- Email: [email protected]
Distraction and driving: Results from a case–control responsibility study of traffic crash injured drivers interviewed at the emergency room
Accident Analysis and Prevention 59 (2013) 588–592
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Distraction and driving: Results from a case–control responsibility study of traffic crash injured drivers interviewed at the emergency room Sarah Bakiri a,b , Cédric Galéra a,b,c , Emmanuel Lagarde a,b , Magali Laborey a,b , Benjamin Contrand a,b , Régis Ribéreau-Gayon a,b , Louis-Rachid Salmi a,b , Catherine Gabaude d , Alexandra Fort d , Bertrand Maury e , Céline Lemercier f , Maurice Cours g , Manuel-Pierre Bouvard c , Ludivine Orriols a,b,∗ a
Univ. Bordeaux, ISPED, Centre INSERM U897 – Épidémiologie-Biostatistique, F-33000 Bordeaux, France INSERM, Équipe Prévention et prise en charge des traumatismes, ISPED, Centre INSERM U897 – Épidémiologie-Biostatistique, F-33000 Bordeaux, France Univ. Bordeaux, Charles Perrens Hospital, Department of Child and Adolescent Psychiatry, F-33000 Bordeaux, France d LESCOT, Institut National de Recherche sur les Transports et leur Sécurité, F-69500 Bron, France e Laboratoire de Mathématiques d’Orsay, Université Paris Sud 11, F-91400 Orsay, France f CLLE-LTC, CNRS, Université le Mirail, F-31000 Toulouse, France g Continental Automotive SAS France, F-31100 Toulouse, France b c
a r t i c l e
i n f o
Article history: Received 14 January 2013 Received in revised form 2 May 2013 Accepted 3 June 2013 Keywords: Distraction Driving
a b s t r a c t Background: Use of cellular phones has been shown to be associated with crashes but many external distractions remain to be studied. Objective: To assess the risk associated with diversion of attention due to unexpected events or secondary tasks at the wheel. Design: Responsibility case–control study. Setting: Adult emergency department of the Bordeaux University Hospital (France) from April 2010 to August 2011. Participants: 955 injured drivers presenting as a result of motor vehicle crash. Main outcome measures: The main outcome variable was responsibility for the crash. Exposures were external distraction, alcohol use, psychotropic medicine use, and sleep deprivation. Potential confounders were sociodemographic and crash characteristics. Results: Beyond classical risk factor found to be associated with responsibility, results showed that distracting events inside the vehicle (picking up an object), distraction due to driver activity (smoking) and distracting events occurring outside were associated with an increased probability of being at fault. These distraction-related factors accounted for 8% of injurious road crashes. Limitations: Retrospective responsibility self-assessment. Conclusions: Diverted attention may carry more risk than expected. Our results are supporting recent research efforts to detect periods of driving vulnerability related to inattention. © 2013 Elsevier Ltd. All rights reserved.
1. Introduction In high-income countries since the 1970s there has been an overall downward trend in road-crash injuries, despite rising motorization. This trend is the result of successive traffic safety policies targeting human risk factors, the development of safer
∗ Corresponding author at: Injury Prevention and Control – ISPED, INSERM U897, Université Bordeaux Segalen, 146 rue Léo-Saignat, Case 11, 33076 Bordeaux Cedex, France. Tel.: +33 05 57 57 15 04; fax: +33 05 57 57 15 04. E-mail address: [email protected] (L. Orriols). 0001-4575/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.aap.2013.06.004
vehicles and the improvement of road design (Peden et al., 2004). Yet, in recent years, the number of lives saved has plateaued. New frontiers must be explored to achieve further progress. Studies based on expert assessment of crash reports noted that driver distraction may be a major cause of road traffic crashes (Wang et al., 1996; Wilson and Stimpson, 2010). Distraction can be defined as the diversion of attention away from activities critical for safe driving, toward a competing activity, which may result in insufficient or no attention to these critical activities (Lee et al., 2008). This excludes inattention due to driver states that may affect performance (bored, sleepy, fatigued, drunk, under the influence of illegal or medicinal drugs, emotionally upset) or due to cognitive
S. Bakiri et al. / Accident Analysis and Prevention 59 (2013) 588–592
workload induced by internal activities (e.g. daydreaming) (Regan et al., 2011). Experimental studies observing participants’ behaviors when driving an instrumented vehicle with induced distracting tasks showed poorer driving performance (Young et al., 2012a,b; Horberry et al., 2006). In actual driving conditions, an unprecedented study that observed, in an unobtrusive way, 100 drivers’ behaviors in naturalistic settings, concluded that secondary-task distraction was a contributing factor in over 22% of all crashes and near crashes (Klauer et al., 2006). These results were, however, mainly based on the analysis of the risk of near crashes and so far there are no available data on the role of these activities in the risk of actual road traffic crashes. Among these secondary tasks, the impact of cell phone use while driving has been extensively investigated in the past few years, showing that it plays a role in about one in ten crashes (Collet et al., 2010; Strayer and Drews, 2007; Strayer et al., 2011), but other sources of distraction have received much less consideration. To assess the risk associated with diversion of attention due to unexpected events or secondary tasks at the wheel, we performed a responsibility case–control study of traffic crash responsibility in drivers involved in injury crashes interviewed at the adult emergency department of Bordeaux University Hospital, France. 2. Materials and methods 2.1. Study design and setting We performed a comparative study of responsibility in a population of patients involved in injurious road traffic crashes. Its basic principle was to compare the frequency of exposures (distracting activities and confounders) between drivers responsible for the crash (cases) and drivers not responsible for the crash (controls), with cases and controls coming from the same source (same period and location of recruitment). The study was conducted in the adult emergency department of the Bordeaux University Hospital (France) which attends urban and rural populations of an area comprising more than 1.4 million people. Patients were recruited from April 2010 to August 2011. Data were collected by trained interviewers (research assistants) through direct interviews conducted with the patient to obtain information about the crash, patient characteristics and potentially distracting tasks at the time of the crash. Informed consent was obtained from all subjects. This study was approved by the French Data Protection Authority (Commission Nationale Informatique et Libertés). 2.2. Participants Patients were eligible for study inclusion if they had been admitted to the emergency department in the previous 72 h for injury linked to a road traffic crash, were aged 18 years or older, drivers in the crash, and able to answer the interviewer (Glasgow Coma Score = 15 at the time of interview, as determined by the attending physician). 1436 patients were assessed for eligibility. Of these, 368 were excluded for ineligibility (not driver n = 93; admission for more than 72 h n = 29; unable to answer n = 246). This led to a total number of eligible patients of 1068. Of these, 57 refused to participate and a further 56 were excluded from the analysis because of incomplete data. The final sample for analysis comprised 955 patients (89% of the 1068 eligible drivers). Mean time between the accident and the interview was: 4 h 34 min (SD = 12 h 58 min). 2.3. Outcome variable: responsibility for the crash Responsibility levels in the crash were determined by a standardized method adapted from the quantitative Robertson and
589
Drummer crash responsibility instrument (Robertson and Drummer). The Robertson and Drummer’s method was validated in several studies assessing the association between responsibility and exposure to drugs (Robertson and Drummer, 1994; Laumon et al., 2005; Orriols et al., 2010; Lowenstein and Koziol-McLain, 2001). The adaptation of the method to the French context has been validated and presented in previous research (Laumon et al., 2005; Orriols et al., 2010). Notably, this method of determining the driver’s crash responsibility was compared with an independent expert responsibility evaluation, achieving fair agreement with a kappa of 0.71 (Laumon et al., 2005). The method takes into consideration for 6 different mitigating factors considered to reduce driver responsibility: road environment, vehicle-related factors, traffic conditions, type of accident, traffic rule obedience and difficulty of the driving task. Compared to the initial method proposed by Drummer, the adapted method does not use 2 items: witness observations and level of fatigue which are inconsistently available in crash police reports in France. For each factor, a score is assigned from 1 (not mitigating, i.e. favorable to driving) to 3 or 4 (mitigating, i.e. not favorable to driving). All 6 scores are subsequently summated into a summary responsibility score. This summary score was then multiplied by 8/6 to be comparable to the 8 factors score proposed by Robertson and Drummer. Higher scores correspond to lower level of responsibility. The allocation of summary scores was: 8–12, responsible; 13–15, contributory; more than 15, not responsible. Drivers who were assigned any degree of crash responsibility were considered to be cases; drivers who were judged not responsible (score of more than 15) served as controls. Sensitivity analyses were performed to assess the robustness of association estimates to the responsibility determination procedure. The responsibility score was modified by eliminating one by one each of 6 mitigating factors that constitute this score, leading to 6 further responsibility scores based on the remaining 5 mitigating factors. Responsibility cut-points were set using the median. The interviewer was blind to the participant responsibility status when using questionnaire sections related to potential distraction because: (1) responsibility score was computed during the analysis step; (2) traffic rule obedience was reported after the distraction section. All information was obtained from participants.
