- Email: [email protected]
Risky driving habits and motor vehicle driver injury
Accident Analysis and Prevention 37 (2005) 619–624
Risky driving habits and motor vehicle driver injury Stephanie Blows a,∗ , Shanthi Ameratunga b , Rebecca Q. Ivers a , Sing Kai Lo a , Robyn Norton a a
The George Institute for International Health, University of Sydney, P.O. Box M201, Missenden Road, Sydney, NSW 2050, Australia b Section of Epidemiology and Biostatistics, School of Population Health, Faculty of Medical & Health Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand Received 23 December 2004; received in revised form 14 March 2005; accepted 14 March 2005
Abstract Risky driving is an important cause of motor vehicle injury, but there is a lack of good epidemiological data in this field, particularly data comparing risky driving in younger drivers to those of other age groups. We examined the relationship between risky driving habits, prior traffic convictions and motor vehicle injury using cross-sectional data amongst 21,893 individuals in New Zealand, including 8029 who were aged 16–24 years. Those who reported frequently racing a motor vehicle for excitement or driving at 20 km/h or more over the speed limit, and those who had received traffic convictions over the past 12 months, were between two and four times more likely to have been injured while driving over the same time period. Driving unlicensed was a risk factor for older but not younger drivers, and driving at 20 km/h or more above the speed limits was a stronger risk factor for younger (<25 years) than older drivers. These results confirm the need for interventions targeting risky driving and suggest that different strategies may be required for different high-risk groups. © 2005 Elsevier Ltd. All rights reserved. Keywords: Risky driving; Traffic conviction; Young driver; Motor vehicle injury
1. Introduction Motor vehicle crashes contribute significantly to the burden of injury and death worldwide and risky driving behaviours, such as drink driving, speeding and non-use of seatbelts, are considered responsible for a significant proportion of this global burden (World Health Organisation, 2003). Other risky driving behaviours such as racing other vehicles for thrills, close following and illegal passing, have also been associated with increased risk in a number of cohort, case control and cross-sectional studies (Evans and Wasielewski, 1982, 1983; Preusser et al., 1991; Centres for Disease Control and Prevention, 1994; Rajalin, 1994; Harrison, 1997; Begg et al., 1999; Bell et al., 2000; Fergusson et al., 2003; Lam, 2003). Risky driving behaviours may be studied as ‘acute’ behaviours (a single incidence of a behaviour that is temporally ∗
Corresponding author. Tel.: +61 2 9993 4585; fax: +61 2 9993 4502. E-mail addresses: [email protected] (S. Blows), [email protected] (S. Ameratunga), [email protected] (R.Q. Ivers), [email protected] (S.K. Lo), [email protected] (R. Norton). 0001-4575/$ – see front matter © 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.aap.2005.03.003
related to the immediate risk of an injury crash) and/or ‘habitual’ (usual or ongoing) risky driving behaviours. Many previous studies have shown that acute risky driving behaviours increase the risk of an injury crash, but several studies have also suggested that people who report ‘habitual’ risky driving and have a history of convictions are also at increased risk (Peck, 1993; Centres for Disease Control and Prevention, 1994; Rajalin, 1994; Bell et al., 2000; Fergusson et al., 2003). Despite this, a recent systematic review of the literature on risky driving and car crash injury found just four epidemiological studies that rigorously examined the association between risk taking and car crash injury (Turner et al., 2004). The authors of this review concluded that although research so far has found a consistent relationship between risk taking and car crash injury, there is a need for more studies of risky behaviour that use injury as an outcome in order to effectively advocate for public health interventions (Turner et al., 2004). Furthermore, although some prior studies have focused on young drivers as a high-risk group (Begg et al., 1999; Fergusson et al., 2003; Lam, 2003; Begg and Langley, 2004), there is a lack of data comparing the effects of risky driving in younger versus older drivers.
620
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624
We examined the associations between habitual risky driving, prior convictions and motor vehicle driver injury using cross-sectional data from the New Zealand Blood Donors’ Health Study (Ameratunga et al., 2002). This ongoing, prospective cohort study aims to examine lifestyle and psychological determinants of serious injury. Baseline data from the study provided an opportunity to examine the cross-sectional associations between risky driving habits and motor vehicle driver injuries in an opportunistic sample of more than 22,000 individuals.
2. Methods The New Zealand Blood Donors’ Health Study is a prospective cohort study of blood donors in the top half of the north island of New Zealand. Recruitment took place between April 1998 and October 1999. The methodology of the study has previously been published elsewhere (Ameratunga et al., 2002). Briefly, individuals aged at least 16 years attending blood donation facilities (including schools, workplaces, shopping centres and mobile sites) in the region were invited to participate in the study. After providing informed consent, participants completed a self-administered baseline questionnaire, including a range of questions covering demographics, drug and alcohol use, smoking habits, risk taking behaviour, social situation, sleep habits, mental health, medical history and injuries sustained in the previous 12 months in a variety of settings including while driving. Participants gave consent for their baseline data collected at recruitment to be linked prospectively to outcome data. This electronic record linkage is expected to occur periodically over the next few decades and will involve confidential matching of baseline data collected at recruitment to the study to the New Zealand Health Information Service’s ‘National Minimum Dataset’. This is a national dataset covering information on all deaths and hospital discharges from public and most private hospitals. Injury information in this dataset, including diagnoses and external causes, is coded according to the International Classification of Diseases (Ministry of Health, 1999). This paper describes cross-sectional associations between self-reported risky driving and injury variables measured in the baseline survey. There were 22,389 participants in the study (81% response rate) and of these, 96% completed the baseline questionnaire. Using specified response categories, participants were asked to report the frequency over the past 12 months of the following risky driving behaviours: driving over the legal limit for alcohol; racing a motor vehicle for excitement; driving more than 20 km per hour (km/h) over the speed limit; not wearing a seatbelt when traveling as a driver or passenger; and driving on a public road without a current license. Different forms of unlicensed driving were not distinguished in this study and the ‘driving unlicensed’ variable included both those who had never held a license and those who held a disqualified or suspended license. Par-
ticipants were also asked to report the number of traffic convictions (excluding parking infringements) they had received over the past 12 months. The outcome variable was measured by asking if the respondent had sustained any injury during the past 12 months while driving a car, van, truck or similar vehicle, where treatment from a doctor was required. Injuries sustained by other occupants of the vehicle, and other road users such as pedestrians, were not measured. We examined the associations between risky driving habits, history of traffic offences and self-reported driver injury by calculating prevalence ratios (PR) and 95% confidence intervals (CI). Potential confounding variables were identified from the road safety literature. These were ethnicity, education level, New Zealand socioeconomic index score (Davis et al., 1999), the ‘CAGE’ score (a screening test for hazardous drinking; Ewing, 1984), drinking frequency in an average week, number of drinks on an average drinking occasion, likelihood of dozing while driving, Epworth sleepiness scale score (a measure of daytime sleepiness; Johns, 1991), frequency of marijuana and other illegal drug use and driving exposure (average number of hours spent driving per week). Of these, only those that were significantly associated with a history of driver injury after adjustment for age and sex were included in the logistic regression models. We tested for trend in association between number of convictions and driver injury using the unadjusted CochranArmitage trend test. Results were stratified by age group to examine the effects of risky driving in younger drivers (those aged less than 25 years) compared to drivers aged 25 years or more.
