Sleepiness among personnel in the Norwegian Air Ambulance Service

To maintain around the clock operations, a large portion of the workforce must be organized in shift work and often works extended hours. Shift work can generally be denoted as work outside normal daytime, and may include work during evening, night, and weekend (National Sleep Foundation 2018). Shift work, including night shift, implies activity and light exposure at a time when inactivity and rest are natural (Akerstedt and Wright 2009). Work during the natural night often influence sleep and functioning and may result in shortened sleep and/or increased sleepiness (Swanson et al. 2011). Studies generally suggest a reduction of sleep duration by approximately 3 h following night shifts (Sallinen and Kecklund 2010), and severe sleepiness has been found during night shift, as this coincides with the nadir for the core body temperature (Härmä et al. 2002; Akerstedt and Wright 2009). Night work is, therefore, associated with impaired performance and occupational accidents (Härmä and Kecklund 2010; Wagstaff and Lie 2011). In addition, working irregular shifts poses a risk for increased sleepiness and sleep debt, indicating an elevated need for recovery (Härmä et al. 2018). Extended working hours have been defined as work exceeding 48 h a week (Harrington 2001). The literature regarding the effect of such working hours on sleep and sleepiness is inconsistent. Some studies suggest that long working hours are associated with increased sleepiness and shortened sleep (Sallinen and Kecklund 2010; Swanson et al. 2011), while other studies conclude that workers adapt quite well (Bjorvatn et al. 2006; Forberg et al. 2010). A review article concluded that extended working hours increased the risk of excessive sleepiness, which further could have implications for performance on the job and injuries (Caruso 2014). Shift work and extended work hours are common in aviation. Many flight operations include long-haul flights and crossing of several time zones. However, these work conditions are quite different from those in the Helicopter Emergency Medical Services (HEMS) characterized by on-demand medical service implying low predictability of workload and work hours. The work schedule and conditions of HEMS crews vary a great deal, both between countries and agencies. Some work 12-, 24-, or 48-h shifts, while others work 7 consecutive days. Furthermore, some live on the base, while others commute home every day when off work (Radstaak et al. 2014; Sallinen et al. 2018). These differences complicate generalization of findings across various operations, schedules, and conditions. In Norway, HEMS crews work 24-h shifts over 7 consecutive days to provide around the clock coverage and operation. However, little is known about the effects this type of shift schedules and these work characteristics have on sleepiness. Sleepiness is established as a risk factor in aviation, and evidence suggests that sleepiness is associated with flying errors and accidents (Goode 2003; Previc et al. 2009). A common distinction is drawn between subjective and objective sleepiness. Objective measures of sleepiness may include measures of sleep tendency (i.e., the Multiple Sleep Latency Test), reduced activation (i.e., pupillometry), or performance deficits (i.e., reaction time). Subjective sleepiness can be subdivided into state measures, which are sensitive to abrupt changes caused, for example, by sleep deprivation, and trait measures, which primarily measures the respondents’ general tendency to experience sleepiness. There are few scientific publications concerning sleepiness in the HEMS, and the few existing findings are ambiguous. Furthermore, the quality of the existing research is often limited due to low response rate (Müller et al. 2014), and low external validity caused by variations in shift systems across countries and agencies (Guyette et al. 2013). In addition, there seems to exist a discrepancy between results resting on measures of subjective sleepiness and those reflecting performance and objective measures of sleepiness (Müller et al. 2014; van Dongen et al. 2003). Furthermore, an appraisal of the fluctuation in sleepiness is important and could have implications for safety and health, for both personnel and patients. Consequently, knowledge about the effects of working long hours and consecutive shifts on sleepiness of pilots and HEMS crew members is thus essential. This current field study was conducted among pilots and HEMS crew members working for the Norwegian Air Ambulance (NAA). The NAA operated 9 out of 12 bases in Norway, including 10 of total 13 helicopters. The specific aim of the present study was to examine sleepiness in pilots and HEMS crew members before, during, and after a work week by administering subjective as well as objective measures of sleepiness. We hypothesized that sleepiness scores (measured with the Accumulated Time with Sleepiness Scale) would be higher during the work week, compared to the week before and after work (H1). Considering the work week, we expected that sleepiness (measured with the Accumulated Time with Sleepiness Scale, the Karolinska Sleepiness Scale, reaction time, and response accuracy) would increase over the work week (H2). We also hypothesized that during a day, the crew would report more sleepiness (measured with the Karolinska Sleepiness Scale) at the end of the wake period (H3). Finally, we expected that the crew members with larger amount of work (number of missions/training sessions and total work time) on duty would report higher sleepiness scores (measured with the Accumulated Time with Sleepiness Scale and the Karolinska Sleepiness Scale) compared to those with medium work amount (H4).

