- Email: [email protected]
Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents
ARTICLE IN PRESS Respiratory Medicine (2006) 100, 1020–1027
Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents Fernanda Pichela, Carlos Zamarro ´na,, ´ R. Rodrı´gueza Fernando Maga ´nb, Jose a
Sleep Unit, Division of Respiratory Medicine, Hospital Clı´nico Universitario, Universidad de Santiago de Compostela, Choupana s/n, 15706-Santiago de Compostela, Spain b Escuela Superior de Ingenierı´a Informa´tica, Universidad de Vigo, Spain Received 5 May 2005; accepted 29 September 2005
KEYWORDS Obstructive sleep apnea; Traffic accidents; Driving simulator; Alertness; Reaction time; Divided attention
Summary The aim of this study was to identify associations between performance on driving simulators, subject sleep complaints, and risk of traffic accidents in a population undergoing OSAS investigation. Patients and Methods: One hundred and twenty-nine consecutive adult patients with a driver’s licence and clinical symptoms of OSA were initially recruited from the hospital waiting list. Each patient was applied a polysomnography and two driving tests. Patients also completed a Basic Nordic Sleep Questionnaire, Epworth Sleepiness Scale and SF-36s quality of life questionnaire. Results: Poor performance in vigilance was associated with alcohol intake (OR 4.41, 95% CI 1.13–17.20, Po0.05), and SF-36 vitality dimension (OR 0.97, 95% CI 0.94–0.99, Po0.05). Poor tracking error was associated with female gender (OR 6.79, 95% CI 1.37–33.65, Po0.05), alcohol intake (OR 3.32, 95% CI 1.03–10.63, Po0.05), and a history of accidents in the previous year (OR 5.84, 95% CI 1.33–25.68, Po0.05). Poor reaction time was only associated with age (OR 1.12, 95% CI 1.03–1.21, Po0.01). When all three performance measures were studied jointly, only reaction time was associated with self-reported dozing while driving (OR 5.39, 95% CI 1.10–26.32, Po0.05), and irresistible tendency to fall asleep was associated with poor tracking error (Po0.05). Conclusions: Performance on driving simulators is associated with some sleep complaints in OSA patients. Although these measures are not directly associated to traffic accidents, they are, in fact, associated to related circumstances such as dozing while driving and falling asleep while driving. & 2005 Elsevier Ltd. All rights reserved.
Corresponding author. Servicio de Neumologı´a, Hospital Clı´nico Universitario de Santiago de Compostela, C/Choupana s/n, 15706
Santiago de Compostela, Spain. E-mail address: [email protected] (C. Zamarro ´n). 0954-6111/$ - see front matter & 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.rmed.2005.09.036
ARTICLE IN PRESS Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents
Introduction A variety of neurobehavioral tests have been used to analyse the impact of sleep disorders.1 Among these, sustained attention and vigilance tests have been applied to OSA. Compared with controls, OSA patients present significant inability to sustain attention2–9 as measured by various tools such as the Four Choice Reaction Time test,10 the Psychomotor Vigilance Test,11 and more recently instruments that incorporate simulated driving scenarios such as the Divided Attention Driving Test.12,13 OSA patients also present altered vigilance as measured by Steer Clears.14 CPAP treatment seems to improve sustained attention performance when evaluated with the Psychomotor Vigilance Test2,3 and Divided Attention Driving Test,15–17 however, the improvement is less pronounced when vigilance is evaluated with Steer Clears.12,18,19 Only a small number of studies have investigated the role of specific OSA factors on the level of sustained attention and vigilance. Moreover, this research has not been conclusive.20–22 The aim of the present study was to identify associations between performance on driving simulators, subject sleep complaints, and risk of traffic accidents in a population undergoing OSAS investigation.
Patients and methods Subjects From December 2003 to September 2004, 129 consecutive adult patients with a driver’s licence and clinical symptoms of OSA were recruited from the hospital waiting list (107 men, 22 women) for a cross-sectional study. Traffic accident information was gathered through a questionnaire and confirmed through Emergency Unit records. As a consequence of the exclusion criteria, 36 patients did not take part in the research for the following reasons: (1) previous treatment for OSA (n ¼ 5), (2) narcolepsy or periodic leg movement (n ¼ 2), (3) chronic intake of sedatives (n ¼ 4), (4) refusal to perform any of the tests (n ¼ 5), (5) driving less than 5000 km in the previous year (n ¼ 11), and (6) the inability to confirm traffic accidents (n ¼ 9). Patients were considered to be positive for alcohol intake if they reported consuming any alcoholic beverage 3 times a day and 3 or more days a week. The Review Board on Human Studies at our institution approved the protocol, and each patient
1021
gave his or her informed consent to participate in the study.
Subjective daytime sleepiness Subjective daytime sleepiness was evaluated by means of the Epworth sleepiness scale (ESS).23,24 This scale consists of eight questions regarding the tendency to fall asleep in situations of differing stimulation. Each question is scored 0–3: The total score can thus vary from 0 (no sleepiness) to 24 (extremely sleepy) with 10 being the upper limit of normal.
Health-related quality of life Health-related quality of life was evaluated with a generic instrument, Medical Outcome Survey— Short Form 36 (SF-36).25,26 Scores of the eight SF36 subscales range from 0 (minimum well-being) to 100 (maximum well-being). The SF-36 addresses the following eight domains: physical functioning, physical problems, emotional problems, mental health, social functioning, energy/vitality, bodily pain and general health perception.
Basic nordic sleep questionnaire The BNSQ is a validated questionnaire27 that explores general complaints about sleep. Questions include habitual total sleep time, sleep latency, sleep quality, snoring, and sleepiness. It ranks symptoms on a frequency basis (i.e. Do you feel excessively sleepy during daytime? 1 ¼ never or less than once per month, 6 ¼ daily or almost daily). An additional question (‘‘Have you suffered from irresistible tendency to fall asleep while driving during the past three months?’’) was included following questionnaire instructions.
Driving questionnaire All patients completed a thorough specific questionnaire regarding their driving history. It includes data about driving habits, years and type of licence, kilometres driven in previous year, and type of vehicle. Subjects were asked if they had been involved in an accident and if so what type, the date and time, the circumstance, and injuries to driver or passengers.
Polysomnography Polysomnography were carried out in our Sleep Unit; usually from midnight to 8 a.m. This
ARTICLE IN PRESS 1022 technique consisted of continuous monitoring using a polygraph (Ultrasom Network, Nicolet, Madison, WI, USA) and included electroencephalogram, electro-oculogram, chin electromyogram, airflow, electrocardiogram and measurement of chest wall movement. The polysomnographic register was analysed in periods of 30 s and during stages 1–4 and REM according to standard criteria. Apnea was defined as the absence of airflow for more than 10 s, and hypopnoea as the reduction of respiratory flow for at least 10 s accompanied by a 4% or more decrease in the saturation of haemoglobin. The average of apnea–hypopnea index (AHI) was calculated in hourly samples of sleep. In this study an AHIX10 was considered as diagnostic of OSA. If the subject had less than 3 h of total sleep, the sleep study was repeated. Patients were classified based on the severity of their polysomnographic data: non-OSA (AHIo10), mild OSA (AHIX10 and o30), and moderate or severe OSA (AHIX30 events/h).
