Objective and subjective measures of sleepiness, and their associations with on-road driving events in shift workers

Authors

  • SUZANNE FTOUNI,

    1. School of Psychology and Psychiatry, Monash University, Clayton, Vic., Australia
    Search for more papers by this author
  • TRACEY L. SLETTEN,

    1. School of Psychology and Psychiatry, Monash University, Clayton, Vic., Australia
    Search for more papers by this author
  • MARK HOWARD,

    1. Institute for Breathing and Sleep, Austin Hospital, Heidelberg, Vic., Australia
    Search for more papers by this author
  • CLARE ANDERSON,

    1. School of Psychology and Psychiatry, Monash University, Clayton, Vic., Australia
    2. Division of Sleep Medicine, Department of Medicine, Brigham and Women’s Hospital, Boston, MA, USA
    3. Division of Sleep Medicine, Department of Medicine, Harvard Medical School, Boston, MA, USA
    Search for more papers by this author
  • MICHAEL G. LENNÉ,

    1. Monash University Accident Research Centre, Monash University, Clayton, Vic., Australia
    Search for more papers by this author
  • STEVEN W. LOCKLEY,

    1. School of Psychology and Psychiatry, Monash University, Clayton, Vic., Australia
    2. Division of Sleep Medicine, Department of Medicine, Brigham and Women’s Hospital, Boston, MA, USA
    3. Division of Sleep Medicine, Department of Medicine, Harvard Medical School, Boston, MA, USA
    Search for more papers by this author
  • SHANTHA M. W. RAJARATNAM

    1. School of Psychology and Psychiatry, Monash University, Clayton, Vic., Australia
    2. Division of Sleep Medicine, Department of Medicine, Brigham and Women’s Hospital, Boston, MA, USA
    3. Division of Sleep Medicine, Department of Medicine, Harvard Medical School, Boston, MA, USA
    Search for more papers by this author

Shantha Rajaratnam, PhD, School of Psychology and Psychiatry, Monash University, Building 17, Wellington Road, Clayton, 3800 Vic., Australia.
Tel.: +61 3 9905 3934;
fax: +61 3 9905 3948;
e-mail: shantha.rajaratnam@monash.edu

Summary

To assess the relationships between sleepiness and the incidence of adverse driving events in nurses commuting to and from night and rotating shifts, 27 rotating and permanent night shift-working nurses were asked to complete daily sleep and duty logs, and wear wrist-activity monitors for 2 weeks (369 driving sessions). During all commutes, ocular measures of drowsiness, including the Johns Drowsiness Scale score, were assessed using the Optalert™ system. Participants self-reported their subjective sleepiness at the beginning and end of each drive, and any events that occurred during the drive. Rotating shift nurses reported higher levels of sleepiness compared with permanent night shift nurses. In both shift-working groups, self-reported sleepiness, drowsiness and drive events were significantly higher during commutes following night shifts compared with commutes before night shifts. Strong associations were found between objective drowsiness and increased odds of driving events during commutes following night shifts. Maximum total blink duration (mean = 7.96 s) during the drive and pre-drive Karolinska Sleepiness Scale (mean = 5.0) were associated with greater incidence of sleep-related events [OR, 5.35 (95% CI, 1.32, 21.60), OR, 1.69 (95% CI, 1.04, 2.73), respectively]. Inattention was strongly associated with a Johns Drowsiness Scale score equal to or above 4.5 [OR, 4.58 (95% CI, 1.26–16.69)]. Hazardous driving events were more likely to occur when drivers had been awake for 16 h or more [OR, 4.50 (95% CI, 1.81, 11.16)]. Under real-world driving conditions, shift-working nurses experience high levels of drowsiness as indicated by ocular measures, which are associated with impaired driving performance following night shift work.

Introduction

Approximately 15–20% of motor vehicle crashes (MVC) are attributed to sleepiness and fatigue in the USA and Australia (Austroads, 2005; Teff, 2010), making it one of the most common preventable causes of death on the roads (Dobbie, 2002; Maclean et al., 2003). Shift workers are overrepresented in sleepiness-related MVCs (Crummy et al., 2008), which often occur during the commute from work following the night shift (Barger et al., 2005; Crummy et al., 2008).

Sleepiness-related MVCs result from a critical combination of sleep deficiency and circadian misalignment (Arendt, 2010; Dijk and Czeisler, 1994). Night shift workers are particularly vulnerable to sleepiness, particularly towards the end of their shift, due to the interaction of circadian (approximately 24 h) and homeostatic factors, which combine in a multiplicative manner. In addition, night shift work is often associated with chronic sleep deficiency and disruption (Ohayon et al., 2002), which further exacerbates cognitive impairment (Cohen et al., 2010), including those skills required to operate a motor vehicle safely (Barger et al., 2005; Lyznicki et al., 1998; Philip et al., 2003). Sleepiness impairs multiple neurobehavioral domains, including sustained attention or vigilance, and also speed and accuracy, working and short-term memory, and reaction time (Alhola and Polo-Kantola, 2007; Bartel et al., 2004; Dinges et al., 1997; Philip et al., 2003) to a degree comparable with alcohol intoxication (Dawson and Reid, 1997). Sleep deficiency increases instances of lane drifting, slows reaction time to on-road events, and increases unintentional changes in speed and jerking motions of the wheel (Lennéet al., 1998; Philip et al., 2005). Consequently, sleepiness in night shift workers is a major risk factor for MVCs (Horne and Reyner, 1995; Lyznicki et al., 1998; Philip et al., 2005), as drivers exhibit more micro-sleeps, periods of inattention, longer and more frequent eye closures, and increased likelihood of falling asleep at the wheel (Akerstedt et al., 2005; Lyznicki et al., 1998).