The sleepiness scores measured by the ATS were lowest in the work week, compared to the weeks at home before and after work. During the work week, the highest sleepiness scores measured by the KSS were reported on the first day of work. However, there was no change during the week in terms of reaction time and response accuracy. Over the course of the day, the highest sleepiness scores measured by the KSS were reported at midnight. Having higher workload was associated with lower sleepiness measured by the KSS compared to medium workload, whereas having longer TWT was associated with higher sleepiness scores measured by the KSS, compared to a medium TWT. The crew members felt less sleepy during their work week, compared to their weeks off (both before and after). Consequently, there was no support for the first hypothesis stating that sleepiness scores would be higher during the work week, compared to the week before and the week after work. We expected opposite findings in line with other studies (Mullins et al. 2014; Akerstedt and Wright 2009). However, the crew in the present study was living at the base during the work week which relieves them of social and domestic obligations. With available base facilities such as separate bedrooms, an exercise room and a living room, the crew likely gets sufficient rest and leisure between the missions and training sessions. This may explain the lower sleepiness scores. Excessive sleepiness during free days has been found in other studies among shift workers (Härmä et al. 2018), suggesting that accumulated sleep deprivation during the work period may become manifest on days off work. However, these studies are only partially comparable to the present due to differences in sample population and work schedule predictability. In addition, the current study’s sample had a minimum of 14-day off between the work weeks. This means that the crew already had at least 1-week off work before the first assessment week in this study. Still, higher sleepiness score compared to the work week was found, although overall sleepiness levels at all 3 weeks were low. This indicates that the slightly higher sleepiness scores in the first week, stems from sources other than work. When at home, the workers have domestic obligations, including children to take care of which could be a possible factor explaining these scores. In compliance with this, Gregory et al. (2010) found that 26% of air medical pilots reported child care as a factor that affected the ability to sleep. Furthermore, some of the crew holds second jobs during their weeks off, which could explain why the sleepiness scores are slightly higher the weeks off duty. Nevertheless, it is important to emphasize that the overall sleepiness scores across all 3 weeks were low considering that the ATS scale range from 0 to 100. Hence, although higher, the sleepiness the week before and after work was not deemed clinically elevated. A comparison between the pilots and the HEMS crew members revealed a somewhat higher score for the pilots on the “difficulty in focusing your eyes” item. As the HEMS crew members are more likely to have a higher workload due to accompanying on the rapid response car, this result seems thus reasonably. However, both the scores were low indicating that neither pilots nor HEMS crew members experienced much sleepiness across the 3 weeks. The crew reported the highest sleepiness scores on the first day at work, compared to the following 6 duty days. Furthermore, the reaction time tests did not change over the course of the work week. Given these results, the second hypothesis must be rejected, postulating that subjective and objective measures of sleepiness would increase during the work week. As the hypothesis suggested, one would expect that the crew members became sleepier over the course of the work week, due to accumulated sleep deprivation caused by shift work and the work load itself (Akerstedt and Kecklund 2005). However, there are other studies, indicating that the workers adapt to shift work during the work period. Bjorvatn et al. (2006) found a decrease in sleepiness scores, both subjective and objective, over a week of night shift offshore. However, these results could be explained by a shift in the circadian rhythm due to the week of night work. Based on the mission log in the present study, it is evident that most of the missions took place during daytime and a circadian alteration is not likely, although this should be investigated in future studies. Despite the fact that the study of Bjorvatn and colleagues comprised oil rig workers who worked a week of night shift followed by a week of day shift, that occupational group still has some similarities to our sample, such as work facilities. The offshore workers live on the oil rig during their work period, and they work shift and have extended work hours. These results could suggest that the work facilities affect sleepiness levels in a positive way during the work period, despite having work schedules that often have been reported to impact sleepiness negatively. One possible explanation for the higher sleepiness scores on the first duty day could also be related to commuting. The majority of the crew reported using a car as a commuter and using 1–6 h to commute. Furthermore, almost all commuted the same day as the shift started. As the shift started at 10.00 in the morning and having up to 6 h of commute by car on the same day, it would imply that the workers needed to wake up very early at the first work day. For this reason, commute length could explain why the workers display higher sleepiness scores on the first duty day. This should receive some attention, as subjective sleepiness is