3. Results Of the 21,893 participants completing the baseline survey, 276 reported a history of motor vehicle driving injury. While a higher proportion of those reporting a history of motor vehicle driving injury, compared with those not reporting such injury, were aged 16–24 years (45.7% versus 36.6%, p = 0.002), no sex differences were observed (p = 0.34). The proportions of study participants who reported engaging in risky driving behaviours, by history of driver injury, are shown in Table 1. Those reporting a history of driver injury tended to report higher levels of driving above the legal limit for alcohol, driving unlicensed, racing for excitement, driving over the speed limit and not wearing a seatbelt, and had more prior convictions than those without a history of driver injury. Missing data for all exposure variables was less than 10%. Table 2 shows the unadjusted and multivariable adjusted prevalence ratios and 95% confidence intervals for associations between risky driving habits, history of convictions and driver injury. After adjustment for age, sex, CAGE score, number of drinks on average drinking occasion, frequency of marijuana use in the past 12 months and driving exposure (average hours per week), exposures significantly associated
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624 Table 1 Number and percentage of participants reporting history of driver injury in the past 12 months, by age, sex, risky driving behaviours and history of driving convictions, New Zealand Blood Donors Health Study, 1998–1999 History of driver injury (n = 276) Age (years) 16–24 25–39 40+ Sex Male Female
No history of driver injury (n = 21,617)
Frequency
(%)
126 73 77
45.7 26.4 27.9
Frequency 7903 5189 8525
(%) 36.6 24.0 39.4
120 156
43.5 56.5
10027 11590
46.4 53.6
Racing for excitement Never 212 Once or twice 18 Several times 34 Always 6 Not applicable 2 Missing 4
76.8 6.5 12.3 2.2 0.7 1.5
18358 1709 724 132 321 373
84.9 7.9 3.4 0.6 1.5 1.7
Driving ≥20 km/h over speed limit Never 70 Once or twice 80 Several times 104 Always 20 Not applicable 0 Missing 2
25.4 29.0 37.7 7.3 0.0 0.7
7424 7061 6032 477 284 339
34.3 32.7 27.9 2.2 1.3 1.6
Wearing a seatbelt Never Once or twice Several times Always Not applicable Missing
10 4 9 248 2 3
3.6 1.5 3.3 89.9 0.7 1.1
595 264 787 19556 93 322
2.8 1.2 3.6 90.5 0.4 1.5
Convictions past 12 months None 193 One 45 Two 14 Three or more 11 Missing 13
69.9 16.3 5.1 4.0 4.7
17868 1726 359 186 1671
81.8 8.0 1.7 0.8 7.7
with driver injury were racing a motor vehicle for excitement (PR 2.4, 95% CI 1.6–3.7), driving at 20 km/h or more over the speed limit (PR 2.5, 95% CI 1.4–4.3), and number of traffic convictions in the past 12 months (one conviction, PR 2.1, 95% CI 1.5–3.0; two convictions PR 2.7, 95% CI 1.5–4.9; three convictions PR 3.4, 95% CI 1.8–6.6). The number of convictions was the only variable for which all levels suggestive of risk were significantly associated with car crash injury, so we tested for trend in this variable; this test was significant (χ32 = 73.6, p < 0.0001). Of those reporting driver injury, 126 (45.7%) were aged less than 25 years and 150 (54.3%) were aged 25 years or older. Table 3 shows the multivariable adjusted prevalence ratios for exposures that were significantly associated with driver injury after stratification by age group. Driving unlicensed was associated with increased risk of injury for drivers aged 25 and over (PR 4.0, 95% CI 1.2–12.8) but not younger
621
Table 2 Prevalence ratios and 95% confidence intervals for associations between risky driving behaviours, history of convictions, and driver injury, New Zealand Blood Donors Health Study, 1998–1999 Unadjusted
Multivariable adjusteda
PR
95% CI
PR
95% CI
Driving while over alcohol limit Never 1.0 Once or twice 1.4 Several times/always 1.4
1.1–1.9 0.8–2.6
1.0 1.2 0.9
0.8–1.6 0.5–1.7
Driving unlicensed Never Once or twice Several times/always
1.0 1.7 2.0
1.0–2.8 1.2–3.4
1.0 1.6 1.3
0.9–2.8 0.7–2.4
Racing for excitement Never Once or twice Several times/always
1.0 1.0 3.9
0.8–1.2 2.8–5.5
1.0 0.8 2.4
0.6–1.0 1.6–3.7
Driving ≥20 km/h over speed limit Never 1.0 Once or twice 1.3 Several times 1.9 Always 4.5
0.9–1.7 1.4–2.6 2.8–7.5
1.0 1.0 1.2 2.5
0.7–1.4 0.8–1.7 1.4–4.3
Not wearing a seatbelt Never Once/twice/several times Always
0.4–1.8 0.5–1.4
1.0 0.8 1.0
0.3–2.0 0.5–2.0
Traffic convictions in past 12 months None 1.0 One 2.4 1.7–3.3 Two 3.6 2.1–6.1 Three or more 5.3 2.9–9.7
1.0 2.1 2.7 3.4
1.5–3.0 1.5–4.9 1.8–6.6
Frequency of risky driving in past 12 months
1.0 0.8 0.9
a Adjusted for age, sex, CAGE score, number of drinks on average drinking occasion, frequency of marijuana use in the past 12 months and driving exposure (average hours per week).