2.4. Exposure to distractors When interviewed, patients were asked to describe distracting events and activities that occurred just before the crash (the event had to be going on at the time of the driving mistake [inappropriate maneuver, failure to detect a threat, etc.] that led to the crash), from a list of potential distracting events and activities including: listening to the radio or music, watching television, cell phone use (specifying whether hand-held or hand-free), conversation, dialing, text messaging, Internet, navigation system use, reading a road map, having a conversation with or listening to passengers, scolding children, arguing, eating, drinking, smoking, picking up an object, putting on make-up, reading, writing, singing, kissing or hugging, and being distracted by an event outside the vehicle. Potential confounders included patient characteristics (age, gender, socio-economic category), crash characteristics (season, time of the day, vehicle type) and self-reported psychotropic medicine use in the preceding week (for anxiety, depression, other nervous disease, sleep, epilepsy). Patients were also asked how many hours they had slept during the last 24 h. They were considered as sleep-deprived if they reported sleeping less than 6 h. Finally, the participants were questioned about alcohol consumption (within the six preceding hours).
590
S. Bakiri et al. / Accident Analysis and Prevention 59 (2013) 588–592
2.5. Comparison of the study sample with data from the national injury crash database To assess how the study sample was representative of injurious road traffic crashes occurring in France, crash and driver characteristics of the study sample were compared with those of the 2010 national database that includes all injury crashes reported by police forces in the country (Amoros et al., 2006). 2.6. Statistical analysis Bivariate and multivariate analyses (based on logistic regression using SAS statistical software package version 9.3 – SAS Institute Inc., Cary, NC, USA) were conducted to determine associations between exposures and the driver’s responsibility for a road traffic crash (responsible versus not responsible). Distracting events or activities were divided into 7 items: listening to music or the radio, any types of phone use, conversation with passenger, navigation system use, distracting events occurring outside the vehicle, distracting events occurring inside the vehicle and potentially distracting the driver’s activity. The association between responsibility and exposures (distracting events or activities alcohol consumption, psychotropic medicine use and number of sleep hours in the past 24 h) and potential confounders or effect modifiers (age, gender, socio-economic category, season, time of the day, vehicle type) was initially investigated using bivariate analysis. These variables were included in the multivariable model if their p-value was less than 20% (Fisher’s exact test). Interactions between independent variables kept in the final model were tested. Attributable fractions were estimated from the adjusted odds ratio (aOR) estimates and the prevalence of risk factors in responsible drivers (Rockhill et al., 1998).
Table 1 Crash and driver characteristics: comparison of study sample and national police force database.
Gender Male Female Age (years) [18–29] [29–54] ≥55 Occupation Worker/farmer/employee Craftsman, tradesman Manager Unemployed Student Season Summer Autumn Winter Spring Vehicle type Four-wheel motorized Bicycle Two-wheel motorized Time of day [6–21] [21–6] Road type Motorway State road County road Local (urban) road Alcohol consumption Yes No
Study sample
National police force database
N
%
N
%
612 393
60.9 39.1
81,287 30,217
72.9 27.1
348 492 165
34.6 49.0 16.4
34,979 55,877 20,648
31.4 50.1 18.5
105 60 600 119 121
10.5 6.0 59.7 11.8 12.0
14,023 4104 32,728 13,800 9300
19.0 5.5 44.2 18.7 12.6
274 261 146 324
27.3 26.0 14.5 32.2
29,946 21,504 21,471 28,583
26.9 28.3 19.3 25.6
528 194 277
52.9 19.4 27.7
78,182 4326 27,094
71.3 4.0 24.7
921 84
91.6 8.4
96,158 15,346
86.2 13.8
31 49 139 778
3.1 4.9 13.9 77.8
8845 7146 37,159 58,354
7.9 6.4 33.3 52.3
67 888
7.0 93.0
9819 101,676
8.8 91.2
3. Results Table 1 compares crash and driver characteristics of the study sample with those of the national police force database. The study sample comprises more drivers being managers, more crashes with a bicycle and that occurred on local urban roads. Based on reports on their crash characteristics, 453 (47%) participants were classified as responsible and 502 (53%) as not responsible. As expected, bivariate analysis showed a higher probability of being responsible for a crash occurring at night, when the driver reported having drunk alcohol within the six preceding hours, reported psychotropic medicine use and had slept less than 6 h in the preceding 24 h (Table 2). Listening to the radio, using a telephone (dialing, text messaging or talking), having a conversation with passenger, and using a navigation system were not associated with a higher probability of being responsible for the crash. After adjustment for the variables found to be associated with responsibility (see Table 3), distraction due to events occurring outside the vehicle, to events occurring inside the vehicle and to driver activity were associated with a strong increase of the probability of being responsible for the crash. Sensitivity analysis showed the same pattern of associations. The attributable fractions for these 3 risk factors were 2%, 3% and 4%, respectively. When grouped together, reported distracting events and activities were estimated to account for 8% of all crashes; 11 out of the 17 reported events inside the vehicle corresponded to a driver who was picking up an object. No significant interaction was found. 4. Discussion Distracting events inside the vehicle (picking up an object), distraction due to driver activity (smoking) and distracting events
occurring outside were associated with an increased probability of being at fault. These distraction-related factors accounted for 8% of injurious road crashes. This is, to our knowledge, the first epidemiological study to assess the risk of road traffic crashes associated with driver distraction. The responsibility case–control design is a real strength of the study as cases and controls share some common characteristics: they were all driving a vehicle. In a previous study on the impact of illegal drug consumption, using the same police national database, but limited to fatal crashes (Laumon et al., 2005), the method used to determine responsibility was approved by an independent expert evaluation of responsibility. Importantly, responsibility levels were computed independently of alcohol and illicit drug use because of their potential interactions with distraction. In addition, as responsibility was determined while analyzing the data, the interviewers were unable to distinguish between cases and controls when interviewing the driver. Another option would be to select controls on the side of the road. This method is however likely to induce measurement bias as potential participants are disturbed by being stopped, generally fearing to be fined by the police. This is likely to lead to spurious responses when asked to report on their activities while driving. The principle of the responsibility analysis is that if a factor contributes to road traffic crash causation, it is expected that it would be overrepresented in the responsible drivers. The method does not, however, capture the risk, for non-responsible drivers, of being unable to avoid a crash. Because distraction may also be involved in such crash mechanism, our attributable fraction estimate may be too low.