The accuracy of subjective assessments of sleepiness is an important factor in the prevention of sleepiness-related MVCs and the development of effective intervention strategies. While increases in subjective sleepiness appear to be associated with more lateral lane deviations and increased number of adverse incidents while driving (Akerstedt et al., 2005; Ingre et al., 2006), findings are inconsistent (Franzen et al., 2008; Tremaine et al., 2010; Van Dongen et al., 2003), with objective levels of impairment often underestimated.

Ocular movements, in particular eye blink parameters such as blink duration, blink reopening time and long eye closure duration, are considered as indicators of drowsiness level (Caffier et al., 2003; Häkkänen et al., 1999). Blinks associated with the onset of sleep or micro-sleeps, often experienced by sleep-deprived individuals, display unique properties (Stern et al., 1984). In a well-rested state the duration of the normal closing and reopening phases of a blink are approximately 150 and 100 ms in duration, respectively (Stern et al., 1984). In a sleep-deprived state, the duration of the closing phase can be >250 ms, the reopening phase could take between 100 and 150 ms, and the duration of eye closure after the closing phase may range from 250 ms to several seconds (Stern et al., 1984). Physiological indicators of sleepiness, such as blink duration and percentage of eyelid closure over the pupil over time (PERCLOS; Bergasa et al., 2006; Singh and Papanikolopoulos, 1999), as well as electroencephalogram (EEG)-derived correlates of sleepiness (Lal and Craig, 2001), have been used as predictors of driving performance (Akerstedt et al., 2005). Crash rate and variability in lane position increase with blink duration (Akerstedt et al., 2005; Anund et al., 2009), particularly following a night shift compared with after a normal night of sleep (Akerstedt et al., 2005). Adverse driving performance is associated with increases in EEG power (4–11 Hz; Horne and Baulk, 2004), eye movement frequency and blink amplitude (Lal and Craig, 2002), and PERCLOS and blink duration (Golz et al., 2010).

While previous studies examining the safety risks associated with shift work have often focused on work-place accidents and injuries, the commute from work is also associated with significant risk (Barger et al., 2005). Few studies to date have assessed drowsiness objectively and continuously in shift workers in a naturalistic setting during driving commutes. The present study aimed to assess objective levels of drowsiness experienced during commutes from night shift work in hospital nurses, in comparison with those experienced during commutes to the night shift, and to examine the associations between objective and subjective measures of drowsiness and self-reported adverse driving events.

Materials and methods

Participants

Twenty-seven healthy, registered nurses were recruited from 11 metropolitan hospitals in Melbourne, Australia, through online newsletters, posting on communal hospital pin boards, and visits to hospitals (Fig. 1). All participants provided written informed consent. The protocol was approved by the Monash University Human Research Ethics Committee and hospital ethical committees. Participants were paid AU$200 for participation.

Figure 1.

 Flow chart of participation in the study.

Participant demographic and driving characteristics are presented in Table 1. Nurses were employed in the following departments: maternity (n = 7, 26%), emergency (n = 6, 22%), pediatrics (n = 4, 15%), oncology (n = 3, 11%), surgery (n = 2, 7%), respiratory care (n = 2, 7%), neuroscience (n = 1, 4%), spinal (n = 1, 4%) and intensive care (n = 1, 4%). All participants were low risk for obstructive sleep apnea based on the Berlin questionnaire (Netzer et al., 1999), although one participant was high risk according to the multivariable apnea risk index (Maislin et al., 1995).

Table 1. Demographic and driving characteristics (n = 27)
  OverallRSWPNSW
  1. PNSW, permanent night shift workers; RSW, rotating shift workers.

  2. *Significant differences between rotating and permanent shift workers (< 0.01).

N n (%)2712 (44.4)15 (55.5)
Males n (%)4 (15)3 (25.0)1 (6.7)
Age (years)*Mean ± SD41.6 ± 12.533.89 ± 12.447.71 ± 8.8
Body mass index (kg m−2)Mean ± SD26.2 ± 4.824.66 ± 4.427.38 ± 4.9
Body mass index ≥30 kg m−2 n (%)6 (22.2)1 (8.3)5 (33.3)
Driving experience (years)Mean ± SD22.9 ± 11.416.40 ± 13.028.12 ± 6.6
Kilometers driven per year*Mean ± SD26 351 ± 12 08429 350 ± 12 01323 953 ± 11 998
Shift work experience (years)Mean ± SD10.6 ± 10.313.26 ± 11.58.43 ± 9.1
Nursing experience (years)Mean ± SD15.5 ± 13.014.70 ± 13.716.13 ± 12.9
Epworth Sleepiness ScoreMean ± SD7.4 ± 3.56.92 ± 3.87.73 ± 3.4

Participants were required to work a minimum of three consecutive night shifts in at least 1 week, with at least three consecutive days off or 3 day or evening shifts preceding night shift work. Night shifts were defined as shifts with at least 6 h on duty between 22:00 and 08:00 hours, and with a maximum scheduled shift length of 12 h. A minimum commute time of 30 min between the hospital and the participant’s home was required.

