Cognitive effects of hospital calls in ... - Wiley Online Library

Acta Anaesthesiol Scand 2014; 58: 177–184 Printed in Singapore. All rights reserved

© 2013 The Acta Anaesthesiologica Scandinavica Foundation. Published by John Wiley & Sons Ltd ACTA ANAESTHESIOLOGICA SCANDINAVICA

doi: 10.1111/aas.12220

Cognitive effects of hospital calls in anaesthesiologists T. Husby1,2, J. Torgersen3 and H. Flaatten1,2

1 Department of Anaesthesiology and Intensive Care Medicine, Haukeland University Hospital, Bergen, Norway, 2Department of Clinical Medicine, University of Bergen, Bergen, Norway and 3The Norwegian Junior Doctors Association, The Norwegian Medical Association, Oslo, Norway

Background: The work hours of Norwegian physicians are under scrutiny because of an increased public focus on patient safety. Ample international research indicate harmful effects of doctor fatigue based on studies on physicians working long weeks and on-call shifts of more than 30 consecutive hours. There is a lack of research on effects relevant for short or intermediate length of work weeks and call shifts. This study intended to study cognitive effects of short or intermediate duration in-hospital calls. Methods: Eighteen anaesthesiology residents working on-call at an operation ward or an intensive care unit at Haukeland University Hospital were invited to participate. Schedules were adapted to allow for two additional experimental shifts. Participants were subjected to Cambridge Neuropsychological Test Automated Battery cognitive testing in a rested state and on three occasions after call. Amount of sleep and self-assessed sleepiness were recorded. Results: Ten residents completed all four tests during 10

D

octors’ working hours and its effect on quality in health care and patient safety have been subject to an increased focus among health professionals, authorities and the public. Research on effects of long work hours has mainly focused on extended on-call shifts (> 24 h). In the United States, studies have shown that working extended hours has negative effects on patient safety and quality of health care and that reducing the duration of in-hospital on-call shifts increases both patient safety and quality of care.1,2 In addition, working extended shifts increases the risk of harmful events for the doctor, such as suffering needle puncture injuries and being involved in car accidents.3,4 In general, night work involves a higher risk of work accidents, even for short duration shifts.5 A majority of studies on physician work hours were performed in countries with a high number of working hours per week and a high number of consecutive work hours, often above 30 h. Looking at doctors that work fewer hours per week, Irish junior

months. Reaction time was longer post-call. It was significantly increased only after the 18 h night call, by 21.1 and 20.5 ms for simple and five-choice reaction time, respectively. Executive function was not significantly altered post-call. Visual memory was improved post-call. Karolinska Sleepiness Score was increased by 3.3 (long day), 2.1 (short night) and 2.5 (long night) points post-call. Conclusion: Reaction times were increased after 18 h night calls and non-significant increases in reaction times were apparent after the other on-call shifts. Self reported sleepiness was increased post-call. We were not able to conclude whether executive function or memory was negatively affected post-call. Accepted for publication 24 September 2013 © 2013 The Acta Anaesthesiologica Scandinavica Foundation. Published by John Wiley & Sons Ltd

doctors working on average 65 h weeks were found to score significantly lower on decision making tests and some, but not all, cognitive tests after an average 32.75 h on-call compared with when rested.6 Less is known of potential negative effects from shorter working weeks and fewer consecutive hours on call. This lack of data has impeded a constructive debate on physician work hours and patient safety in countries where an intermediate (16–18 h) length of on-call shifts is the standard. During the last decade, most in-hospital on-calls in Norway have been reduced from long (> 24 h) to intermediate duration shifts because of demands from both the Norwegian Medical Association and the authorities. Still, doctors’ working hours are debated in the public. Media have claimed that harmful events in public hospitals can be linked to physicians working frequent intermediate duration on-call shifts, which usually include nights. Cognitive function is known to be affected by sleep deprivation and has frequently been tested to

177 bs_bs_banner

T. Husby et al. Hour 6

12

18

0

6

12

18

0

12

6

18

0

6

12

18

test

Long day

test

Short night

test

Long night

Fig. 1. Overview of the three different types of call shift, illustrating the relation to offduty hours and regular clinic hours before and after call.

Regular work Call

assess the effect of both sleep deprivation and long working hours.6–8 We have not found studies assessing the potential negative effect of working intermediate shifts compared with short shifts (< 10 h). The aim of this study was to assess some key domains of cognitive performance in residents in an academic anaesthesia and intensive care department after various lengths of in-hospital on-calls. Our hypothesis was that the assessed cognitive performance would not be negatively affected by postcall fatigue.