drivers. Racing for excitement was associated with a significant increase in risk for both drivers aged 25 and over (PR 3.8, 95% CI 1.8–8.0) and those under 25 (PR 2.3, 95% CI 1.4–3.7). Driving at 20 km/h or more over the speed limit was associated with an increased risk for younger drivers (PR 3.4, 95% CI 1.6–7.0); the increased risk was not significant in older drivers (PR 2.0, 95% CI 0.7–5.7). Higher numbers of convictions in the past 12 months was also significantly associated with increased risk of injury in both age groups, although the increased risk was not significant for drivers aged 25 and over with three or more convictions. This may reflect the smaller number of events in this age group and resulting imprecision of estimates.
4. Discussion We found that individuals who reported frequently engaging in risky driving behaviours over the past 12 months were between two and four times more likely to have been injured while driving during the same time period, compared to individuals who reported infrequently or never engaging in these
622
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624
Table 3 Multivariable-adjusted prevalence ratiosa (95% CI) for driver injury, for younger (<25 years) and older (≥25 years) drivers, New Zealand Blood Donors Health Study, 1998–1999b Driver <25 years
Driver ≥25 years
PR
95% CI
PR
95% CI
Driving unlicensed Never Once or twice Several times/always
1.0 1.6 1.1
0.9–2.9 0.6–2.1
1.0 1.9 4.0
0.3–14.0 1.2–12.8
Racing for excitement Never Once or twice Several times/always
1.0 0.9 2.3
0.7–1.2 1.4–3.7
1.0 0.4 3.8
0.2–1.1 1.8–8.0
Driving ≥20 km/h over speed limit Never 1.0 Once or twice 1.2 Several times 1.6 Always 3.4
0.6–2.2 0.9–2.8 1.6–7.0
1.0 1.0 1.2 2.0
0.7–1.6 0.7–1.8 0.7–5.7
Traffic convictions in past 12 months None 1.0 One 2.2 Two 3.3 Three or more 4.1
1.3–3.8 1.5–7.3 1.9–8.9
1.0 1.9 2.4 2.3
1.2–3.0 1.0–5.5 0.5–9.2
Frequency of risky driving in past 12 months
a
Adjusted for sex, CAGE score, number of drinks on average drinking occasion, frequency of marijuana use in the past 12 months and driving exposure (average hours per week). b Prevalence ratios are only reported for those exposures containing categories that were significantly associated with injury after stratification by age group.
behaviours. The specific behaviours that were associated with driver injury in these data were racing a motor vehicle for excitement, driving at 20 km/h or more over the speed limit, and a higher number of traffic convictions. We also found evidence that some of these associations were different for younger and older drivers. Specifically, driving unlicensed was a risk factor for older but not younger drivers and driving at 20 km/h or more above the speed limits was a stronger risk factor for younger than older drivers. These findings have some limitations. We asked only one question about each form of risky driving, and we cannot be sure that this was a reliable and valid measure. Many of these behaviours are illegal, and as with all self-reported data, we cannot exclude the risk of socially desirable responses. This may mean we have underestimated the magnitude of the association between risky driving and injury. However, we used standardised questions of risky driving based on previous studies and noted a wide distribution in responses across the levels of risk. The measure of driver injury did not distinguish between levels of injury severity, beyond the requirement for treatment from a doctor, and we did not measure or control for acute driving-related factors in the analyses. In this cross-sectional analysis, although the temporal relationships between risky driving, convictions and driver injury cannot be determined with certainty, it seems unlikely that having a driver injury would increase risky driving behaviour. On the other hand, the direction of the association
between having a driver injury and traffic convictions seems less certain. No doubt in some cases, the injury incident may have also resulted in a traffic conviction and/or loss of a driver license. This was a study of volunteer blood donors, who are unlikely to be representative of the general population, and for this reason we did not estimate population prevalence of exposures or outcomes. As published previously, however, the study population represented a broad cross-section of society and substantial heterogeneity in the distribution of potential risk factors for driver injury, making it possible to determine the associations of interest (Ameratunga et al., 2002). This comparatively large study was also able to assess a range of both risky driving behaviours and potential confounding variables and included a high proportion of young drivers, allowing useful comparisons to be made across age groups. The findings from this study support those from several previous studies that have also found that risky driving habits are associated with driver injury. The Christchurch Health and Development Study, a birth cohort of 1265 children born in 1977, found that the risk of a crash increased significantly with the number of risky driving behaviours reported, including speeding, drink driving, close following, not wearing a seatbelt, racing and running red lights (Fergusson et al., 2003). Risky driving behaviours as predictors of injury have also been investigated in another prospective cohort study of young New Zealanders, the Dunedin Multidisciplinary Health and Development Study (Begg and Langley, 2001; Begg et al., 2003). Like ours, this study did not find a significant association between habitual seatbelt use and car crash injury (Begg and Langley, 2000). A population based case–control study conducted in the United States also examined associations between habitual risky driving and car crash injury (Centres for Disease Control and Prevention, 1994). Driving more than 20 miles/h over the speed limit, passing a car in a no-passing zone, and taking risks for fun while driving were associated with an increased risk of injury. Other risky driving habits, including non-use of seatbelts, close following behind other drivers, passing in no-passing zones, taking risks in traffic for fun and running red lights, have also been found to be associated with increased risk of injury (Centres for Disease Control and Prevention, 1994; Bell et al., 2000; Norris et al., 2000; Fergusson et al., 2003). Previous studies have also found an association between convictions and crash risk. A population based case–control study in Finland examined fatal crashes and a random selection of license holders as controls (Rajalin, 1994). This study found that fatally crashed male drivers were significantly more likely to have traffic offences on their driving record after controlling for age and an estimate of miles driven by the driver annually. A literature review of variables associated with increased crash risk concluded that prior traffic offences were a good predictor of subsequent crash involvement (Peck, 1993). Our results, in combination with these studies, thus provide strong evidence for an association between risky driving habits, previous convictions and driver injury.