All nurses worked night shifts scheduled from 21:00 to 07:30 hours. Rotating shift nurses worked day shifts from approximately 07:00 to 15:30 hours, and evening shifts from 13:00 to 21:30 hours.

Procedure

Participants completed questionnaires assessing demographics, employment, shift work schedules, driving experience, medication and sleep history. Data were collected for 2 weeks based on each participant’s shift schedule, such that data collection included at least three consecutive days off or day shifts followed by at least three consecutive night shifts. Participants were instructed to maintain normal sleep patterns and work behavior during this period, with no restrictions on caffeine consumption, napping or other behaviors.

Measurements

Actigraphy, sleep log and duty log

Wrist actigraphy (Actiwatch-L; Respironics, Bend, OR, USA) was recorded continuously to assess rest–activity patterns (Cole et al., 1992). A self-report sleep log was also completed each day to record sleep and wake times, sleep quality (rated on a visual scale from 0 = ‘worse sleep for a long time’ to 10 = ‘best sleep for a long time’), and time and duration of any naps taken. At the end of each work shift, participants reported their scheduled and actual work hours in a duty log developed for the study.

Driving log

A self-report driving log was developed for the study, in which participants recorded the start and end times of each commute to or from work (pre-shift and post-shift drives, respectively), subjective sleepiness on the modified, 10-point Karolinska Sleepiness Scale (KSS; 1 = ‘extremely alert’ to 10 = ‘extremely sleepy, cannot keep awake’; Barrett et al., 2004) immediately prior to and following each drive (pre-drive and post-drive, respectively), self-reported traffic density (0 = ‘virtually no traffic’ to 12 = ‘heavy traffic’), complexity/demand of the drive (0 = ‘no more complex/demanding’ to 8 = ‘very complex/demanding’), road surface conditions during the commute, and driving events (near-misses, crashes, falling asleep at a stop light, missing a turn, hitting the rumble strips, driving through a stop light, braked sharply, lack of awareness, swerved violently, shouting at another person, resting your eyes, being distracted, pulled over for a nap, fixation on a internal/external object). Participants were also asked to report retrospectively any consumption of alcohol, caffeine or medication in the 12 h prior to each drive.

Oculography

Participants’ eye and eyelid movements were monitored by infrared reflectance (IR) oculography (Optalert™, Melbourne, Australia; Johns et al., 2007) to record drowsiness levels continuously during commutes. Optalert™ is based on the principle that while people are drowsy the muscle groups controlling eye and eyelid movements are inhibited by the central nervous system (Johns et al., 2007; Stern et al., 1994). IR transducers fitted inside spectacle frames are positioned towards the eye to measure the relative velocity of the opening and closing of the eyelid and blink durations. A combination of oculometric variables are used to calculate a driver’s level of drowsiness in real time, providing a minute-to-minute Johns Drowsiness Scale (JDS) rating (see Johns et al., 2007). JDS is a continuous scale with scores ranging from 0 to 10 (very alert to very drowsy, respectively). The commercially available system is designed to emit auditory warnings when drivers reach a JDS score of 4.5–4.9 (cautionary level of drowsiness), and a score of 5.0 or above (critical level of drowsiness), associated with an increased risk of severe lane excursions on a driving simulator (Johns et al., 2007; Stephan et al., 2006). These warnings were disabled in the present research study, such that the system was used purely to monitor the drowsy state. The Optalert™ system also provides measurements of other ocular variables, including: percentage of long eye closures (when the eyes are closed for longer than 10 ms in each recorded minute); the percentage of time that the eyes are deemed closed in each minute; and mean total blink durations (duration of the closing, closed and reopening phases of each blink) in each minute.

Participants were asked to wear the Optalert™ glasses during each commute to and from all work shifts, allowing up to a 5-min baseline recording before commencing each drive.

Data analysis

Sleep/wake times were determined from sleep logs and checked for accuracy against actigraphic estimates of sleep onset and offset times (Actiware; Philips Respironics). To determine mean sleep/wake times during night shifts, only shifts that were preceded or followed by a scheduled night shift were included in the analysis to ensure sleep/wake times were reflective of sleep habits during night shifts only. The same process was applied to calculate sleep/wake times during days off, and during day and evening shifts, with either a day off or respective shift type preceding the shifts included in the analysis. Within-subjects analysis of variance (anova) was used to examine differences in sleep/wake times in nurses working rotating shifts and permanent night shifts. anova was used to examine differences in sleep/wake times, total sleep time, sleep quality and sleep efficiency between rotating and permanent night shift worker groups, and also differences in objective and subjective measures of drowsiness and sleepiness between groups. KSS scores were dichotomized to assess the rates of self-reported sleepiness levels above 5 (‘some signs of sleepiness’ to ‘extremely sleepy, cannot keep awake’). JDS scores were dichotomized to assess the incidence of scores at or above 4.5 (cautionary level). The number of hours awake before starting a drive was dichotomized to assess the incidence of cumulative time awake >16 h. Chi square tests were used to assess differences in dichotomized variables.