Methods Study context and sample Haukeland University Hospital is a secondary and tertiary health-care facility, among the largest hospitals in the country with 950 somatic beds, providing health services to approximately 1.1 million inhabitants of Western Norway. This study focuses on residents working in the two on-call duty lines considered to be the busiest in our department, namely the operating rooms (ORs) and the intensive care units (ICUs). In our experience, the two duty lines are similar in workload and sleep deprivation. A regular work schedule for residents on-call in-hospital (depending on area of responsibility) is a total of seven night shifts and two weekend day shifts per nine (OR) or 11 (ICU) weeks. The average number of working hours per week is 43. The residents start their shift at 15:00 h. and go home at 08:30 h next morning, working in total 18 consecutive hours. Of the 20 anaesthesiology trainees currently assigned to either the OR or the ICU in-hospital on-call duty, 18 were asked to participate, and they all signed an informed consent form. Two trainees who were soon to be rotated to other duties were not

178

asked to participate. Participants were not screened for health conditions that theoretically may interfere with cognitive test performance. They were considered fit for inclusion as long as they were fit for on-call duty by their own and department opinion.

Test procedure Participants were subjected to cognitive testing on four occasions. First, a baseline test session early on a regular clinic day after a normal night’s sleep (a minimum 6 h of sleep was set as a requirement). The remaining test sessions were executed after three different on-call shifts (Fig. 1) but with the same area of responsibility (OR or ICU); one after the traditional long afternoon/night shift (18 h including night), one after a long day shift (16 h without night) and one after a short night shift (10 h). The latter two types of shifts were introduced for the benefit of this study in order to investigate two possible alternatives to the traditional 18 h shift. See Fig. 2 for an overview of the test schedule. The postcall test sessions were executed immediately after hand-off and were scheduled with minimum 60-day intervals between sessions to minimise learning effects. All cognitive testings were performed by the same investigator (TH). Because of the limited time available and the required 60-day interval between test sessions, a fully randomised test setup was not feasible. The participants were however always tested in a rested state before the post-call test sessions were scheduled.

Data collection The following data were recorded: date and time for testing, date of birth, gender, area of responsibility (OR or ICU), experience in training (months), amount of sleep before test (h), and subjective rate of sleepiness (1–9) according to the Karolinska Sleepiness Scale (KSS).9

Cognitive effects of hospital calls > 60 days

CANTAB session 1 Rested (0-test)

MOT PAL SOC RTI

> 60 days

> 60 days

CANTAB session 4

CANTAB session 2

CANTAB session 3

Post call (16d, 10n or 18n)



MOT PAL SOC RTI

MOT PAL SOC RTI

MOT PAL SOC RTI

Dropout: n=1

Dropout: n=2

Dropout: n=5

Fig. 2. Test schedule. Participants were tested four times, with a planned > 60-day interval (median 72). The Cambridge Neuropsychological Test Automated Battery (CANTAB) test battery consisted of four tests: Motor Screening Test (MOT), Paired Associates Learning (PAL), Stockings of Cambridge (SOC) and Reaction Time (RTI). 16d, long day shift; 10n, short night shift; 18n, long night shift.

Cognitive function assessment was achieved using the Cambridge Neuropsychological Test Automated Battery (CANTAB) system. This system utilises a tablet touch-screen computer and allows us to assemble neuropsychological test batteries for addressing relevant cognitive domains. We chose the following four tests based on previous investigated cognitive domains in doctors post-call.6,8 Motor Screening Test (MOT) was a screening test for inability to cooperate with the apparatus, testing vision and basic motor function. The subject was to touch coloured crosses on the screen. Paired Associates Learning (PAL) was a test of episodic and visual memory and learning. A set of hidden figures on the screen were uncovered briefly, and the subject was to recall its position on the screen. Difficulty increased with the number of figures presented (two to eight figures). Stockings of Cambridge (SOC) was a test of executive function, working memory and strategic planning. The subject was to rearrange a set of coloured circles to correspond with a given template. Difficulty level increased as the number of minimum moves needed to complete the task rose from two to five. Reaction Time (RTI) was an attention and reaction time test, and measured speed of response and movement in a single and five-choice paradigms. Accuracy was also recorded. The subject was to hold a button on a press pad. After a stimulus on the screen, the hand was removed from the pad to touch the screen as quickly as possible.* *http://www.camcog.com/cantab-tests.asp [Accessed 11 June 2013]