Vigilance testing—Steer Clears Vigilance was assessed with Steer Clears,12 a computer program involving a monotonous task. Subjects were given instructions and a 1-min practice test before the actual test. Based on the number of obstacles hit,4 patients were divided into two groups: normal within the mean (0.89) and one SD (1.10); otherwise, poor performance.
Divided attention steering simulator DASS-2Ds After 5-min of instruction and practice, all subjects performed a 20-min steering attention test on DASS-2Ds (SimDrive Divided Attention Steering Simulator 2D, Stowood Scientific Instruments, Oxford, UK). The test combines a primary task, keeping the cursor within a box on the screen, which is made more difficult by automatic movement of the cursor, and a secondary task, pressing a button on the steering wheel when a target number (‘‘3’’) appears in one of the corners of the screen. The main outcome variables from the test are tracking error (standard deviation of time outside of the box) and reaction time (the lag in responding to the target number). Cut-off points for normal performance on the DASSs test were determined based on Turkington et al.,20 that is, a tracking error o2 (standard deviation) and a reaction time o2 s.
F. Pichel et al.
Follow-up protocol All patients completed questionnaires at inclusion. A study-team member who was not aware of the patient’s AHI status and who did not participate in any other aspects of clinical management conducted all patient interviews. The above tests were all carried out on the same day (between 12:00 and 2:00 p.m.) and in the same sequence.
Data analysis Continuous variables were expressed as the mean (7standard deviation) or median [inter-quartile range] in the case of non-normal distributed data. The t-test, two tailed, was applied to test differences between performance levels; non-normally distributed variables were compared using Mann–Whitney test. Normality assumption was checked by means of Kolmogorov–Smirnov statistic with signification level of Lilliefors (if the group sample size 450) or by means of Shapiro–Wilks statistic in the other cases. Confidence intervals were set at 95%. Regarding the SF-36 questionnaire, a control group from the community population (CP) was included. Data from study group and control groupcorrected for age and sex. Independent associations between possible factors and performance result were studied using binary logistic regression models. On the other hand, relationships between results obtained by using Likert questions and the three variables to study as factors were analysed by searching ordinal regression models. Association between simulator test results was explored using Spearman analysis correlation. The w2-test was used to compare categorical and ordinal data. Statistical significance was accepted at Po0.05. All analyses were developed using the Statistical Package for Social Sciences (SPSSs, ver. 11.0; SPSS Inc., Chicago, IL, USA).
Results A diagnosis of OSA was confirmed in 77 (82.8%) out of a total of 93 subjects (78 males and 15 females). The group had a mean age of 50.8710.7 years, BMI of 30.175.3 kg/m2 and the AHI was 37.2723.4. Categorized according to severity, 16 subjects (17.2%) presented AHIp10; 25 subjects (26.9%) presented 10oAHIp30; and 52 subject (55.9%) presented AHI430.
ARTICLE IN PRESS Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents The presence of co-morbidities was observed in 38 patients (21 presented mainly arterial hypertension, 5 hiatus hernia, 4 COPD, 3 depression and 5 presented others). All of the patients were medically stable at the time of their initial evaluation. Alcohol intake was considered to be positive in 63 subjects (67.7%). In the previous year, 21 (22.6%) suffered traffic accidents as confirmed by emergency room records. On the vigilance test, 81 patients registered normal performance and 12 patients registered poor performance. With respect to tracking error on the sustained attention test, 50 patients had normal performance and 43 had poor performance. With respect to reaction time on the sustained attention test, 17 patients had normal performance and 76 poor performances. In Table 1, the column headed ‘‘vigilance’’ compares the anthropometrics, clinical variables, and SF-36 dimension scores of subjects with normal vs. poor performance on Steer Clears. Significant difference was only found in the energy/vitality dimension of the SF-36. In addition, the columns headed ‘‘Tracking Error’’ and ‘‘Reaction Time’’ compare subjects with normal vs. poor performance on the DASSs test. In terms of tracking, poor performers had significantly lower SF-36 general health perception values, while in terms of reaction time, poor performers were signifi-
1023
cantly older and had more years of driving experience. Correlations were found between vigilance and tracking error (r ¼ 0:359; Po0.001), as well as tracking error and reaction time (r ¼ 0:512; Po0.001). On the other hand, no significant correlation was found between vigilance and reaction time. The results of applying a regression model to data from the Steer Clears test on vigilance are shown in Table 2. According to the Wald criterion, a high score on reported habitual alcohol intake and a low score on the SF-36 energy/vitality dimension were significantly associated with poor performance. Using this model, 98.8% of those who had normal performance on Steer Clears were correctly classified, but only 16.7% of those with poor performance were correctly classified. The overall classification was 88.2%. This indicates that vigilance may be influenced by factors which are not strictly related to OSA. Table 3 shows the results when tracking error was the dependent variable. Gender-female, alcohol intake and history of accidents in the previous year were significantly associated with poor tracking on the driving simulator. Using this model, 57.1% of patients with poor performance were correctly classified and 85.7% of those with normal performance were correctly
Table 1 Anthropometrical, clinical and SF-36 dimension scores in subjects with normal and poor performance in vigilance, tracking error, and reaction time. Vigilance (Steer Clears)
Tracking error (DASS 2Ds)
Reaction time (DASS 2Ds)
Normal performance (n ¼ 81)
Poor performance (n ¼ 12)
Normal performance (n ¼ 50)
Poor performance (n ¼ 43)
Normal performance (n ¼ 17)
Poor performance (n ¼ 76)
70 (86) 50 [16] 29.4 [5.1] 11 [7.0] 28 [11] 29493 (30955) 38.0 [45.7]
8 (66) 48 [22] 33.1 [14.7] 10 [4.0] 26 [15] 22677 (31639) 42.0 [44.0]
47 (94) 48 [16.5] 29.1 [4.3] 11 [6.0] 30 [43.7] 29550 (31284) 35.1 (23.9)
31 (72) 51 [16.0] 31.9 [7.7] 10 [6.5] 42 [48.6] 27557 (30801) 40.4 (22.7)
15 (88) 41 [12.0] 29.4 [8.0] 10 [2.0] 25 [11.5] 27375 (18586) 29.4 (25.8)
63 (83) 52 [14.5] 30.0 [5.8] 10 [7.0] 30 [10.5] 29043 (33228) 39.0 (22.6)
SF-36 dimensions Energy/Vitality 55 [42] General Health 62.4 (23.3) Perception
30 [22] 48.8 (24.1)
52.5 [45] 72 [40.5]
50 [35] 57 [32]
45 [47] 72 [39]
50 [42] 62 [42]
Male (%) Age (yr) BMI (kg/m2) ESS score Years driving km/yr AHI
BMI, Body mass index; ESS, Epworth sleepiness scale; AHI, Apnea-hipopnea index. Mean (SD), Median [IQR]. Values showing significant differences between performance groups are indicated. Po0.01. Po0.05.