Drive event variables were categorized into four event types (sleep-related, inattention, hazardous and violations; adapted from Reason et al., 1990). The Mantel–Haenszel chi square test using a within-subject comparison (participant as covariate) was used to examine the odds of reporting events in pre-shift compared with post-shift drives. The case-crossover study design eliminated the need to account for potential confounders (Barger et al., 2005). Chi square tests were used to assess differences in reported event rates between rotating and permanent night shift workers. To calculate odds ratios (OR) of reporting categorized events with objective levels of drowsiness and subjective sleepiness [hours awake since last major sleep, cumulative hours awake ≥16 h, duration of prior sleep, pre- and post-drive KSS score, pre- and post-drive KSS score ≥5, mean JDS score, maximum JDS score, JDS scores ≥4.5, cumulative JDS score during drive, JDS score change from baseline (pre-shift JDS score), mean percentage of long eye closures, maximum percentage of long eye closures, mean percentage of time with eyes closed, maximum percentage of time with eyes closed, mean total blink duration and maximum total blink duration], logistic regression was conducted for each category, controlling for sex, age, body mass index (BMI), drive duration, shift type and cycle type. All data are reported as mean ± SD. ORs are reported with 95% confidence intervals (CI). ORs for continuous variables indicate the change in odds for an increase of one standard deviation. A P-value of <0.05 (two-sided) was considered statistically significant. Statistical analysis was conducted with SPSS version 18.0 (SPSS, Chicago, IL, USA).

Results

Compliance

A total of 192 shifts was reported in the duty logs, with subjective data provided for 187 pre-shift drives and 182 post-shift drives. Of these drives, 64.7% of pre-shift and 63.7% of post-shift drives also had Optalert™ recorded successfully. Of the 393 study days, 365 (92.9%) days of sleep logs were available and used in the analysis, and 344 (87.5%) days of actigraphy data were available and used in the analysis.

Sleep

Permanent night shift workers slept significantly less following nights shifts (5.6 h) as compared with days off (8.4 h; Table 2). Rotating shift workers slept most following evening shifts (8.1 h), but significantly less following day shifts (6.2 h) and night shifts (6.2 h; Table 2).

Table 2. Sleep/wake characteristics according to shift type (mean ± SD)
 Rotating shift workers (n = 12)Permanent night shift workers (n = 15)
Days offDay shiftsEvening shiftsNight shiftDays offNight shift
  1. Night shift sleep/wake times significantly different to all other sleep times (< 0.01).

  2. No significant interaction between total sleep time and shift worker cycle type.

  3. *Significantly different to respective night shift (< 0.03).

  4. Significantly different sleep/wake times compared with permanent night shift workers (< 0.01).

Sleep log
 Bed time23:24 ± 00:5923:14 ± 00:5923:11 ± 01:1408:51 ± 00:2622:50 ± 01:1110:33 ± 01:26
 Sleep onset23:47 ± 00:4523:27 ± 00:5823:45 ± 00:5908:58 ± 00:2422:55 ± 00:4710:51 ± 01:31
 Sleep offset07:44 ± 01:1705:31 ± 00:2207:48 ± 01:1815:05 ± 01:2107:08 ± 00:5816:28 ± 02:22
 Rise time08:44 ± 00:4805:40 ± 00:1608:45 ± 00:5515:57 ± 01:2207:36 ± 00:5417:02 ± 02:12
 Total sleep time7.95 ± 1.306.15 ± 1.168.05 ± 0.87*6.19 ± 1.248.35 ± 1.18*5.63 ± 1.38
 Self-rated sleep quality5.62 ± 1.345.04 ± 1.254.87 ± 1.674.99 ± 0.785.47 ± 1.855.31 ± 1.53
Wrist actigraphy
 Sleep onset23:59 ± 00:4523:32 ± 01:0023:37 ± 01:0509:01 ± 00:3422:50 ± 00:5510:37 ± 01:07
 Sleep offset08:14 ± 00:5205:28 ± 00:1908:18 ± 00:5615:31 ± 01:2207:19 ± 01:0216:41 ± 02:14
 Total sleep time8.40 ± 1.066.02 ± 1.178.72 ± 1.75*6.58 ± 1.358.63 ± 1.20*6.02 ± 1.44
 Sleep efficiency80.20 ± 5.5081.66 ± 3.55*77.00 ± 8.6478.81 ± 7.9784.12 ± 6.8383.15 ± 9.15

Traffic and light conditions during drives to and from night shifts

More traffic was reported during the post-night shift drive (M = 6.00, SD = 2.81) compared with the pre-night shift drive (M = 4.64, SD = 2.29; = 0.001). A trend was observed for demand of the drive to be perceived as higher during post-night shift drives (M = 2.31, SD = 1.91) compared with pre-night shift drives (M = 1.78, SD = 1.75; = 0.066). All pre-night shift drives occurred between 20:15 and 20:52 hours, and all post-night shift drives occurred between 07:44 and 08:29 hours. Therefore, all pre-night shift drives were in darkness.

Subjective sleepiness and objective drowsiness during drives to and from night shifts

All measures of sleepiness and drowsiness were significantly higher during post-night shift drives compared with pre-night shift drives (Fig. 2; Fig. 3; Table 3). Although rotating and permanent night shift workers were found to be generally consistent in terms of increased sleepiness and drowsiness from pre-night shift drives to post-night shift drives, rotating shift workers showed significantly higher levels of subjective sleepiness (KSS) and JDS scores compared with permanent night shift workers before/during the drive to their night shifts (pre-drive KSS: 4.21 ± 1.49 versus 3.56 ± 2.04, < 0.05; post-drive KSS: 4.67 ± 1.62 versus 3.68 ± 2.12, < 0.05; maximum JDS score: 2.39 ± 1.31 versus 1.81 ± 1.21, < 0.05). For the drive home after the night shift, rotating shift workers reported higher KSS scores than permanent night shift workers (pre-drive KSS: 6.08 ± 1.74 versus 4.56 ± 2.04, < 0.05; post-drive KSS: 7.29 ± 1.81 versus 5.30 ± 2.27, < 0.05), but no differences were observed in post-night shift JDS scores or any other objective measures of drowsiness (> 0.05).