Each test reported a wide range of results. We chose to investigate the results previously shown to have the best test-retest reliability.10

Statistics and ethics Results were analysed using IBM SPSS Statistics Version 19–20 (SPSS Inc., Chicago, IL, USA). The results suggested an approximately normal distribution, and paired samples t-test was used for comparing CANTAB results from different shifts and for comparing KSS scores before and after shifts. Because we planned a large number of comparisons (different tests and different lengths of on-call), we estimated sample size for paired difference for some of the comparisons. We aimed to detect a 20-ms change in reaction time based on other studies on reaction time under varying conditions.8,11–14 With a probability of type 1 error of 0.05 and power 0.8, the number needed to find a difference in simple reaction time of 20 ms with estimated standard deviation of 24 was n = 14 participants. The study was approved by the Regional Ethics Committee (REK Vest, Bergen, Norway. Protocol number 2011/801. Date of approval 6 May, 2011). Signed informed consents were obtained from all participants.

Results All eligible 18 physicians (13 male and five female) were included in the study. The average age was 35 (range 31–48 years), and their median experience in clinical anaesthesiology and intensive care was 51 months (range 8–88 months). Testing was conducted in the period June 2011 to March 2012, and

179

T. Husby et al. Table 1 Obtained sleep during shifts. Karolinska Sleepiness Scale (KSS) before and after different shifts.

Sleep during shift (h) KSS before shift KSS after shift Means comparison preand post-call KSS, P

Long day (16d)

Short night (10n)

Long night (18n)

Mean (SD) range

Mean (SD) range

Mean (SD) range

0 (0) 0 3.21 (1.37) 1–5 6.07 (1.07) 4–7 < 0.01

1.53 (1.06) 0–3 3.93 (1.53) 1–6 6.33 (1.18) 4–8 < 0.01

1.38 (1.70) 0–6.5 3.54 (1.71) 1–7 6.54 (1.98) 3–9 0.03

Range 1–9 (‘very rested’ to ‘very sleepy’). SD, standard deviation. 16d, long day shift; 10n, short night shift; 18n, long night shift.

10 participants completed all four scheduled test sessions. Among the remaining eight, five participants completed three sessions, two completed two sessions and one completed one. The reasons for dropout were change of work place (n = 1), reassignment to a different on-call shift (n = 4), leave of absence (n = 2) and sick leave (n = 1). One of the 60 sessions in total was performed after an 18 h night shift at the Ob&Gyn department instead of one of the two intended departments because of reassignment. This was allowed because this particular night had been busy and no sleep was obtained throughout the shift. The average time between two tests on the same individual was 80 days (50–171, median 72). Self-reported sleep and sleepiness are displayed in Table 1. CANTAB test results from the baseline and post-call tests are given in Table 2 and Fig. 3. Comparisons between different types of shifts are displayed in Table 3.

RTI Reaction time was increased significantly after the long (18 h) night call compared with baseline on the simple RTI test. Increased reaction time was also observed after the long day (21.8 ms) and the short night (17.5 ms), although not significantly different from baseline. No significant difference was found between performances on the three different postcall tests. On the five-choice RTI, all post-call reaction times were increased compared with baseline but significantly different only after the long night shift. There were no significant differences between post-call tests. Movement times were not significantly different, nor were accuracy scores.

180

MOT and SOC When comparing baseline and post-call results, no differences were apparent on the MOT or the SOC tests. Neither were any differences found between post-call tests.

PAL PAL test results were improved on most post-call tests compared with the initial baseline test. The best results were achieved on the post-16-hour-day call test.

Discussion In this study on the cognitive impact of various on-call shifts for anaesthesiologists in training, we found that reaction times were increased after the traditional long night call compared with the baseline but with no significant difference between shifts of different duration. No significant differences were found between baseline and post-call or between different types of shifts on the SOC (executive functions) test. Post-call results were improved on the PAL (episodic and visual memory) test compared with baseline. Self-reported sleepiness was increased post-call for all shift durations compared with when rested. Our study differs from most studies in this field in the respect that we have focused on shortand intermediate-length on-call shifts. Sixteen- to eighteen-hour shifts are gradually becoming the standard, at least in the first post-graduate years of medical practice. Our study has also evaluated alternative shift lengths, addressing the general notion that shorter shifts will reduce the negative effects of fatigue and night work. The length of shifts will be a subject for discussion in the future also, given the multiple factors that may affect quality of health care, e.g., loss of continuity and efficiency owing to

Time is measured in milliseconds (ms). Error and accuracy are expressed in numeric score. P-levels have been adjusted using the Holm-Bonferroni procedure. The number of compared results is expressed as n, which is the number of participants who completed that particular shift. *P < 0.05. CANTAB, Cambridge Neuropsychological Test Automated Battery; RTI, Reaction Time; SD, standard deviation; SOC, Stockings of Cambridge; 16d, long day shift; 10n, short night shift; 18n, long night shift.