ARTICLE IN PRESS 1024 Table 2
F. Pichel et al. Logistic regression analysis vigilance on the driving simulator. B
Alcohol Energy/vitality Age
Table 3
Wald
1.484 0.034 0.060
4.571 5.530 2.729
OR
0.033 0.019 0.099
95% CI (OR)
4.411 0.967 1.062
Lower
Upper
1.132 0.940 0.989
17.195 0.994 1.140
Logistic regression analysis of tracking error on the driving simulator. B
Gender (female) Alcohol Accident
Table 4
P-value
Wald
1.915 1.198 1.764
P-value
5.497 4.061 5.441
OR
0.019 0.044 0.020
95% CI (OR)
6.787 3.315 5.835
Lower
Upper
1.369 1.033 1.326
33.652 10.633 25.682
Ordinal regression model for irresistible tendency to fall asleep while driving during the past 3 months. Estimate
Fall asleep driving
[2] [3] [4] [5] [6]
Only one time Sometimes Many times Almost always Always driving
Measurements
Vigilance Tracking error Reaction time
Wald
P-value
95% CI Lower
Upper
0.103 0.971 2.288 3.265 4.169
0.020 1.770 8.908 15.776 20.015
0.887 0.183 0.003 0.000 0.000
1.310 0.459 0.785 1.654 2.343
1.515 2.402 3.790 4.876 5.996
0.258 1.066 0.003
0.117 5.066 0.000
0.733 0.024 0.996
1.223 1.138 1.103
1.739 1.994 1.097
Category [1] ¼ ‘‘never or less than once per month’’ taken as reference. Last category [6] ¼ ‘‘Always when driving’’. Po0.05.
classified. Overall, 73.8% of subjects were correctly classified. Age was independently associated with reaction time (OR ¼ 1.12, 95% CI (1.03–1.21), Po0.01). Using this model, 95.6% of patients with poor performance were correctly classified and only 18.8% of those with normal performance were correctly classified. Overall, 81.8% of subjects were correctly classified. Poor reaction time was independently associated with dozing while driving during the previous year [OR ¼ 5.38, 95%CI (1.10–26.32), Po0.05]. Using this model, 92.2% of patients who had reported dozing while driving were correctly classified but only 9.1% of those who did not report dozing while driving were correctly classified. Overall, 59.5% of subjects were correctly classified.
As Table 4 shows, tendency to fall asleep while driving was predicted by tracking error. [Wald’s statistic ¼ 5.06, 95% CI (1.14–1.99), Po0.05]. OSA, at any level of severity, did not seem related to any of the performance measures on the simulators. That is to say, none of the tests used was able to identify the presence/absence of OSA. Nor was any relation found between OSA and the parameters on the driving questionnaire. None of the performance variables were associated to Likert variables, although the relation between snoring and vigilance, and reaction time were nearly significant. In the regression analysis, none of the tests studied (vigilance with Steer Clear, reaction time and tracking error with DASS), were significantly related to traffic accidents, however, the tracking
ARTICLE IN PRESS Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents error test was very nearly significant in the logistic regression analysis (P ¼ 0.094).
Discussion This study shows that alcohol consumption, age, female gender, history of accidents in the previous year and health related quality of life have significant influence on performance in driving tests. OSA has been shown to be associated with an increased risk of road traffic accidents. Predicting the driving ability and risk of road traffic accidents in an individual with OSA is complex. Moreover, it is difficult to identify what specific factors influence the ability to respond appropriately. On-road testing is the gold standard for determining driving ability, but it is difficult to apply and, so, simple computer-based driving simulators have been used. Various authors have shown that OSA patients have impaired driving performance which improves following treatment with nasal continuous positive airway pressure.4,12–17,28–30 In the present study, alcohol intake, SF-36 vitality dimension and age were significantly associated with poor performance on vigilance test. Hack et al. examined steering simulator performance in apnoeic patients as compared to non-apnoeic drivers who consumed alcohol (below legal limits) or were sleep deprived. The authors conclude that alcohol mainly influences cognition or motor skills, while sleep deprivation mainly alters vigilance; thus, demonstrating that patient groups may present different performance patterns on driving simulators.17 Regarding the SF-36 vitality dimension, diminishing scores in the poor performance group are probably related to OSA-related factors. We have used control group data from another study in order to compare our findings.31 OSA patients present significantly lower values in three dimensions: physical functioning, role limitation due to physical problems and energy/vitality. Our results with respect to age may reflect a lack of ability to participate in a ‘‘computer game’’ irrespective driving ability on the part of older subjects. Female gender, alcohol intake and history of accidents in the previous year were significantly associated to poor tracking error and only age was independently associated to poor performance in reaction time. Other authors’ findings regarding gender are similar to ours, but the reasons are not clear. In fact, some studies show that women perform better at tasks involving hand/eye coordination.32 Nevertheless, Rizzo found that women
1025
presented greater tracking error that may have been due to their greater reported discomfort and likelihood to drop out than men.33 The history of accidents in the previous year was significantly associated, in this model, with poor tracking error performance on the driving simulator. However, Turkington et al. found (using a 3D more complex simulator, DADS) no relationship between poor performance on the simulator and accident history.20 Sleepiness increases reaction time and hinders the ability to stop.34 Also, performance on attention-based tasks has been found to decline with sleepiness.20 Our study shows that poor reaction time on the driving simulator predicted reports of dozing while driving. In the Turkington study, Epworth Sleepiness Scale results correlated with ‘‘near miss accidents.’’ This contrasted with our own results and those reported by Masa et al.35 which found no such significant relationship. We have found no correlation between OSA severity and simulator performance, thus, confirming the observation in other studies.8 Nevertheless, other authors have found that, as a group, patients with OSA do not perform as well as matched controls on a simulator. These authors consider that the substantial overlap confirms the influence of other factors.8,13,14,16,36 Our methodology presents some limitations. The fact that a large number of patients refused to take part in the study may have caused some selection bias. The unfounded fear that poor performance could lead to loss of driving license may have caused some subjects to be more reluctant to participate. Five patients refused to perform either simulator (2 of them cited visual difficulties and 3 lack of time). Secondly, we used a questionnaire to obtain information about driving history and accidents. If used alone, reliance on a questionnaire poses a risk of recall bias, however, in our study; information gathered from questionnaires was confirmed with information from Health Department records. It has been shown that patients with OSA are reluctant to report accidents and may, in fact, under report symptoms.20 We have considered this to be the best line of attack option because the data is readily available and because the options used in other similar studies, such as information from police, vehicle licensing authorities, or insurers may also involve underestimation due the fact that not all accidents are reported and they provide no information on near misses. The practice run that our patients performed may not have been sufficient. One study14 has
ARTICLE IN PRESS 1026 suggested that at least three 5 min practice runs are required to abolish any learning effect on the simulator, whereas others have used practice times ranging from 1 to 12 min.4,16 It was felt that a longer practice run may have further reduced recruitment and limited future applicability in routine clinical practice. All subjects performed the practice runs as recommended by the supplier, that is, 1- and 5-min for Steer Clears and DASSs, respectively. This study shows that performance on a driving simulator is influenced by a number of factors that also play a role in traffic accidents. None of the tests studied (vigilance with Steer Clear, reaction time and tracking error with DASS); either independently or alone, proved to be significantly related to traffic accidents, however, the tracking error test was very nearly significant. Nevertheless, the above tests were found to be significantly associated with factors that play a role in poor driving. Namely, poor performance in vigilance was associated with alcohol intake and age, while reaction time and tracking error were significantly associated with tendency to sleep while driving. Our findings indicate that it is difficult to draw a definitive conclusion about real-world driving ability for the general population based on driving simulator tests. This is because real-world driving is essentially different from the simulators we have used and because our study group (OSA patients) has special characteristics. Nevertheless, driving simulators are attractive because they offer the possibility of controlling certain skills in a way that is difficult to do with on-road testing. Moreover, driving simulators are becoming increasingly realistic. We conclude that performance on driving simulators is associated with some sleep complaints in OSA patients. Although these measures are not directly associated to traffic accident, they are, in fact, associated to related circumstances such as dozing while driving and falling asleep while driving.
