Self-reported Drowsiness and Safety Outcomes While Driving After an Extended Duration Work Shift in Trainee Physicians

Self-reported Drowsiness and Safety Outcomes While Driving After an Extended Duration Work Shift... Abstract Study Objectives Extended duration (≥24 hours) work shifts (EDWSs) are associated with increased risk of motor vehicle crashes, and awareness of any impairment has important implications on legal accountability for any adverse driving outcome. The extent to which adverse driving events were preceded by predrive self-reported sleepiness was evaluated in medical residents after an EDWS. Methods Sixteen resident physicians (10 females; 29.2 ± 2.0 years) working EDWS were monitored when driving on their commute to and from the hospital (438 drives). Sleep and work hours were obtained from daily logs, and adverse driving outcomes were captured from a driving log completed at the end of each drive. Self-reported sleepiness (Karolinska Sleepiness Scale; KSS) and objective drowsiness were captured using a time-stamped, hand-held device and infra-red reflectance oculography. Results Self-reported sleepiness and objective indices of drowsiness were positively correlated, and both were elevated following EDWSs. Compared with the commute to work, EDWSs were associated with more than double the self-reported adverse outcomes when driving home, significantly higher than drives to or from the day shift at comparable times of day. EDWSs more than tripled the odds of reporting sleep-related, inattentive, hazardous, or violation-driving events. The number and type of adverse event was predicted by the predrive KSS level and in a dose-dependent manner. Conclusions Driving after an EDWS puts resident physicians/drivers and other road users at avoidable and unnecessary risk. Demonstrating self-reported sleepiness at the beginning of the drive is associated with adverse outcomes has serious implications on the legal accountability for driving when drowsy. drowsiness, sleepiness, driving, safety, residents, shift work Statement of Significance Driving while drowsy represents a significant risk to fatal or serious car crashes and is attributed to approximately 20% of motor vehicle crashes. One strategy for minimizing the risk of drowsy driving is for drivers to acknowledge their level of drowsiness and to take evasive action. The extent to which drivers can accurately self-assess their drowsiness has been of continued debate in the field, however. Our study provides clear evidence that, before driving home from a 24-hour shift, drivers are aware of their drowsiness level which predicts subsequent on-road adverse driving events. Drivers should therefore proactively evaluate their drowsiness level prior to driving and not expose themselves and others to the significant, and avoidable, risk of drowsy driving. Introduction In the United States, medical and surgical trainee physicians beyond their first year of residency are permitted to work for up to 28 hours continuously and are routinely scheduled to work extended duration work shifts (EDWSs) of 24 hours or more twice per week [1]. Since 2011, such marathon shifts were not permitted for first-year residents (“interns”), but the Accreditation Council for Graduate Medical Education (ACGME) proposes to reintroduce them in 2017 [2]. Although these EDWSs present a significant risk to patient safety due to increased attentional failures [3], risk of errors on clinical task [4], and higher rates of serious medical errors [5, 6], they also negatively affect the health and safety of the trainee physicians themselves [7]. Although physician risk may involve increased rates of depression [8] or needle stick injury [9], the most documented risk for trainee physicians is the risk of a motor vehicle crash [10–13]. It is well documented that these EDWSs negatively affect the safety of trainee physicians by increasing the risk of motor vehicle crashes on the commute home [10–13]. For instance, trainee physicians working a heavy call schedule exhibit deficits on a simulated driving task comparable to performance at 0.04%–0.05% blood alcohol concentration [14] due to a combination of excessive time awake, overnight work, and chronic sleep deficiency. Notably, the odds of a trainee physician being involved in a motor vehicle crash or near-crash after an EDWS are increased by 2.3 and 5.9, respectively, compared with a nonextended shift [10]. These motor vehicle (near) crashes may not be directly attributed to having fallen asleep (although previous studies suggest 49% of trainee physicians admitted to having falling asleep while driving [11], compared with 11%–31% of the nonmedical drivers [see Ref. 15]). Instead, and based on laboratory-evidence, other causal factors for drowsy crashes may include enhanced distractibility [16, 17], reduced vigilance [18], and poor decision making [19]. Despite the prevalence of driving while drowsy in trainee physicians, little is known about the impact of EDWSs on safety outcomes while driving. Alleviating crash risk in resident physicians (and all road users) requires a multifaceted approach. Firstly, understanding causal factors of drowsy driving crashes is essential to informing intervention strategies for reducing crash risk. Secondly, and perhaps most importantly, drivers should be able to determine when they have reached a level of drowsiness associated with increased risk of adverse driving outcomes. Although it is now largely accepted that drivers are typically aware when they have reached a level of sleepiness that is dangerous for driving [20–24], they often continue to drive [25]. Ideally then, drivers should actively evaluate whether they are safe to drive before commencing the journey. To the best of our knowledge, no study has examined self-reported sleepiness at the onset of the drive and the utility of this approach to predict the risk of adverse events on the subsequent drive. Our study objective was therefore to first characterize adverse driving outcomes and examine the prevalence of objective drowsiness and self-reported sleepiness in trainee physicians driving to and from the hospital, and secondly to examine the extent to which a predrive self-assessment of sleepiness accurately predicts adverse driving outcomes. Methods Participants Sixteen resident physicians (10 females; 29.2 ± 2.0 years; range 24.5–33.4 years; Table 1) from seven hospitals in the Boston metropolitan area, and who were experienced drivers (licenced >2 years; drove >50 miles per week) driving >15 minutes alone to and from work, provided a total of 438 driving sessions (Figure 1). Trainee physicians were recruited from four clinical disciplines including internal medicine, surgery, anaesthesiology, and pediatric medicine. Advertisements were placed around the main teaching hospitals in the Boston metropolitan region and emailed to interns and residents via the Committee of Interns and Residents. Interested individuals (n = 192) contacted the study coordinator (CA) and were provided with additional information about the study. For those interested in the study (n = 69), an initial phone interview determined work and driving commitments to ensure that they were within the limits for inclusion criteria. Participants, who commuted to and from the hospital, worked EDWSs, and with a roster that was consistent with the requirements of the study (i.e., a period of noncall plus six consecutive EDWSs), were invited to take part in the full study. Participant demographic, driving experience, residency experience, and presence of sleep or medical problems are reported in Table 1. The study was approved by the Human Research Committee of Partners Healthcare and all participants provided written informed consent and were remunerated for their involvement. Table 1. Demographics and Work/Sleep Parameters in Trainee Physicians Variable        Range  N    16  —  Males  n (%)  6 (37.5%)  —  Age (y)  Mean ± SD  29.2 ± 2.0  24.6–33.4 y  Year of residency  n (%)       First    9 (56.3%)     Second    4 (25%)     Third    1 (6.3%)     Fourth    2 (12.5%)    Clinical disciplines  n (%)       Internal medicine    12 (75%)     Surgery    2 (12.5%)     Anaesthesiology    1 (6.25%)     Pediatric medicine    1 (6.25%)    Body mass index (kg m−2)  Mean ± SD  24.2 ± 3.0  18.3–30.8  Body mass index >30  n (%)  1 (6.25%)  —  Diagnosed illness/disorder  n (%)a  6 (37.5%)  —  Epworth Sleepiness Score  Mean ± SD  10.8 ± 4.8  3–18  Sleep need  Mean ± SD  7.5 ± 1.0  6–10 h  Fall asleep events in past month, when  Telephone  Mean ± SD (%)b  1.4 ± 1.8 (62.5%)  0–6  Meeting  Mean ± SD (%)b  0.07 ± 0.25 (12.5%)  0–1  Traffic  Mean ± SD (%)b  0.33 ± 0.49 (37.5%)  0–1  Driving  Mean ± SD (%)b  0.87 ± 1.77 (31.25%)  0–5  Driving experience (y)  Mean ± SD  11.6 ± 2.3  6-14 y  Miles driven per week  Mean ± SD  105.8 ± 56.0  50–250 miles  Commute time  Mean ± SD  26.7 ± 10.3  15.8–56.8 min  Duration of residency training (mo)  Mean ± SD  18.6 ± 19.5  5–80 mo  Extended work shift duration  Mean ± SD (d.h)  28.3 ± 1.85    Day shift duration  Mean ± SD (d.h)  9.62 ± 2.6    Sleep duration/24-h period   Prestudy (with napsc)  Mean ± SD (d.h)  9.11 ± 0.81     Prestudy (without naps)  Mean ± SD (d.h)  7.9 ± 0.86     Preday shift (with naps)  Mean ± SD (d.h)  8.43 ± 2.17     Preday shift (without naps)  Mean ± SD (d.h)  6.67 ± 0.79     Pre-extended shift (with naps)  Mean ± SD (d.h)  6.97 ± 1.33     Pre-extended shift (without naps)  Mean ± SD (d.h)  7.45 ± 1.42     During extended shift (with naps)  Mean ± SD (d.h)  1.88 ± 1.99     During extended shift (without naps)  Mean ± SD (d.h)  2.32 ± 1.96    Variable        Range  N    16  —  Males  n (%)  6 (37.5%)  —  Age (y)  Mean ± SD  29.2 ± 2.0  24.6–33.4 y  Year of residency  n (%)       First    9 (56.3%)     Second    4 (25%)     Third    1 (6.3%)     Fourth    2 (12.5%)    Clinical disciplines  n (%)       Internal medicine    12 (75%)     Surgery    2 (12.5%)     Anaesthesiology    1 (6.25%)     Pediatric medicine    1 (6.25%)    Body mass index (kg m−2)  Mean ± SD  24.2 ± 3.0  18.3–30.8  Body mass index >30  n (%)  1 (6.25%)  —  Diagnosed illness/disorder  n (%)a  6 (37.5%)  —  Epworth Sleepiness Score  Mean ± SD  10.8 ± 4.8  3–18  Sleep need  Mean ± SD  7.5 ± 1.0  6–10 h  Fall asleep events in past month, when  Telephone  Mean ± SD (%)b  1.4 ± 1.8 (62.5%)  0–6  Meeting  Mean ± SD (%)b  0.07 ± 0.25 (12.5%)  0–1  Traffic  Mean ± SD (%)b  0.33 ± 0.49 (37.5%)  0–1  Driving  Mean ± SD (%)b  0.87 ± 1.77 (31.25%)  0–5  Driving experience (y)  Mean ± SD  11.6 ± 2.3  6-14 y  Miles driven per week  Mean ± SD  105.8 ± 56.0  50–250 miles  Commute time  Mean ± SD  26.7 ± 10.3  15.8–56.8 min  Duration of residency training (mo)  Mean ± SD  18.6 ± 19.5  5–80 mo  Extended work shift duration  Mean ± SD (d.h)  28.3 ± 1.85    Day shift duration  Mean ± SD (d.h)  9.62 ± 2.6    Sleep duration/24-h period   Prestudy (with napsc)  Mean ± SD (d.h)  9.11 ± 0.81     Prestudy (without naps)  Mean ± SD (d.h)  7.9 ± 0.86     Preday shift (with naps)  Mean ± SD (d.h)  8.43 ± 2.17     Preday shift (without naps)  Mean ± SD (d.h)  6.67 ± 0.79     Pre-extended shift (with naps)  Mean ± SD (d.h)  6.97 ± 1.33     Pre-extended shift (without naps)  Mean ± SD (d.h)  7.45 ± 1.42     During extended shift (with naps)  Mean ± SD (d.h)  1.88 ± 1.99     During extended shift (without naps)  Mean ± SD (d.h)  2.32 ± 1.96    aCurrent illnesses reported included high blood pressure (unmedicated), depression (medicated), gastrointestinal problems (medicated [n = 2] and unmedicated [n = 1]), and insomnia (unmedicated). b% with an affirmative response for OSA or fall asleep event in the last month. cnap defined as sleep <20 minutes in duration. View Large Figure 1. View largeDownload slide Flow chart of physicians’ participation. For those screened to fit within our shift and drive requirements, 75% were not eligible for the study largely due to not driving and/or living too close to the hospital (55%). Of those screened, 25% consented to take part, resulting in 23% completing the study. Figure 1. View largeDownload slide Flow chart of physicians’ participation. For those screened to fit within our shift and drive requirements, 75% were not eligible for the study largely due to not driving and/or living too close to the hospital (55%). Of those screened, 25% consented to take part, resulting in 23% completing the study. Design We monitored trainee residents during the commute to and from work, for six consecutive EDWSs and all intervening non-EDWSs. Using a repeated measures design, we compared self-reported sleepiness, objective drowsiness, and self-reported driving outcomes when driving to and from an EDWS and non-EDWS. Shift Schedules In the United States, a trainee physician (beyond their second year of residency) can be scheduled to work EDWSs of 24–28 consecutive hours (24 + 4) within a 80 to 88–hour work week, averaged over 4 weeks [26]. Our study included trainee physicians working EDWSs every third (Q3), fourth (Q4), or fifth (Q5) shift. In between individuals were scheduled to work a “noncall” shift (~06:00 am until ~03:00–05:00 pm) or have a day off. The study ended after six EDWSs, which were completed over 3 to 5 weeks. Methods During baseline assessments, we obtained information on participants’ demographics, shift schedules, driving experience, and trait levels of daytime sleepiness using the Epworth Sleepiness Scale [27]. For 1 week prior to study, participants worked on a noncall rotation without night duty or EDWS (n = 11) or were on a scheduled week of vacation (n = 5). This ensured adequate opportunity for rest prior to taking part in the study. During this prestudy phase, hours of work and rest, and sleep outcomes were monitored via actigraphy (Motionlogger, Ambulatory Monitoring, Inc., Ardsley, NY) and daily sleep and work logs. During the main study, we obtained data relating to sleep, work hours, and driving outcomes over six consecutive EDWSs and all intermediary day shifts (as per our previous studies [3, 5, 28]), using validated daily sleep [3] and work logs and driving logs completed after each commute to and from the hospital (Supplementary Material 1). Sleep diaries comprised 14 questions on sleep timing, duration, disturbance, and alcohol or medication use. Participants also completed the start and end dates and times of any shift worked in the last 24 hours. The driving log contained 18 items including information on the type and duration of commute, incidence and information of near-crashes or actual crash events, adverse driving events, countermeasure to sleepiness actions, and other information relating to the drive (i.e., traffic density, weather, and consumption of caffeine/alcohol). Participants selected as many driving events and/or countermeasure actions as applicable. The diary is available in Supplementary Material. From the driving log, self-reported driving events were categorized according to criteria adapted from the work of Reason [21, 29]: (1) sleep-related (i.e., “falling asleep at a stop light”), (2) inattention (i.e., “being distracted”), (iii) hazardous (i.e., “swerving violently”), and (iv) violation (i.e., “driving through a stop light”). Objective measures of drowsiness (Johns Drowsiness Score; JDS) were recorded while driving using infrared reflectance oculography (Optalert™ Drowsiness Measurement System, Sleep Diagnostics Pty Ltd., Melbourne, Australia [30, 31]). This approach captures information from involuntary eye blinks, such as amplitude/velocity ratio, long eye closures, and blink duration. The ocular parameters are then combined via a proprietary algorithm to derive the JDS, which compared with other ocular parameters from which the composite JDS is derived, is a more sensitive marker of drowsiness [32], and has been previously associated with driving performance in real on road driving [21]. Participants completed an electronic, time-stamped Karolinska Sleepiness Scale (KSS) [33] at the beginnng and end of each commute. Data Analysis We categorized all data to either a “day shift”—for shifts starting at 06:00–07:00 am of <16 hours duration—or an “EDWS”—for shifts ≥24 hours duration (referred to as EDWS throughout). Commutes to the hospital are referred to as “preshift drive” and the commute home as “postshift drive.” We averaged data within an individual for each pre- and postshift drive for day shifts and extended shifts separately, and compared using repeated measures two-way ANOVA (shift*commute). To examine the likelihood of reporting each of the four driving event categories ((1) sleep-related, (2) inattention, (3) hazardous, and (4) violation events) during preshift drives compared with postshift drives, we employed a chi square relative risk test using a within-subject comparison for day shift and EDWSs separately. We obtained KSS scores from PDA (Personal Digital Assistants) devices logged at the beginning and end of each commute. Where there were missing KSS data on the PDA devices, these were replaced with scores reported in the driving log at the beginning and end of each drive. A total of 56 (13%) KSS scores were derived from driving logs. Taking all data, driving log and PDA-derived KSS scores were highly similar (KSS start of drive, r = 0.84, p < .001; KSS end of drive, r = 0.86, p < .001). To examine the impact of shift, we averaged KSS within subject and then within-shift (day shifts and extended work shifts) for pre- and postshift drives and compared using a two-way (2 × 2) ANOVA (shift*commute). For objective drowsiness, we first removed any commute without 10 consecutive data points (minutes) from the analysis due to reduced reliability of the output. We then visually inspected all data for artifact. Each EDWS was paired with a preceding or subsequent day shift within each individual for comparison and removed any participant without a matching day and EDWS (n = 8 remaining). We then used a two-way (2*2) ANOVA (shift*commute) to examine the impact of shift on ocular-derived drowsiness outcomes. To verify self-reported outcomes, we examined the association between self-reported sleepiness and objective markers of drowsiness using Pearson correlations. To evaluate the predictive capacity of predrive KSS, we performed binary logistic regression where KSS was the predictor variable (continuous) and self-reported adverse driving event the outcome variable. This analysis was conducted separately for each event category. Goodness of fit was determined by Omnibus Tests of Model Coefficients. In addition, to assess predictive capacity of KSS for subsequent adverse events, we constructed receiver–operator characteristic (ROC) curves and performed area under the curve (AUC) analyses. Pretest probability was estimated empirically from the data, which was p = .41 for all events, p = .23 for sleep-related events, and p = .17 for inattentive events. All cost-ratios were set at 1.0. These values were used to determine optimal cutoff points. Results Data Obtained Fourteen trainee physicians completed the full study (six EDWSs) and two physicians completed only four EDWSs (study finished early due to sick leave [n = 1] and poor weather conditions [n = 1]). We obtained data from 438 drives: 218 (49.8%) commuting to the hospital and 220 (50.2%) commuting home after the work shift and 241 (55.0%) for day shifts and 197 (45.0%) for EDWSs. Of the 438 drives monitored with driving logs, 289 were also monitored with the objective drowsiness detection system (67%). Although individual reasons for lost oculography data were not reported by participants, these included a failure of the technology to record, a loss of battery power, or participants not recording the drive. Following visual quality checks of the data, only 182/289 (63%) were deemed of sufficient quality for analysis. Although the majority of drives were rejected for multiple signal artifact, individual reasons for rejection included a biphasic signal (n = 54), clipping of the signal (n = 14), blinks not clearly identifiable (n = 95), noise in the signal (n = 42), an inverted signal (n = 10), and a small signal (n = 93). We then omitted a further 20 drives due to failing to meet the criteria of at least 10 consecutive minutes of monitoring (6.9%). The final number of drives containing objective drowsiness data was 162 (56%) and was equally distributed across day and EDWS (87/46.3% vs. 75/53.7%, respectively; z = 0.858, p = .39) and pre- and postshift drives (73/45.1% vs. 89/54.9%, respectively; z = 1.109, p = .27). Sleep and Work Scheduling Sleep and work-scheduling data can be seen in Table 1. For an EDWS, the proportion of individuals with an average KSS score of 6 or more (some signs of sleepiness) on the drive home was 75%, compared with 12% of the drive to work. The proportion of individuals with an average score of 7+ (moderate-severe sleepiness) was 44% on the drive home, compared with 12% on the drive to work. Self-reported sleepiness was elevated by 46% on the commute home after an EDWS (p < .0001), compared with only 1.26% after a day shift. Likewise, objective measures of drowsiness were also elevated on the drive home, but only after an EDWS (p = .012). Here, JDS increased by 48% when driving home after an EDWS, compared with a 25% decrease after a day shift. Predrive self-reported sleepiness assessments and subsequent objective drowsiness levels were highly correlated (r = 0.81, p = .003) (Figure 2). Figure 2. View largeDownload slide Self-reported sleepiness and objective drowsiness while driving to and from the hospital on day versus EDWSs: self-reported ratings of sleepiness [(A) Day shift: 3.98 ± 0.42 to 4.03 ± 0.40; EDWS: 4.08 ± 0.41 to 5.97 ± 0.40] and objective ocular-derived measures of drowsiness [(B) 1.05 ± 0.23 to 0.79 ± 0.15; EDWS: 0.97 ± 0.12 to 1.43 ± 0.31] were elevated when driving home after an EDWS. Self-reported sleepiness prior to the drive was positive associated with subsequent objective measures of drowsiness (C). Figure 2. View largeDownload slide Self-reported sleepiness and objective drowsiness while driving to and from the hospital on day versus EDWSs: self-reported ratings of sleepiness [(A) Day shift: 3.98 ± 0.42 to 4.03 ± 0.40; EDWS: 4.08 ± 0.41 to 5.97 ± 0.40] and objective ocular-derived measures of drowsiness [(B) 1.05 ± 0.23 to 0.79 ± 0.15; EDWS: 0.97 ± 0.12 to 1.43 ± 0.31] were elevated when driving home after an EDWS. Self-reported sleepiness prior to the drive was positive associated with subsequent objective measures of drowsiness (C). There was a three-fold increase in the average number of self-reported adverse driving events on the drive home from an EDWS compared with driving to work (1.1 ± 0.78 versus 0.37 ± 0.49 events/drive, p < .005, respectively), much more than the increase associated with driving to and from a day shift (0.35 ± 0.53 versus 0.59 ± 0.63 events/drive, respectively; Figure 3). Compared with driving to work, the odds (odds ratio [OR], 95% confidence interval [CI]) of reporting an adverse event on the drive from an EDWS was 3.90 (1.73–8.80, p = .001) for sleep-related events, 3.58 (1.72–7.47, p < .0001) for inattention events, 3.32 (1.16–9.53, p = .02) for hazardous events, and 5.39 (1.15–25.29, p = .018) for violation events. In comparison, after working a day shift, there were no increased odds of self-reporting an adverse event (p > .1) (Table 2). Table 2. The Odds Ratio of a Drive Postshift Containing a Specific Driving Event Compared With the Drive Preshift: Day Shifts and EDWSs     Day shift  Extended work shift  Preshifta  Postshiftb  χ2  OR  95% CI  p  Preshiftc  Postshiftd  χ2  OR  95% CI  p  Sleep- related  Resting the eyes  31 (25.8%)  31 (25.6%)  0.00  0.99  0.56–1.76  .97  9 (9.2%)  28 (28.3%)  11.78  3.9  1.73–8.8  .001  Asleep stop light  Inattention  Fixation  11 (9.2%)  16 (13.2%)  1.00  1.51  0.67–3.41  .32  12 (12.2%)  34 (33.3%)  12.43  3.58  1.72–7.47  <.0001  Lack awareness  Distracted  Missed turn  Hazardous  Braked sharply  9 (7.5%)  13 (10.7%)  0.76  1.49  0.61–3.62  .38  5 (5.1%)  15 (15.2%)  5.45  3.32  1.16–9.53  .02  Hit rumble strip  Swerved violently  Near Miss  Violation  Shouting  1 (0.8%)  5 (4.1%)  2.7e  5.13  0.59–44.58  .10  2 (2.0%)  10 (10.1%)  5.59e  5.39  1.15–25.29  .018  Driving through stop light      Day shift  Extended work shift  Preshifta  Postshiftb  χ2  OR  95% CI  p  Preshiftc  Postshiftd  χ2  OR  95% CI  p  Sleep- related  Resting the eyes  31 (25.8%)  31 (25.6%)  0.00  0.99  0.56–1.76  .97  9 (9.2%)  28 (28.3%)  11.78  3.9  1.73–8.8  .001  Asleep stop light  Inattention  Fixation  11 (9.2%)  16 (13.2%)  1.00  1.51  0.67–3.41  .32  12 (12.2%)  34 (33.3%)  12.43  3.58  1.72–7.47  <.0001  Lack awareness  Distracted  Missed turn  Hazardous  Braked sharply  9 (7.5%)  13 (10.7%)  0.76  1.49  0.61–3.62  .38  5 (5.1%)  15 (15.2%)  5.45  3.32  1.16–9.53  .02  Hit rumble strip  Swerved violently  Near Miss  Violation  Shouting  1 (0.8%)  5 (4.1%)  2.7e  5.13  0.59–44.58  .10  2 (2.0%)  10 (10.1%)  5.59e  5.39  1.15–25.29  .018  Driving through stop light  aData are based on 120 drives. bData are based on 121 drives. cData are based on 98 drives. dData are based on 99 drives. eFishers exact test reported when number of events falls below 5. View Large Figure 3. View largeDownload slide Adverse drive events while driving to and from the hospital on day shifts versus EDWSs. Compared with preshift, more adverse events were reported postshift for the EDWS (0.37 ± 0.49 and 1.1 ± 0.78). This was less evident for day shift (0.35 ± 0.53 and 0.59 ± 0.63), despite being at a comparable time of day. Refer (A) and (B) for group and individual level data, respectively. Importantly, increasing KSS was associated with an increased average number of adverse driving events in a dose response manner, at the group (C) and individual level (D). Figure 3. View largeDownload slide Adverse drive events while driving to and from the hospital on day shifts versus EDWSs. Compared with preshift, more adverse events were reported postshift for the EDWS (0.37 ± 0.49 and 1.1 ± 0.78). This was less evident for day shift (0.35 ± 0.53 and 0.59 ± 0.63), despite being at a comparable time of day. Refer (A) and (B) for group and individual level data, respectively. Importantly, increasing KSS was associated with an increased average number of adverse driving events in a dose response manner, at the group (C) and individual level (D). Although no crashes occurred, 10 near-crash events were reported (2.3% of total drives) from 6/16 trainee physicians (38%): one from a day shift, one while commuting to an EDWS, and eight while driving home after an EDWS (8.2% of commutes from EDWSs). Descriptive evaluation of these near-crash events can be seen in Table 3. Table 3. Descriptive Data for Near-Crash Events Shift  Pre- or post shift  Time of day  Amount of sleep in previous 24 h  Time since awake (h)  Description  Day  Postshift  17:25  6  10  “Other car went through a red light”  Extended  Preshift  06:25  8  1  “Fast moving car in left lane, tried to change lanes but nearly rear-ended”  Extended  Postshift  14:30  2  9  “I was merging into traffic and dozed off and almost hit another car”  Extended  Postshift  12:50  0  30  “Eyes closed on freeway, veared 1/4 way into adjacent lane, had it been a busier time of day would have probably hit someone”  Extended  Postshift  12:25  0.5  36  “Other than 30 min, been awake 36 h…Clear day, someone ran a red light in Boston and I had to slam on brakes to avoid”  Extendeda  Postshift  13:45  0  30  “Had four near misses involving nodding off while driving”  Extended  Postshift  12:45  8  7  “Driving in wet but then another car crossed over into the lane without signaling I was able to slow down in time”  Shift  Pre- or post shift  Time of day  Amount of sleep in previous 24 h  Time since awake (h)  Description  Day  Postshift  17:25  6  10  “Other car went through a red light”  Extended  Preshift  06:25  8  1  “Fast moving car in left lane, tried to change lanes but nearly rear-ended”  Extended  Postshift  14:30  2  9  “I was merging into traffic and dozed off and almost hit another car”  Extended  Postshift  12:50  0  30  “Eyes closed on freeway, veared 1/4 way into adjacent lane, had it been a busier time of day would have probably hit someone”  Extended  Postshift  12:25  0.5  36  “Other than 30 min, been awake 36 h…Clear day, someone ran a red light in Boston and I had to slam on brakes to avoid”  Extendeda  Postshift  13:45  0  30  “Had four near misses involving nodding off while driving”  Extended  Postshift  12:45  8  7  “Driving in wet but then another car crossed over into the lane without signaling I was able to slow down in time”  aDescription of four events in one commute home. View Large Predrive KSS was a significant predictor of subsequent driving events in a dose-dependent manner (Figure 3). For each one-point increase in predrive KSS, there was a 2.39 times increase in the odds of reporting a sleep-related adverse event during the subsequent drive (95% CI: 1.98–2.89, p < .0001). The odds of reporting an inattention or hazardous event also increased per each single-point KSS rise [1.33 (1.13–1.56) and 1.20 (1.06–1.36), respectively] (Table 4). Receiver–operator curves demonstrated that a predrive KSS rating of ≥6 (“some signs of sleepiness”) had 91% sensitivity and 69% specificity for predicting a sleep-related adverse driving event (AUC = 0.86; p < .001). Table 4. Binary Logistic Regression Examining the Predictive Capacity of Predrive KSS on Subsequent Adverse Driving Events   Goodness of fit  Binary logistic regression parameters  χ2  Sig.  B(S.E.)  Wald  df  Exp(B)  Sig.  Upper  Sleep-related  145.5  <.0001  0.87 (0.10)  82.06  1  2.39  <.0001  1.98–2.89  Inattention  8.8  .003  0.18 (0.06)  8.53  1  1.20  .003  1.06–1.36  Hazardous  12.87  <.0001  0.28 (0.08)  11.57  1  1.33  .001  1.13–1.56  Violation  0.01  .92  −0.01 (0.11)  0.01  1  0.99  .025  0.80–1.23    Goodness of fit  Binary logistic regression parameters  χ2  Sig.  B(S.E.)  Wald  df  Exp(B)  Sig.  