1.00 1.00 1.00 9.2 (2.3) 5–12 91 (150.3) 0–389 772.8 (1034.4) 0–3362 1.00 1.00 1.00 8.7 (1.9) 5–11 19 (73.6) 0–285 414.4 (656) 0–2382 9.4 (1.8) 6–12 0 (0) 0–0 461.5 (547.3) 0–1961 9.3 (1.5) 6–11 61.6 (131.2) 0–517.5 519.9 (587.4) 0–1959.7

1.00 0.82 1.00

0.04* 0.25 1.00 0.03* 1.00 1.00 278.8 (33.4) 228–342 260 (71.9) 164–364 14.8 (0.4) 14–15 324.3 (31) 279–385 278.5 (75.6) 164–392 14.8 (0.4) 14–15 0.29 1.00 1.00 0.30 1.00 1.00 275.1 (39.9) 217–364 264.9 (81.2) 133–403 14.7 (0.6) 13–15 317.4 (36.9) 270–392 298.7 (92.5) 168–469 14.7 (0.8) 12–15 0.24 1.00 1.00 1.00 1.00 1.00 279.4 (36.8) 210–336 242 (73.7) 141–411 14.7 (0.5) 14–15 311.3 (32.9) 261–362 268.2 (91.5) 148–452 14.9 (0.5) 13–15

n = 14 n = 18

257.6 (23.1) 219.7–300.2 248.4 (73.6) 157.8–431.9 14.7 (0.6) 13–15 303.8 (24.4) 260.9–371.6 287 (75.2) 186.9–459.3 14.9 (0.3) 14–15

RTI Simple reaction time Simple movement time Simple accuracy score Five-choice reaction time Five-choice movement time Five-choice accuracy score SOC Problems solved, min. moves M. subseq. thinking time, 2 moves M. Subseq. thinking time, 5 moves

Mean (SD) range

n = 13 n = 15

Mean (SD) range Mean (SD) range Mean (SD) range

P (0–16d) 16 h day shift (16d) 0-test (0) CANTAB test/measure

CANTAB results, means comparisons (paired t-test) between baseline (0-test) and different shifts.

Table 2

10 h night shift (10n)

P (0–10n)


P (0–18n)

Cognitive effects of hospital calls

frequent transfer of care, change in staff requirements and family-related consequences for the physicians. A limitation in our study is primarily the suspected learning effect from repeated cognitive testing on the same healthy individuals, which may bias the findings towards better results on post-call tests, and may thereby hide actual performance deterioration. This is important for at least some cognitive domains such as executive function and may also affect temporal lobe function tests if tests are not optimally executed. The participants showed obvious improvement on PAL tests with repeated testing despite evidence of post-call fatigue on KSS scores, so the PAL results are not of use to this study. This learning effect was more profound than expected because the PAL test-retest correlation previously has been found to be good (0.68–0.86).10 We believe that the greatest learning effect takes place between the first and second test session. Therefore, a pre-baseline familiarisation† where the subject is allowed to try the test equipment and resolve misunderstandings would probably decrease learning effects. Another improvement would be to follow a strict standardised task description instead of the free verbal instruction given by the investigator prior to our testing. In general, attention and reaction time results may be biased by acquisition of a speed/accuracy tradeoff strategy.10 However, accuracy scores were hardly any different in the RTI test sessions, and any such effect is therefore unlikely to have occurred in this study. Another limitation is that the study group is small. While the size of our resident staff is relatively big by national standards, we limited inclusion to residents with comparable workloads and sleep (OR and ICU). The dropout rate was high. We had limited time available to perform four tests on each individual with 60-day intervals. A less ambitious test plan, e.g., a comparison between a rested state and only one type of shift would increase the strength of our results by reducing the dropout rate. The fact that this study is not randomised, has a small sample size and that some tests may be biased by learning effects increases the risk of a type II error. It is therefore not possible to conclude with certainty that executive function is unaffected by the three types of on-call shifts investigated in our study, although our results indicate so.