Figure 2.

 Representative data for one participant from the rotating shift work group. (a) Raster plot of worked shifts, sleep, and pre- and post-shift drives. Black bars: sleep; gray bars: time in bed; diagonal patterned bars: work; small black bars: drive. (b–d) Johns Drowsiness Scale (JDS) scores during commutes to and from three consecutive night shifts (Days 7–9). Open symbols: pre-shift drive plots; closed symbols: post-shift drive plots. Solid reference line: critical drowsiness warning (JDS ≥ 5); broken reference line: cautionary warning (JDS ≥ 4.5).

Figure 3.

 Subjective and objective measures of sleepiness during commutes to and from night shift work. Drives were separated into first and second halves. Open symbols: pre-shift drive plots; closed symbols: post-shift drive plots. Solid bars represent work. Higher values indicate increased levels of drowsiness. JDS, Johns Drowsiness Scale.

Table 3. Subjective and objective measures of sleepiness and drowsiness for night shifts, across rotating and permanent night shift workers
  Pre-shift drivePost-shift drive P
  1. JDS, Johns Drowsiness Scale; KSS, Karolinska Sleepiness Scale.

Subjective measures N (drives)130130 
 Pre-drive KSS scoreMean ± SD3.77 ± 1.865.00 ± 2.08<0.001
 Post-drive KSS scoreMean ± SD4.02 ± 2.005.96 ± 2.32<0.001
 Mean KSS scoreMean ± SD3.89 ± 1.905.48 ± 2.12<0.001
 Pre-drive KSS score ≥5 n (N) %58 (149) 38.995(146) 65.1<0.001
 Post-drive KSS score ≥5 n (N) %61(140) 43.6110 (144) 76.4<0.001
Duration of prior wakefulness N (drives)149149 
 Hours awakeMean ± SD4.31 ± 3.3115.79 ± 3.32<0.001
 Incidence of cumulative hours awake (≥16 h) n (N) %0 (154) 058 (149) 38.9<0.001
Ocular measures N (drives)9393 
 Mean % of long eye closures per minuteMean ± SD0.09 ± 0.180.27 ± 0.44<0.001
 Maximum % of long eye closures per minuteMean ± SD1.27 ± 2.074.40 ± 7.69<0.001
 Mean % of time with eyes closed per minuteMean ± SD0.50 ± 0.800.94 ± 1.22<0.001
 Maximum % of time with eyes closed per minuteMean ± SD4.13 ± 7.107.82 ± 10.21<0.001
 Mean total blink duration per minute (ms)Mean ± SD283.63 ± 79.24326.05 ± 155.700.004
 Maximum total blink duration per minute (ms)Mean ± SD527.15 ± 315.52796.30 ± 981.490.010
JDS measures N (drives)9191 
 Mean JDS scoreMean ± SD1.19 ± 0.911.46 ± 1.030.004
 Maximum JDS scoreMean ± SD2.08 ± 1.282.76 ± 1.55<0.001
 Cumulative JDS score over the driveMean ± SD32.54 ± 28.2448.08 ± 41.99<0.001
 Incidence of JDS ≥ 4.5 n (N) %5 (101) 5.014 (94) 14.90.017

On-road events during post-shift commutes compared with pre-shift commutes

Overall, participants had greater odds of reporting driving events during the post-shift commute compared with the pre-shift commute (Fig. 4; Table 4). Post-shift drives were strongly associated with hazardous driving events, with 8.05 greater odds compared with the pre-shift commute. Inattention and sleep-related events were also associated with higher odds during the post-shift drive (OR, 3.50; OR, 35.20, respectively).

Figure 4.

 Dichotomized drive events reported during commutes to and from all worked shifts. Drive events were categorized into four event types (sleep-related, inattention, hazardous, and violations; as shown in Table 4). Percentages represent the proportion of reported events during all pre-shift and post-shift drives, with 95% confidence intervals for proportions. Asterisks indicate significant differences (< 0.001) between pre- and post-shift drives.

Table 4. Drive events reported during drives to and from all worked shifts
CategoryVariableTotal no. reported eventsDichotomized eventsχ2OR95% CI P
Pre-shift drivePost-shift drivePre-shift drivePost-shift drive
  1. Variables were dichotomized (reported or not reported in each drive) into the respective four categories. Mantel–Haenszel chi-square test: drive events reported during pre-shift and post-shift drives (all shifts; within-subjects comparison: participant as covariate).