Upper  Sleep-related  145.5  <.0001  0.87 (0.10)  82.06  1  2.39  <.0001  1.98–2.89  Inattention  8.8  .003  0.18 (0.06)  8.53  1  1.20  .003  1.06–1.36  Hazardous  12.87  <.0001  0.28 (0.08)  11.57  1  1.33  .001  1.13–1.56  Violation  0.01  .92  −0.01 (0.11)  0.01  1  0.99  .025  0.80–1.23  View Large Discussion We examined self-reported and objective measures of drowsiness in US trainee physicians when driving to and from the hospital after day shifts (<12 hours) and EDWSs (>24 hours). In our study, driving home after working an EDWS was associated with an approximately 40% increase in self-reported sleepiness at the start of the drive home, elevated levels of objectively measured drowsiness while driving, and a greater number of adverse driving events. Self-reported sleepiness at the beginning of the drive was associated with objective drowsiness during the drive, and both the number and type of adverse driving events reported. Currently, resident and attending physicians are permitted to work >24-hour shifts [1], which have adverse implications for the physician’s safety while driving [10]. Half of trainee physicians admit to having fallen asleep while driving [11], compared with 11%–31% of typical drivers not working EDWS [15]; this is consistent with our observation of nearly four times greater odds of reporting a sleep-related adverse driving event on the drive home after an EDWS. Sleep deficiency also manifests in other behaviors, such as increased distractibility [16, 17], aggression and irritability [34], and poor decision making [19]. Consistent with this wider scope of drowsy driving symptomology, trainee physicians driving home after an EDWS in the current study had 3–5 times greater odds of reporting an inattentive, hazardous, or violation-driving event. Although inattention was comprised of drivers reporting being fixated, missing the turn, being distracted, and having a lack of awareness, our findings of increased inattention were largely due to being distracted and having a lack of awareness. These data are in agreement with our previous simulated driving study that showed drivers were more distracted when drowsy. Importantly, this inattentive behavior was associated with an increased likelihood of the car leaving the driving lane [17]. Three quarters of trainee physicians reported a KSS of 6+ (“some signs of sleepiness”), whereas 44% reported a KSS of 7+ (moderate–severe sleepiness), prior to driving home after an EDWS. Given that the levels of sleepiness reported by the interns and residents in our study was similar to that reported elsewhere for studies of comparable demographic [35] and design [23], there is no evidence to suggest that our sample represents an unusual population. Our data revealed a strong dose-dependent association between the predrive KSS and the incidence of sleep-related adverse events reported during the subsequent drive home after an EDWS. These data suggest that trainee physicians’ self-assessment of their sleepiness before driving can be a powerful and reliable warning indicator of driving impairment. This is consistent with our previous study assessing on-road driving impairment following a night shift; although night shift workers reported an average KSS of 6 at the onset of the drive, the drive was associated with increasing KSS, use of emergency braking manoeuvres (on 38% of postnight shift drives, PNSD), a failure to adequately remain in control over the vehicle (44% of PNSD), and the potential for falling asleep at the wheel [23]. A KSS score of 6 or more should therefore be considered a warning indicator for high risk of drowsy driving and should prompt individuals to avail themselves of alternative transportation or to deploy appropriate countermeasures such as sleep at the hospital before driving home (with adequate time upon wakening to dissipate any sleep inertia [36]), or caffeine [37]. Although alternative transport was available to the physicians in the current study through hospital-funded taxi-cab schemes, only two participants reported taking a taxi during the study suggesting as follows: (1) despite being aware of feeling sleepy, there was a failure to associate this with the risk of adverse safety outcomes while driving, similar to the impairment of risk assessment under the influence of alcohol [38] or (2) inadequate education or reinforcement/encouragement to use the service. Despite trainee physicians reporting elevated levels of sleepiness after an EDWS, they continue to drive with adverse safety outcomes for themselves and other road users. Much debate exists on the utility of self-reported sleepiness in reducing motor vehicle crash risk. Previous studies showing a lack of agreement between self-reported and objective indices of sleepiness have focused on either individuals in a laboratory-based, nondriving scenarios (e.g., Ref. 39) or examined sleepiness levels just prior to a driving event (when introspection into sleepiness is likely to be low) [40]. Laboratory-based simulator studies have consistently argued that self-reported sleepiness is associated with either objective indices of drowsiness [22, 41–44] or impaired driving performance [22, 44]. Although driving simulator studies have provided much evidence for the agreement between self-reported and objective measures, real on-road studies provide the most ecological valid evidence for evaluating the association between self-reported sleepiness and objective physiological (e.g., slow eye closures) and behavioral (e.g., adverse driving event) indices of drowsiness. These studies have demonstrated consistently that drivers are aware of increasing sleepiness while driving, which occurs in parallel with increased adverse outcomes [20, 21, 23, 45, 46].. Of relevance to our study, self-reported ratings prior to the drive has been associated with subsequent driving events previously. For instance, in sleep-restricted individuals driving at night on a French highway, increasing self-reported sleepiness prior to the drive was associated with more lane crossing in the subsequent drive [46], our study of night shift workers found a predrive KSS of 6 at was associated with adverse driving outcomes (emergency braking manoeuvres required, failure to adequately remain in control over the vehicle, and falling asleep at the wheel) [23], and our study of Australian nurses driving home after a night shift revealed a higher reported pre-drive KSS was associated with an increased risk of a sleep-related event (e.g., falling asleep at a stop light) during the subsequent drive [21]. Despite this awareness, drivers often start, or continue to, drive [25], possibly due to a desire to reach the destination or not being able to find a suitable and safe place to stop and rest if the drive has already commenced. Ideally then, drivers should evaluate their sleepiness prior to commencing the journey. Our study is informative in that it examines sleepiness ratings prior to starting the drive and thus provides an opportunity for evasive action. Habitually reflecting on sleepiness prior to driving a motor vehicle may therefore provide an opportunity to take action and ultimately alleviate accident risk. Our study has a number of limitations. Firstly, as participants were not selected, randomly prior concerns about sleepiness and work hours may have motivated individuals to volunteer. Although our study data were collected from 2009–2011 when the ACGME were working toward their recommendation to abolish EDWS for PGY-1 residents, enquiring about attitudes toward shifts and sleepiness may have helped us to understand whether individuals felt strongly about work hours and driving, and to understand if our participants were motivated to participate by those feelings. It should be noted however that of those who expressed an interest to participate (i.e., phone interview), 42% did not drive and 12% did not work overnight or have extended work shifts, suggesting that strong attitudes toward “marathon shifts” or drowsy driving was not a major factor in choosing to volunteer. Furthermore, as described above, self-reported sleepiness ratings were very similar to a recent independent study of 124 interns and residents, suggesting that our samples was not dominated by interns with higher-than-normal levels of sleepiness [35]. Secondly, the self-report nature of the adverse driving outcomes may not reflect actual adverse events due to report bias. It is unlikely, however, that this would have affected the objective infrared reflectance oculography which was correlated with the self-reported data. Furthermore, although self-reported outcomes may be perceived as unreliable, previous studies have shown self-reported measures of driver performance to be associated with actual driver performance as assessed by objective methods (i.e., speed) or a trained assessor [47] and, of significance, crash risk [48]. Thirdly, our sample size was small and was not powered to detect crash or near-crash events. Rather, this study was designed to examine associations between sleepiness and adverse driving events. Fourthly, we lost a large amount of data from the infrared oculography device due to missing or reduced quality data. Given the nature of the study, and the work commitments and priorities of the study physicians, it is not unexpected that data loss may occur. As the study took place in built-up areas which require more movement and monitoring than that associated with highway driving (i.e., long stretches of straight road, largely comprised of looking ahead), signals can be lost due to excess head movements affecting the ocular signal. The importance of cleaning the data when using these types of devices in field settings is high. Finally, our “control” drives likely involved high levels of drowsiness, thus reflecting lower odds ratios for adverse driving events. For instance, odds ratios for adverse driving events on the commute home were lower in our study compared with those reported in the study of night shift working Australian nurses [21]. This is likely due to (1) the preshift drives in our study (which serve as a comparative baseline for odds ratio) occurred during an adverse circadian phase (~06:00 am), whereas the Australian nurses drove to work during the wake maintenance zone (~07:00 pm), and (2) the resident physicians in our study probably had higher levels of chronic sleep restriction due to repeated shifts, which subsequently affects sleep-related outcomes [28], whereas the Australian nurses worked a maximum of three night shifts. Although EDWSs were originally designed for residents working, sleeping and residing in the hospital [34], modern day trainees commuting to and from the hospital are exposing themselves and others to an elevated risk of drowsy driving. Higher levels of predrive sleepiness, the objective confirmation of drowsiness while driving, and the high rate of self-reported adverse events measured during the drive home from EDWS suggest that those working these shifts may knowingly place themselves and other road users at an unnecessary risk of drowsy driving crashes. Our data confirm that these risks are predictable—drivers were able to recognize their own elevated sleepiness levels before driving home and those assessments predicted the risk of a subsequent adverse driving event with relatively high sensitivity and specificity. This has serious implications on the legal accountability of the driver for adverse outcomes on the drive [49]. It is essential, therefore, that larger-scale studies in real, on-road environments should examine this wider issue in physicians and other shift worker populations. Our data are newly relevant, given that the ACGME have recently reinstated EWDS for postgraduate year 1 (PGY-1) resident physicians [2]. Although EDWSs are permitted, our data suggest that trainee physicians, and all other road users driving during this period considered at risk for drowsiness, should consciously evaluate their sleepiness level prior to each and every drive and, if they suspect that they are impaired, seek alternative transportation and not expose themselves and others to the significant, and avoidable, risk of drowsy driving. Ethical Approval The study was approved by the Human Research Committee of Partners Healthcare and all participants provided written informed consent and were remunerated for their involvement. Supplementary Material Supplementary material is available at SLEEP online. Funding This study was supported by a President and Fellows of Harvard College seed funding award presented to CA, and equipment was funded by a Defense University Research Instrumentation Program (DURIP) award FA9550-07-1-0242 from the Air Force Office of Scientific Research. The content is solely the responsibility of the authors and does not represent the official views of the funding bodies. Author Contributions All authors have made substantial contributions to the work presented and have approved the final version of the manuscript. CA, CAC, and SWL designed the study; CA conducted the study, analyzed and interpreted the data, and wrote the manuscript; SF assisted in the analysis and interpretation of the data; JMR provided technological oversight of the OptalertTM measurements and data output; SWR, CAC, and SWL contributed to data analysis, interpretation of data, and edited the manuscript. All authors approved the final manuscript. Disclosure Statement CA has no conflicts of interests related to the results reported in this paper report. In the interest of full disclosure, she has received a research award/prize from Sanofi-Aventis; contract research support from VicRoads, Rio Tinto Coal Australia, and Tontine/Pacific Brands; and lecturing fees from Brown Medical School/Rhode Island Hospital, Ausmed, Healthmed, and TEVA. In addition, she has served as a consultant through her institution to the Rail, Bus, and Tram Union, the Transport Accident Commission (TAC), and the National Transportation Committee (NTC). She has also served as an expert witness and/or consultant in relation to fatigue and drowsy driving. CA is a Theme Leader in the Cooperative Research Centre for Alertness, Safety, and Productivity. SF serves as a Project Leader in the Cooperative Research Centre for Alertness, Safety, and Productivity. JMR reports that he has served as a consultant through his institution to Takeda Pharmaceuticals North America and Cephalon and has received consulting fees from Vanda Pharmaceuticals. SMWR reports that he has served as a consultant through his institution to Vanda Pharmaceuticals, Philips Respironics, EdanSafe, The Australian Workers’ Union, National Transport Commission, Transport Accident Commission, and New South Wales Department of Education and Communities; has through his institution received research grants and/or unrestricted educational grants from Vanda Pharmaceuticals, Shell, Teva Pharmaceuticals, Rio Tinto, Seeing Machines, 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 served as an expert witness and/or consultant to shift work organizations. SMWR also serves as a Program Leader in the Cooperative Research Centre for Alertness, Safety, and Productivity. CAC has received