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624
We also found differences in the association between some risky driving habits and injury according to the driver’s age. Although one prior study has suggested that younger drivers killed in crashes are more likely than drivers of other ages to have been driving in a risky manner (Zhang et al., 1998), we did not find any previously published research suggesting that the effect of risky driving differed by age. This finding is therefore a potentially important contribution of the current study and may have significant implications. For example, lack of experience in young drivers may mean they are less likely to exercise safe judgment when driving at high speeds. However, it should be noted that only a small number of people reported ‘always’ speeding, and that the point estimate of the prevalence ratio for older drivers also suggest an effect although this was not significant. This may be a worthwhile area for future research. The finding that unlicensed driving is a risk factor for injury crashes in older but not young drivers is also unexpected. One possible explanation might be that more young people may drive unlicensed before they are old enough to gain a license, making this a more normative behaviour and less an indicator of risky driving in this age group, whereas older drivers who drive without a valid license represent a more risky driving group. However, we did not distinguish between drivers who had never held a license and those whose license was disqualified or suspended, and there may be age differences between these two groups. Although habitual risk takers may be difficult to identify, this is an important at-risk group amongst whom interventions to reduce car crash injury should be considered. Those whose risk taking had resulted in a conviction were a highrisk group in our study. These drivers may be specifically targeted with health promotion strategies such as education and community awareness campaigns in collaboration with organisations such as the police force. This strategy, if proven to be effective, may also be considered for other identifiable groups such as older unlicensed drivers. Our results also suggest that speed-related behaviours are of particular concern, especially in young drivers. Potentially useful strategies to counter this may include increased enforcement and higher penalties for risky driving, including loss of license. Unfortunately risky driving behaviours, especially speeding, are frequently reinforced by popular culture (Henley, 2004) and for this reason, behaviour change may not be achieved by health promotion strategies alone. Legislative approaches, such as modifications to Graduated Licensing Systems, may be effective in newly licensed drivers. Popular culture itself can also be a worthwhile target for interventions. For example, a recent Australian voluntary code of practice for motor vehicle advertising has been developed, stating that advertisements for motor vehicles should not portray any form of unsafe or reckless driving or speeding (Federal Chamber of Automotive Industries, 2004), although its success to date appears to be limited by the voluntary nature of the code. This study confirms previous evidence that individuals who are habitually risky drivers and have a history of convictions are more likely to experience driver injury. The
623
results suggest that speed-related behaviours are especially risky, particularly in young drivers, and that older unlicensed drivers are also a high-risk group. The cohort component of this study will prospectively examine the risk for serious injury and death in this population.
Acknowledgements The New Zealand Blood Donors’ Health Study was funded by the Health Research Council of New Zealand and its conduct supported by the New Zealand Blood Service. We acknowledge with gratitude the participants in this study and the contributions of Dr. Gary Whitlock (Research Fellow), Di Smith (Project Manager) and the collaborators involved with the study.
References Ameratunga, S., Norton, R.N., et al., 2002. The New Zealand Blood Donors’ Health Study: baseline findings of a large prospective cohort study of injury. Inj. Prev. 8, 66–69. Begg, D.J., Langley, J.D., 2000. Seat-belt use and related behaviors among young adults. J. Saf. Res. 31 (4), 211–220. Begg, D.J., Langley, J.D., 2001. Changes in risky driving behaviour from age 21 to 26 years. J. Saf. Res. 32, 491–499. Begg, D.J., Langley, J.D., 2004. Identifying predictors of persistent nonalcohol or drug-related risky driving behaviours among a cohort of young adults. Accid. Anal. Prev. 36, 1067–1071. Begg, D.J., Langley, J.D., et al., 2003. Identifying factors that predict persistent driving after drinking, and driving after using cannabis among young adults. Accid. Anal. Prev. 35, 669–675. Begg, D.J., Langley, J.D., et al., 1999. A longitudinal study of lifestyle factors as predictors of injuries and crashes among young adults. Accid. Anal. Prev. 31, 1–11. Bell, N.S., Amoroso, P.J., et al., 2000. Self-reported risk-taking behaviors and hospitalization for motor vehicle injury among active duty army personnel. Am. J. Prev. Med. 18 (Suppl. 3), 85–95. Centres for Disease Control and Prevention, 1993. From the Centers for Disease Control and Prevention. Risky driving behaviors among teenagers—Gwinnett County, Georgia. JAMA 272 (11), 844–845. Davis, P., McLeod, K., et al., 1999. The New Zealand Socioeconomic Index: develping and validating an occupationally-derived indicator of socio-economic status. Aust. N.Z. J. Public Health 23 (1), 27–33. Evans, L., Wasielewski, P., 1982. Do accident-involved drivers exhibit riskier everyday driving behaviour? Accid. Anal. Prev. 14 (1), 57–64. Evans, L., Wasielewski, P., 1983. Risky driving related to driver and vehicle characteristics. Accid. Anal. Prev. 15 (2), 121–136. Ewing, J.A., 1984. Detecting alcoholism: the CAGE questionnaire. JAMA 252, 1905–1907. Federal Chamber of Automotive Industries (2004). Advertising for motor vehicles: voluntary code of practice. Australia, Advertising Standards Bureau. Fergusson, D.M., Swain-Campbell, N., et al., 2003. Risky driving behaviour in young people: prevalence, personal characteristics and traffic accidents. Aust. N.Z. J. Public Health 27 (3), 337–342. Harrison, W.A., 1997. An exploratory investigation of the crash involvement of disqualified drivers and motorcyclists. J. Saf. Res. 28 (3), 213–219. Henley, N., 2004. Social marketing: ‘selling’ injury prevention. In: McClure, R., Stevenson, M., McEvoy, S. (Eds.), The Scientific Basis of Injury Prevention and Control. IP Communications, Melbourne.