Sleep-related eventResting your eyes6376 (3.2%)37 (20.3%)31.7935.206.74, 183.80<0.001
Pulled over for a nap02      
Fell asleep at a stop light02      
InattentionFixation on internal/external object91622 (11.8%)48 (26.4%)15.043.501.79, 6.57<0.001
Lack of awareness631      
Being distracted1113      
Hazardous driving eventBraked sharply6199 (4.8%)37 (20.3%)22.618.053.17, 20.47<0.001
Hit roadside rumble strips322      
Swerved violently13      
Missed your turn01      
Near-miss19      
ViolationCrash008 (4.3%)13 (7.1%)1.112.050.071, 5.96>0.05
Drove through a stop light01      
Shouting at another person812      

In both rotating and permanent night shift workers, Mantel–Haenszel chi square tests showed significant differences between pre- and post-night shift drives in the number of sleep-related [rotating n = 4 (4.0%) pre-shift versus n = 21 (21.4%) post-shift, χ2 = 13.63, < 0.001; permanent n = 2 (2.3%) pre-shift versus n = 16 (19.0%) post-shift, χ2 = 12.73, < 0.001] and hazardous [rotating n = 8 (8.0%) pre-shift versus n = 20 (20.4%) post-shift, χ2 = 6.28, = 0.010; permanent n = 1 (1.1%) pre-shift versus n = 17 (20.2%) post-shift, χ2 = 16.53, < 0.001] driving events. While rotating shift workers did not show significant differences in inattentive drive events between pre- and post-shift drives [n = 11 (11.0%) pre-shift versus n = 18 (18.4%) post-shift, χ2 = 2.15, = 0.163], permanent night shift workers reported significantly more inattentive drive events during the post-shift drive compared with the pre-shift drive [n = 11 (12.6%) pre-shift versus n = 30 (35.7%) post-shift, χ2 = 12.48, = 0.001].

Relationship between drive events and sleepiness measures

Logistic regression analysis was performed on post-shift drives for each drive event category, adjusting for sex, age, BMI, drive duration, and cycle and shift type. The primary measures were deemed as pre-drive KSS scores, maximum blink durations per minute, incidence of JDS scores at or above 4.5, and cumulative hours awake at or above 16 h (as shown in bold in Table 5). Ocular measures of drowsiness and sleepiness measures were associated with greater odds of reporting drive events during commutes following worked shifts. Higher reported pre-drive KSS scores were associated with an increased risk in sleep-related events during the post-shift drive [OR, 1.69 (95% CI, 1.04, 2.73)]. Higher maximum blink durations recorded during the drive were also strongly associated with increased odds of reporting a sleep-related event [OR, 5.35 (95% CI, 1.32, 21.60)]. JDS score of 4.5 or above (cautionary warning) during a post-shift drive was associated with greater odds of reporting inattention during the drive [OR, 4.58 (95% CI, 1.26–16.69)]. Finally, participants who drove after 16 h or more of continuous wakefulness were more likely to report experiencing a hazardous driving event [OR, 4.50 (95% CI, 1.81, 11.16)].

Table 5. Associations between sleepiness measures and on-road driving events during commutes post-shift
 UnadjustedAdjusted*
OR95% CIOR95% CI
  1. JDS, Johns Drowsiness Scale; KSS, Karolinska Sleepiness Scale.

  2. Odds ratios (OR) of the associations between sleepiness/drowsiness measures and drive events reported during post-shift commutes. ORs for continuous variables indicate the change in odds for an increase of one standard deviation. All adjusted ORs significant to P < 0.05.

  3. *Controlling for sex, age, body mass index, drive duration, cycle type (rotating or permanent night shift) and shift type (day shift, evening shift, night shift). For each drive event category, only significant sleepiness variables are reported. Sleepiness variables examined were hours awake since last major sleep, cumulative hours awake >16 h, past sleep duration, pre- and post-drive KSS score, pre- and post-drive KSS score above 5, mean JDS score, maximum JDS score, JDS scores equal to or >4.5, cumulative JDS score of drive, JDS score change from baseline (pre-shift JDS score), mean percentage of long eye closures, maximum percentage of long eye closures, mean percentage time with eyes closed, maximum percentage time with eyes closed, mean total blink duration and maximum total blink duration.

  4. Baseline was measured as the maximum JDS score during the pre-shift drive.

Sleep-related events
 Pre-drive KSS score1.591.06, 2.391.69 1.04, 2.73
 Maximum blink duration per minute2.141.02, 4.485.35 1.32, 21.60
 Maximum % time spent with eyes closed per minute1.340.90, 2.001.831.11, 2.90
 Maximum % long eye closures per minute1.781.08, 2.932.561.35, 4.91
 Maximum JDS score change from baseline1.240.92, 1.661.521.06, 2.19
 Mean KSS score1.721.11, 2.711.951.15, 3.32
Inattention
 JDS (<4.5, ≥4.5)3.631.36, 9.654.58 1.26, 16.69
 Maximum JDS score1.501.02, 2.211.701.00, 2.88
 Mean JDS score1.661.17, 2.371.781.09, 2.88
 Cumulative JDS score1.340.92, 2.302.221.52, 3.46
 Mean % time with eyes closed per minute1.801.21, 2.672.381.31, 4.32
 Minutes above 4.51.471.06, 2.021.581.05, 2.38
Hazardous driving event
 Cumulative hours awake (<16, ≥16)4.712.19, 10.174.50 1.81, 11.16
 Hours awake before drive1.751.25, 2.451.881.25, 2.84
 Mean KSS1.460.96, 2.231.771.00, 3.12
Violations
 Hours awake before drive1.360.84, 2.211.821.03, 3.20