†http://www.camcog.com/faqs.asp [Accessed 28 October 2013]

181

T. Husby et al.

Fig. 3. CANTAB results boxplot, rested and post call. RTI, Reaction time. SOC, Stockings of Cambridge. *P < 0.05 with paired t-test (compared with rested state), adjusted using the Holm-Bonferroni procedure.

A future study should ideally include a larger homogenous population, randomised to baseline or post-call testing (with each individual tested only once), or with additional preventive measures to counter a learning effect. The participants in this study represent diversity in age and clinical experience that is quite common among the in-hospital call residents at a university hospital department in Norway. We do not view this mix as a confounding factor but rather the typical Norwegian situation. Studies have shown that all physicians are vulnerable to sleep deprivation and circadian rhythm disruption regardless of experience.6,13,15 We consider the average self reported sleep of 1–2 h representative of these areas of responsibility

182

and conclude that acute sleep deprivation was present at the time of post night call testing. Several studies have found a negative effect of prolonged wakefulness on attention.11,12,16 Gander et al. found that longer time since waking was associated with faster median reaction time on the day shift and slower median reaction time on the night shifts, consistent with circadian rhythm variation. Further, that a greater acute sleep loss was associated with poorer performance on night shifts on the 10% slowest reaction times.8 This correlates with our findings. Although only the 18 h night-call reaction time tests were significantly worse than baseline (mean difference 21.2 ms), all post-call reaction times were longer than baseline. This suggests that long shifts like the 16 h day call may cause an

Cognitive effects of hospital calls Table 3 CANTAB results after and means comparison between different shifts. CANTAB test/measure

16 h day shift (16d)



Mean RTI Simple reaction time Simple movement time Simple accuracy score Five-choice reaction time Five-choice movement time Five-choice accuracy score SOC Problems solved in minimum moves Mean subsequent thinking time, 2 moves Mean subsequent thinking time, 5 moves

Comparisons 10n–16d

18n–16d

18n–10n

P

279.4 242.0 14.7 311.3 268.2 14.9

275.1 264.9 14.7 317.4 298.7 14.7

278.8 260.0 14.8 324.3 278.5 14.8

1.00 0.65 1.00 1.00 1.00 1.00

1.00 0.46 1.00 1.00 1.00 1.00

1.00 1.00 1.00 1.00 1.00 1.00

9.4 0.0 461.5

8.7 19.0 414.4

9.2 91.0 772.8

1.00 1.00 1.00

1.00 0.49 1.00

1.00 0.57 0.32

Time is milliseconds (ms). Error and accuracy are expressed in numerical score. P-levels have been adjusted using the HolmBonferroni procedure. CANTAB, Cambridge Neuropsychological Test Automated Battery; RTI, Reaction Time; SOC, Stockings of Cambridge; 16d, long day shift; 10n, short night shift; 18n, long night shift.

increase in reaction time even before the circadian rhythm has taken a full downturn on factors known to influence alertness, such as serum melatonin level. The fatigue after 16 consecutive work hours may be sufficient to experience cognitive performance impairment, without the full circadian rhythm disruption. On the SOC (executive function) test, our participants did not perform significantly different between rested state and post-call. Frontal lobe and executive function tests such as SOC are generally expected to have a lower reliability on repeated testing because the participants acquire strategies that will help them improve their scores on repeated testing.10 In our study, there is a possibility that actual post-call fatigue may have been counterbalanced by a learning effect, although other research groups that have evaluated executive function of residents after call have found similar results.6,17 Landrigan et al. have previously established that a considerable part of the diagnostic and medication errors in an ICU are due to fatigue.1 Flinn and Armstrong found negative effects of fatigue on clinical decision-making tests and some cognitive tests in a group of Irish junior doctors. This is, in their opinion, evidence that there is a relationship between cognitive performance and medical errors in clinical decision making.6 Medical procedures are apparently less influenced by sleep deprivation and circadian rhythm disruption.1,18 Cognitive function measurement is therefore, in our opinion, partly useful in terms of assessing clinically relevant effects of fatigue. Still, to our knowledge, a direct correla-

tion between cognitive performance and patient care has not been documented. The clinical implications of the measured increase in reaction time postcall are uncertain, but attention is by default a vital part of the on-call physician’s functions. Several research groups suggest that reduced attention or increased reaction time is likely to affect the clinical performance of physicians.8,11,16

Conclusion We found that the 18 h night call reduces the level of performance on the reaction time test compared with a rested state but did not find a significant difference between three different types of on-call shifts. We cannot conclude for certain whether short or intermediate duration shifts have a negative influence on executive function or memory. Long day call increases self-reported sleepiness, as do short or intermediate duration night calls.