consulting fees from or served as a paid member of scientific advisory boards for Bose Corporation, Boston Celtics, Columbia River Bar Pilots, Institute of Digital Media and Child Development, Klarman Family Foundation, Quest Diagnostics, Inc., Vanda Pharmaceuticals, and V-Watch/PPRS. He has also received education/research support from Cephalon Inc., Mary Ann & Stanley Snider via Combined Jewish Philanthropies, Optum, Philips Respironics, Inc., ResMed Foundation, San Francisco Bar Pilots, Schneider Inc., and Sysco. He has received lecture fees from American Academy of Sleep Medicine (AADSM), CurtCo Media Labs LLC, Global Council on Brain Health/AARP, Hawaii Sleep Health and Wellness Foundation, National Sleep Foundation, University of Michigan, University of Washington, and Zurich Insurance Company, Ltd. The Sleep and Health Education Program of the Harvard Medical School Division of Sleep Medicine (which CAC directs) has received Educational Grant funding from Cephalon, Inc., Jazz Pharmaceuticals, Takeda Pharmaceuticals, Teva Pharmaceuticals Industries Ltd., Sanofi-Aventis, Inc., Sepracor, Inc., and Wake Up Narcolepsy. CAC is the incumbent of an endowed professorship provided to Harvard University by Cephalon, Inc. and holds a number of process patents in the field of sleep/circadian rhythms (e.g., photic resetting of the human circadian pacemaker). Since 1985, he has served as an expert on various legal and technical cases related to sleep and/or circadian rhythms including those involving the following commercial entities: Bombardier, Inc., Continental Airlines, FedEx, Greyhound, and United Parcel Service (UPS). CAC owns or owned an equity interest in Somnus Therapeutics, Inc. and Vanda Pharmaceuticals. He received royalties from McGraw Hill and Koninklijke Philips Electronics, N.V. for the Actiwatch-2 and Actiwatch-Spectrum devices. CAC’s interests were reviewed and managed by Brigham and Women’s Hospital and Partners HealthCare in accordance with their conflict of interest policies. SWL has no conflicts of interests related to the research or results reported in this paper. In the interests of full disclosure, commercial interests from the last 3 years (2014–2017) are listed below. SWL has received consulting fees from the Atlanta Falcons, Atlanta Hawks, Carbon Limiting Technologies Ltd. on behalf of PhotoStar LED, Perceptive Advisors, Serrado Capital, and SlingshOT Insights and has current consulting contracts with Akili Interactive, Consumer Sleep Solutions, Delos Living LLC, Headwaters Inc., Hintsa Performance AG, Light Cognitive, Mental Workout, Pegasus Capital Advisors LP, PlanLED, OpTerra Energy Services, Inc., and Wyle Integrated Science and Engineering. SWL has received unrestricted equipment gifts from Biological Illuminations LLC, Bionetics Corporation, and F.LUX Software LLC; has equity in iSLEEP, Pty; advance author payment and/or royalties from Oxford University Press; honoraria plus travel, accommodation, and/or meals for invited seminars, conference presentations or teaching from BHP Billiton, Estee Lauder, Lightfair, Informa Exhibitions (USGBC), Teague; and travel, accommodation, and/or meals only (no honoraria) for invited seminars, conference presentations, or teaching from FASEB, Hintsa Performance AG, Lightfair, and USGBC. SWL has completed investigator-initiated research grants from Biological Illumination LLC and Vanda Pharmaceuticals, Inc. and has an ongoing investigator initiated grant from F. Lux Software LLC; completed service agreements from Rio Tinto Iron Ore and Vanda Pharmaceuticals Inc.; and completed three sponsor-initiated clinical research contracts from Vanda Pharmaceuticals, Inc. SWL holds a process patent for “Systems and methods for determining and/or controlling sleep quality,” which is assigned to the Brigham and Women’s Hospital per Hospital policy. SWL has also served as a paid expert on behalf of several public bodies on arbitrations related to sleep, light, circadian rhythms, and/or work hours for City of Brantford, Canada, and legal proceedings related to light, sleep, and health. SWL is also a Program Leader for the CRC for Alertness, Safety, and Productivity, Australia. Notes Work Performed: Department of Medicine and Neurology, Division of Sleep and Circadian Disorders, Brigham and Women’s Hospital and the Department of Medicine, Division of Sleep Medicine, Harvard Medical School, Boston, MA Acknowledgments We thank the resident physicians who took part in the study and the participating hospitals in the Boston Metropolitan area. We also thank the Committee of Interns and Residents for help with advertising the study to members. We also thank Robert Kilpatrick for assistance in pilot testing of the OptalertTM equipment. We also acknowledge Jason Sullivan and Kate Crowley for helping with data storage and transfer. References 1. Nasca TJ, Day SH, Amis ESJr; ACGME Duty Hour Task Force. The new recommendations on duty hours from the ACGME Task Force. N Engl J Med . 2010; 363( 2): e3. Google Scholar CrossRef Search ADS PubMed  2. Accreditation Council for Graduate Medical Education (ACGME). ACGME Common Program Requirements. Section VI. Proposed Major Revisions. 2016 [accessed 2016 21/12/2016]; Available from: http://www.acgme.org/Portals/0/PFAssets/ReviewandComment/CPR_SectionVI_ChangesTracked.pdf 3. Lockley SW, Cronin JW, Evans EEet al.  ; Harvard Work Hours, Health and Safety Group. Effect of reducing interns’ weekly work hours on sleep and attentional failures. N Engl J Med . 2004; 351( 18): 1829– 1837. Google Scholar CrossRef Search ADS PubMed  4. Philibert I. Sleep loss and performance in residents and nonphysicians: a meta-analytic examination. Sleep . 2005; 28( 11): 1392– 1402. Google Scholar CrossRef Search ADS PubMed  5. Landrigan CP, Rothschild JM, Cronin JWet al.   Effect of reducing interns’ work hours on serious medical errors in intensive care units. N Engl J Med . 2004; 351( 18): 1838– 1848. Google Scholar CrossRef Search ADS PubMed  6. Baldwin DCJr, Daugherty SR, Tsai R, Scotti MJJr. A national survey of residents’ self-reported work hours: thinking beyond specialty. Acad Med . 2003; 78( 11): 1154– 1163. Google Scholar CrossRef Search ADS PubMed  7. Veasey S, Rosen R, Barzansky B, Rosen I, Owens J. Sleep loss and fatigue in residency training: a reappraisal. JAMA . 2002; 288( 9): 1116– 1124. Google Scholar CrossRef Search ADS PubMed  8. Goebert D, Thompson D, Takeshita Jet al.   Depressive symptoms in medical students and residents: a multischool study. Acad Med . 2009; 84( 2): 236– 241. Google Scholar CrossRef Search ADS PubMed  9. Ayas NT, Barger LK, Cade BEet al.   Extended work duration and the risk of self-reported percutaneous injuries in interns. JAMA . 2006; 296( 9): 1055– 1062. Google Scholar CrossRef Search ADS PubMed  10. Barger LK, Cade BE, Ayas NTet al.  ; Harvard Work Hours, Health, and Safety Group. Extended work shifts and the risk of motor vehicle crashes among interns. N Engl J Med . 2005; 352( 2): 125– 134. Google Scholar CrossRef Search ADS PubMed  11. Marcus CL, Loughlin GM. Effect of sleep deprivation on driving safety in housestaff. Sleep . 1996; 19( 10): 763– 766. Google Scholar CrossRef Search ADS PubMed  12. Wendt JR, Yen LJ. The resident by moonlight: a misguided missile. JAMA . 1988; 259( 1): 43– 44. Google Scholar CrossRef Search ADS PubMed  13. Steele MT, Ma OJ, Watson WA, Thomas HAJr, Muelleman RL. The occupational risk of motor vehicle collisions for emergency medicine residents. Acad Emerg Med . 1999; 6( 10): 1050– 1053. Google Scholar CrossRef Search ADS PubMed  14. Arnedt JT, Owens J, Crouch M, Stahl J, Carskadon MA. Neurobehavioral performance of residents after heavy night call vs after alcohol ingestion. JAMA . 2005; 294( 9): 1025– 1033. Google Scholar CrossRef Search ADS PubMed  15. MacLean AW, Davies DR, Thiele K. The hazards and prevention of driving while sleepy. Sleep Med Rev . 2003; 7( 6): 507– 521. Google Scholar CrossRef Search ADS PubMed  16. Anderson C, Horne JA. Sleepiness enhances distraction during a monotonous task. Sleep . 2006; 29( 4): 573– 576. Google Scholar CrossRef Search ADS PubMed  17. Anderson C, Horne JA. Driving drowsy also worsens driver distraction. Sleep Med . 2013; 14( 5): 466– 468. Google Scholar CrossRef Search ADS PubMed  18. Durmer JS, Dinges DF. Neurocognitive consequences of sleep deprivation. Semin Neurol . 2005; 25( 1): 117– 129. Google Scholar CrossRef Search ADS PubMed  19. Harrison Y, Horne JA. The impact of sleep deprivation on decision making: a review. J Exp Psychol Appl . 2000; 6( 3): 236– 249. Google Scholar CrossRef Search ADS PubMed  20. Akerstedt T, Hallvig D, Anund A, Fors C, Schwarz J, Kecklund G. Having to stop driving at night because of dangerous sleepiness–awareness, physiology and behaviour. J Sleep Res . 2013; 22( 4): 380– 388. Google Scholar CrossRef Search ADS PubMed  21. Ftouni S, Sletten TL, Howard Met al.   Objective and subjective measures of sleepiness, and their associations with on-road driving events in shift workers. J Sleep Res . 2013; 22( 1): 58– 69. Google Scholar CrossRef Search ADS PubMed  22. Howard ME, Jackson ML, Berlowitz Det al.   Specific sleepiness symptoms are indicators of performance impairment during sleep deprivation. Accid Anal Prev . 2014; 62: 1– 8. Google Scholar CrossRef Search ADS PubMed  23. Lee ML, Howard ME, Horrey WJet al.   High risk of near-crash driving events following night-shift work. Proc Natl Acad Sci U S A . 2016; 113( 1): 176– 181. Google Scholar CrossRef Search ADS PubMed  24. Watling CN, Armstrong KA, Smith SS, Wilson A. The on-road experiences and awareness of sleepiness in a sample of Australian highway drivers: A roadside driver sleepiness study. Traffic Inj Prev . 2016; 17( 1): 24– 30. Google Scholar CrossRef Search ADS PubMed  25. Nabi H, Guéguen A, Chiron M, Lafont S, Zins M, Lagarde E. Awareness of driving while sleepy and road traffic accidents: prospective study in GAZEL cohort. BMJ . 2006; 333( 7558): 75. Google Scholar CrossRef Search ADS PubMed  26. Ulmer C, Wolman D, Johns M. Committee on Optimizing Graduate Medical Trainee (Resident) Hours and Work Schedules to Improve Patient Safety for the Institute of Medicine. Resident Duty Hours: Enhancing Sleep, Supervision, and Safety . Washington, DC: The National Academies Press; 2008. 27. Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep . 1991; 14( 6): 540– 545. Google Scholar CrossRef Search ADS PubMed  28. Anderson C, Sullivan JP, Flynn-Evans EE, Cade BE, Czeisler CA, Lockley SW. Deterioration of neurobehavioral performance in resident physicians during repeated exposure to extended duration work shifts. Sleep . 2012; 35( 8): 1137– 1146. Google Scholar PubMed  29. Reason J, Manstead A, Stradling S, Baxter J, Campbell K. Errors and violations on the roads: a real distinction? Ergonomics . 1990; 33( 10-11): 1315– 1332. Google Scholar CrossRef Search ADS PubMed  30. Johns MW, Chapman R, Crowley K, Tucker A. A new method for assessing the risks of drowsiness while driving. Somnologie . 2008; 12(1): 66– 74. Google Scholar CrossRef Search ADS   31. Johns MW, Tucker A, Chapman R, Crowley K, Michael N. Monitoring eye and eyelid movements by infrared reflectance oculography to measure drowsiness in drivers. Somnologie . 2007; 11(4): 234– 42. Google Scholar CrossRef Search ADS   32. Anderson C, Chang AM, Sullivan JP, Ronda JM, Czeisler CA. Assessment of drowsiness based on ocular parameters detected by infrared reflectance oculography. J Clin Sleep Med . 2013; 9( 9): 907– 20, 920A. Google Scholar PubMed  33. Akerstedt T, Gillberg M. Subjective and objective sleepiness in the active individual. Int J Neurosci . 1990; 52( 1-2): 29– 37. Google Scholar CrossRef Search ADS PubMed  34. Kamphuis J, Meerlo P, Koolhaas JM, Lancel M. Poor sleep as a potential causal factor in aggression and violence. Sleep Med . 2012; 13( 4): 327– 334. Google Scholar CrossRef Search ADS PubMed  35. Basner M, Dinges DF, Shea JAet al.   Sleep and alertness in medical interns and residents. An observational study on the role of extended shifts. Sleep . 2017; 40( 4):zsx027. 36. Achermann P, Werth E, Dijk DJ, Borbely AA. Time course of sleep inertia after nighttime and daytime sleep episodes. Arch Ital Biol . 1995; 134( 1): 109– 119. Google Scholar PubMed  37. Horne JA, Reyner LA. Counteracting driver sleepiness: effects of napping, caffeine, and placebo. Psychophysiology . 1996; 33( 3): 306– 309. Google Scholar CrossRef Search ADS PubMed  38. Mongrain S, Standing L. Impairment of cognition, risk-taking, and self-perception by alcohol. Percept Mot Skills . 1989; 69( 1): 199– 210. Google Scholar CrossRef Search ADS PubMed  39. Van Dongen HP, Maislin G, Mullington JM, Dinges DF. The cumulative cost of additional wakefulness: dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep . 2003; 26( 2): 117– 126. Google Scholar CrossRef Search ADS PubMed  40. Kaplan KA, Itoi A, Dement WC. Awareness of sleepiness and ability to predict sleep onset: can drivers avoid falling asleep at the wheel? Sleep Med . 2007; 9( 1): 71– 79. Google Scholar CrossRef Search ADS PubMed  41. Filtness AJ, Anund A, Fors C, Ahlström C, Akerstedt T, Kecklund G. Sleep-related eye symptoms and their potential for identifying driver sleepiness. J Sleep Res . 2014; 23( 5): 568– 575. Google Scholar CrossRef Search ADS PubMed  42. Alvaro PK, Jackson ML, Berlowitz DJ, Swann P, Howard ME. Prolonged eyelid closure episodes during sleep deprivation in professional drivers. J Clin Sleep Med . 2016; 12( 8): 1099– 1103. Google Scholar CrossRef Search ADS PubMed  43. Jackson ML, Kennedy GA, Clarke Cet al.   The utility of automated measures of ocular metrics for detecting driver drowsiness during extended wakefulness. Accid Anal Prev . 2016; 87: 127– 133. Google Scholar CrossRef Search ADS PubMed  44. Horne JA, Baulk SD. Awareness of sleepiness when driving. Psychophysiology . 2004; 41( 1): 161– 165. Google Scholar CrossRef Search ADS PubMed  45. Sagaspe P, Taillard J, Akerstedt Tet al.   Extended driving impairs nocturnal driving performances. PLoS One . 2008; 3( 10): e3493. Google Scholar CrossRef Search ADS PubMed  46. Philip P, Sagaspe P, Moore Net al.   Fatigue, sleep restriction and driving performance. Accid Anal Prev . 2005; 37( 3): 473– 478. Google Scholar CrossRef Search ADS PubMed  47. West R, French D, Kemp R, Elander J. Direct observation of driving, self reports of driver behaviour, and accident involvement. Ergonomics . 1993; 36( 5): 557– 567. Google Scholar CrossRef Search ADS PubMed  48. Parker D, Reason JT, Manstead ASR, Stradling SG. Driving errors, driving violations and accident involvement. Ergonomics . 1995; 38( 5): 1036– 1048. Google Scholar CrossRef Search ADS PubMed  49. Czeisler CA. The Gordon Wilson Lecture: work hours, sleep and patient safety in residency training. Trans Am Clin Climatol Assoc . 2006; 117: 159– 188. Google Scholar PubMed  © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png SLEEP Oxford University Press

Self-reported Drowsiness and Safety Outcomes While Driving After an Extended Duration Work Shift in Trainee Physicians

Loading next page...