Acknowledgements This work was supported by grants from the Health Research Fund (FIS 01/0634) and XUGA (PGIDT99PXI90201A).
References 1. Weaver TE. Outcome measurement in sleep medicine practice and research. Part 2: Assessment of neurobehavioral performance and mood. Sleep Med Rev 2001;5:223–36.
F. Pichel et al. 2. Be´dard MA, Montplaisir J, Malo J, et al. Persistent neuropsycological deficits and vigilance impairment in sleep apnea syndrome after treatment with continuous positive airways pressure (CPAP). J Clin Exp Neuropsychol 1993;15: 330–41. 3. Kribbs NB, Pack AL, Kline LR, et al. Effects of one night without nasal CPAP treatment on sleep and sleepiness in patients with obstructive sleep apnea. Am Rev Respir Dis 1993;147:1162–8. 4. Findley L, Unverzagt M, Guchu R, et al. Vigilance and automobile accidents in patients with sleep apnea or narcolepsy. Chest 1995;108:619–24. 5. George CF, Boudreau AC, Smiley A. Comparison of simulated driving performance in narcolepsy and sleep apnea patients. Sleep 1996;19:711–7. 6. Chugh DK, Weaver TE, Dinges DF. Psychomotor vigilance performance in sleep apnea patients compared to patients presenting with snoring with apnea. Sleep 1998;21(Suppl):159. 7. Dinges D, Maislin G, Staley B, et al. Sleepiness and neurobehavioral functioning in relation to apnea severity in a cohort of commercial motor vehicle operaters. Sleep 1998;21(Suppl):83. 8. Barbe ´ F, Perica ´s J, Mun ˜oz A, et al. Automobile accidents in patients with sleep apnea syndrome. An epidemiological and mechanistic study. Am J Respir Crit Care Med 1998;158: 18–22. 9. Mun ˜oz A, Mayoralas LR, Barbe F, et al. Long-term effects of CPAP on daytime functioning in patients with sleep apnoea syndrome. Eur Respir 2000;15:676–81. 10. Wilkinson RT, Houghton D. Portable four-choice reaction time test with magnetic tape memory. Behav Res Methods Instrum Comput 1995;7:441–6. 11. Dinges DF, Powell JW. Microcomputer analysis of performance on a portable, simple visual RT task during sustained operations. Behav Res Methods Instrum Comput 1985;17: 652–5. 12. George CF, Boudreau AC, Smiley A. Simulated driving performance in patients with obstructive sleep apnea. Am J Respir Crit Care Med 1996;154:175–81. 13. Juniper M, Hack MA, George CF, et al. Steering simulation performance in patients with obstructive sleep apnoea and matched control subjects. Eur Respir J 2000;15:590–5. 14. Findley LJ, Fabrizio MJ, Knight H, et al. Driving simulator performance in patients with sleep apnea. Am Rev Respir Dis 1989;140:529–30. 15. George CF, Boudreau AC, Smiley A. Effects of nasal CPAP on simulated driving performance in patients with obstructive sleep apnoea. Thorax 1997;52:648–53. 16. Hack M, Davies RJ, Mullins R, et al. Randomised prospective parallel trial of therapeutic versus subtherapeutic nasal continuous positive airway pressure on simulated steering performance in patients with obstructive sleep apnoea. Thorax 2000;55:224–31. 17. Hack MA, Choi SJ, Vijayapalan P, et al. Comparison of the effects of sleep deprivation, alcohol and obstructive sleep apnoea (OSA) on simulated steering performance. Respir Med 2001;95:594–601. 18. Findley LJ, Levinson MP, Bonnie RJ. Driving performance and automobile accidents in patients with sleep apnea. Clin Chest Med 1992;13:427–35. 19. Douglas NJ. Systematic review of the efficacy of nasal CPAP. Thorax 1998;53:414–5. 20. Turkington PM, Sircar M, Allgar V, et al. Relationship between obstructive sleep apnoea, driving simulator performance, and risk of road traffic accidents. Thorax 2001;56:800–5.
ARTICLE IN PRESS Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents 21. Telakivi T, Partinen M, Hublin C, et al. Obstructive sleep apnea syndrome among city bus drivers: diagnostic implications and driving capacity. Sleep 2002;25:A–225. 22. Risser MR, Ware JC, Freeman FG. Driving simulation with EEG monitoring in normal and obstructive sleep apnea patients. Sleep 2000;23:393–8. 23. Johns M. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 1991;14:54–5. 24. Johns MW. Sleepiness in different situations measured by the Epworth sleepiness scale. Sleep 1994;17:703–10. 25. Ware Jr JE, Snow KK, Kosinski M, et al. SF-36 Health survey manual and interpretation guide. Boston: The Health Institute, New England Medical Center; 1993. 26. Alonso J, Regidor E, Barrio G, et al. [Population reference values of the Spanish version of the Health Questionnaire SF36] Valores poblacionales de referencia de la version espan ˜ola del Cuestionario de Salud SF-36. Med Clin (Barc) 1998;111:410–6. 27. Partinen M, Gislason T. Basic Nordic Sleep Questionnaire (BNSQ): a quantitated measure of subjective sleep complaints. J Sleep Res 1995;4(Suppl 1):150–5. 28. American Thoracic Society. Sleep apnea, sleepiness,and driving risk. Am J Respir Crit Care Med 1994;50:1463–73. 29. Haraldsson PO, Carenfelt C, Laurell H, et al. Acta Otolaryngol. (Stockholm) 1990;110:136–40. 30. Jenkinson C, Davies R, Mullins R, et al. Randomised prospective parallel trial of therapeutic nasal continuous
31.
32.
33.
34. 35.
36.
1027
positive airway pressure (NCPAP) against sub-therapeutic NCPAP for obstructive sleep apnoea. The Lancet 1999;353: 2100–5. Pichel F, Zamarro ´n C, Maga ´n F, et al. Health-related quality of life in patients with obstructive sleep apnea: effects of long-term positive airway pressure treatment. Respir Med 2004;98:968–76. Schueneman AL, Pickleman J, Freeark RJ. Age, gender, lateral dominance, and prediction of operative skill among general surgery residents. Surgery 1985;98: 506–15. Rizzo M, Sheffield RA. Demographic and driving performance factors in simulator adaptation syndrome. In Proceedings of the second international driving symposium on human factors in driver assessment, training and vehicle design, Iowa, IA, 2003. ppc.uiowa.edu/driving-assessment/2003/ Summaries/Downloads/Final_Papers/PDF/46_Rizzoformat. pdf. Date last accessed: February 2, 2005, 2003. Dinges D. An overview of sleepiness and accidents. J Sleep Res 1995;4:4–14. Masa JF, Rubio M, Findley LJ. Habitually sleepy drivers have a high frequency of automobile crashes associated with respiratory disorders during sleep. Am J Respir Crit Care Med 2000;162:1407–12. Findley LJ, Suratt PM, Dinges DF. Time-on-task decrements in steer clear performance of patients with sleep apnea and narcolepsy. Sleep 1999;22:804–9.
Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents Fernanda Pichela, Carlos Zamarro ´na,, ´ R. Rodrı´gueza Fernando Maga ´nb, Jose a
Sleep Unit, Division of Respiratory Medicine, Hospital Clı´nico Universitario, Universidad de Santiago de Compostela, Choupana s/n, 15706-Santiago de Compostela, Spain b Escuela Superior de Ingenierı´a Informa´tica, Universidad de Vigo, Spain Received 5 May 2005; accepted 29 September 2005
KEYWORDS Obstructive sleep apnea; Traffic accidents; Driving simulator; Alertness; Reaction time; Divided attention
Summary The aim of this study was to identify associations between performance on driving simulators, subject sleep complaints, and risk of traffic accidents in a population undergoing OSAS investigation. Patients and Methods: One hundred and twenty-nine consecutive adult patients with a driver’s licence and clinical symptoms of OSA were initially recruited from the hospital waiting list. Each patient was applied a polysomnography and two driving tests. Patients also completed a Basic Nordic Sleep Questionnaire, Epworth Sleepiness Scale and SF-36s quality of life questionnaire. Results: Poor performance in vigilance was associated with alcohol intake (OR 4.41, 95% CI 1.13–17.20, Po0.05), and SF-36 vitality dimension (OR 0.97, 95% CI 0.94–0.99, Po0.05). Poor tracking error was associated with female gender (OR 6.79, 95% CI 1.37–33.65, Po0.05), alcohol intake (OR 3.32, 95% CI 1.03–10.63, Po0.05), and a history of accidents in the previous year (OR 5.84, 95% CI 1.33–25.68, Po0.05). Poor reaction time was only associated with age (OR 1.12, 95% CI 1.03–1.21, Po0.01). When all three performance measures were studied jointly, only reaction time was associated with self-reported dozing while driving (OR 5.39, 95% CI 1.10–26.32, Po0.05), and irresistible tendency to fall asleep was associated with poor tracking error (Po0.05). Conclusions: Performance on driving simulators is associated with some sleep complaints in OSA patients. Although these measures are not directly associated to traffic accidents, they are, in fact, associated to related circumstances such as dozing while driving and falling asleep while driving. & 2005 Elsevier Ltd. All rights reserved.
Corresponding author. Servicio de Neumologı´a, Hospital Clı´nico Universitario de Santiago de Compostela, C/Choupana s/n, 15706
Santiago de Compostela, Spain. E-mail address: [email protected] (C. Zamarro ´n). 0954-6111/$ - see front matter & 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.rmed.2005.09.036
ARTICLE IN PRESS Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents
Introduction A variety of neurobehavioral tests have been used to analyse the impact of sleep disorders.1 Among these, sustained attention and vigilance tests have been applied to OSA. Compared with controls, OSA patients present significant inability to sustain attention2–9 as measured by various tools such as the Four Choice Reaction Time test,10 the Psychomotor Vigilance Test,11 and more recently instruments that incorporate simulated driving scenarios such as the Divided Attention Driving Test.12,13 OSA patients also present altered vigilance as measured by Steer Clears.14 CPAP treatment seems to improve sustained attention performance when evaluated with the Psychomotor Vigilance Test2,3 and Divided Attention Driving Test,15–17 however, the improvement is less pronounced when vigilance is evaluated with Steer Clears.12,18,19 Only a small number of studies have investigated the role of specific OSA factors on the level of sustained attention and vigilance. Moreover, this research has not been conclusive.20–22 The aim of the present study was to identify associations between performance on driving simulators, subject sleep complaints, and risk of traffic accidents in a population undergoing OSAS investigation.
Patients and methods Subjects From December 2003 to September 2004, 129 consecutive adult patients with a driver’s licence and clinical symptoms of OSA were recruited from the hospital waiting list (107 men, 22 women) for a cross-sectional study. Traffic accident information was gathered through a questionnaire and confirmed through Emergency Unit records. As a consequence of the exclusion criteria, 36 patients did not take part in the research for the following reasons: (1) previous treatment for OSA (n ¼ 5), (2) narcolepsy or periodic leg movement (n ¼ 2), (3) chronic intake of sedatives (n ¼ 4), (4) refusal to perform any of the tests (n ¼ 5), (5) driving less than 5000 km in the previous year (n ¼ 11), and (6) the inability to confirm traffic accidents (n ¼ 9). Patients were considered to be positive for alcohol intake if they reported consuming any alcoholic beverage 3 times a day and 3 or more days a week. The Review Board on Human Studies at our institution approved the protocol, and each patient
1021
gave his or her informed consent to participate in the study.
Subjective daytime sleepiness Subjective daytime sleepiness was evaluated by means of the Epworth sleepiness scale (ESS).23,24 This scale consists of eight questions regarding the tendency to fall asleep in situations of differing stimulation. Each question is scored 0–3: The total score can thus vary from 0 (no sleepiness) to 24 (extremely sleepy) with 10 being the upper limit of normal.
Health-related quality of life Health-related quality of life was evaluated with a generic instrument, Medical Outcome Survey— Short Form 36 (SF-36).25,26 Scores of the eight SF36 subscales range from 0 (minimum well-being) to 100 (maximum well-being). The SF-36 addresses the following eight domains: physical functioning, physical problems, emotional problems, mental health, social functioning, energy/vitality, bodily pain and general health perception.
Basic nordic sleep questionnaire The BNSQ is a validated questionnaire27 that explores general complaints about sleep. Questions include habitual total sleep time, sleep latency, sleep quality, snoring, and sleepiness. It ranks symptoms on a frequency basis (i.e. Do you feel excessively sleepy during daytime? 1 ¼ never or less than once per month, 6 ¼ daily or almost daily). An additional question (‘‘Have you suffered from irresistible tendency to fall asleep while driving during the past three months?’’) was included following questionnaire instructions.
Driving questionnaire All patients completed a thorough specific questionnaire regarding their driving history. It includes data about driving habits, years and type of licence, kilometres driven in previous year, and type of vehicle. Subjects were asked if they had been involved in an accident and if so what type, the date and time, the circumstance, and injuries to driver or passengers.
Polysomnography Polysomnography were carried out in our Sleep Unit; usually from midnight to 8 a.m. This
ARTICLE IN PRESS 1022 technique consisted of continuous monitoring using a polygraph (Ultrasom Network, Nicolet, Madison, WI, USA) and included electroencephalogram, electro-oculogram, chin electromyogram, airflow, electrocardiogram and measurement of chest wall movement. The polysomnographic register was analysed in periods of 30 s and during stages 1–4 and REM according to standard criteria. Apnea was defined as the absence of airflow for more than 10 s, and hypopnoea as the reduction of respiratory flow for at least 10 s accompanied by a 4% or more decrease in the saturation of haemoglobin. The average of apnea–hypopnea index (AHI) was calculated in hourly samples of sleep. In this study an AHIX10 was considered as diagnostic of OSA. If the subject had less than 3 h of total sleep, the sleep study was repeated. Patients were classified based on the severity of their polysomnographic data: non-OSA (AHIo10), mild OSA (AHIX10 and o30), and moderate or severe OSA (AHIX30 events/h).
Vigilance testing—Steer Clears Vigilance was assessed with Steer Clears,12 a computer program involving a monotonous task. Subjects were given instructions and a 1-min practice test before the actual test. Based on the number of obstacles hit,4 patients were divided into two groups: normal within the mean (0.89) and one SD (1.10); otherwise, poor performance.