S. Bakiri et al. / Accident Analysis and Prevention 59 (2013) 588–592 Table 2 Bivariate responsibility analysis. Participants Gender 580 Male Female 375 Age (years) 335 [18–29] [29–54] 461 ≥55 159 Occupation 101 Worker/farmer/employee Craftsman, tradesman 57 Manager 580 101 Unemployed 116 Student Season 258 Summer Autumn 245 137 Winter Spring 315 Vehicle type 503 Four-wheel motorized Bicycle 188 264 Two-wheel motorized Time of day (h) [6–21] 881 [21–6] 74 Alcohol consumption 888 No Yes 67 Psychotropic medicine use No 854 Yes 101 Sleep hours 851 Hours of sleep > 6 Hours of sleep < 6 104 Listening to music or the radio 694 No Yes 261 Phone use 947 No Yes 8 Had a conversation with passenger No 911 Yes 44 Navigation system use 945 No Yes 10 Distracting event outside the vehicle 936 No Yes 19 Distracting event inside the vehiclea No 938 Yes 17 Distraction related to driver’s activityb No 929 Yes 26 Passenger 860 Absence Adults 60 35 Children
Responsible drivers % (n)
p-value (N2 )
49.1 (285) 44.8 (168)
0.19
51.9 (174) 43.6 (201) 49.1 (78)
0.06
50.5 (51) 45.6 (26) 46.2 (268) 44.6 (45) 54.3 (63)
0.5
50.8 (131) 46.9 (115) 35.8 (49) 50.2 (158)
0.02
45.9 (231) 51.1 (96) 47.7 (126)
0.4
45.4 (400) 71.6 (53)
<0.0001
45.3 (402) 76.1 (51)
<0.0001
45.9 (392) 60.4 (61)
0.006
45.0 (383) 67.3 (70)
<0.0001
46.1 (320) 51.0 (133)
0.19
47.2 (447) 75.0 (6)
0.16
48.0 (437) 36.4 (16)
0.13
47.3 (447) 60.0 (6)
0.53
46.9 (439) 73.7 (14)
0.021
46.7 (438) 88.2 (15)
0.0007
46.4 (431) 84.6 (22)
0.0001
48.4 (416) 33.3 (20) 48.6 (17)
0.08
a 11 picking up an object (10 responsible); 2 having an argument with passenger (1 responsible); and 4 scolding a child (4 responsible). b 14 smoking (11 responsible); all other activities were reported by one or two drivers: tuning the radio, smelling a flower, drinking, putting on makeup, eating, blowing one’s nose, looking in a mirror, and readjusting a garment.
Another strength of the study is the availability of several potential confounders: alcohol consumption, sleep deprivation, psychotropic medicine use, and age. However, a number of confounders may not have been taken into account, such as illegal drug consumption, mileage exposure, lighting and risky driving behaviors.
591
Table 3 Multivariate responsibility analysis performed by logistic regression (N = 955 participants). Adjusted OR Gender 1.00 Male 0.82 Female Age (years) [18–29] 1.40 [29–54] 1.00 ≥55 1.18 Vehicle type Four-wheel motorized 1.00 Bicycle 1.21 Two-wheel motorized 1.06 Time of day (hours) [6–21] 1.00 [21–6] 1.94 Alcohol consumption No 1.00 Yes 2.45 Psychotropic medicine use No 1.00 Yes 1.66 Sleep hours Hours of sleep > 6 1.00 Hours of sleep <6 2.06 Distracting event outside the vehicle 1.00 No Yes 3.27 Distracting event inside the vehiclea No 1.00 Yes 7.25 Distraction due to the driver’s activityb No 1.00 Yes 4.80
95% CI
n
[0.62–1.10]
580 375
[1.04–1.90]
Attributable fraction (%)
[0.80–1.73]
335 461 159
[0.85–1.73] [0.76–1.47]
503 188 264
[1.10–3.43]
881 74
[1.30–4.58]
888 67
[1.06–2.62]
854 101
[1.31–3.23]
851 104
– [1.15–9.35]
936 19
2
– [1.57–33.44]
938 17
3
– [1.58–14.60]
929 26
4
a 11 picking up an object (10 responsible); 2 having an argument with passenger (1 responsible); 4 scolding a child (4 responsible). b 14 smoking (11 responsible); all other activities were reported by one or two drivers: tuning the radio, smelling a flower, drinking, putting on makeup, eating, blowing one’s nose, looking in a mirror, and readjusting a garment.