624
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624
Johns, M., 1991. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 14, 54–55. Lam, L.T., 2003. Factors associated with fatal and injurious car crash among learner drivers in New South Wales, Australia. Accid. Anal. Prev. 35 (3), 333–340. Ministry of Health, 1999. Our Health, Our Future: The health of New Zealanders. New Zealand Ministry of Health, Wellington. Norris, F.H., Matthews, B.A., et al., 2000. Characterological, situational, and behavioral risk factors for motor vehicle accidents: a prospective examination. Accid. Anal. Prev. 32 (4), 505– 515. Peck, R.C., 1993. The identification of multiple accident correlates in high risk drivers with specific emphasis on the role of age, experience
and prior traffic violation frequency. Alcohol, Drugs, Driving 9, 145– 166. Preusser, D.F., Williams, A.F., et al., 1991. Characteristics of belted and unbelted drivers. Accid. Anal. Prev. 23 (6), 475–482. Rajalin, S., 1994. The connection between risky driving and involvement in fatal accidents. Accid. Anal. Prev. 26 (5), 555–562. Turner, C., McClure, R., et al., 2004. Injury and risk-taking behaviour—a systematic review. Accid. Anal. Prev. 36 (1), 93–101. World Health Organisation, 2003. Facts About Injuries: Road Traffic Injuries. World Health Organisation, Geneva. Zhang, J., Fraser, S., et al., 1998. Age-specific patterns of factors related to fatal motor vehicle crashes: focus on young and elderly drivers. Public Health 112, 289–295.
Risky driving habits and motor vehicle driver injury Stephanie Blows a,∗ , Shanthi Ameratunga b , Rebecca Q. Ivers a , Sing Kai Lo a , Robyn Norton a a
The George Institute for International Health, University of Sydney, P.O. Box M201, Missenden Road, Sydney, NSW 2050, Australia b Section of Epidemiology and Biostatistics, School of Population Health, Faculty of Medical & Health Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand Received 23 December 2004; received in revised form 14 March 2005; accepted 14 March 2005
Abstract Risky driving is an important cause of motor vehicle injury, but there is a lack of good epidemiological data in this field, particularly data comparing risky driving in younger drivers to those of other age groups. We examined the relationship between risky driving habits, prior traffic convictions and motor vehicle injury using cross-sectional data amongst 21,893 individuals in New Zealand, including 8029 who were aged 16–24 years. Those who reported frequently racing a motor vehicle for excitement or driving at 20 km/h or more over the speed limit, and those who had received traffic convictions over the past 12 months, were between two and four times more likely to have been injured while driving over the same time period. Driving unlicensed was a risk factor for older but not younger drivers, and driving at 20 km/h or more above the speed limits was a stronger risk factor for younger (<25 years) than older drivers. These results confirm the need for interventions targeting risky driving and suggest that different strategies may be required for different high-risk groups. © 2005 Elsevier Ltd. All rights reserved. Keywords: Risky driving; Traffic conviction; Young driver; Motor vehicle injury
1. Introduction Motor vehicle crashes contribute significantly to the burden of injury and death worldwide and risky driving behaviours, such as drink driving, speeding and non-use of seatbelts, are considered responsible for a significant proportion of this global burden (World Health Organisation, 2003). Other risky driving behaviours such as racing other vehicles for thrills, close following and illegal passing, have also been associated with increased risk in a number of cohort, case control and cross-sectional studies (Evans and Wasielewski, 1982, 1983; Preusser et al., 1991; Centres for Disease Control and Prevention, 1994; Rajalin, 1994; Harrison, 1997; Begg et al., 1999; Bell et al., 2000; Fergusson et al., 2003; Lam, 2003). Risky driving behaviours may be studied as ‘acute’ behaviours (a single incidence of a behaviour that is temporally ∗
Corresponding author. Tel.: +61 2 9993 4585; fax: +61 2 9993 4502. E-mail addresses: [email protected] (S. Blows), [email protected] (S. Ameratunga), [email protected] (R.Q. Ivers), [email protected] (S.K. Lo), [email protected] (R. Norton). 0001-4575/$ – see front matter © 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.aap.2005.03.003
related to the immediate risk of an injury crash) and/or ‘habitual’ (usual or ongoing) risky driving behaviours. Many previous studies have shown that acute risky driving behaviours increase the risk of an injury crash, but several studies have also suggested that people who report ‘habitual’ risky driving and have a history of convictions are also at increased risk (Peck, 1993; Centres for Disease Control and Prevention, 1994; Rajalin, 1994; Bell et al., 2000; Fergusson et al., 2003). Despite this, a recent systematic review of the literature on risky driving and car crash injury found just four epidemiological studies that rigorously examined the association between risk taking and car crash injury (Turner et al., 2004). The authors of this review concluded that although research so far has found a consistent relationship between risk taking and car crash injury, there is a need for more studies of risky behaviour that use injury as an outcome in order to effectively advocate for public health interventions (Turner et al., 2004). Furthermore, although some prior studies have focused on young drivers as a high-risk group (Begg et al., 1999; Fergusson et al., 2003; Lam, 2003; Begg and Langley, 2004), there is a lack of data comparing the effects of risky driving in younger versus older drivers.
620
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624
We examined the associations between habitual risky driving, prior convictions and motor vehicle driver injury using cross-sectional data from the New Zealand Blood Donors’ Health Study (Ameratunga et al., 2002). This ongoing, prospective cohort study aims to examine lifestyle and psychological determinants of serious injury. Baseline data from the study provided an opportunity to examine the cross-sectional associations between risky driving habits and motor vehicle driver injuries in an opportunistic sample of more than 22,000 individuals.