Discussion

This study demonstrated significant increases in self-reported sleepiness and objectively measured drowsiness during commutes following night shift work in nurses working both rotating and permanent night shifts. Evident across both rotating and permanent night shift workers was the significant increase in unsafe driving events experienced following a shift. Shift workers were over eight times more likely to experience hazardous driving events, including hitting roadside rumble strips and having a near-miss incident, during commutes home following an 8–10-h shift, compared with during their drive to work. These events were more frequent (OR, 4.50) when the shift worker had been awake for over 16 h. Inattention while driving was three times more likely to occur on the drive home following a shift as compared with driving to the shift. Ocular measures of drowsiness were related to adverse driving events. In particular, increases in ocular markers of drowsiness during the commute home were associated with more sleep-related events, such as resting the eyes and falling asleep at a stop light, in particular longer maximum blink duration, which was associated with fivefold increased odds of such an event. Furthermore, drowsiness levels above the cautionary warning level (JDS ≥ 4.5) were related to episodes of inattention while driving.

Differences were evident between rotating and permanent night shift workers in sleep/wake times during a sequence of night shifts, reported sleepiness levels and reported events during commutes. Both groups reported similar total sleep durations during a sequence of night shifts and instances of cumulative hours awake >16 h before commuting home. Rotating shift workers, however, reported higher levels of sleepiness during commutes to and from night shift work compared with permanent night shift workers, consistent with previous reports (Muecke, 2005). There are several possible reasons for these differences. First, rotating shift workers in this sample were significantly younger than the permanent night shift worker group. Based on the previous finding that neurobehavioral deficits due to sleep loss are greater in younger people (Duffy et al., 2009), rotating shift workers in our sample may have a greater vulnerability to the impact of sleep loss inherent in shift work. Furthermore, rotating shift schedules can result in more chronic sleep loss and circadian misalignment as a result of the relatively slow rate of adaptation to changing shifts in rotating cycles (Pilcher et al., 2000).

Ocular measures of drowsiness were higher during the drive home following night shift work compared with the drive to work. This finding supports previous research demonstrating that shift workers experience greater levels of sleepiness, fatigue and drowsiness following night shift as a result of the combination of excessive hours of wakefulness, chronic sleep restriction, poor quality of sleep and circadian misalignment (Arendt, 2010; Barger et al., 2009). Increased blink duration, which reflects higher sleepiness (Caffier et al., 2003; Häkkänen et al., 1999), was observed in post-night shift drives compared with pre-night shift drives. This is of great concern due to the association between long blink duration and obstruction of visual input, leading to lapses in performance (Anderson et al., 2010). In the driving context, at their most tired (i.e. mean maximum long eye closure per minute), drivers’ eyes were closed for over 7% of each minute, which for a 30-min drive travelling at 60 km h−1 translates to a total of approximately 2.1 km with eyes closed over the span of the drive.

Reported adverse driving events occurred more frequently during the drive after work. Furthermore, increases in objective measures of drowsiness were associated with increased odds of such adverse drive events. Sleep-related driving events, such as ‘resting your eyes’ or ‘falling asleep at the stop light’, were associated mainly with ocular measures, such as maximum blink duration, percentage of time with eyes closed and long eye closures. Events reflecting inattention, including ‘a lack of awareness’ and ‘being distracted’, were associated mainly with JDS scores of 4.5 and above. Hazardous driving events, such as ‘hitting the roadside rumble strips’ or a ‘near-miss’, were more likely to be reported with increased time spent awake prior to the drive, particularly when the duration of wakefulness was >16 h. These findings suggest that increased sleepiness during a drive home after a work shift is significantly associated with impaired driving performance.

Whereas associations between objective sleepiness during the drive and driving events were observed, pre-drive subjective ratings of sleepiness did not reliably predict such driving events. One possible explanation for these findings is that participants may not be aware of the full extent of their sleepiness (Van Dongen et al., 2003). Divergence between objective and subjective measures of sleepiness was also reflected in the finding that despite significantly higher subjective sleepiness levels reported during the post-shift drives in rotating shift workers compared with permanent night shift workers, objective drowsiness levels did not differ between the groups.

This study is limited by the relatively small sample of nurses (n = 27), although it is noted that this sample provided a total of 369 driving sessions, and also use of self-reported driving assessments. Future studies should aim to assess driving performance in real-time, such as lane deviations, with continuous objective measures of drowsiness. Furthermore, due to the limited sample size and the exploratory nature of the study, we did not adjust our analysis for multiple comparisons. The findings should therefore be interpreted with caution. Furthermore, we found that self-assessed traffic density was higher and driving demands showed a trend towards being higher during the post-night shift drives compared with the pre-night shift drives. Traffic conditions and ambient lighting are reported to impact on drowsiness while driving. Previous work suggests that monotonous drives with little traffic and lower task demand are more likely to promote sleepiness while driving (Reyner et al., 2010), while laboratory studies show that increased light exposure promotes alertness (Phipps-Nelson et al., 2003). As such, higher driving demands and light intensity levels experienced during post-shift (daytime) drives would increase alertness compared with the low ambient light levels experienced during the pre-shift (night) drives. In contrast to the predictions made from these findings, a naturalistic study reported a slightly increased risk of MVC in daylight compared with dark, and that the majority of the drowsiness-related crashes occurred in heavy traffic (Klauer et al., 2006). The interactions between drowsiness, traffic density and lighting conditions do not, however, detract from the key findings of our study, that drowsiness levels are higher in post-shift drives and that this is associated with adverse driving events.