The lack of association with listening to music is consistent with results from simulation-based studies (Consiglio et al., 2003; Strayer and Johnston, 2001). Results from the naturalistic driving study based on observation of 100 instrumented cars evidenced a range of driver activities associated with a risk of near-crashes. The most important one was picking up an object, with an OR of 8.82 [2.50–31.2], a result highly consistent with our data as this event was also strongly associated with the risk of being responsible in our study. Also consistent with our results, the same 100-car naturalistic driving study showed that looking at an event outside the vehicle and putting on makeup were also found to be associated with near-crashes. Reading at the wheel and dialing are the two distracting activities cited from the 100-car study that did not show up. This is, however, not surprising as the study was conducted in the US, with a significant proportion of mileages driven on rural roads, while 78% of the crashes from our sample happened on urban roads, where such behaviors, which require long off-road glances, are less likely. One unexpected result of our study was that cell phone use was rarely reported (by 0.8% of participants) and was not associated with responsibility. We were however unable to assess correctly the association as the number of drivers reporting cell phone use was very small. The impact of cell phone use on the risk of road traffic crashes has been well documented (Collet et al., 2010), including in case–control studies (Sagberg, 2001). The risk of road traffic crashes is increased by a factor of 3–4 when the driver is using a cell phone. The last survey conducted in 2011 in France reported that 2.0% of drivers of passenger vehicles use a hand-held phone at
592
S. Bakiri et al. / Accident Analysis and Prevention 59 (2013) 588–592
the wheel (ONISR, 2011). Our results may be explained by the fear of reporting an illegal practice, but could also stem from a simple sampling fluctuation. Some limitations should be taken into account when interpreting the results. The use of a Glasgow coma score of 15 excludes drivers with more severe head injuries, which limits the generalizability of the findings. Retrospective responsibility self-assessment might have influenced the reports of distraction, because of incomplete recall or desirability bias toward the interviewer. As mentioned earlier, this may be the case for the use of cell phones at the wheel, which is prohibited in France. The questionnaire was, however, administered in a medical setting where the police are not supposed to be present and participants were informed and provided with a signed consent form that included sentences explaining that their reports would be kept anonymous. Another possible limitation is that drivers who believed themselves to be at fault might be more likely to report distracting activities that would explain their role in the crash. Another limitation is the urban environment from which the participants originated. Our sample did, however, include crashes that occurred on motorways, state and county roads. Driver distraction appeared to be associated with an increased risk of being at fault in a road traffic crash, suggesting a potential causative role. Our data led us to estimate that 8% of injurious crashes could have been attributed to distraction. Picking up an object and smoking were the most prevalent distractors. The impact of these activities has hardly been addressed in the literature (Koppel et al., 2011; Brison, 1990). The small number of drivers concerned in our sample leads us to recommend further study of these particular issues, as potential prevention implications are obvious. More generally, there is a lack of studies on the overall impact in road safety of inattention, due to distraction, sleepiness or mind wandering. Several recent research programs have undertaken to find new methods to assist the driver by detecting periods of inattention (Kircher and Ahlstrom, 2010; Liang et al., 2007). One of the strategies that are being considered is to develop adaptive mitigation systems which would provide assistance and warning based on the state of the driver, assessed using real-time information collected by sensors and/or driving data analyzers. Funding The research was supported by an ANR (French National Research Agency) grant (ANR-09-VTT-04, project ATLAS). Competing interests All authors declare no financial relationships with any organization that might have an interest in the submitted work in the previous three years; and no other relationships or activities that could appear to have influenced the submitted work. References Amoros, E., Martin, J.L., Laumon, B., 2006. Under-reporting of road crash casualties in France. Accident; Analysis and Prevention 38 (July (4)), 627–635.
Brison, R.J., 1990. Risk of automobile accidents in cigarette smokers. Canadian Journal of Public Health 81 (March–April (2)), 102–106. Collet, C., Guillot, A., Petit, C., 2010. Phoning while driving I: a review of epidemiological, psychological, behavioural and physiological studies. Ergonomics 53 (5), 589–601. Consiglio, W., Driscoll, P., Witte, M., Berg, W.P., 2003. Effect of cellular telephone conversations and other potential interference on reaction time in a braking response. Accident; Analysis and Prevention 35 (July (4)), 495–500. Horberry, T., Anderson, J., Regan, M.A., Triggs, T.J., Brown, J., 2006. Driver distraction: the effects of concurrent in-vehicle tasks, road environment complexity and age on driving performance. Accident; Analysis and Prevention 38 (January (1)), 185–191. Kircher, K., Ahlstrom, C., 2010. Predicting visual distraction using driving performance data. Annals of Advances in Automotive Medicine 54, 333–342. Klauer, S.G., Dingus, T.A., Neale, V.L., Sudweeks, J.D., Ramsey, D.J., 2006. The impact of driver inattention on near-crash/crash risk: an analysis using the 100-car Naturalistic Driving Study data. Virginia Tech Transportation Institute, DOT HS 810 594. Koppel, S., Charlton, J., Kopinathan, C., Taranto, D., 2011. Are child occupants a significant source of driving distraction? Accident; Analysis and Prevention 43 (May (3)), 1236–1244. Laumon, B., Gadegbeku, B., Martin, J.L., Biecheler, M.B., 2005. Cannabis intoxication and fatal road crashes in France: population based case–control study. British Medical Journal 331 (7529), 1371. Lee, J.D., Young, K.L., Regan, M.A., 2008. Defining driver distgraction. In: Regan, M.A., Lee, J.D., Young, K.L. (Eds.), Driver Distraction: Theory, Effects, and Mitigation. CRC Press, Taylor & Francis Group, Boca Raton, pp. 31–40. Liang, Y., Reyes, M.L., Lee, J.D., 2007. Real-time detection of driver cognitive distraction using support vector machines. IEEE Transactions on Intelligent Transportation System 8 (2), 340–350. Lowenstein, S.R., Koziol-McLain, J., 2001. Drugs and traffic crash responsibility: a study of injured motorists in Colorado. The Journal of Trauma 50 (February (2)), 313–320. ONISR (Observatoire national interministériel de sécurité routière), 2011. La sécurité routière en France. Bilan de l‘année 2012, 66–67. Orriols, L., Delorme, B., Gadegbeku, B., et al., 2010. Prescription medicines and the risk of road traffic crashes: a French registry-based study. PLoS Medicine 7 (11), e1000366, http://dx.doi.org/10.1371/journal.pmed.1000366. Peden, M., Scurfield, R., Sleet, D., et al., 2004. World Report on Road Traffic Injury Prevention. World Health Organization, WHO, Geneva. Regan, M.A., Hallett, C., Gordon, C.P., 2011. Driver distraction and driver inattention: definition, relationship and taxonomy. Accident; Analysis and Prevention 43 (September (5)), 1771–1781. Robertson, M.D., Drummer, O.H., 1994. Responsibility analysis: a methodology to study the effects of drugs in driving. Accident; Analysis and Prevention 26 (April (2)), 243–247. Rockhill, B., Newman, B., Weinberg, C., 1998. Use and misuse of population attributable fractions. American Journal of Public Health 88 (January (1)), 15–19. Sagberg, F., 2001. Accident risk of car drivers during mobile telephone use. International Journal of Vehicle Design 26, 57–69. Strayer, D.L., Drews, F.A., 2007. Multi-tasking in the automobile. In: Kramer, A., Wiegmann, D., Kirlik, A. (Eds.), Applied Attention: From Theory to Practice. Oxford University Press, pp. 121–133. Strayer, D.L., Johnston, W.A., 2001. Driven to distraction: dual-task studies of simulated driving and conversing on a cellular telephone. Psychological Science 12 (November (6)), 462–466. Strayer, D.L., Watson, J.M., Drews, F.A., 2011. Cognitive distraction while multitasking in the automobile. In: Ross, B. (Ed.), The Psychology of Learning and Motivation, vol. 54. Academic Press, Burlington, pp. 29–58. Wang, J.S., Knipling, R.R., Goodman, M.J., October 7–9, 1996. The role of driver inattention in crashes; new statistics from the 1995 crashworthiness data system. In: Paper presented at the 20th Annual Proceedings of the Association for the Advancement of Automotive Medicine, Vancouver, British Columbia. Wilson, F.A., Stimpson, J.P., 2010. Trends in fatalities from distracted driving in the United States, 1999 to 2008. American Journal of Public Health 100 (November (11)), 2213–2219. Young, K.L., Salmon, P.M., Cornelissen, M., 2012a. Distraction-induced driving error: An on-road examination of the errors made by distracted and undistracted drivers. Accident; Analysis and Prevention (June (22)) [Epub ahead of print]. Young, K.L., Mitsopoulos-Rubens, E., Rudin-Brown, C.M., Lenne, M.G., 2012b. The effects of using a portable music player on simulated driving performance and task-sharing strategies. Applied Ergonomics 43 (July (4)), 738–746.