Acknowledgements The study team is thankful for the efforts, positive attitude and patience of all participants, and for the comments and suggestions of Professor Anne Berit Guttormsen and Dr. Reidar Kvåle. Conflict of interest. During the planning and start of this study, co-author Dr. Torgersen was an employee at the Department of Anaesthesia and Intensive Care Medicine at Haukeland University Hospital. He later became the elected Chair of the Norwegian Junior Doctors Association. His current position has not influenced the results or conclusions of this study. None of the authors have stated any relevant conflicts of interest regarding their involvement in this study. Funding: The study was funded by the Norwegian Medical Association Fund for Quality Improvement and Patient Safety

183

T. Husby et al. after application presenting the study protocol. The fund imposed no obligations or restrictions on any part of this study.

References 1. Landrigan CP, Rothschild JM, Cronin JW, Kaushal R, Burdick E, Katz JT, Lilly CM, Stone PH, Lockley SW, Bates DW, Czeisler CA. Effect of reducing interns’ work hours on serious medical errors in intensive care units. N Engl J Med 2004; 351: 1838–48. 2. Levine AC, Adusumilli J, Landrigan CP. Effects of reducing or eliminating resident work shifts over 16 hours: a systematic review. Sleep 2010; 33: 1043–53. 3. Barger LK, Cade BE, Ayas NT, Cronin JW, Rosner B, Speizer FE, Czeisler CA, 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: 125–34. 4. Ayas NT, Barger LK, Cade BE, Hashimoto DM, Rosner B, Cronin JW, Speizer FE, Czeisler CA. Extended work duration and the risk of self-reported percutaneous injuries in interns. JAMA 2006; 296: 1055–62. 5. Folkard S, Lombardi DA, Tucker PT. Shiftwork: safety, sleepiness and sleep. Ind Health 2005; 43: 20–3. 6. Flinn F, Armstrong C. Junior doctors’ extended work hours and the effects on their performance: the Irish case. Int J Qual Health Care 2011; 23: 210–7. 7. Dawson D, Reid K. Fatigue, alcohol and performance impairment. Nature 1997; 388: 235. 8. Gander P, Millar M, Webster C, Merry A. Sleep loss and performance of anaesthesia trainees and specialists. Chronobiol Int 2008; 25: 1077–91. 9. Kaida K, Takahashi M, Åkerstedt T, Nakata A, Otsuka Y, Haratani T, Fukasawa K. Validation of the Karolinska sleepiness scale against performance and EEG variables. Clin Neurophysiol 2006; 117: 1574–81. 10. Lowe C, Rabbitt P. Test/re-test reliability of the CANTAB and ISPOCD neuropsychological batteries: theoretical and practical issues. Neuropsychologia 1998; 36: 915–23.

184

11. 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: 1025–33. 12. Saxena AD, George CFP. Sleep and motor performance in on-call internal medicine residents. Sleep 2005; 28: 1386–91. 13. Chang LC, Mahoney JJ III, Raty SR, Ortiz J, Apodaca S, La Garza De R II. Neurocognitive effects following an overnight call shift on faculty anesthesiologists. Acta Anaesthesiol Scand 2013; 57: 1051–7. 14. Irwin C, Leveritt M, Shum D, Desbrow B. The effects of dehydration, moderate alcohol consumption, and rehydration on cognitive functions. Alcohol 2013; 47: 203–13. 15. Rothschild JM, Keohane CA, Rogers S, Gardner R, Lipsitz SR, Salzberg CA, Yu T, Yoon CS, Williams DH, Wien MF, Czeisler CA, Bates DW, Landrigan CP. Risks of complications by attending physicians after performing nighttime procedures. JAMA 2009; 302: 1565–72. 16. Bartel P. Attention and working memory in resident anaesthetists after night duty: group and individual effects. Occup Environ Med 2004; 61: 167–70. 17. Mak SK, Spurgeon P. The effects of acute sleep deprivation on performance of medical residents in a regional hospital: prospective study. Hong Kong Med J 2004; 10: 14–20. 18. Hayter MA, Friedman Z, Katznelson R, Hanlon JG, Borges B, Naik VN. Effect of sleep deprivation on labour epidural catheter placement. Br J Anaesth 2010; 104: 619–27.

Address: Thomas Husby Department of Anaesthesiology and Intensive Care Medicine Haukeland University Hospital Jonas Lies vei 65 5021 Bergen Norway e-mail: [email protected]