Divided attention steering simulator DASS-2Ds After 5-min of instruction and practice, all subjects performed a 20-min steering attention test on DASS-2Ds (SimDrive Divided Attention Steering Simulator 2D, Stowood Scientific Instruments, Oxford, UK). The test combines a primary task, keeping the cursor within a box on the screen, which is made more difficult by automatic movement of the cursor, and a secondary task, pressing a button on the steering wheel when a target number (‘‘3’’) appears in one of the corners of the screen. The main outcome variables from the test are tracking error (standard deviation of time outside of the box) and reaction time (the lag in responding to the target number). Cut-off points for normal performance on the DASSs test were determined based on Turkington et al.,20 that is, a tracking error o2 (standard deviation) and a reaction time o2 s.
F. Pichel et al.
Follow-up protocol All patients completed questionnaires at inclusion. A study-team member who was not aware of the patient’s AHI status and who did not participate in any other aspects of clinical management conducted all patient interviews. The above tests were all carried out on the same day (between 12:00 and 2:00 p.m.) and in the same sequence.
Data analysis Continuous variables were expressed as the mean (7standard deviation) or median [inter-quartile range] in the case of non-normal distributed data. The t-test, two tailed, was applied to test differences between performance levels; non-normally distributed variables were compared using Mann–Whitney test. Normality assumption was checked by means of Kolmogorov–Smirnov statistic with signification level of Lilliefors (if the group sample size 450) or by means of Shapiro–Wilks statistic in the other cases. Confidence intervals were set at 95%. Regarding the SF-36 questionnaire, a control group from the community population (CP) was included. Data from study group and control groupcorrected for age and sex. Independent associations between possible factors and performance result were studied using binary logistic regression models. On the other hand, relationships between results obtained by using Likert questions and the three variables to study as factors were analysed by searching ordinal regression models. Association between simulator test results was explored using Spearman analysis correlation. The w2-test was used to compare categorical and ordinal data. Statistical significance was accepted at Po0.05. All analyses were developed using the Statistical Package for Social Sciences (SPSSs, ver. 11.0; SPSS Inc., Chicago, IL, USA).
Results A diagnosis of OSA was confirmed in 77 (82.8%) out of a total of 93 subjects (78 males and 15 females). The group had a mean age of 50.8710.7 years, BMI of 30.175.3 kg/m2 and the AHI was 37.2723.4. Categorized according to severity, 16 subjects (17.2%) presented AHIp10; 25 subjects (26.9%) presented 10oAHIp30; and 52 subject (55.9%) presented AHI430.
ARTICLE IN PRESS Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents The presence of co-morbidities was observed in 38 patients (21 presented mainly arterial hypertension, 5 hiatus hernia, 4 COPD, 3 depression and 5 presented others). All of the patients were medically stable at the time of their initial evaluation. Alcohol intake was considered to be positive in 63 subjects (67.7%). In the previous year, 21 (22.6%) suffered traffic accidents as confirmed by emergency room records. On the vigilance test, 81 patients registered normal performance and 12 patients registered poor performance. With respect to tracking error on the sustained attention test, 50 patients had normal performance and 43 had poor performance. With respect to reaction time on the sustained attention test, 17 patients had normal performance and 76 poor performances. In Table 1, the column headed ‘‘vigilance’’ compares the anthropometrics, clinical variables, and SF-36 dimension scores of subjects with normal vs. poor performance on Steer Clears. Significant difference was only found in the energy/vitality dimension of the SF-36. In addition, the columns headed ‘‘Tracking Error’’ and ‘‘Reaction Time’’ compare subjects with normal vs. poor performance on the DASSs test. In terms of tracking, poor performers had significantly lower SF-36 general health perception values, while in terms of reaction time, poor performers were signifi-
1023
cantly older and had more years of driving experience. Correlations were found between vigilance and tracking error (r ¼ 0:359; Po0.001), as well as tracking error and reaction time (r ¼ 0:512; Po0.001). On the other hand, no significant correlation was found between vigilance and reaction time. The results of applying a regression model to data from the Steer Clears test on vigilance are shown in Table 2. According to the Wald criterion, a high score on reported habitual alcohol intake and a low score on the SF-36 energy/vitality dimension were significantly associated with poor performance. Using this model, 98.8% of those who had normal performance on Steer Clears were correctly classified, but only 16.7% of those with poor performance were correctly classified. The overall classification was 88.2%. This indicates that vigilance may be influenced by factors which are not strictly related to OSA. Table 3 shows the results when tracking error was the dependent variable. Gender-female, alcohol intake and history of accidents in the previous year were significantly associated with poor tracking on the driving simulator. Using this model, 57.1% of patients with poor performance were correctly classified and 85.7% of those with normal performance were correctly
Table 1 Anthropometrical, clinical and SF-36 dimension scores in subjects with normal and poor performance in vigilance, tracking error, and reaction time. Vigilance (Steer Clears)
Tracking error (DASS 2Ds)
Reaction time (DASS 2Ds)
Normal performance (n ¼ 81)
Poor performance (n ¼ 12)
Normal performance (n ¼ 50)
Poor performance (n ¼ 43)
Normal performance (n ¼ 17)
Poor performance (n ¼ 76)
70 (86) 50 [16] 29.4 [5.1] 11 [7.0] 28 [11] 29493 (30955) 38.0 [45.7]
8 (66) 48 [22] 33.1 [14.7] 10 [4.0] 26 [15] 22677 (31639) 42.0 [44.0]
47 (94) 48 [16.5] 29.1 [4.3] 11 [6.0] 30 [43.7] 29550 (31284) 35.1 (23.9)
31 (72) 51 [16.0] 31.9 [7.7] 10 [6.5] 42 [48.6] 27557 (30801) 40.4 (22.7)
15 (88) 41 [12.0] 29.4 [8.0] 10 [2.0] 25 [11.5] 27375 (18586) 29.4 (25.8)
63 (83) 52 [14.5] 30.0 [5.8] 10 [7.0] 30 [10.5] 29043 (33228) 39.0 (22.6)
SF-36 dimensions Energy/Vitality 55 [42] General Health 62.4 (23.3) Perception
30 [22] 48.8 (24.1)
52.5 [45] 72 [40.5]
50 [35] 57 [32]
45 [47] 72 [39]
50 [42] 62 [42]
Male (%) Age (yr) BMI (kg/m2) ESS score Years driving km/yr AHI
BMI, Body mass index; ESS, Epworth sleepiness scale; AHI, Apnea-hipopnea index. Mean (SD), Median [IQR]. Values showing significant differences between performance groups are indicated. Po0.01. Po0.05.