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Distraction and driving: Results from a case–control responsibility study of traffic crash injured drivers interviewed at the emergency room Sarah Bakiri a,b , Cédric Galéra a,b,c , Emmanuel Lagarde a,b , Magali Laborey a,b , Benjamin Contrand a,b , Régis Ribéreau-Gayon a,b , Louis-Rachid Salmi a,b , Catherine Gabaude d , Alexandra Fort d , Bertrand Maury e , Céline Lemercier f , Maurice Cours g , Manuel-Pierre Bouvard c , Ludivine Orriols a,b,∗ a
Univ. Bordeaux, ISPED, Centre INSERM U897 – Épidémiologie-Biostatistique, F-33000 Bordeaux, France INSERM, Équipe Prévention et prise en charge des traumatismes, ISPED, Centre INSERM U897 – Épidémiologie-Biostatistique, F-33000 Bordeaux, France Univ. Bordeaux, Charles Perrens Hospital, Department of Child and Adolescent Psychiatry, F-33000 Bordeaux, France d LESCOT, Institut National de Recherche sur les Transports et leur Sécurité, F-69500 Bron, France e Laboratoire de Mathématiques d’Orsay, Université Paris Sud 11, F-91400 Orsay, France f CLLE-LTC, CNRS, Université le Mirail, F-31000 Toulouse, France g Continental Automotive SAS France, F-31100 Toulouse, France b c
a r t i c l e
i n f o
Article history: Received 14 January 2013 Received in revised form 2 May 2013 Accepted 3 June 2013 Keywords: Distraction Driving
a b s t r a c t Background: Use of cellular phones has been shown to be associated with crashes but many external distractions remain to be studied. Objective: To assess the risk associated with diversion of attention due to unexpected events or secondary tasks at the wheel. Design: Responsibility case–control study. Setting: Adult emergency department of the Bordeaux University Hospital (France) from April 2010 to August 2011. Participants: 955 injured drivers presenting as a result of motor vehicle crash. Main outcome measures: The main outcome variable was responsibility for the crash. Exposures were external distraction, alcohol use, psychotropic medicine use, and sleep deprivation. Potential confounders were sociodemographic and crash characteristics. Results: Beyond classical risk factor found to be associated with responsibility, results showed that distracting events inside the vehicle (picking up an object), distraction due to driver activity (smoking) and distracting events occurring outside were associated with an increased probability of being at fault. These distraction-related factors accounted for 8% of injurious road crashes. Limitations: Retrospective responsibility self-assessment. Conclusions: Diverted attention may carry more risk than expected. Our results are supporting recent research efforts to detect periods of driving vulnerability related to inattention. © 2013 Elsevier Ltd. All rights reserved.
1. Introduction In high-income countries since the 1970s there has been an overall downward trend in road-crash injuries, despite rising motorization. This trend is the result of successive traffic safety policies targeting human risk factors, the development of safer
∗ Corresponding author at: Injury Prevention and Control – ISPED, INSERM U897, Université Bordeaux Segalen, 146 rue Léo-Saignat, Case 11, 33076 Bordeaux Cedex, France. Tel.: +33 05 57 57 15 04; fax: +33 05 57 57 15 04. E-mail address: [email protected] (L. Orriols). 0001-4575/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.aap.2013.06.004
vehicles and the improvement of road design (Peden et al., 2004). Yet, in recent years, the number of lives saved has plateaued. New frontiers must be explored to achieve further progress. Studies based on expert assessment of crash reports noted that driver distraction may be a major cause of road traffic crashes (Wang et al., 1996; Wilson and Stimpson, 2010). Distraction can be defined as the diversion of attention away from activities critical for safe driving, toward a competing activity, which may result in insufficient or no attention to these critical activities (Lee et al., 2008). This excludes inattention due to driver states that may affect performance (bored, sleepy, fatigued, drunk, under the influence of illegal or medicinal drugs, emotionally upset) or due to cognitive
S. Bakiri et al. / Accident Analysis and Prevention 59 (2013) 588–592
workload induced by internal activities (e.g. daydreaming) (Regan et al., 2011). Experimental studies observing participants’ behaviors when driving an instrumented vehicle with induced distracting tasks showed poorer driving performance (Young et al., 2012a,b; Horberry et al., 2006). In actual driving conditions, an unprecedented study that observed, in an unobtrusive way, 100 drivers’ behaviors in naturalistic settings, concluded that secondary-task distraction was a contributing factor in over 22% of all crashes and near crashes (Klauer et al., 2006). These results were, however, mainly based on the analysis of the risk of near crashes and so far there are no available data on the role of these activities in the risk of actual road traffic crashes. Among these secondary tasks, the impact of cell phone use while driving has been extensively investigated in the past few years, showing that it plays a role in about one in ten crashes (Collet et al., 2010; Strayer and Drews, 2007; Strayer et al., 2011), but other sources of distraction have received much less consideration. To assess the risk associated with diversion of attention due to unexpected events or secondary tasks at the wheel, we performed a responsibility case–control study of traffic crash responsibility in drivers involved in injury crashes interviewed at the adult emergency department of Bordeaux University Hospital, France. 2. Materials and methods 2.1. Study design and setting We performed a comparative study of responsibility in a population of patients involved in injurious road traffic crashes. Its basic principle was to compare the frequency of exposures (distracting activities and confounders) between drivers responsible for the crash (cases) and drivers not responsible for the crash (controls), with cases and controls coming from the same source (same period and location of recruitment). The study was conducted in the adult emergency department of the Bordeaux University Hospital (France) which attends urban and rural populations of an area comprising more than 1.4 million people. Patients were recruited from April 2010 to August 2011. Data were collected by trained interviewers (research assistants) through direct interviews conducted with the patient to obtain information about the crash, patient characteristics and potentially distracting tasks at the time of the crash. Informed consent was obtained from all subjects. This study was approved by the French Data Protection Authority (Commission Nationale Informatique et Libertés). 2.2. Participants Patients were eligible for study inclusion if they had been admitted to the emergency department in the previous 72 h for injury linked to a road traffic crash, were aged 18 years or older, drivers in the crash, and able to answer the interviewer (Glasgow Coma Score = 15 at the time of interview, as determined by the attending physician). 1436 patients were assessed for eligibility. Of these, 368 were excluded for ineligibility (not driver n = 93; admission for more than 72 h n = 29; unable to answer n = 246). This led to a total number of eligible patients of 1068. Of these, 57 refused to participate and a further 56 were excluded from the analysis because of incomplete data. The final sample for analysis comprised 955 patients (89% of the 1068 eligible drivers). Mean time between the accident and the interview was: 4 h 34 min (SD = 12 h 58 min). 2.3. Outcome variable: responsibility for the crash Responsibility levels in the crash were determined by a standardized method adapted from the quantitative Robertson and
589
Drummer crash responsibility instrument (Robertson and Drummer). The Robertson and Drummer’s method was validated in several studies assessing the association between responsibility and exposure to drugs (Robertson and Drummer, 1994; Laumon et al., 2005; Orriols et al., 2010; Lowenstein and Koziol-McLain, 2001). The adaptation of the method to the French context has been validated and presented in previous research (Laumon et al., 2005; Orriols et al., 2010). Notably, this method of determining the driver’s crash responsibility was compared with an independent expert responsibility evaluation, achieving fair agreement with a kappa of 0.71 (Laumon et al., 2005). The method takes into consideration for 6 different mitigating factors considered to reduce driver responsibility: road environment, vehicle-related factors, traffic conditions, type of accident, traffic rule obedience and difficulty of the driving task. Compared to the initial method proposed by Drummer, the adapted method does not use 2 items: witness observations and level of fatigue which are inconsistently available in crash police reports in France. For each factor, a score is assigned from 1 (not mitigating, i.e. favorable to driving) to 3 or 4 (mitigating, i.e. not favorable to driving). All 6 scores are subsequently summated into a summary responsibility score. This summary score was then multiplied by 8/6 to be comparable to the 8 factors score proposed by Robertson and Drummer. Higher scores correspond to lower level of responsibility. The allocation of summary scores was: 8–12, responsible; 13–15, contributory; more than 15, not responsible. Drivers who were assigned any degree of crash responsibility were considered to be cases; drivers who were judged not responsible (score of more than 15) served as controls. Sensitivity analyses were performed to assess the robustness of association estimates to the responsibility determination procedure. The responsibility score was modified by eliminating one by one each of 6 mitigating factors that constitute this score, leading to 6 further responsibility scores based on the remaining 5 mitigating factors. Responsibility cut-points were set using the median. The interviewer was blind to the participant responsibility status when using questionnaire sections related to potential distraction because: (1) responsibility score was computed during the analysis step; (2) traffic rule obedience was reported after the distraction section. All information was obtained from participants.