The effects of shift work on sleepiness and driving have implications for public safety policy and legal liability (Rajaratnam and Arendt, 2001). Shift workers behind the wheel after night shifts are exposed to known risks not only to themselves but to the general public. A previous study reported that the risks of documented MVCs and near-misses were significantly increased during the drive home following extended work shifts (≥24 h) in first-year postgraduate medical interns (Barger et al., 2005). We suggest, based on the findings of our study, that education programs be developed and implemented in occupational settings that employ night shift workers. These programs should highlight the significant risk of sleepiness-related driving impairment and MVC risk, particularly during commutes following the night shift. In addition, employers should provide alternative means of transportation to employees who have worked night shifts as part of a comprehensive, evidence-based fatigue management program.

In conclusion, this study demonstrated an increased likelihood of adverse driving events following night shift work, which are associated with objective measures of drowsiness. Hazardous driving events were more frequent after being awake for more than 16 h, reflecting the impact of extended wakefulness on driving performance in night shift workers. The effectiveness of automated warning systems based on objective monitoring of drowsiness needs to be evaluated.

Acknowledgements

Financial support for this study was provided by the Monash University, Faculty of Medicine, Nursing & Health Sciences Strategic Grant Scheme and VicRoads. Optalert™ equipment was provided by Optalert™ Pty Ltd. We would also like to thank Dr Matthew Naughton for contribution to study grant.

Disclosure of conflicts of interest

Dr Rajaratnam reports that he has served as a consultant through his institution to Vanda Pharmaceuticals, Philips Respironics, EdanSafe, The Australian Workers’ Union, Rail, Bus and Tram Union, and National Transport Commission, and has through his institution received research grants and/or unrestricted educational grants from Vanda Pharmaceuticals, Takeda Pharmaceuticals North America, Philips Lighting, Philips Respironics, Cephalon and ResMed Foundation, and reimbursements for conference travel expenses from Vanda Pharmaceuticals. His institution has received equipment donations or other support from Optalert™, Compumedics, and Tyco Healthcare. He has also served as an expert witness and/or consultant to shift work organizations.

Ms Ftouni reports her institution has received equipment donations or other support from Optalert™ and Compumedics.

Dr Anderson has served as consultant to the Rail, Bus and Tram Union through an agreement between Monash University and the Rail, Bus and Tram Union. She has also received research support from VicRoads, and research funds from Sanofi-Aventis. She has received lecturing fees from Brown Medical School/Rhode Island Hospital and Ausmed.

Dr Sletten reports her institution has received equipment donations or other support from Optalert™ and Compumedics.

Dr Lockley reports that he received two investigator-initiated research grants from the ResMed Foundation and an unrestricted equipment gift from ResMed Inc, in support of the studies described in this article; receiving consulting fees from Apollo Lighting, Naturebright, Sound Oasis, and Wyle Integrated Science and Engineering, and federally funded projects at Brigham and Women’s Hospital, Thomas Jefferson University, and Warwick Medical School; lecture fees from Takeda Pharmaceuticals North America, I Slept Great/Euforma, LLC and Emergency Social Services Association Conference, UK; unrestricted equipment gifts from Philips Lighting and Bionetics Corporation; an unrestricted monetary gift to support research from Swinburne University of Technology, Australia; a fellowship gift from Optalert, Pty Ltd, Melbourne, Australia; advance author payment and royalties from Oxford University Press, and honoraria from Servier Inc for writing an article for Dialogues in Clinical Neuroscience and from AMO Inc, for writing an educational monograph, neither of which refer to the companies’ products; honoraria or travel and accommodation support for invited seminars, conference presentations or teaching from the Second International Symposium on the Design of Artificial Environments, Eighth International Conference on Managing Fatigue, American Academy of Sleep Medicine, American Society for Photobiology, Apollo Lighting, Bar Harbor Chamber of Commerce, Bassett Research Institute, Canadian Sleep Society, Committee of Interns and Residents, Coney Island Hospital, FASEB, Harvard University, Illinois Coalition for Responsible Outdoor Lighting, International Graduate School of Neuroscience, Japan National Institute of Occupational Safety and Health, Lightfair, National Research Council Canada, New York Academy of Sciences, North East Sleep Society, Ontario Association of Fire Chiefs, Philips Lighting, Thomas Jefferson University, University of Montreal, University of Tsukuba, University of Vermont College of Medicine, Utica College, Vanda Pharmaceuticals, Velux, Warwick Medical School, Woolcock Institute of Medical Research, and Wyle Integrated Science and Engineering (NASA); investigator-initiated research grants from Respironics Inc, Philips Lighting, Apollo Lighting and Alcon Inc; and a service agreement and sponsor-initiated research contract from Vanda Pharmaceuticals. Dr Lockley also holds a process patent for the use of short-wavelength light for resetting the human circadian pacemaker and improving alertness and performance, which is assigned to the Brigham and Women’s Hospital per Hospital policy and has received revenue from a patent on the use of short-wavelength light, which is assigned to the University of Surrey. Dr Lockley has also served as a paid expert witness on behalf of two public bodies on arbitration panels related to sleep, circadian rhythms and work hours.

Dr Howard has received an unrestricted research grant from ResMed Foundation and equipment on loan from Sleep Diagnostics for research.

Dr Lenné reports receiving through his institution research grants to undertake human factors research for Wesfarmers Limited and Australian Airexpress.

Ancillary