ARTICLE IN PRESS 1024 Table 2
F. Pichel et al. Logistic regression analysis vigilance on the driving simulator. B
Alcohol Energy/vitality Age
Table 3
Wald
1.484 0.034 0.060
4.571 5.530 2.729
OR
0.033 0.019 0.099
95% CI (OR)
4.411 0.967 1.062
Lower
Upper
1.132 0.940 0.989
17.195 0.994 1.140
Logistic regression analysis of tracking error on the driving simulator. B
Gender (female) Alcohol Accident
Table 4
P-value
Wald
1.915 1.198 1.764
P-value
5.497 4.061 5.441
OR
0.019 0.044 0.020
95% CI (OR)
6.787 3.315 5.835
Lower
Upper
1.369 1.033 1.326
33.652 10.633 25.682
Ordinal regression model for irresistible tendency to fall asleep while driving during the past 3 months. Estimate
Fall asleep driving
[2] [3] [4] [5] [6]
Only one time Sometimes Many times Almost always Always driving
Measurements
Vigilance Tracking error Reaction time
Wald
P-value
95% CI Lower
Upper
0.103 0.971 2.288 3.265 4.169
0.020 1.770 8.908 15.776 20.015
0.887 0.183 0.003 0.000 0.000
1.310 0.459 0.785 1.654 2.343
1.515 2.402 3.790 4.876 5.996
0.258 1.066 0.003
0.117 5.066 0.000
0.733 0.024 0.996
1.223 1.138 1.103
1.739 1.994 1.097
Category [1] ¼ ‘‘never or less than once per month’’ taken as reference. Last category [6] ¼ ‘‘Always when driving’’. Po0.05.
classified. Overall, 73.8% of subjects were correctly classified. Age was independently associated with reaction time (OR ¼ 1.12, 95% CI (1.03–1.21), Po0.01). Using this model, 95.6% of patients with poor performance were correctly classified and only 18.8% of those with normal performance were correctly classified. Overall, 81.8% of subjects were correctly classified. Poor reaction time was independently associated with dozing while driving during the previous year [OR ¼ 5.38, 95%CI (1.10–26.32), Po0.05]. Using this model, 92.2% of patients who had reported dozing while driving were correctly classified but only 9.1% of those who did not report dozing while driving were correctly classified. Overall, 59.5% of subjects were correctly classified.
As Table 4 shows, tendency to fall asleep while driving was predicted by tracking error. [Wald’s statistic ¼ 5.06, 95% CI (1.14–1.99), Po0.05]. OSA, at any level of severity, did not seem related to any of the performance measures on the simulators. That is to say, none of the tests used was able to identify the presence/absence of OSA. Nor was any relation found between OSA and the parameters on the driving questionnaire. None of the performance variables were associated to Likert variables, although the relation between snoring and vigilance, and reaction time were nearly significant. In the regression analysis, none of the tests studied (vigilance with Steer Clear, reaction time and tracking error with DASS), were significantly related to traffic accidents, however, the tracking
ARTICLE IN PRESS Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents error test was very nearly significant in the logistic regression analysis (P ¼ 0.094).
Discussion This study shows that alcohol consumption, age, female gender, history of accidents in the previous year and health related quality of life have significant influence on performance in driving tests. OSA has been shown to be associated with an increased risk of road traffic accidents. Predicting the driving ability and risk of road traffic accidents in an individual with OSA is complex. Moreover, it is difficult to identify what specific factors influence the ability to respond appropriately. On-road testing is the gold standard for determining driving ability, but it is difficult to apply and, so, simple computer-based driving simulators have been used. Various authors have shown that OSA patients have impaired driving performance which improves following treatment with nasal continuous positive airway pressure.4,12–17,28–30 In the present study, alcohol intake, SF-36 vitality dimension and age were significantly associated with poor performance on vigilance test. Hack et al. examined steering simulator performance in apnoeic patients as compared to non-apnoeic drivers who consumed alcohol (below legal limits) or were sleep deprived. The authors conclude that alcohol mainly influences cognition or motor skills, while sleep deprivation mainly alters vigilance; thus, demonstrating that patient groups may present different performance patterns on driving simulators.17 Regarding the SF-36 vitality dimension, diminishing scores in the poor performance group are probably related to OSA-related factors. We have used control group data from another study in order to compare our findings.31 OSA patients present significantly lower values in three dimensions: physical functioning, role limitation due to physical problems and energy/vitality. Our results with respect to age may reflect a lack of ability to participate in a ‘‘computer game’’ irrespective driving ability on the part of older subjects. Female gender, alcohol intake and history of accidents in the previous year were significantly associated to poor tracking error and only age was independently associated to poor performance in reaction time. Other authors’ findings regarding gender are similar to ours, but the reasons are not clear. In fact, some studies show that women perform better at tasks involving hand/eye coordination.32 Nevertheless, Rizzo found that women
1025
presented greater tracking error that may have been due to their greater reported discomfort and likelihood to drop out than men.33 The history of accidents in the previous year was significantly associated, in this model, with poor tracking error performance on the driving simulator. However, Turkington et al. found (using a 3D more complex simulator, DADS) no relationship between poor performance on the simulator and accident history.20 Sleepiness increases reaction time and hinders the ability to stop.34 Also, performance on attention-based tasks has been found to decline with sleepiness.20 Our study shows that poor reaction time on the driving simulator predicted reports of dozing while driving. In the Turkington study, Epworth Sleepiness Scale results correlated with ‘‘near miss accidents.’’ This contrasted with our own results and those reported by Masa et al.35 which found no such significant relationship. We have found no correlation between OSA severity and simulator performance, thus, confirming the observation in other studies.8 Nevertheless, other authors have found that, as a group, patients with OSA do not perform as well as matched controls on a simulator. These authors consider that the substantial overlap confirms the influence of other factors.8,13,14,16,36 Our methodology presents some limitations. The fact that a large number of patients refused to take part in the study may have caused some selection bias. The unfounded fear that poor performance could lead to loss of driving license may have caused some subjects to be more reluctant to participate. Five patients refused to perform either simulator (2 of them cited visual difficulties and 3 lack of time). Secondly, we used a questionnaire to obtain information about driving history and accidents. If used alone, reliance on a questionnaire poses a risk of recall bias, however, in our study; information gathered from questionnaires was confirmed with information from Health Department records. It has been shown that patients with OSA are reluctant to report accidents and may, in fact, under report symptoms.20 We have considered this to be the best line of attack option because the data is readily available and because the options used in other similar studies, such as information from police, vehicle licensing authorities, or insurers may also involve underestimation due the fact that not all accidents are reported and they provide no information on near misses. The practice run that our patients performed may not have been sufficient. One study14 has
ARTICLE IN PRESS 1026 suggested that at least three 5 min practice runs are required to abolish any learning effect on the simulator, whereas others have used practice times ranging from 1 to 12 min.4,16 It was felt that a longer practice run may have further reduced recruitment and limited future applicability in routine clinical practice. All subjects performed the practice runs as recommended by the supplier, that is, 1- and 5-min for Steer Clears and DASSs, respectively. This study shows that performance on a driving simulator is influenced by a number of factors that also play a role in traffic accidents. None of the tests studied (vigilance with Steer Clear, reaction time and tracking error with DASS); either independently or alone, proved to be significantly related to traffic accidents, however, the tracking error test was very nearly significant. Nevertheless, the above tests were found to be significantly associated with factors that play a role in poor driving. Namely, poor performance in vigilance was associated with alcohol intake and age, while reaction time and tracking error were significantly associated with tendency to sleep while driving. Our findings indicate that it is difficult to draw a definitive conclusion about real-world driving ability for the general population based on driving simulator tests. This is because real-world driving is essentially different from the simulators we have used and because our study group (OSA patients) has special characteristics. Nevertheless, driving simulators are attractive because they offer the possibility of controlling certain skills in a way that is difficult to do with on-road testing. Moreover, driving simulators are becoming increasingly realistic. We conclude that performance on driving simulators is associated with some sleep complaints in OSA patients. Although these measures are not directly associated to traffic accident, they are, in fact, associated to related circumstances such as dozing while driving and falling asleep while driving.