associated with an increase in automobile accidents (Bioulac et al. 2017). Another possible explanation to our findings is that phase delay due to late bedtime and rise time might have occurred during the preceding weekend (Yang et al. 2001). Nevertheless, the subjective sleepiness scores were all distributed on the lower part of the scale; thus, statistically significant results must be interpreted with prudence. The practical meaning of this result should also be considered with caution, as the difference from the remaining duty days was small. The crew members showed no evidence of increased sleepiness over the course of the work week, as measured with reaction time tests. Interestingly, there was no increase on the first duty day despite having higher subjective sleepiness scores. In accordance with other studies, this could suggest that alertness is maintained, by keeping low response time and high response accuracy, despite reporting subjective sleepiness (Cullip et al. 2014; Thomas et al. 2006). The previous studies report a decrease in performance due to sleepiness (Myers et al. 2017), while others report no difference in sleepiness and/or performance despite working long shifts (Amann et al. 2014; Guyette et al. 2013). The results from the current study provide support for the latter findings. However, it is worth mentioning that the test could activate the participants by being a distinct new component in their work environment and explain the lack of change in reaction time and response accuracy. Over the course of a duty day, the crew members did report changes in levels of sleepiness. The highest sleepiness scores were not surprisingly reported at midnight, significantly different from the other timepoints of the day. Therefore, the third hypothesis, stating that crew would experience increased sleepiness at the end of the wake period, was supported. Still, the average sleepiness scores were low and distributed between the ‘very alert’ and ‘neither sleepy nor alert’ step of the scale. The distribution of the sleepiness scores resembles an oscillation in sleepiness that follows the circadian rhythm, rather than sleepiness due to work schedule (Borbély et al. 2016). The previous studies on shift work have indicated that night work and morning work are associated with sleepiness during the day, and that rotational work, rather than fixed work, is associated with higher sleepiness (Thun et al. 2016). Furthermore, a study on air plane operations found that sleepiness levels increased after flight duty (Yen et al. 2009). The work characteristics of the air ambulance service involve both night work and early morning work, often in a rotational manner, which make the present study relevant. An interpretation of the present result is that the work schedule did not affect the sleepiness score over the course of a duty day. Two variables were created based on the mission log: workload (total number of missions and training sessions during the work week), and TWT (total time in hours and minutes spent actively working during the 7-day shift). Both were made categorical and based on tertiles. Having higher workload was associated with lower sleepiness scores compared to having medium workload. In contrast, having higher TWT was associated with higher sleepiness scores compared to medium TWT. This gives partial support for the fourth hypothesis, postulating that the crew members with larger amount of work (both workload and TWT) would have higher sleepiness scores compared to those with medium work amount. These results may indicate that the activation related to more missions reduced sleepiness levels, while the activation related to longer missions did not. This is in accordance with other studies, where higher sleepiness occurred on the longest missions during a day (Amann et al. 2014; Powell et al. 2008). Nevertheless, again, the sleepiness was overall low, indicating that the crew and pilots were sufficiently alert.

Overall, the results suggest that the crew members experienced low levels of state sleepiness despite having 7-day shift on the base and unpredictable working hours. Furthermore, the trait measure of sleepiness revealed low levels (ESS = 7.1), compared to the Norwegian male population, where the mean ESS score is 7.4 (Pallesen et al. 2007). The results obtained on the state measures of sleepiness were all within the non-pathological/non-problematic level. This is in agreement with studies from other fields that share similar work schedules and arrangements, including offshore workers, tunnel workers, and construction workers (Bjorvatn et al. 2006; Forberg et al. 2010; Persson et al. 2006).

The present study revealed that the overall sleepiness scores were low during all 3 test weeks. When comparing the 3 weeks, the lowest sleepiness levels were found for the work week. There was a small difference across work days, in which subjective sleepiness scores were highest the first duty day. No change in the reaction time tests was evident during the work week. The crew members reported being most sleepy at midnight, compared to all the other timepoints over the course of a duty day. Regarding workload and TWT, having larger workload was associated with lower sleepiness scores, while having higher TWT was associated with higher sleepiness score, both compared to the medium category. To the author’s knowledge, this study is unique in being one of the first field studies on this occupational group that included both subjective and objective measures of sleepiness over an extended period. Overall, our findings indicate that the work setting and schedules for this particular occupational group do not seem to negatively affect the sleepiness levels.