 
/lp/ou_press/self-reported-drowsiness-and-safety-outcomes-while-driving-after-an-jqUX0yzgJC
Publisher
Oxford University Press
Copyright
© Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail journals.permissions@oup.com.
ISSN
0161-8105
eISSN
1550-9109
D.O.I.
10.1093/sleep/zsx195
Publisher site
See Article on Publisher Site

Abstract

Abstract Study Objectives Extended duration (≥24 hours) work shifts (EDWSs) are associated with increased risk of motor vehicle crashes, and awareness of any impairment has important implications on legal accountability for any adverse driving outcome. The extent to which adverse driving events were preceded by predrive self-reported sleepiness was evaluated in medical residents after an EDWS. Methods Sixteen resident physicians (10 females; 29.2 ± 2.0 years) working EDWS were monitored when driving on their commute to and from the hospital (438 drives). Sleep and work hours were obtained from daily logs, and adverse driving outcomes were captured from a driving log completed at the end of each drive. Self-reported sleepiness (Karolinska Sleepiness Scale; KSS) and objective drowsiness were captured using a time-stamped, hand-held device and infra-red reflectance oculography. Results Self-reported sleepiness and objective indices of drowsiness were positively correlated, and both were elevated following EDWSs. Compared with the commute to work, EDWSs were associated with more than double the self-reported adverse outcomes when driving home, significantly higher than drives to or from the day shift at comparable times of day. EDWSs more than tripled the odds of reporting sleep-related, inattentive, hazardous, or violation-driving events. The number and type of adverse event was predicted by the predrive KSS level and in a dose-dependent manner. Conclusions Driving after an EDWS puts resident physicians/drivers and other road users at avoidable and unnecessary risk. Demonstrating self-reported sleepiness at the beginning of the drive is associated with adverse outcomes has serious implications on the legal accountability for driving when drowsy. drowsiness, sleepiness, driving, safety, residents, shift work Statement of Significance Driving while drowsy represents a significant risk to fatal or serious car crashes and is attributed to approximately 20% of motor vehicle crashes. One strategy for minimizing the risk of drowsy driving is for drivers to acknowledge their level of drowsiness and to take evasive action. The extent to which drivers can accurately self-assess their drowsiness has been of continued debate in the field, however. Our study provides clear evidence that, before driving home from a 24-hour shift, drivers are aware of their drowsiness level which predicts subsequent on-road adverse driving events. Drivers should therefore proactively evaluate their drowsiness level prior to driving and not expose themselves and others to the significant, and avoidable, risk of drowsy driving. Introduction In the United States, medical and surgical trainee physicians beyond their first year of residency are permitted to work for up to 28 hours continuously and are routinely scheduled to work extended duration work shifts (EDWSs) of 24 hours or more twice per week [1]. Since 2011, such marathon shifts were not permitted for first-year residents (“interns”), but the Accreditation Council for Graduate Medical Education (ACGME) proposes to reintroduce them in 2017 [2]. Although these EDWSs present a significant risk to patient safety due to increased attentional failures [3], risk of errors on clinical task [4], and higher rates of serious medical errors [5, 6], they also negatively affect the health and safety of the trainee physicians themselves [7]. Although physician risk may involve increased rates of depression [8] or needle stick injury [9], the most documented risk for trainee physicians is the risk of a motor vehicle crash [10–13]. It is well documented that these EDWSs negatively affect the safety of trainee physicians by increasing the risk of motor vehicle crashes on the commute home [10–13]. For instance, trainee physicians working a heavy call schedule exhibit deficits on a simulated driving task comparable to performance at 0.04%–0.05% blood alcohol concentration [14] due to a combination of excessive time awake, overnight work, and chronic sleep deficiency. Notably, the odds of a trainee physician being involved in a motor vehicle crash or near-crash after an EDWS are increased by 2.3 and 5.9, respectively, compared with a nonextended shift [10]. These motor vehicle (near) crashes may not be directly attributed to having fallen asleep (although previous studies suggest 49% of trainee physicians admitted to having falling asleep while driving [11], compared with 11%–31% of the nonmedical drivers [see Ref. 15]). Instead, and based on laboratory-evidence, other causal factors for drowsy crashes may include enhanced distractibility [16, 17], reduced vigilance [18], and poor decision making [19]. Despite the prevalence of driving while drowsy in trainee physicians, little is known about the impact of EDWSs on safety outcomes while driving. Alleviating crash risk in resident physicians (and all road users) requires a multifaceted approach. Firstly, understanding causal factors of drowsy driving crashes is essential to informing intervention strategies for reducing crash risk. Secondly, and perhaps most importantly, drivers should be able to determine when they have reached a level of drowsiness associated with increased risk of adverse driving outcomes. Although it is now largely accepted that drivers are typically aware when they have reached a level of sleepiness that is dangerous for driving [20–24], they often continue to drive [25]. Ideally then, drivers should actively evaluate whether they are safe to drive before commencing the journey. To the best of our knowledge, no study has examined self-reported sleepiness at the onset of the drive and the utility of this approach to predict the risk of adverse events on the subsequent drive. Our study objective was therefore to first characterize adverse driving outcomes and examine the prevalence of objective drowsiness and self-reported sleepiness in trainee physicians driving to and from the hospital, and secondly to examine the extent to which a predrive self-assessment of sleepiness accurately predicts adverse driving outcomes. Methods Participants Sixteen resident physicians (10 females; 29.2 ± 2.0 years; range 24.5–33.4 years; Table 1) from seven hospitals in the Boston metropolitan area, and who were experienced drivers (licenced >2 years; drove >50 miles per week) driving >15 minutes alone to and from work, provided a total of 438 driving sessions (Figure 1). Trainee physicians were recruited from four clinical disciplines including internal medicine, surgery, anaesthesiology, and pediatric medicine. Advertisements were placed around the main teaching hospitals in the Boston metropolitan region and emailed to interns and residents via the Committee of Interns and Residents. Interested individuals (n = 192) contacted the study coordinator (CA) and were provided with additional information about the study. For those interested in the study (n = 69), an initial phone interview determined work and driving commitments to ensure that they were within the limits for inclusion criteria. Participants, who commuted to and from the hospital, worked EDWSs, and with a roster that was consistent with the requirements of the study (i.e., a period of noncall plus six consecutive EDWSs), were invited to take part in the full study. Participant demographic, driving experience, residency experience, and presence of sleep or medical problems are reported in Table 1. The study was approved by the Human Research Committee of Partners Healthcare and all participants provided written informed consent and were remunerated for their involvement. Table 1. Demographics and Work/Sleep Parameters in Trainee Physicians Variable        Range  N    16  —  Males  n (%)  6 (37.5%)  —  Age (y)  Mean ± SD  29.2 ± 2.0  24.6–33.4 y  Year of residency  n (%)       First    9 (56.3%)     Second    4 (25%)     Third    1 (6.3%)     Fourth    2 (12.5%)    Clinical disciplines  n (%)       Internal medicine    12 (75%)     Surgery    2 (12.5%)     Anaesthesiology    1 (6.25%)     Pediatric medicine    1 (6.25%)    Body mass index (kg m−2)  Mean ± SD  24.2 ± 3.0  18.3–30.8  Body mass index >30  n (%)  1 (6.25%)  —  Diagnosed illness/disorder  n (%)a  6 (37.5%)  —  Epworth Sleepiness Score  Mean ± SD  10.8 ± 4.8  3–18  Sleep need  Mean ± SD  7.5 ± 1.0  6–10 h  Fall asleep events in past month, when  Telephone  Mean ± SD (%)b  1.4 ± 1.8 (62.5%)  0–6  Meeting  Mean ± SD (%)b  0.07 ± 0.25 (12.5%)  0–1  Traffic  Mean ± SD (%)b  0.33 ± 0.49 (37.5%)  0–1  Driving  Mean ± SD (%)b  0.87 ± 1.77 (31.25%)  0–5  Driving experience (y)  Mean ± SD  11.6 ± 2.3  6-14 y  Miles driven per week  Mean ± SD  105.8 ± 56.0  50–250 miles  Commute time  Mean ± SD  26.7 ± 10.3  15.8–56.8 min  Duration of residency training (mo)  Mean ± SD  18.6 ± 19.5  5–80 mo  Extended work shift duration  Mean ± SD (d.h)  28.3 ± 1.85    Day shift duration  Mean ± SD (d.h)  9.62 ± 2.6    Sleep duration/24-h period   Prestudy (with napsc)  Mean ± SD (d.h)  9.11 ± 0.81     Prestudy (without naps)  Mean ± SD (d.h)  7.9 ± 0.86     Preday shift (with naps)  Mean ± SD (d.h)  8.43 ± 2.17     Preday shift (without naps)  Mean ± SD (d.h)  6.67 ± 0.79     Pre-extended shift (with naps)  Mean ± SD (d.h)  6.97 ± 1.33     Pre-extended shift (without naps)  Mean ± SD (d.h)  7.45 ± 1.42     During extended shift (with naps)  Mean ± SD (d.h)  1.88 ± 1.99     During extended shift (without naps)  Mean ± SD (d.h)  2.32 ± 1.96    Variable        Range  N    16  —  Males  n (%)  6 (37.5%)  —  Age (y)  Mean ± SD  29.2 ± 2.0  24.6–33.4 y  Year of residency  n (%)       First    9 (56.3%)     Second    4 (25%)     Third    1 (6.3%)     Fourth    2 (12.5%)    Clinical disciplines  n (%)       Internal medicine    12 (75%)     Surgery    2 (12.5%)     Anaesthesiology    1 (6.25%)     Pediatric medicine    1 (6.25%)    Body mass index (kg m−2)  Mean ± SD  24.2 ± 3.0  18.3–30.8  Body mass index >30  n (%)  1 (6.25%)  —  Diagnosed illness/disorder  n (%)a  6 (37.5%)  —  Epworth Sleepiness Score  Mean ± SD  10.8 ± 4.8  3–18  Sleep need  Mean ± SD  7.5 ± 1.0  6–10 h  Fall asleep events in past month, when  Telephone  Mean ± SD (%)b  1.4 ± 1.8 (62.5%)  0–6  Meeting  Mean ± SD (%)b  0.07 ± 0.25 (12.5%)  0–1  Traffic  Mean ± SD (%)b  0.33 ± 0.49 (37.5%)  0–1  Driving  Mean ± SD (%)b  0.87 ± 1.77 (31.25%)  0–5  Driving experience (y)  Mean ± SD  11.6 ± 2.3  6-14 y  Miles driven per week  Mean ± SD  105.8 ± 56.0  50–250 miles  Commute time  Mean ± SD  26.7 ± 10.3  15.8–56.8 min  Duration of residency training (mo)  Mean ± SD  18.6 ± 19.5  5–80 mo  Extended work shift duration  Mean ± SD (d.h)  28.3 ± 1.85    Day shift duration  Mean ± SD (d.h)  9.62 ± 2.6    Sleep duration/24-h period   Prestudy (with napsc)  Mean ± SD (d.h)  9.11 ± 0.81     Prestudy (without naps)  Mean ± SD (d.h)  7.9 ± 0.86     Preday shift (with naps)  Mean ± SD (d.h)  8.43 ± 2.17     Preday shift (without naps)  Mean ± SD (d.h)  6.67 ± 0.79     Pre-extended shift (with naps)  Mean ± SD (d.h)  6.97 ± 1.33     Pre-extended shift (without naps)  Mean ± SD (d.h)  7.45 ± 1.42     During extended shift (with naps)  Mean ± SD (d.h)  1.88 ± 1.99     During extended shift (without naps)  Mean ± SD (d.h)  2.32 ± 1.96    aCurrent illnesses reported included high blood pressure (unmedicated), depression (medicated), gastrointestinal problems (medicated [n = 2] and unmedicated [n = 1]), and insomnia (unmedicated). b% with an affirmative response for OSA or fall asleep event in the last month. cnap defined as sleep <20 minutes in duration. View Large Figure 1. View largeDownload slide Flow chart of physicians’ participation. For those screened to fit within our shift and drive requirements, 75% were not eligible for the study largely due to not driving and/or living too close to the hospital (55%). Of those screened, 25% consented to take part, resulting in 23% completing the study. Figure 1. View largeDownload slide Flow chart of physicians’ participation. For those screened to fit within our shift and drive requirements, 75% were not eligible for the study largely due to not driving and/or living too close to the hospital (55%). Of those screened, 25% consented to take part, resulting in 23% completing the study. Design We monitored trainee residents during the commute to and from work, for six consecutive EDWSs and all intervening non-EDWSs. Using a repeated measures design, we compared self-reported sleepiness, objective drowsiness, and self-reported driving outcomes when driving to and from an EDWS and non-EDWS. Shift Schedules In the United States, a trainee physician (beyond their second year of residency) can be scheduled to work EDWSs of 24–28 consecutive hours (24 + 4) within a 80 to 88–hour work week, averaged over 4 weeks [26]. Our study included trainee physicians working EDWSs every third (Q3), fourth (Q4), or fifth (Q5) shift. In between individuals were scheduled to work a “noncall” shift (~06:00 am until ~03:00–05:00 pm) or have a day off. The study ended after six EDWSs, which were completed over 3 to 5 weeks. Methods During baseline assessments, we obtained information on participants’ demographics, shift schedules, driving experience, and trait levels of daytime sleepiness using the Epworth Sleepiness Scale [27]. For 1 week prior to study, participants worked on a noncall rotation without night duty or EDWS (n = 11) or were on a scheduled week of vacation (n = 5). This ensured adequate opportunity for rest prior to taking part in the study. During this prestudy phase, hours of work and rest, and sleep outcomes were monitored via actigraphy (Motionlogger, Ambulatory Monitoring, Inc., Ardsley, NY) and daily sleep and work logs. During the main study, we obtained data relating to sleep, work hours, and driving outcomes over six consecutive EDWSs and all intermediary day shifts (as per our previous studies [3, 5, 28]), using validated daily sleep [3] and work logs and driving logs completed after each commute to and from the hospital (Supplementary Material 1). Sleep diaries comprised 14 questions on sleep timing, duration, disturbance, and alcohol or medication use. Participants also completed the start and end dates and times of any shift worked in the last 24 hours. The driving log contained 18 items including information on the type and duration of commute, incidence and information of near-crashes or actual crash events, adverse driving events, countermeasure to sleepiness actions, and other information relating to the drive (i.e., traffic density, weather, and consumption of caffeine/alcohol). Participants selected as many driving events and/or countermeasure actions as applicable. The diary is available in Supplementary Material. From the driving log, self-reported driving events were categorized according to criteria adapted from the