2.4. Exposure to distractors When interviewed, patients were asked to describe distracting events and activities that occurred just before the crash (the event had to be going on at the time of the driving mistake [inappropriate maneuver, failure to detect a threat, etc.] that led to the crash), from a list of potential distracting events and activities including: listening to the radio or music, watching television, cell phone use (specifying whether hand-held or hand-free), conversation, dialing, text messaging, Internet, navigation system use, reading a road map, having a conversation with or listening to passengers, scolding children, arguing, eating, drinking, smoking, picking up an object, putting on make-up, reading, writing, singing, kissing or hugging, and being distracted by an event outside the vehicle. Potential confounders included patient characteristics (age, gender, socio-economic category), crash characteristics (season, time of the day, vehicle type) and self-reported psychotropic medicine use in the preceding week (for anxiety, depression, other nervous disease, sleep, epilepsy). Patients were also asked how many hours they had slept during the last 24 h. They were considered as sleep-deprived if they reported sleeping less than 6 h. Finally, the participants were questioned about alcohol consumption (within the six preceding hours).
590
S. Bakiri et al. / Accident Analysis and Prevention 59 (2013) 588–592
2.5. Comparison of the study sample with data from the national injury crash database To assess how the study sample was representative of injurious road traffic crashes occurring in France, crash and driver characteristics of the study sample were compared with those of the 2010 national database that includes all injury crashes reported by police forces in the country (Amoros et al., 2006). 2.6. Statistical analysis Bivariate and multivariate analyses (based on logistic regression using SAS statistical software package version 9.3 – SAS Institute Inc., Cary, NC, USA) were conducted to determine associations between exposures and the driver’s responsibility for a road traffic crash (responsible versus not responsible). Distracting events or activities were divided into 7 items: listening to music or the radio, any types of phone use, conversation with passenger, navigation system use, distracting events occurring outside the vehicle, distracting events occurring inside the vehicle and potentially distracting the driver’s activity. The association between responsibility and exposures (distracting events or activities alcohol consumption, psychotropic medicine use and number of sleep hours in the past 24 h) and potential confounders or effect modifiers (age, gender, socio-economic category, season, time of the day, vehicle type) was initially investigated using bivariate analysis. These variables were included in the multivariable model if their p-value was less than 20% (Fisher’s exact test). Interactions between independent variables kept in the final model were tested. Attributable fractions were estimated from the adjusted odds ratio (aOR) estimates and the prevalence of risk factors in responsible drivers (Rockhill et al., 1998).
Table 1 Crash and driver characteristics: comparison of study sample and national police force database.
Gender Male Female Age (years) [18–29] [29–54] ≥55 Occupation Worker/farmer/employee Craftsman, tradesman Manager Unemployed Student Season Summer Autumn Winter Spring Vehicle type Four-wheel motorized Bicycle Two-wheel motorized Time of day [6–21] [21–6] Road type Motorway State road County road Local (urban) road Alcohol consumption Yes No
Study sample
National police force database
N
%
N
%
612 393
60.9 39.1
81,287 30,217
72.9 27.1
348 492 165
34.6 49.0 16.4
34,979 55,877 20,648
31.4 50.1 18.5
105 60 600 119 121
10.5 6.0 59.7 11.8 12.0
14,023 4104 32,728 13,800 9300
19.0 5.5 44.2 18.7 12.6
274 261 146 324
27.3 26.0 14.5 32.2
29,946 21,504 21,471 28,583
26.9 28.3 19.3 25.6
528 194 277
52.9 19.4 27.7
78,182 4326 27,094
71.3 4.0 24.7
921 84
91.6 8.4
96,158 15,346
86.2 13.8
31 49 139 778
3.1 4.9 13.9 77.8
8845 7146 37,159 58,354
7.9 6.4 33.3 52.3
67 888
7.0 93.0
9819 101,676
8.8 91.2
3. Results Table 1 compares crash and driver characteristics of the study sample with those of the national police force database. The study sample comprises more drivers being managers, more crashes with a bicycle and that occurred on local urban roads. Based on reports on their crash characteristics, 453 (47%) participants were classified as responsible and 502 (53%) as not responsible. As expected, bivariate analysis showed a higher probability of being responsible for a crash occurring at night, when the driver reported having drunk alcohol within the six preceding hours, reported psychotropic medicine use and had slept less than 6 h in the preceding 24 h (Table 2). Listening to the radio, using a telephone (dialing, text messaging or talking), having a conversation with passenger, and using a navigation system were not associated with a higher probability of being responsible for the crash. After adjustment for the variables found to be associated with responsibility (see Table 3), distraction due to events occurring outside the vehicle, to events occurring inside the vehicle and to driver activity were associated with a strong increase of the probability of being responsible for the crash. Sensitivity analysis showed the same pattern of associations. The attributable fractions for these 3 risk factors were 2%, 3% and 4%, respectively. When grouped together, reported distracting events and activities were estimated to account for 8% of all crashes; 11 out of the 17 reported events inside the vehicle corresponded to a driver who was picking up an object. No significant interaction was found. 4. Discussion Distracting events inside the vehicle (picking up an object), distraction due to driver activity (smoking) and distracting events
occurring outside were associated with an increased probability of being at fault. These distraction-related factors accounted for 8% of injurious road crashes. This is, to our knowledge, the first epidemiological study to assess the risk of road traffic crashes associated with driver distraction. The responsibility case–control design is a real strength of the study as cases and controls share some common characteristics: they were all driving a vehicle. In a previous study on the impact of illegal drug consumption, using the same police national database, but limited to fatal crashes (Laumon et al., 2005), the method used to determine responsibility was approved by an independent expert evaluation of responsibility. Importantly, responsibility levels were computed independently of alcohol and illicit drug use because of their potential interactions with distraction. In addition, as responsibility was determined while analyzing the data, the interviewers were unable to distinguish between cases and controls when interviewing the driver. Another option would be to select controls on the side of the road. This method is however likely to induce measurement bias as potential participants are disturbed by being stopped, generally fearing to be fined by the police. This is likely to lead to spurious responses when asked to report on their activities while driving. The principle of the responsibility analysis is that if a factor contributes to road traffic crash causation, it is expected that it would be overrepresented in the responsible drivers. The method does not, however, capture the risk, for non-responsible drivers, of being unable to avoid a crash. Because distraction may also be involved in such crash mechanism, our attributable fraction estimate may be too low.