2. Methods The New Zealand Blood Donors’ Health Study is a prospective cohort study of blood donors in the top half of the north island of New Zealand. Recruitment took place between April 1998 and October 1999. The methodology of the study has previously been published elsewhere (Ameratunga et al., 2002). Briefly, individuals aged at least 16 years attending blood donation facilities (including schools, workplaces, shopping centres and mobile sites) in the region were invited to participate in the study. After providing informed consent, participants completed a self-administered baseline questionnaire, including a range of questions covering demographics, drug and alcohol use, smoking habits, risk taking behaviour, social situation, sleep habits, mental health, medical history and injuries sustained in the previous 12 months in a variety of settings including while driving. Participants gave consent for their baseline data collected at recruitment to be linked prospectively to outcome data. This electronic record linkage is expected to occur periodically over the next few decades and will involve confidential matching of baseline data collected at recruitment to the study to the New Zealand Health Information Service’s ‘National Minimum Dataset’. This is a national dataset covering information on all deaths and hospital discharges from public and most private hospitals. Injury information in this dataset, including diagnoses and external causes, is coded according to the International Classification of Diseases (Ministry of Health, 1999). This paper describes cross-sectional associations between self-reported risky driving and injury variables measured in the baseline survey. There were 22,389 participants in the study (81% response rate) and of these, 96% completed the baseline questionnaire. Using specified response categories, participants were asked to report the frequency over the past 12 months of the following risky driving behaviours: driving over the legal limit for alcohol; racing a motor vehicle for excitement; driving more than 20 km per hour (km/h) over the speed limit; not wearing a seatbelt when traveling as a driver or passenger; and driving on a public road without a current license. Different forms of unlicensed driving were not distinguished in this study and the ‘driving unlicensed’ variable included both those who had never held a license and those who held a disqualified or suspended license. Par-
ticipants were also asked to report the number of traffic convictions (excluding parking infringements) they had received over the past 12 months. The outcome variable was measured by asking if the respondent had sustained any injury during the past 12 months while driving a car, van, truck or similar vehicle, where treatment from a doctor was required. Injuries sustained by other occupants of the vehicle, and other road users such as pedestrians, were not measured. We examined the associations between risky driving habits, history of traffic offences and self-reported driver injury by calculating prevalence ratios (PR) and 95% confidence intervals (CI). Potential confounding variables were identified from the road safety literature. These were ethnicity, education level, New Zealand socioeconomic index score (Davis et al., 1999), the ‘CAGE’ score (a screening test for hazardous drinking; Ewing, 1984), drinking frequency in an average week, number of drinks on an average drinking occasion, likelihood of dozing while driving, Epworth sleepiness scale score (a measure of daytime sleepiness; Johns, 1991), frequency of marijuana and other illegal drug use and driving exposure (average number of hours spent driving per week). Of these, only those that were significantly associated with a history of driver injury after adjustment for age and sex were included in the logistic regression models. We tested for trend in association between number of convictions and driver injury using the unadjusted CochranArmitage trend test. Results were stratified by age group to examine the effects of risky driving in younger drivers (those aged less than 25 years) compared to drivers aged 25 years or more.
3. Results Of the 21,893 participants completing the baseline survey, 276 reported a history of motor vehicle driving injury. While a higher proportion of those reporting a history of motor vehicle driving injury, compared with those not reporting such injury, were aged 16–24 years (45.7% versus 36.6%, p = 0.002), no sex differences were observed (p = 0.34). The proportions of study participants who reported engaging in risky driving behaviours, by history of driver injury, are shown in Table 1. Those reporting a history of driver injury tended to report higher levels of driving above the legal limit for alcohol, driving unlicensed, racing for excitement, driving over the speed limit and not wearing a seatbelt, and had more prior convictions than those without a history of driver injury. Missing data for all exposure variables was less than 10%. Table 2 shows the unadjusted and multivariable adjusted prevalence ratios and 95% confidence intervals for associations between risky driving habits, history of convictions and driver injury. After adjustment for age, sex, CAGE score, number of drinks on average drinking occasion, frequency of marijuana use in the past 12 months and driving exposure (average hours per week), exposures significantly associated