work of Reason [21, 29]: (1) sleep-related (i.e., “falling asleep at a stop light”), (2) inattention (i.e., “being distracted”), (iii) hazardous (i.e., “swerving violently”), and (iv) violation (i.e., “driving through a stop light”). Objective measures of drowsiness (Johns Drowsiness Score; JDS) were recorded while driving using infrared reflectance oculography (Optalert™ Drowsiness Measurement System, Sleep Diagnostics Pty Ltd., Melbourne, Australia [30, 31]). This approach captures information from involuntary eye blinks, such as amplitude/velocity ratio, long eye closures, and blink duration. The ocular parameters are then combined via a proprietary algorithm to derive the JDS, which compared with other ocular parameters from which the composite JDS is derived, is a more sensitive marker of drowsiness [32], and has been previously associated with driving performance in real on road driving [21]. Participants completed an electronic, time-stamped Karolinska Sleepiness Scale (KSS) [33] at the beginnng and end of each commute. Data Analysis We categorized all data to either a “day shift”—for shifts starting at 06:00–07:00 am of <16 hours duration—or an “EDWS”—for shifts ≥24 hours duration (referred to as EDWS throughout). Commutes to the hospital are referred to as “preshift drive” and the commute home as “postshift drive.” We averaged data within an individual for each pre- and postshift drive for day shifts and extended shifts separately, and compared using repeated measures two-way ANOVA (shift*commute). To examine the likelihood of reporting each of the four driving event categories ((1) sleep-related, (2) inattention, (3) hazardous, and (4) violation events) during preshift drives compared with postshift drives, we employed a chi square relative risk test using a within-subject comparison for day shift and EDWSs separately. We obtained KSS scores from PDA (Personal Digital Assistants) devices logged at the beginning and end of each commute. Where there were missing KSS data on the PDA devices, these were replaced with scores reported in the driving log at the beginning and end of each drive. A total of 56 (13%) KSS scores were derived from driving logs. Taking all data, driving log and PDA-derived KSS scores were highly similar (KSS start of drive, r = 0.84, p < .001; KSS end of drive, r = 0.86, p < .001). To examine the impact of shift, we averaged KSS within subject and then within-shift (day shifts and extended work shifts) for pre- and postshift drives and compared using a two-way (2 × 2) ANOVA (shift*commute). For objective drowsiness, we first removed any commute without 10 consecutive data points (minutes) from the analysis due to reduced reliability of the output. We then visually inspected all data for artifact. Each EDWS was paired with a preceding or subsequent day shift within each individual for comparison and removed any participant without a matching day and EDWS (n = 8 remaining). We then used a two-way (2*2) ANOVA (shift*commute) to examine the impact of shift on ocular-derived drowsiness outcomes. To verify self-reported outcomes, we examined the association between self-reported sleepiness and objective markers of drowsiness using Pearson correlations. To evaluate the predictive capacity of predrive KSS, we performed binary logistic regression where KSS was the predictor variable (continuous) and self-reported adverse driving event the outcome variable. This analysis was conducted separately for each event category. Goodness of fit was determined by Omnibus Tests of Model Coefficients. In addition, to assess predictive capacity of KSS for subsequent adverse events, we constructed receiver–operator characteristic (ROC) curves and performed area under the curve (AUC) analyses. Pretest probability was estimated empirically from the data, which was p = .41 for all events, p = .23 for sleep-related events, and p = .17 for inattentive events. All cost-ratios were set at 1.0. These values were used to determine optimal cutoff points. Results Data Obtained Fourteen trainee physicians completed the full study (six EDWSs) and two physicians completed only four EDWSs (study finished early due to sick leave [n = 1] and poor weather conditions [n = 1]). We obtained data from 438 drives: 218 (49.8%) commuting to the hospital and 220 (50.2%) commuting home after the work shift and 241 (55.0%) for day shifts and 197 (45.0%) for EDWSs. Of the 438 drives monitored with driving logs, 289 were also monitored with the objective drowsiness detection system (67%). Although individual reasons for lost oculography data were not reported by participants, these included a failure of the technology to record, a loss of battery power, or participants not recording the drive. Following visual quality checks of the data, only 182/289 (63%) were deemed of sufficient quality for analysis. Although the majority of drives were rejected for multiple signal artifact, individual reasons for rejection included a biphasic signal (n = 54), clipping of the signal (n = 14), blinks not clearly identifiable (n = 95), noise in the signal (n = 42), an inverted signal (n = 10), and a small signal (n = 93). We then omitted a further 20 drives due to failing to meet the criteria of at least 10 consecutive minutes of monitoring (6.9%). The final number of drives containing objective drowsiness data was 162 (56%) and was equally distributed across day and EDWS (87/46.3% vs. 75/53.7%, respectively; z = 0.858, p = .39) and pre- and postshift drives (73/45.1% vs. 89/54.9%, respectively; z = 1.109, p = .27). Sleep and Work Scheduling Sleep and work-scheduling data can be seen in Table 1. For an EDWS, the proportion of individuals with an average KSS score of 6 or more (some signs of sleepiness) on the drive home was 75%, compared with 12% of the drive to work. The proportion of individuals with an average score of 7+ (moderate-severe sleepiness) was 44% on the drive home, compared with 12% on the drive to work. Self-reported sleepiness was elevated by 46% on the commute home after an EDWS (p < .0001), compared with only 1.26% after a day shift. Likewise, objective measures of drowsiness were also elevated on the drive home, but only after an EDWS (p = .012). Here, JDS increased by 48% when driving home after an EDWS, compared with a 25% decrease after a day shift. Predrive self-reported sleepiness assessments and subsequent objective drowsiness levels were highly correlated (r = 0.81, p = .003) (Figure 2). Figure 2. View largeDownload slide Self-reported sleepiness and objective drowsiness while driving to and from the hospital on day versus EDWSs: self-reported ratings of sleepiness [(A) Day shift: 3.98 ± 0.42 to 4.03 ± 0.40; EDWS: 4.08 ± 0.41 to 5.97 ± 0.40] and objective ocular-derived measures of drowsiness [(B) 1.05 ± 0.23 to 0.79 ± 0.15; EDWS: 0.97 ± 0.12 to 1.43 ± 0.31] were elevated when driving home after an EDWS. Self-reported sleepiness prior to the drive was positive associated with subsequent objective measures of drowsiness (C). Figure 2. View largeDownload slide Self-reported sleepiness and objective drowsiness while driving to and from the hospital on day versus EDWSs: self-reported ratings of sleepiness [(A) Day shift: 3.98 ± 0.42 to 4.03 ± 0.40; EDWS: 4.08 ± 0.41 to 5.97 ± 0.40] and objective ocular-derived measures of drowsiness [(B) 1.05 ± 0.23 to 0.79 ± 0.15; EDWS: 0.97 ± 0.12 to 1.43 ± 0.31] were elevated when driving home after an EDWS. Self-reported sleepiness prior to the drive was positive associated with subsequent objective measures of drowsiness (C). There was a three-fold increase in the average number of self-reported adverse driving events on the drive home from an EDWS compared with driving to work (1.1 ± 0.78 versus 0.37 ± 0.49 events/drive, p < .005, respectively), much more than the increase associated with driving to and from a day shift (0.35 ± 0.53 versus 0.59 ± 0.63 events/drive, respectively; Figure 3). Compared with driving to work, the odds (odds ratio [OR], 95% confidence interval [CI]) of reporting an adverse event on the drive from an EDWS was 3.90 (1.73–8.80, p = .001) for sleep-related events, 3.58 (1.72–7.47, p < .0001) for inattention events, 3.32 (1.16–9.53, p = .02) for hazardous events, and 5.39 (1.15–25.29, p = .018) for violation events. In comparison, after working a day shift, there were no increased odds of self-reporting an adverse event (p > .1) (Table 2). Table 2. The Odds Ratio of a Drive Postshift Containing a Specific Driving Event Compared With the Drive Preshift: Day Shifts and EDWSs     Day shift  Extended work shift  Preshifta  Postshiftb  χ2  OR  95% CI  p  Preshiftc  Postshiftd  χ2  OR  95% CI  p  Sleep- related  Resting the eyes  31 (25.8%)  31 (25.6%)  0.00  0.99  0.56–1.76  .97  9 (9.2%)  28 (28.3%)  11.78  3.9  1.73–8.8  .001  Asleep stop light  Inattention  Fixation  11 (9.2%)  16 (13.2%)  1.00  1.51  0.67–3.41  .32  12 (12.2%)  34 (33.3%)  12.43  3.58  1.72–7.47  <.0001  Lack awareness  Distracted  Missed turn  Hazardous  Braked sharply  9 (7.5%)  13 (10.7%)  0.76  1.49  0.61–3.62  .38  5 (5.1%)  15 (15.2%)  5.45  3.32  1.16–9.53  .02  Hit rumble strip  Swerved violently  Near Miss  Violation  Shouting  1 (0.8%)  5 (4.1%)  2.7e  5.13  0.59–44.58  .10  2 (2.0%)  10 (10.1%)  5.59e  5.39  1.15–25.29  .018  Driving through stop light      Day shift  Extended work shift  Preshifta  Postshiftb  χ2  OR  95% CI  p  Preshiftc  Postshiftd  χ2  OR  95% CI  p  Sleep- related  Resting the eyes  31 (25.8%)  31 (25.6%)  0.00  0.99  0.56–1.76  .97  9 (9.2%)  28 (28.3%)  11.78  3.9  1.73–8.8  .001  Asleep stop light  Inattention  Fixation  11 (9.2%)  16 (13.2%)  1.00  1.51  0.67–3.41  .32  12 (12.2%)  34 (33.3%)  12.43  3.58  1.72–7.47  <.0001  Lack awareness  Distracted  Missed turn  Hazardous  Braked sharply  9 (7.5%)  13 (10.7%)  0.76  1.49  0.61–3.62  .38  5 (5.1%)  15 (15.2%)  5.45  3.32  1.16–9.53  .02  Hit rumble strip  Swerved violently  Near Miss  Violation  Shouting  1 (0.8%)  5 (4.1%)  2.7e  5.13  0.59–44.58  .10  2 (2.0%)  10 (10.1%)  5.59e  5.39  1.15–25.29  .018  Driving through stop light  aData are based on 120 drives. bData are based on 121 drives. cData are based on 98 drives. dData are based on 99 drives. eFishers exact test reported when number of events falls below 5. View Large Figure 3. View largeDownload slide Adverse drive events while driving to and from the hospital on day shifts versus EDWSs. Compared with preshift, more adverse events were reported postshift for the EDWS (0.37 ± 0.49 and 1.1 ± 0.78). This was less evident for day shift (0.35 ± 0.53 and 0.59 ± 0.63), despite being at a comparable time of day. Refer (A) and (B) for group and individual level data, respectively. Importantly, increasing KSS was associated with an increased average number of adverse driving events in a dose response manner, at the group (C) and individual level (D). Figure 3. View largeDownload slide Adverse drive events while driving to and from the hospital on day shifts versus EDWSs. Compared with preshift, more adverse events were reported postshift for the EDWS (0.37 ± 0.49 and 1.1 ± 0.78). This was less evident for day shift (0.35 ± 0.53 and 0.59 ± 0.63), despite being at a comparable time of day. Refer (A) and (B) for group and individual level data, respectively. Importantly, increasing KSS was associated with an increased average number of adverse driving events in a dose response manner, at the group (C) and individual level (D). Although no crashes occurred, 10 near-crash events were reported (2.3% of total drives) from 6/16 trainee physicians (38%): one from a day shift, one while commuting to an EDWS, and eight while driving home after an EDWS (8.2% of commutes from EDWSs). Descriptive evaluation of these near-crash events can be seen in Table 3. Table 3. Descriptive Data for Near-Crash Events Shift  Pre- or post shift  Time of day  Amount of sleep in previous 24 h  Time since awake (h)  Description  Day  Postshift  17:25  6  10  “Other car went through a red light”  Extended  Preshift  06:25  8  1  “Fast moving car in left lane, tried to change lanes but nearly rear-ended”  Extended  Postshift  14:30  2  9  “I was merging into traffic and dozed off and almost hit another car”  Extended  Postshift  12:50  0  30  “Eyes closed on freeway, veared 1/4 way into adjacent lane, had it been a busier time of day would have probably hit someone”  Extended  Postshift  12:25  0.5  36  “Other than 30 min, been awake 36 h…Clear day, someone ran a red light in Boston and I had to slam on brakes to avoid”  Extendeda  Postshift  13:45  0  30  “Had four near misses involving nodding off while driving”  Extended  Postshift  12:45  8  7  “Driving in wet but then another car crossed over into the lane without signaling I was able to slow down in time”  Shift  Pre- or post shift  Time of day  Amount of sleep in previous 24 h  Time since awake (h)  Description  Day  Postshift  17:25  6  10  “Other car went through a red light”  Extended  Preshift  06:25  8  1  “Fast moving car in left lane, tried to change lanes but nearly rear-ended”  Extended  Postshift  14:30  2  9  “I was merging into traffic and dozed off and almost hit another car”  Extended  Postshift  12:50  0  30  “Eyes closed on freeway, veared 1/4 way into adjacent lane, had it been a busier time of day would have probably hit someone”  Extended  Postshift  12:25  0.5  36  “Other than 30 min, been awake 36 h…Clear day, someone ran a red light in Boston and I had to slam on brakes to avoid”  Extendeda  Postshift  13:45  0  30  “Had four near misses involving nodding off while driving”  Extended  Postshift  12:45  8  7  “Driving in wet but then another car crossed over into the lane without signaling I was able to slow down in time”  aDescription of four events in one commute home. View Large Predrive KSS was a significant predictor of subsequent driving events in a dose-dependent manner (Figure 3). For each one-point increase in predrive KSS, there was a 2.39 times increase in the odds of reporting a sleep-related adverse event during the subsequent drive (95% CI: 1.98–2.89, p < .0001). The odds of reporting an inattention or hazardous event also increased per each single-point KSS rise [1.33 (1.13–1.56) and 1.20 (1.06–1.36), respectively] (Table 4). Receiver–operator curves demonstrated that a predrive KSS rating of ≥6 (“some signs of sleepiness”) had 91% sensitivity and 69% specificity for predicting a sleep-related adverse driving event (AUC = 0.86; p < .001). Table 4. Binary Logistic Regression Examining the Predictive Capacity of Predrive KSS on Subsequent Adverse Driving Events   Goodness of fit  Binary logistic regression parameters  χ2  Sig.  B(S.E.)  Wald  df  Exp(B)  Sig.  Upper  Sleep-related  145.5  <.0001  0.87 (0.10)  82.06  1  2.39  <.0001  1.98–2.89  Inattention  8.8  .003  0.18 (0.06)  8.53  1  1.20  .003  1.06–1.36  Hazardous  12.87  <.0001  0.28 (0.08)  11.57  1  1.33  .001  1.13–1.56  Violation  0.01  .92  −0.01 (0.11)  0.01  1  0.99  .025  0.80–1.23    Goodness of fit  Binary logistic regression parameters  χ2  Sig.  B(S.E.)  Wald  df  Exp(B)  Sig.  