S. Bakiri et al. / Accident Analysis and Prevention 59 (2013) 588–592 Table 2 Bivariate responsibility analysis. Participants Gender 580 Male Female 375 Age (years) 335 [18–29] [29–54] 461 ≥55 159 Occupation 101 Worker/farmer/employee Craftsman, tradesman 57 Manager 580 101 Unemployed 116 Student Season 258 Summer Autumn 245 137 Winter Spring 315 Vehicle type 503 Four-wheel motorized Bicycle 188 264 Two-wheel motorized Time of day (h) [6–21] 881 [21–6] 74 Alcohol consumption 888 No Yes 67 Psychotropic medicine use No 854 Yes 101 Sleep hours 851 Hours of sleep > 6 Hours of sleep < 6 104 Listening to music or the radio 694 No Yes 261 Phone use 947 No Yes 8 Had a conversation with passenger No 911 Yes 44 Navigation system use 945 No Yes 10 Distracting event outside the vehicle 936 No Yes 19 Distracting event inside the vehiclea No 938 Yes 17 Distraction related to driver’s activityb No 929 Yes 26 Passenger 860 Absence Adults 60 35 Children
Responsible drivers % (n)
p-value (N2 )
49.1 (285) 44.8 (168)
0.19
51.9 (174) 43.6 (201) 49.1 (78)
0.06
50.5 (51) 45.6 (26) 46.2 (268) 44.6 (45) 54.3 (63)
0.5
50.8 (131) 46.9 (115) 35.8 (49) 50.2 (158)
0.02
45.9 (231) 51.1 (96) 47.7 (126)
0.4
45.4 (400) 71.6 (53)
<0.0001
45.3 (402) 76.1 (51)
<0.0001
45.9 (392) 60.4 (61)
0.006
45.0 (383) 67.3 (70)
<0.0001
46.1 (320) 51.0 (133)
0.19
47.2 (447) 75.0 (6)
0.16
48.0 (437) 36.4 (16)
0.13
47.3 (447) 60.0 (6)
0.53
46.9 (439) 73.7 (14)
0.021
46.7 (438) 88.2 (15)
0.0007
46.4 (431) 84.6 (22)
0.0001
48.4 (416) 33.3 (20) 48.6 (17)
0.08
a 11 picking up an object (10 responsible); 2 having an argument with passenger (1 responsible); and 4 scolding a child (4 responsible). b 14 smoking (11 responsible); all other activities were reported by one or two drivers: tuning the radio, smelling a flower, drinking, putting on makeup, eating, blowing one’s nose, looking in a mirror, and readjusting a garment.
Another strength of the study is the availability of several potential confounders: alcohol consumption, sleep deprivation, psychotropic medicine use, and age. However, a number of confounders may not have been taken into account, such as illegal drug consumption, mileage exposure, lighting and risky driving behaviors.
591
Table 3 Multivariate responsibility analysis performed by logistic regression (N = 955 participants). Adjusted OR Gender 1.00 Male 0.82 Female Age (years) [18–29] 1.40 [29–54] 1.00 ≥55 1.18 Vehicle type Four-wheel motorized 1.00 Bicycle 1.21 Two-wheel motorized 1.06 Time of day (hours) [6–21] 1.00 [21–6] 1.94 Alcohol consumption No 1.00 Yes 2.45 Psychotropic medicine use No 1.00 Yes 1.66 Sleep hours Hours of sleep > 6 1.00 Hours of sleep <6 2.06 Distracting event outside the vehicle 1.00 No Yes 3.27 Distracting event inside the vehiclea No 1.00 Yes 7.25 Distraction due to the driver’s activityb No 1.00 Yes 4.80
95% CI
n
[0.62–1.10]
580 375
[1.04–1.90]
Attributable fraction (%)
[0.80–1.73]
335 461 159
[0.85–1.73] [0.76–1.47]
503 188 264
[1.10–3.43]
881 74
[1.30–4.58]
888 67
[1.06–2.62]
854 101
[1.31–3.23]
851 104
– [1.15–9.35]
936 19
2
– [1.57–33.44]
938 17
3
– [1.58–14.60]
929 26
4
a 11 picking up an object (10 responsible); 2 having an argument with passenger (1 responsible); 4 scolding a child (4 responsible). b 14 smoking (11 responsible); all other activities were reported by one or two drivers: tuning the radio, smelling a flower, drinking, putting on makeup, eating, blowing one’s nose, looking in a mirror, and readjusting a garment.