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624 Table 1 Number and percentage of participants reporting history of driver injury in the past 12 months, by age, sex, risky driving behaviours and history of driving convictions, New Zealand Blood Donors Health Study, 1998–1999 History of driver injury (n = 276) Age (years) 16–24 25–39 40+ Sex Male Female
No history of driver injury (n = 21,617)
Frequency
(%)
126 73 77
45.7 26.4 27.9
Frequency 7903 5189 8525
(%) 36.6 24.0 39.4
120 156
43.5 56.5
10027 11590
46.4 53.6
Racing for excitement Never 212 Once or twice 18 Several times 34 Always 6 Not applicable 2 Missing 4
76.8 6.5 12.3 2.2 0.7 1.5
18358 1709 724 132 321 373
84.9 7.9 3.4 0.6 1.5 1.7
Driving ≥20 km/h over speed limit Never 70 Once or twice 80 Several times 104 Always 20 Not applicable 0 Missing 2
25.4 29.0 37.7 7.3 0.0 0.7
7424 7061 6032 477 284 339
34.3 32.7 27.9 2.2 1.3 1.6
Wearing a seatbelt Never Once or twice Several times Always Not applicable Missing
10 4 9 248 2 3
3.6 1.5 3.3 89.9 0.7 1.1
595 264 787 19556 93 322
2.8 1.2 3.6 90.5 0.4 1.5
Convictions past 12 months None 193 One 45 Two 14 Three or more 11 Missing 13
69.9 16.3 5.1 4.0 4.7
17868 1726 359 186 1671
81.8 8.0 1.7 0.8 7.7
with driver injury were racing a motor vehicle for excitement (PR 2.4, 95% CI 1.6–3.7), driving at 20 km/h or more over the speed limit (PR 2.5, 95% CI 1.4–4.3), and number of traffic convictions in the past 12 months (one conviction, PR 2.1, 95% CI 1.5–3.0; two convictions PR 2.7, 95% CI 1.5–4.9; three convictions PR 3.4, 95% CI 1.8–6.6). The number of convictions was the only variable for which all levels suggestive of risk were significantly associated with car crash injury, so we tested for trend in this variable; this test was significant (χ32 = 73.6, p < 0.0001). Of those reporting driver injury, 126 (45.7%) were aged less than 25 years and 150 (54.3%) were aged 25 years or older. Table 3 shows the multivariable adjusted prevalence ratios for exposures that were significantly associated with driver injury after stratification by age group. Driving unlicensed was associated with increased risk of injury for drivers aged 25 and over (PR 4.0, 95% CI 1.2–12.8) but not younger
621
Table 2 Prevalence ratios and 95% confidence intervals for associations between risky driving behaviours, history of convictions, and driver injury, New Zealand Blood Donors Health Study, 1998–1999 Unadjusted
Multivariable adjusteda
PR
95% CI
PR
95% CI
Driving while over alcohol limit Never 1.0 Once or twice 1.4 Several times/always 1.4
1.1–1.9 0.8–2.6
1.0 1.2 0.9
0.8–1.6 0.5–1.7
Driving unlicensed Never Once or twice Several times/always
1.0 1.7 2.0
1.0–2.8 1.2–3.4
1.0 1.6 1.3
0.9–2.8 0.7–2.4
Racing for excitement Never Once or twice Several times/always
1.0 1.0 3.9
0.8–1.2 2.8–5.5
1.0 0.8 2.4
0.6–1.0 1.6–3.7
Driving ≥20 km/h over speed limit Never 1.0 Once or twice 1.3 Several times 1.9 Always 4.5
0.9–1.7 1.4–2.6 2.8–7.5
1.0 1.0 1.2 2.5
0.7–1.4 0.8–1.7 1.4–4.3
Not wearing a seatbelt Never Once/twice/several times Always
0.4–1.8 0.5–1.4
1.0 0.8 1.0
0.3–2.0 0.5–2.0
Traffic convictions in past 12 months None 1.0 One 2.4 1.7–3.3 Two 3.6 2.1–6.1 Three or more 5.3 2.9–9.7
1.0 2.1 2.7 3.4
1.5–3.0 1.5–4.9 1.8–6.6
Frequency of risky driving in past 12 months
1.0 0.8 0.9
a Adjusted for age, sex, CAGE score, number of drinks on average drinking occasion, frequency of marijuana use in the past 12 months and driving exposure (average hours per week).
drivers. Racing for excitement was associated with a significant increase in risk for both drivers aged 25 and over (PR 3.8, 95% CI 1.8–8.0) and those under 25 (PR 2.3, 95% CI 1.4–3.7). Driving at 20 km/h or more over the speed limit was associated with an increased risk for younger drivers (PR 3.4, 95% CI 1.6–7.0); the increased risk was not significant in older drivers (PR 2.0, 95% CI 0.7–5.7). Higher numbers of convictions in the past 12 months was also significantly associated with increased risk of injury in both age groups, although the increased risk was not significant for drivers aged 25 and over with three or more convictions. This may reflect the smaller number of events in this age group and resulting imprecision of estimates.
4. Discussion We found that individuals who reported frequently engaging in risky driving behaviours over the past 12 months were between two and four times more likely to have been injured while driving during the same time period, compared to individuals who reported infrequently or never engaging in these
622
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624
Table 3 Multivariable-adjusted prevalence ratiosa (95% CI) for driver injury, for younger (<25 years) and older (≥25 years) drivers, New Zealand Blood Donors Health Study, 1998–1999b Driver <25 years
Driver ≥25 years
PR
95% CI
PR
95% CI
Driving unlicensed Never Once or twice Several times/always
1.0 1.6 1.1
0.9–2.9 0.6–2.1
1.0 1.9 4.0
0.3–14.0 1.2–12.8
Racing for excitement Never Once or twice Several times/always
1.0 0.9 2.3
0.7–1.2 1.4–3.7
1.0 0.4 3.8
0.2–1.1 1.8–8.0
Driving ≥20 km/h over speed limit Never 1.0 Once or twice 1.2 Several times 1.6 Always 3.4
0.6–2.2 0.9–2.8 1.6–7.0
1.0 1.0 1.2 2.0
0.7–1.6 0.7–1.8 0.7–5.7
Traffic convictions in past 12 months None 1.0 One 2.2 Two 3.3 Three or more 4.1
1.3–3.8 1.5–7.3 1.9–8.9
1.0 1.9 2.4 2.3
1.2–3.0 1.0–5.5 0.5–9.2
Frequency of risky driving in past 12 months
a
Adjusted for sex, CAGE score, number of drinks on average drinking occasion, frequency of marijuana use in the past 12 months and driving exposure (average hours per week). b Prevalence ratios are only reported for those exposures containing categories that were significantly associated with injury after stratification by age group.
