of all employees. Additionally, use of employer- or registry data in futures studies, as opposed to self-report, could both ease time demands on the participants and be a source of objective, high quality information on how shift working nurses are affected by a BDLE. Such sources could be used to collect information on e.g., sickness absence, other types of leave, health care resource use, and medical or mental health diagnoses. They will not, however, be suitable to investigate individual experiences of day-to-day life, and important information on e.g., personal experiences in and of the work environment, levels of presenteeism (i.e., reduced productivity while at work), impairment in general activities outside of work, or subclinical symptoms of medical or mental health conditions.