Upper  Sleep-related  145.5  <.0001  0.87 (0.10)  82.06  1  2.39  <.0001  1.98–2.89  Inattention  8.8  .003  0.18 (0.06)  8.53  1  1.20  .003  1.06–1.36  Hazardous  12.87  <.0001  0.28 (0.08)  11.57  1  1.33  .001  1.13–1.56  Violation  0.01  .92  −0.01 (0.11)  0.01  1  0.99  .025  0.80–1.23  View Large Discussion We examined self-reported and objective measures of drowsiness in US trainee physicians when driving to and from the hospital after day shifts (<12 hours) and EDWSs (>24 hours). In our study, driving home after working an EDWS was associated with an approximately 40% increase in self-reported sleepiness at the start of the drive home, elevated levels of objectively measured drowsiness while driving, and a greater number of adverse driving events. Self-reported sleepiness at the beginning of the drive was associated with objective drowsiness during the drive, and both the number and type of adverse driving events reported. Currently, resident and attending physicians are permitted to work >24-hour shifts [1], which have adverse implications for the physician’s safety while driving [10]. Half of trainee physicians admit to having fallen asleep while driving [11], compared with 11%–31% of typical drivers not working EDWS [15]; this is consistent with our observation of nearly four times greater odds of reporting a sleep-related adverse driving event on the drive home after an EDWS. Sleep deficiency also manifests in other behaviors, such as increased distractibility [16, 17], aggression and irritability [34], and poor decision making [19]. Consistent with this wider scope of drowsy driving symptomology, trainee physicians driving home after an EDWS in the current study had 3–5 times greater odds of reporting an inattentive, hazardous, or violation-driving event. Although inattention was comprised of drivers reporting being fixated, missing the turn, being distracted, and having a lack of awareness, our findings of increased inattention were largely due to being distracted and having a lack of awareness. These data are in agreement with our previous simulated driving study that showed drivers were more distracted when drowsy. Importantly, this inattentive behavior was associated with an increased likelihood of the car leaving the driving lane [17]. Three quarters of trainee physicians reported a KSS of 6+ (“some signs of sleepiness”), whereas 44% reported a KSS of 7+ (moderate–severe sleepiness), prior to driving home after an EDWS. Given that the levels of sleepiness reported by the interns and residents in our study was similar to that reported elsewhere for studies of comparable demographic [35] and design [23], there is no evidence to suggest that our sample represents an unusual population. Our data revealed a strong dose-dependent association between the predrive KSS and the incidence of sleep-related adverse events reported during the subsequent drive home after an EDWS. These data suggest that trainee physicians’ self-assessment of their sleepiness before driving can be a powerful and reliable warning indicator of driving impairment. This is consistent with our previous study assessing on-road driving impairment following a night shift; although night shift workers reported an average KSS of 6 at the onset of the drive, the drive was associated with increasing KSS, use of emergency braking manoeuvres (on 38% of postnight shift drives, PNSD), a failure to adequately remain in control over the vehicle (44% of PNSD), and the potential for falling asleep at the wheel [23]. A KSS score of 6 or more should therefore be considered a warning indicator for high risk of drowsy driving and should prompt individuals to avail themselves of alternative transportation or to deploy appropriate countermeasures such as sleep at the hospital before driving home (with adequate time upon wakening to dissipate any sleep inertia [36]), or caffeine [37]. Although alternative transport was available to the physicians in the current study through hospital-funded taxi-cab schemes, only two participants reported taking a taxi during the study suggesting as follows: (1) despite being aware of feeling sleepy, there was a failure to associate this with the risk of adverse safety outcomes while driving, similar to the impairment of risk assessment under the influence of alcohol [38] or (2) inadequate education or reinforcement/encouragement to use the service. Despite trainee physicians reporting elevated levels of sleepiness after an EDWS, they continue to drive with adverse safety outcomes for themselves and other road users. Much debate exists on the utility of self-reported sleepiness in reducing motor vehicle crash risk. Previous studies showing a lack of agreement between self-reported and objective indices of sleepiness have focused on either individuals in a laboratory-based, nondriving scenarios (e.g., Ref. 39) or examined sleepiness levels just prior to a driving event (when introspection into sleepiness is likely to be low) [40]. Laboratory-based simulator studies have consistently argued that self-reported sleepiness is associated with either objective indices of drowsiness [22, 41–44] or impaired driving performance [22, 44]. Although driving simulator studies have provided much evidence for the agreement between self-reported and objective measures, real on-road studies provide the most ecological valid evidence for evaluating the association between self-reported sleepiness and objective physiological (e.g., slow eye closures) and behavioral (e.g., adverse driving event) indices of drowsiness. These studies have demonstrated consistently that drivers are aware of increasing sleepiness while driving, which occurs in parallel with increased adverse outcomes [20, 21, 23, 45, 46].. Of relevance to our study, self-reported ratings prior to the drive has been associated with subsequent driving events previously. For instance, in sleep-restricted individuals driving at night on a French highway, increasing self-reported sleepiness prior to the drive was associated with more lane crossing in the subsequent drive [46], our study of night shift workers found a predrive KSS of 6 at was associated with adverse driving outcomes (emergency braking manoeuvres required, failure to adequately remain in control over the vehicle, and falling asleep at the wheel) [23], and our study of Australian nurses driving home after a night shift revealed a higher reported pre-drive KSS was associated with an increased risk of a sleep-related event (e.g., falling asleep at a stop light) during the subsequent drive [21]. Despite this awareness, drivers often start, or continue to, drive [25], possibly due to a desire to reach the destination or not being able to find a suitable and safe place to stop and rest if the drive has already commenced. Ideally then, drivers should evaluate their sleepiness prior to commencing the journey. Our study is informative in that it examines sleepiness ratings prior to starting the drive and thus provides an opportunity for evasive action. Habitually reflecting on sleepiness prior to driving a motor vehicle may therefore provide an opportunity to take action and ultimately alleviate accident risk. Our study has a number of limitations. Firstly, as participants were not selected, randomly prior concerns about sleepiness and work hours may have motivated individuals to volunteer. Although our study data were collected from 2009–2011 when the ACGME were working toward their recommendation to abolish EDWS for PGY-1 residents, enquiring about attitudes toward shifts and sleepiness may have helped us to understand whether individuals felt strongly about work hours and driving, and to understand if our participants were motivated to participate by those feelings. It should be noted however that of those who expressed an interest to participate (i.e., phone interview), 42% did not drive and 12% did not work overnight or have extended work shifts, suggesting that strong attitudes toward “marathon shifts” or drowsy driving was not a major factor in choosing to volunteer. Furthermore, as described above, self-reported sleepiness ratings were very similar to a recent independent study of 124 interns and residents, suggesting that our samples was not dominated by interns with higher-than-normal levels of sleepiness [35]. Secondly, the self-report nature of the adverse driving outcomes may not reflect actual adverse events due to report bias. It is unlikely, however, that this would have affected the objective infrared reflectance oculography which was correlated with the self-reported data. Furthermore, although self-reported outcomes may be perceived as unreliable, previous studies have shown self-reported measures of driver performance to be associated with actual driver performance as assessed by objective methods (i.e., speed) or a trained assessor [47] and, of significance, crash risk [48]. Thirdly, our sample size was small and was not powered to detect crash or near-crash events. Rather, this study was designed to examine associations between sleepiness and adverse driving events. Fourthly, we lost a large amount of data from the infrared oculography device due to missing or reduced quality data. Given the nature of the study, and the work commitments and priorities of the study physicians, it is not unexpected that data loss may occur. As the study took place in built-up areas which require more movement and monitoring than that associated with highway driving (i.e., long stretches of straight road, largely comprised of looking ahead), signals can be lost due to excess head movements affecting the ocular signal. The importance of cleaning the data when using these types of devices in field settings is high. Finally, our “control” drives likely involved high levels of drowsiness, thus reflecting lower odds ratios for adverse driving events. For instance, odds ratios for adverse driving events on the commute home were lower in our study compared with those reported in the study of night shift working Australian nurses [21]. This is likely due to (1) the preshift drives in our study (which serve as a comparative baseline for odds ratio) occurred during an adverse circadian phase (~06:00 am), whereas the Australian nurses drove to work during the wake maintenance zone (~07:00 pm), and (2) the resident physicians in our study probably had higher levels of chronic sleep restriction due to repeated shifts, which subsequently affects sleep-related outcomes [28], whereas the Australian nurses worked a maximum of three night shifts. Although EDWSs were originally designed for residents working, sleeping and residing in the hospital [34], modern day trainees commuting to and from the hospital are exposing themselves and others to an elevated risk of drowsy driving. Higher levels of predrive sleepiness, the objective confirmation of drowsiness while driving, and the high rate of self-reported adverse events measured during the drive home from EDWS suggest that those working these shifts may knowingly place themselves and other road users at an unnecessary risk of drowsy driving crashes. Our data confirm that these risks are predictable—drivers were able to recognize their own elevated sleepiness levels before driving home and those assessments predicted the risk of a subsequent adverse driving event with relatively high sensitivity and specificity. This has serious implications on the legal accountability of the driver for adverse outcomes on the drive [49]. It is essential, therefore, that larger-scale studies in real, on-road environments should examine this wider issue in physicians and other shift worker populations. Our data are newly relevant, given that the ACGME have recently reinstated EWDS for postgraduate year 1 (PGY-1) resident physicians [2]. Although EDWSs are permitted, our data suggest that trainee physicians, and all other road users driving during this period considered at risk for drowsiness, should consciously evaluate their sleepiness level prior to each and every drive and, if they suspect that they are impaired, seek alternative transportation and not expose themselves and others to the significant, and avoidable, risk of drowsy driving. Ethical Approval The study was approved by the Human Research Committee of Partners Healthcare and all participants provided written informed consent and were remunerated for their involvement. Supplementary Material Supplementary material is available at SLEEP online. Funding This study was supported by a President and Fellows of Harvard College seed funding award presented to CA, and equipment was funded by a Defense University Research Instrumentation Program (DURIP) award FA9550-07-1-0242 from the Air Force Office of Scientific Research. The content is solely the responsibility of the authors and does not represent the official views of the funding bodies. Author Contributions All authors have made substantial contributions to the work presented and have approved the final version of the manuscript. CA, CAC, and SWL designed the study; CA conducted the study, analyzed and interpreted the data, and wrote the manuscript; SF assisted in the analysis and interpretation of the data; JMR provided technological oversight of the OptalertTM measurements and data output; SWR, CAC, and SWL contributed to data analysis, interpretation of data, and edited the manuscript. All authors approved the final manuscript. Disclosure Statement CA has no conflicts of interests related to the results reported in this paper report. In the interest of full disclosure, she has received a research award/prize from Sanofi-Aventis; contract research support from VicRoads, Rio Tinto Coal Australia, and Tontine/Pacific Brands; and lecturing fees from Brown Medical School/Rhode Island Hospital, Ausmed, Healthmed, and TEVA. In addition, she has served as a consultant through her institution to the Rail, Bus, and Tram Union, the Transport Accident Commission (TAC), and the National Transportation Committee (NTC). She has also served as an expert witness and/or consultant in relation to fatigue and drowsy driving. CA is a Theme Leader in the Cooperative Research Centre for Alertness, Safety, and Productivity. SF serves as a Project Leader in the Cooperative Research Centre for Alertness, Safety, and Productivity. JMR reports that he has served as a consultant through his institution to Takeda Pharmaceuticals North America and Cephalon and has received consulting fees from Vanda Pharmaceuticals. SMWR reports that he has served as a consultant through his institution to Vanda Pharmaceuticals, Philips Respironics, EdanSafe, The Australian Workers’ Union, National Transport Commission, Transport Accident Commission, and New South Wales Department of Education and Communities; has through his institution received research grants and/or unrestricted educational grants from Vanda Pharmaceuticals, Shell, Teva Pharmaceuticals, Rio Tinto, Seeing Machines, 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 served as an expert witness and/or consultant to shift work organizations. SMWR also serves as a Program Leader in the