The lack of association with listening to music is consistent with results from simulation-based studies (Consiglio et al., 2003; Strayer and Johnston, 2001). Results from the naturalistic driving study based on observation of 100 instrumented cars evidenced a range of driver activities associated with a risk of near-crashes. The most important one was picking up an object, with an OR of 8.82 [2.50–31.2], a result highly consistent with our data as this event was also strongly associated with the risk of being responsible in our study. Also consistent with our results, the same 100-car naturalistic driving study showed that looking at an event outside the vehicle and putting on makeup were also found to be associated with near-crashes. Reading at the wheel and dialing are the two distracting activities cited from the 100-car study that did not show up. This is, however, not surprising as the study was conducted in the US, with a significant proportion of mileages driven on rural roads, while 78% of the crashes from our sample happened on urban roads, where such behaviors, which require long off-road glances, are less likely. One unexpected result of our study was that cell phone use was rarely reported (by 0.8% of participants) and was not associated with responsibility. We were however unable to assess correctly the association as the number of drivers reporting cell phone use was very small. The impact of cell phone use on the risk of road traffic crashes has been well documented (Collet et al., 2010), including in case–control studies (Sagberg, 2001). The risk of road traffic crashes is increased by a factor of 3–4 when the driver is using a cell phone. The last survey conducted in 2011 in France reported that 2.0% of drivers of passenger vehicles use a hand-held phone at
592
S. Bakiri et al. / Accident Analysis and Prevention 59 (2013) 588–592
the wheel (ONISR, 2011). Our results may be explained by the fear of reporting an illegal practice, but could also stem from a simple sampling fluctuation. Some limitations should be taken into account when interpreting the results. The use of a Glasgow coma score of 15 excludes drivers with more severe head injuries, which limits the generalizability of the findings. Retrospective responsibility self-assessment might have influenced the reports of distraction, because of incomplete recall or desirability bias toward the interviewer. As mentioned earlier, this may be the case for the use of cell phones at the wheel, which is prohibited in France. The questionnaire was, however, administered in a medical setting where the police are not supposed to be present and participants were informed and provided with a signed consent form that included sentences explaining that their reports would be kept anonymous. Another possible limitation is that drivers who believed themselves to be at fault might be more likely to report distracting activities that would explain their role in the crash. Another limitation is the urban environment from which the participants originated. Our sample did, however, include crashes that occurred on motorways, state and county roads. Driver distraction appeared to be associated with an increased risk of being at fault in a road traffic crash, suggesting a potential causative role. Our data led us to estimate that 8% of injurious crashes could have been attributed to distraction. Picking up an object and smoking were the most prevalent distractors. The impact of these activities has hardly been addressed in the literature (Koppel et al., 2011; Brison, 1990). The small number of drivers concerned in our sample leads us to recommend further study of these particular issues, as potential prevention implications are obvious. More generally, there is a lack of studies on the overall impact in road safety of inattention, due to distraction, sleepiness or mind wandering. Several recent research programs have undertaken to find new methods to assist the driver by detecting periods of inattention (Kircher and Ahlstrom, 2010; Liang et al., 2007). One of the strategies that are being considered is to develop adaptive mitigation systems which would provide assistance and warning based on the state of the driver, assessed using real-time information collected by sensors and/or driving data analyzers. Funding The research was supported by an ANR (French National Research Agency) grant (ANR-09-VTT-04, project ATLAS). Competing interests All authors declare no financial relationships with any organization that might have an interest in the submitted work in the previous three years; and no other relationships or activities that could appear to have influenced the submitted work. References Amoros, E., Martin, J.L., Laumon, B., 2006. Under-reporting of road crash casualties in France. Accident; Analysis and Prevention 38 (July (4)), 627–635.
Brison, R.J., 1990. Risk of automobile accidents in cigarette smokers. Canadian Journal of Public Health 81 (March–April (2)), 102–106. Collet, C., Guillot, A., Petit, C., 2010. Phoning while driving I: a review of epidemiological, psychological, behavioural and physiological studies. Ergonomics 53 (5), 589–601. Consiglio, W., Driscoll, P., Witte, M., Berg, W.P., 2003. Effect of cellular telephone conversations and other potential interference on reaction time in a braking response. Accident; Analysis and Prevention 35 (July (4)), 495–500. Horberry, T., Anderson, J., Regan, M.A., Triggs, T.J., Brown, J., 2006. Driver distraction: the effects of concurrent in-vehicle tasks, road environment complexity and age on driving performance. Accident; Analysis and Prevention 38 (January (1)), 185–191. Kircher, K., Ahlstrom, C., 2010. Predicting visual distraction using driving performance data. Annals of Advances in Automotive Medicine 54, 333–342. Klauer, S.G., Dingus, T.A., Neale, V.L., Sudweeks, J.D., Ramsey, D.J., 2006. The impact of driver inattention on near-crash/crash risk: an analysis using the 100-car Naturalistic Driving Study data. Virginia Tech Transportation Institute, DOT HS 810 594. Koppel, S., Charlton, J., Kopinathan, C., Taranto, D., 2011. Are child occupants a significant source of driving distraction? Accident; Analysis and Prevention 43 (May (3)), 1236–1244. Laumon, B., Gadegbeku, B., Martin, J.L., Biecheler, M.B., 2005. Cannabis intoxication and fatal road crashes in France: population based case–control study. British Medical Journal 331 (7529), 1371. Lee, J.D., Young, K.L., Regan, M.A., 2008. Defining driver distgraction. In: Regan, M.A., Lee, J.D., Young, K.L. (Eds.), Driver Distraction: Theory, Effects, and Mitigation. CRC Press, Taylor & Francis Group, Boca Raton, pp. 31–40. Liang, Y., Reyes, M.L., Lee, J.D., 2007. Real-time detection of driver cognitive distraction using support vector machines. IEEE Transactions on Intelligent Transportation System 8 (2), 340–350. Lowenstein, S.R., Koziol-McLain, J., 2001. Drugs and traffic crash responsibility: a study of injured motorists in Colorado. The Journal of Trauma 50 (February (2)), 313–320. ONISR (Observatoire national interministériel de sécurité routière), 2011. La sécurité routière en France. Bilan de l‘année 2012, 66–67. Orriols, L., Delorme, B., Gadegbeku, B., et al., 2010. Prescription medicines and the risk of road traffic crashes: a French registry-based study. PLoS Medicine 7 (11), e1000366, http://dx.doi.org/10.1371/journal.pmed.1000366. Peden, M., Scurfield, R., Sleet, D., et al., 2004. World Report on Road Traffic Injury Prevention. World Health Organization, WHO, Geneva. Regan, M.A., Hallett, C., Gordon, C.P., 2011. Driver distraction and driver inattention: definition, relationship and taxonomy. Accident; Analysis and Prevention 43 (September (5)), 1771–1781. Robertson, M.D., Drummer, O.H., 1994. Responsibility analysis: a methodology to study the effects of drugs in driving. Accident; Analysis and Prevention 26 (April (2)), 243–247. Rockhill, B., Newman, B., Weinberg, C., 1998. Use and misuse of population attributable fractions. American Journal of Public Health 88 (January (1)), 15–19. Sagberg, F., 2001. Accident risk of car drivers during mobile telephone use. International Journal of Vehicle Design 26, 57–69. Strayer, D.L., Drews, F.A., 2007. Multi-tasking in the automobile. In: Kramer, A., Wiegmann, D., Kirlik, A. (Eds.), Applied Attention: From Theory to Practice. Oxford University Press, pp. 121–133. Strayer, D.L., Johnston, W.A., 2001. Driven to distraction: dual-task studies of simulated driving and conversing on a cellular telephone. Psychological Science 12 (November (6)), 462–466. Strayer, D.L., Watson, J.M., Drews, F.A., 2011. Cognitive distraction while multitasking in the automobile. In: Ross, B. (Ed.), The Psychology of Learning and Motivation, vol. 54. Academic Press, Burlington, pp. 29–58. Wang, J.S., Knipling, R.R., Goodman, M.J., October 7–9, 1996. The role of driver inattention in crashes; new statistics from the 1995 crashworthiness data system. In: Paper presented at the 20th Annual Proceedings of the Association for the Advancement of Automotive Medicine, Vancouver, British Columbia. Wilson, F.A., Stimpson, J.P., 2010. Trends in fatalities from distracted driving in the United States, 1999 to 2008. American Journal of Public Health 100 (November (11)), 2213–2219. Young, K.L., Salmon, P.M., Cornelissen, M., 2012a. Distraction-induced driving error: An on-road examination of the errors made by distracted and undistracted drivers. Accident; Analysis and Prevention (June (22)) [Epub ahead of print]. Young, K.L., Mitsopoulos-Rubens, E., Rudin-Brown, C.M., Lenne, M.G., 2012b. The effects of using a portable music player on simulated driving performance and task-sharing strategies. Applied Ergonomics 43 (July (4)), 738–746.