Acknowledgements This work was supported by grants from the Health Research Fund (FIS 01/0634) and XUGA (PGIDT99PXI90201A).
References 1. Weaver TE. Outcome measurement in sleep medicine practice and research. Part 2: Assessment of neurobehavioral performance and mood. Sleep Med Rev 2001;5:223–36.
F. Pichel et al. 2. Be´dard MA, Montplaisir J, Malo J, et al. Persistent neuropsycological deficits and vigilance impairment in sleep apnea syndrome after treatment with continuous positive airways pressure (CPAP). J Clin Exp Neuropsychol 1993;15: 330–41. 3. Kribbs NB, Pack AL, Kline LR, et al. Effects of one night without nasal CPAP treatment on sleep and sleepiness in patients with obstructive sleep apnea. Am Rev Respir Dis 1993;147:1162–8. 4. Findley L, Unverzagt M, Guchu R, et al. Vigilance and automobile accidents in patients with sleep apnea or narcolepsy. Chest 1995;108:619–24. 5. George CF, Boudreau AC, Smiley A. Comparison of simulated driving performance in narcolepsy and sleep apnea patients. Sleep 1996;19:711–7. 6. Chugh DK, Weaver TE, Dinges DF. Psychomotor vigilance performance in sleep apnea patients compared to patients presenting with snoring with apnea. Sleep 1998;21(Suppl):159. 7. Dinges D, Maislin G, Staley B, et al. Sleepiness and neurobehavioral functioning in relation to apnea severity in a cohort of commercial motor vehicle operaters. Sleep 1998;21(Suppl):83. 8. Barbe ´ F, Perica ´s J, Mun ˜oz A, et al. Automobile accidents in patients with sleep apnea syndrome. An epidemiological and mechanistic study. Am J Respir Crit Care Med 1998;158: 18–22. 9. Mun ˜oz A, Mayoralas LR, Barbe F, et al. Long-term effects of CPAP on daytime functioning in patients with sleep apnoea syndrome. Eur Respir 2000;15:676–81. 10. Wilkinson RT, Houghton D. Portable four-choice reaction time test with magnetic tape memory. Behav Res Methods Instrum Comput 1995;7:441–6. 11. Dinges DF, Powell JW. Microcomputer analysis of performance on a portable, simple visual RT task during sustained operations. Behav Res Methods Instrum Comput 1985;17: 652–5. 12. George CF, Boudreau AC, Smiley A. Simulated driving performance in patients with obstructive sleep apnea. Am J Respir Crit Care Med 1996;154:175–81. 13. Juniper M, Hack MA, George CF, et al. Steering simulation performance in patients with obstructive sleep apnoea and matched control subjects. Eur Respir J 2000;15:590–5. 14. Findley LJ, Fabrizio MJ, Knight H, et al. Driving simulator performance in patients with sleep apnea. Am Rev Respir Dis 1989;140:529–30. 15. George CF, Boudreau AC, Smiley A. Effects of nasal CPAP on simulated driving performance in patients with obstructive sleep apnoea. Thorax 1997;52:648–53. 16. Hack M, Davies RJ, Mullins R, et al. Randomised prospective parallel trial of therapeutic versus subtherapeutic nasal continuous positive airway pressure on simulated steering performance in patients with obstructive sleep apnoea. Thorax 2000;55:224–31. 17. Hack MA, Choi SJ, Vijayapalan P, et al. Comparison of the effects of sleep deprivation, alcohol and obstructive sleep apnoea (OSA) on simulated steering performance. Respir Med 2001;95:594–601. 18. Findley LJ, Levinson MP, Bonnie RJ. Driving performance and automobile accidents in patients with sleep apnea. Clin Chest Med 1992;13:427–35. 19. Douglas NJ. Systematic review of the efficacy of nasal CPAP. Thorax 1998;53:414–5. 20. Turkington PM, Sircar M, Allgar V, et al. Relationship between obstructive sleep apnoea, driving simulator performance, and risk of road traffic accidents. Thorax 2001;56:800–5.
ARTICLE IN PRESS Sustained attention measurements in obstructive sleep apnea and risk of traffic accidents 21. Telakivi T, Partinen M, Hublin C, et al. Obstructive sleep apnea syndrome among city bus drivers: diagnostic implications and driving capacity. Sleep 2002;25:A–225. 22. Risser MR, Ware JC, Freeman FG. Driving simulation with EEG monitoring in normal and obstructive sleep apnea patients. Sleep 2000;23:393–8. 23. Johns M. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 1991;14:54–5. 24. Johns MW. Sleepiness in different situations measured by the Epworth sleepiness scale. Sleep 1994;17:703–10. 25. Ware Jr JE, Snow KK, Kosinski M, et al. SF-36 Health survey manual and interpretation guide. Boston: The Health Institute, New England Medical Center; 1993. 26. Alonso J, Regidor E, Barrio G, et al. [Population reference values of the Spanish version of the Health Questionnaire SF36] Valores poblacionales de referencia de la version espan ˜ola del Cuestionario de Salud SF-36. Med Clin (Barc) 1998;111:410–6. 27. Partinen M, Gislason T. Basic Nordic Sleep Questionnaire (BNSQ): a quantitated measure of subjective sleep complaints. J Sleep Res 1995;4(Suppl 1):150–5. 28. American Thoracic Society. Sleep apnea, sleepiness,and driving risk. Am J Respir Crit Care Med 1994;50:1463–73. 29. Haraldsson PO, Carenfelt C, Laurell H, et al. Acta Otolaryngol. (Stockholm) 1990;110:136–40. 30. Jenkinson C, Davies R, Mullins R, et al. Randomised prospective parallel trial of therapeutic nasal continuous
31.
32.
33.
34. 35.
36.
1027
positive airway pressure (NCPAP) against sub-therapeutic NCPAP for obstructive sleep apnoea. The Lancet 1999;353: 2100–5. Pichel F, Zamarro ´n C, Maga ´n F, et al. Health-related quality of life in patients with obstructive sleep apnea: effects of long-term positive airway pressure treatment. Respir Med 2004;98:968–76. Schueneman AL, Pickleman J, Freeark RJ. Age, gender, lateral dominance, and prediction of operative skill among general surgery residents. Surgery 1985;98: 506–15. Rizzo M, Sheffield RA. Demographic and driving performance factors in simulator adaptation syndrome. In Proceedings of the second international driving symposium on human factors in driver assessment, training and vehicle design, Iowa, IA, 2003. ppc.uiowa.edu/driving-assessment/2003/ Summaries/Downloads/Final_Papers/PDF/46_Rizzoformat. pdf. Date last accessed: February 2, 2005, 2003. Dinges D. An overview of sleepiness and accidents. J Sleep Res 1995;4:4–14. Masa JF, Rubio M, Findley LJ. Habitually sleepy drivers have a high frequency of automobile crashes associated with respiratory disorders during sleep. Am J Respir Crit Care Med 2000;162:1407–12. Findley LJ, Suratt PM, Dinges DF. Time-on-task decrements in steer clear performance of patients with sleep apnea and narcolepsy. Sleep 1999;22:804–9.