Cooperative Research Centre for Alertness, Safety, and Productivity. CAC has received consulting fees from or served as a paid member of scientific advisory boards for Bose Corporation, Boston Celtics, Columbia River Bar Pilots, Institute of Digital Media and Child Development, Klarman Family Foundation, Quest Diagnostics, Inc., Vanda Pharmaceuticals, and V-Watch/PPRS. He has also received education/research support from Cephalon Inc., Mary Ann & Stanley Snider via Combined Jewish Philanthropies, Optum, Philips Respironics, Inc., ResMed Foundation, San Francisco Bar Pilots, Schneider Inc., and Sysco. He has received lecture fees from American Academy of Sleep Medicine (AADSM), CurtCo Media Labs LLC, Global Council on Brain Health/AARP, Hawaii Sleep Health and Wellness Foundation, National Sleep Foundation, University of Michigan, University of Washington, and Zurich Insurance Company, Ltd. The Sleep and Health Education Program of the Harvard Medical School Division of Sleep Medicine (which CAC directs) has received Educational Grant funding from Cephalon, Inc., Jazz Pharmaceuticals, Takeda Pharmaceuticals, Teva Pharmaceuticals Industries Ltd., Sanofi-Aventis, Inc., Sepracor, Inc., and Wake Up Narcolepsy. CAC is the incumbent of an endowed professorship provided to Harvard University by Cephalon, Inc. and holds a number of process patents in the field of sleep/circadian rhythms (e.g., photic resetting of the human circadian pacemaker). Since 1985, he has served as an expert on various legal and technical cases related to sleep and/or circadian rhythms including those involving the following commercial entities: Bombardier, Inc., Continental Airlines, FedEx, Greyhound, and United Parcel Service (UPS). CAC owns or owned an equity interest in Somnus Therapeutics, Inc. and Vanda Pharmaceuticals. He received royalties from McGraw Hill and Koninklijke Philips Electronics, N.V. for the Actiwatch-2 and Actiwatch-Spectrum devices. CAC’s interests were reviewed and managed by Brigham and Women’s Hospital and Partners HealthCare in accordance with their conflict of interest policies. SWL has no conflicts of interests related to the research or results reported in this paper. In the interests of full disclosure, commercial interests from the last 3 years (2014–2017) are listed below. SWL has received consulting fees from the Atlanta Falcons, Atlanta Hawks, Carbon Limiting Technologies Ltd. on behalf of PhotoStar LED, Perceptive Advisors, Serrado Capital, and SlingshOT Insights and has current consulting contracts with Akili Interactive, Consumer Sleep Solutions, Delos Living LLC, Headwaters Inc., Hintsa Performance AG, Light Cognitive, Mental Workout, Pegasus Capital Advisors LP, PlanLED, OpTerra Energy Services, Inc., and Wyle Integrated Science and Engineering. SWL has received unrestricted equipment gifts from Biological Illuminations LLC, Bionetics Corporation, and F.LUX Software LLC; has equity in iSLEEP, Pty; advance author payment and/or royalties from Oxford University Press; honoraria plus travel, accommodation, and/or meals for invited seminars, conference presentations or teaching from BHP Billiton, Estee Lauder, Lightfair, Informa Exhibitions (USGBC), Teague; and travel, accommodation, and/or meals only (no honoraria) for invited seminars, conference presentations, or teaching from FASEB, Hintsa Performance AG, Lightfair, and USGBC. SWL has completed investigator-initiated research grants from Biological Illumination LLC and Vanda Pharmaceuticals, Inc. and has an ongoing investigator initiated grant from F. Lux Software LLC; completed service agreements from Rio Tinto Iron Ore and Vanda Pharmaceuticals Inc.; and completed three sponsor-initiated clinical research contracts from Vanda Pharmaceuticals, Inc. SWL holds a process patent for “Systems and methods for determining and/or controlling sleep quality,” which is assigned to the Brigham and Women’s Hospital per Hospital policy. SWL has also served as a paid expert on behalf of several public bodies on arbitrations related to sleep, light, circadian rhythms, and/or work hours for City of Brantford, Canada, and legal proceedings related to light, sleep, and health. SWL is also a Program Leader for the CRC for Alertness, Safety, and Productivity, Australia. Notes Work Performed: Department of Medicine and Neurology, Division of Sleep and Circadian Disorders, Brigham and Women’s Hospital and the Department of Medicine, Division of Sleep Medicine, Harvard Medical School, Boston, MA Acknowledgments We thank the resident physicians who took part in the study and the participating hospitals in the Boston Metropolitan area. We also thank the Committee of Interns and Residents for help with advertising the study to members. We also thank Robert Kilpatrick for assistance in pilot testing of the OptalertTM equipment. We also acknowledge Jason Sullivan and Kate Crowley for helping with data storage and transfer. References 1. Nasca TJ, Day SH, Amis ESJr; ACGME Duty Hour Task Force. The new recommendations on duty hours from the ACGME Task Force. N Engl J Med . 2010; 363( 2): e3. Google Scholar CrossRef Search ADS PubMed  2. Accreditation Council for Graduate Medical Education (ACGME). ACGME Common Program Requirements. Section VI. Proposed Major Revisions. 2016 [accessed 2016 21/12/2016]; Available from: http://www.acgme.org/Portals/0/PFAssets/ReviewandComment/CPR_SectionVI_ChangesTracked.pdf 3. Lockley SW, Cronin JW, Evans EEet al.  ; Harvard Work Hours, Health and Safety Group. Effect of reducing interns’ weekly work hours on sleep and attentional failures. N Engl J Med . 2004; 351( 18): 1829– 1837. Google Scholar CrossRef Search ADS PubMed  4. Philibert I. Sleep loss and performance in residents and nonphysicians: a meta-analytic examination. Sleep . 2005; 28( 11): 1392– 1402. Google Scholar CrossRef Search ADS PubMed  5. Landrigan CP, Rothschild JM, Cronin JWet al.   Effect of reducing interns’ work hours on serious medical errors in intensive care units. N Engl J Med . 2004; 351( 18): 1838– 1848. Google Scholar CrossRef Search ADS PubMed  6. Baldwin DCJr, Daugherty SR, Tsai R, Scotti MJJr. A national survey of residents’ self-reported work hours: thinking beyond specialty. Acad Med . 2003; 78( 11): 1154– 1163. Google Scholar CrossRef Search ADS PubMed  7. Veasey S, Rosen R, Barzansky B, Rosen I, Owens J. Sleep loss and fatigue in residency training: a reappraisal. JAMA . 2002; 288( 9): 1116– 1124. Google Scholar CrossRef Search ADS PubMed  8. Goebert D, Thompson D, Takeshita Jet al.   Depressive symptoms in medical students and residents: a multischool study. Acad Med . 2009; 84( 2): 236– 241. Google Scholar CrossRef Search ADS PubMed  9. Ayas NT, Barger LK, Cade BEet al.   Extended work duration and the risk of self-reported percutaneous injuries in interns. JAMA . 2006; 296( 9): 1055– 1062. Google Scholar CrossRef Search ADS PubMed  10. Barger LK, Cade BE, Ayas NTet al.  ; Harvard Work Hours, Health, and Safety Group. Extended work shifts and the risk of motor vehicle crashes among interns. N Engl J Med . 2005; 352( 2): 125– 134. Google Scholar CrossRef Search ADS PubMed  11. Marcus CL, Loughlin GM. Effect of sleep deprivation on driving safety in housestaff. Sleep . 1996; 19( 10): 763– 766. Google Scholar CrossRef Search ADS PubMed  12. Wendt JR, Yen LJ. The resident by moonlight: a misguided missile. JAMA . 1988; 259( 1): 43– 44. Google Scholar CrossRef Search ADS PubMed  13. Steele MT, Ma OJ, Watson WA, Thomas HAJr, Muelleman RL. The occupational risk of motor vehicle collisions for emergency medicine residents. Acad Emerg Med . 1999; 6( 10): 1050– 1053. Google Scholar CrossRef Search ADS PubMed  14. Arnedt JT, Owens J, Crouch M, Stahl J, Carskadon MA. Neurobehavioral performance of residents after heavy night call vs after alcohol ingestion. JAMA . 2005; 294( 9): 1025– 1033. Google Scholar CrossRef Search ADS PubMed  15. MacLean AW, Davies DR, Thiele K. The hazards and prevention of driving while sleepy. Sleep Med Rev . 2003; 7( 6): 507– 521. Google Scholar CrossRef Search ADS PubMed  16. Anderson C, Horne JA. Sleepiness enhances distraction during a monotonous task. Sleep . 2006; 29( 4): 573– 576. Google Scholar CrossRef Search ADS PubMed  17. Anderson C, Horne JA. Driving drowsy also worsens driver distraction. Sleep Med . 2013; 14( 5): 466– 468. Google Scholar CrossRef Search ADS PubMed  18. Durmer JS, Dinges DF. Neurocognitive consequences of sleep deprivation. Semin Neurol . 2005; 25( 1): 117– 129. Google Scholar CrossRef Search ADS PubMed  19. Harrison Y, Horne JA. The impact of sleep deprivation on decision making: a review. J Exp Psychol Appl . 2000; 6( 3): 236– 249. Google Scholar CrossRef Search ADS PubMed  20. Akerstedt T, Hallvig D, Anund A, Fors C, Schwarz J, Kecklund G. Having to stop driving at night because of dangerous sleepiness–awareness, physiology and behaviour. J Sleep Res . 2013; 22( 4): 380– 388. Google Scholar CrossRef Search ADS PubMed  21. Ftouni S, Sletten TL, Howard Met al.   Objective and subjective measures of sleepiness, and their associations with on-road driving events in shift workers. J Sleep Res . 2013; 22( 1): 58– 69. Google Scholar CrossRef Search ADS PubMed  22. Howard ME, Jackson ML, Berlowitz Det al.   Specific sleepiness symptoms are indicators of performance impairment during sleep deprivation. Accid Anal Prev . 2014; 62: 1– 8. Google Scholar CrossRef Search ADS PubMed  23. Lee ML, Howard ME, Horrey WJet al.   High risk of near-crash driving events following night-shift work. Proc Natl Acad Sci U S A . 2016; 113( 1): 176– 181. Google Scholar CrossRef Search ADS PubMed  24. Watling CN, Armstrong KA, Smith SS, Wilson A. The on-road experiences and awareness of sleepiness in a sample of Australian highway drivers: A roadside driver sleepiness study. Traffic Inj Prev . 2016; 17( 1): 24– 30. Google Scholar CrossRef Search ADS PubMed  25. Nabi H, Guéguen A, Chiron M, Lafont S, Zins M, Lagarde E. Awareness of driving while sleepy and road traffic accidents: prospective study in GAZEL cohort. BMJ . 2006; 333( 7558): 75. Google Scholar CrossRef Search ADS PubMed  26. Ulmer C, Wolman D, Johns M. Committee on Optimizing Graduate Medical Trainee (Resident) Hours and Work Schedules to Improve Patient Safety for the Institute of Medicine. Resident Duty Hours: Enhancing Sleep, Supervision, and Safety . Washington, DC: The National Academies Press; 2008. 27. Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep . 1991; 14( 6): 540– 545. Google Scholar CrossRef Search ADS PubMed  28. Anderson C, Sullivan JP, Flynn-Evans EE, Cade BE, Czeisler CA, Lockley SW. Deterioration of neurobehavioral performance in resident physicians during repeated exposure to extended duration work shifts. Sleep . 2012; 35( 8): 1137– 1146. Google Scholar PubMed  29. Reason J, Manstead A, Stradling S, Baxter J, Campbell K. Errors and violations on the roads: a real distinction? Ergonomics . 1990; 33( 10-11): 1315– 1332. Google Scholar CrossRef Search ADS PubMed  30. Johns MW, Chapman R, Crowley K, Tucker A. A new method for assessing the risks of drowsiness while driving. Somnologie . 2008; 12(1): 66– 74. Google Scholar CrossRef Search ADS   31. Johns MW, Tucker A, Chapman R, Crowley K, Michael N. Monitoring eye and eyelid movements by infrared reflectance oculography to measure drowsiness in drivers. Somnologie . 2007; 11(4): 234– 42. Google Scholar CrossRef Search ADS   32. Anderson C, Chang AM, Sullivan JP, Ronda JM, Czeisler CA. Assessment of drowsiness based on ocular parameters detected by infrared reflectance oculography. J Clin Sleep Med . 2013; 9( 9): 907– 20, 920A. Google Scholar PubMed  33. Akerstedt T, Gillberg M. Subjective and objective sleepiness in the active individual. Int J Neurosci . 1990; 52( 1-2): 29– 37. Google Scholar CrossRef Search ADS PubMed  34. Kamphuis J, Meerlo P, Koolhaas JM, Lancel M. Poor sleep as a potential causal factor in aggression and violence. Sleep Med . 2012; 13( 4): 327– 334. Google Scholar CrossRef Search ADS PubMed  35. Basner M, Dinges DF, Shea JAet al.   Sleep and alertness in medical interns and residents. An observational study on the role of extended shifts. Sleep . 2017; 40( 4):zsx027. 36. Achermann P, Werth E, Dijk DJ, Borbely AA. Time course of sleep inertia after nighttime and daytime sleep episodes. Arch Ital Biol . 1995; 134( 1): 109– 119. Google Scholar PubMed  37. Horne JA, Reyner LA. Counteracting driver sleepiness: effects of napping, caffeine, and placebo. Psychophysiology . 1996; 33( 3): 306– 309. Google Scholar CrossRef Search ADS PubMed  38. Mongrain S, Standing L. Impairment of cognition, risk-taking, and self-perception by alcohol. Percept Mot Skills . 1989; 69( 1): 199– 210. Google Scholar CrossRef Search ADS PubMed  39. Van Dongen HP, Maislin G, Mullington JM, Dinges DF. The cumulative cost of additional wakefulness: dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep . 2003; 26( 2): 117– 126. Google Scholar CrossRef Search ADS PubMed  40. Kaplan KA, Itoi A, Dement WC. Awareness of sleepiness and ability to predict sleep onset: can drivers avoid falling asleep at the wheel? Sleep Med . 2007; 9( 1): 71– 79. Google Scholar CrossRef Search ADS PubMed  41. Filtness AJ, Anund A, Fors C, Ahlström C, Akerstedt T, Kecklund G. Sleep-related eye symptoms and their potential for identifying driver sleepiness. J Sleep Res . 2014; 23( 5): 568– 575. Google Scholar CrossRef Search ADS PubMed  42. Alvaro PK, Jackson ML, Berlowitz DJ, Swann P, Howard ME. Prolonged eyelid closure episodes during sleep deprivation in professional drivers. J Clin Sleep Med . 2016; 12( 8): 1099– 1103. Google Scholar CrossRef Search ADS PubMed  43. Jackson ML, Kennedy GA, Clarke Cet al.   The utility of automated measures of ocular metrics for detecting driver drowsiness during extended wakefulness. Accid Anal Prev . 2016; 87: 127– 133. Google Scholar CrossRef Search ADS PubMed  44. Horne JA, Baulk SD. Awareness of sleepiness when driving. Psychophysiology . 2004; 41( 1): 161– 165. Google Scholar CrossRef Search ADS PubMed  45. Sagaspe P, Taillard J, Akerstedt Tet al.   Extended driving impairs nocturnal driving performances. PLoS One . 2008; 3( 10): e3493. Google Scholar CrossRef Search ADS PubMed  46. Philip P, Sagaspe P, Moore Net al.   Fatigue, sleep restriction and driving performance. Accid Anal Prev . 2005; 37( 3): 473– 478. Google Scholar CrossRef Search ADS PubMed  47. West R, French D, Kemp R, Elander J. Direct observation of driving, self reports of driver behaviour, and accident involvement. Ergonomics . 1993; 36( 5): 557– 567. Google Scholar CrossRef Search ADS PubMed  48. Parker D, Reason JT, Manstead ASR, Stradling SG. Driving errors, driving violations and accident involvement. Ergonomics . 1995; 38( 5): 1036– 1048. Google Scholar CrossRef Search ADS PubMed  49. Czeisler CA. The Gordon Wilson Lecture: work hours, sleep and patient safety in residency training. Trans Am Clin Climatol Assoc . 2006; 117: 159– 188. Google Scholar PubMed  © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail journals.permissions@oup.com.

Journal

SLEEPOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off