behaviours. The specific behaviours that were associated with driver injury in these data were racing a motor vehicle for excitement, driving at 20 km/h or more over the speed limit, and a higher number of traffic convictions. We also found evidence that some of these associations were different for younger and older drivers. Specifically, driving unlicensed was a risk factor for older but not younger drivers and driving at 20 km/h or more above the speed limits was a stronger risk factor for younger than older drivers. These findings have some limitations. We asked only one question about each form of risky driving, and we cannot be sure that this was a reliable and valid measure. Many of these behaviours are illegal, and as with all self-reported data, we cannot exclude the risk of socially desirable responses. This may mean we have underestimated the magnitude of the association between risky driving and injury. However, we used standardised questions of risky driving based on previous studies and noted a wide distribution in responses across the levels of risk. The measure of driver injury did not distinguish between levels of injury severity, beyond the requirement for treatment from a doctor, and we did not measure or control for acute driving-related factors in the analyses. In this cross-sectional analysis, although the temporal relationships between risky driving, convictions and driver injury cannot be determined with certainty, it seems unlikely that having a driver injury would increase risky driving behaviour. On the other hand, the direction of the association
between having a driver injury and traffic convictions seems less certain. No doubt in some cases, the injury incident may have also resulted in a traffic conviction and/or loss of a driver license. This was a study of volunteer blood donors, who are unlikely to be representative of the general population, and for this reason we did not estimate population prevalence of exposures or outcomes. As published previously, however, the study population represented a broad cross-section of society and substantial heterogeneity in the distribution of potential risk factors for driver injury, making it possible to determine the associations of interest (Ameratunga et al., 2002). This comparatively large study was also able to assess a range of both risky driving behaviours and potential confounding variables and included a high proportion of young drivers, allowing useful comparisons to be made across age groups. The findings from this study support those from several previous studies that have also found that risky driving habits are associated with driver injury. The Christchurch Health and Development Study, a birth cohort of 1265 children born in 1977, found that the risk of a crash increased significantly with the number of risky driving behaviours reported, including speeding, drink driving, close following, not wearing a seatbelt, racing and running red lights (Fergusson et al., 2003). Risky driving behaviours as predictors of injury have also been investigated in another prospective cohort study of young New Zealanders, the Dunedin Multidisciplinary Health and Development Study (Begg and Langley, 2001; Begg et al., 2003). Like ours, this study did not find a significant association between habitual seatbelt use and car crash injury (Begg and Langley, 2000). A population based case–control study conducted in the United States also examined associations between habitual risky driving and car crash injury (Centres for Disease Control and Prevention, 1994). Driving more than 20 miles/h over the speed limit, passing a car in a no-passing zone, and taking risks for fun while driving were associated with an increased risk of injury. Other risky driving habits, including non-use of seatbelts, close following behind other drivers, passing in no-passing zones, taking risks in traffic for fun and running red lights, have also been found to be associated with increased risk of injury (Centres for Disease Control and Prevention, 1994; Bell et al., 2000; Norris et al., 2000; Fergusson et al., 2003). Previous studies have also found an association between convictions and crash risk. A population based case–control study in Finland examined fatal crashes and a random selection of license holders as controls (Rajalin, 1994). This study found that fatally crashed male drivers were significantly more likely to have traffic offences on their driving record after controlling for age and an estimate of miles driven by the driver annually. A literature review of variables associated with increased crash risk concluded that prior traffic offences were a good predictor of subsequent crash involvement (Peck, 1993). Our results, in combination with these studies, thus provide strong evidence for an association between risky driving habits, previous convictions and driver injury.
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624
We also found differences in the association between some risky driving habits and injury according to the driver’s age. Although one prior study has suggested that younger drivers killed in crashes are more likely than drivers of other ages to have been driving in a risky manner (Zhang et al., 1998), we did not find any previously published research suggesting that the effect of risky driving differed by age. This finding is therefore a potentially important contribution of the current study and may have significant implications. For example, lack of experience in young drivers may mean they are less likely to exercise safe judgment when driving at high speeds. However, it should be noted that only a small number of people reported ‘always’ speeding, and that the point estimate of the prevalence ratio for older drivers also suggest an effect although this was not significant. This may be a worthwhile area for future research. The finding that unlicensed driving is a risk factor for injury crashes in older but not young drivers is also unexpected. One possible explanation might be that more young people may drive unlicensed before they are old enough to gain a license, making this a more normative behaviour and less an indicator of risky driving in this age group, whereas older drivers who drive without a valid license represent a more risky driving group. However, we did not distinguish between drivers who had never held a license and those whose license was disqualified or suspended, and there may be age differences between these two groups. Although habitual risk takers may be difficult to identify, this is an important at-risk group amongst whom interventions to reduce car crash injury should be considered. Those whose risk taking had resulted in a conviction were a highrisk group in our study. These drivers may be specifically targeted with health promotion strategies such as education and community awareness campaigns in collaboration with organisations such as the police force. This strategy, if proven to be effective, may also be considered for other identifiable groups such as older unlicensed drivers. Our results also suggest that speed-related behaviours are of particular concern, especially in young drivers. Potentially useful strategies to counter this may include increased enforcement and higher penalties for risky driving, including loss of license. Unfortunately risky driving behaviours, especially speeding, are frequently reinforced by popular culture (Henley, 2004) and for this reason, behaviour change may not be achieved by health promotion strategies alone. Legislative approaches, such as modifications to Graduated Licensing Systems, may be effective in newly licensed drivers. Popular culture itself can also be a worthwhile target for interventions. For example, a recent Australian voluntary code of practice for motor vehicle advertising has been developed, stating that advertisements for motor vehicles should not portray any form of unsafe or reckless driving or speeding (Federal Chamber of Automotive Industries, 2004), although its success to date appears to be limited by the voluntary nature of the code. This study confirms previous evidence that individuals who are habitually risky drivers and have a history of convictions are more likely to experience driver injury. The
623
results suggest that speed-related behaviours are especially risky, particularly in young drivers, and that older unlicensed drivers are also a high-risk group. The cohort component of this study will prospectively examine the risk for serious injury and death in this population.
Acknowledgements The New Zealand Blood Donors’ Health Study was funded by the Health Research Council of New Zealand and its conduct supported by the New Zealand Blood Service. We acknowledge with gratitude the participants in this study and the contributions of Dr. Gary Whitlock (Research Fellow), Di Smith (Project Manager) and the collaborators involved with the study.
References Ameratunga, S., Norton, R.N., et al., 2002. The New Zealand Blood Donors’ Health Study: baseline findings of a large prospective cohort study of injury. Inj. Prev. 8, 66–69. Begg, D.J., Langley, J.D., 2000. Seat-belt use and related behaviors among young adults. J. Saf. Res. 31 (4), 211–220. Begg, D.J., Langley, J.D., 2001. Changes in risky driving behaviour from age 21 to 26 years. J. Saf. Res. 32, 491–499. Begg, D.J., Langley, J.D., 2004. Identifying predictors of persistent nonalcohol or drug-related risky driving behaviours among a cohort of young adults. Accid. Anal. Prev. 36, 1067–1071. Begg, D.J., Langley, J.D., et al., 2003. Identifying factors that predict persistent driving after drinking, and driving after using cannabis among young adults. Accid. Anal. Prev. 35, 669–675. Begg, D.J., Langley, J.D., et al., 1999. A longitudinal study of lifestyle factors as predictors of injuries and crashes among young adults. Accid. Anal. Prev. 31, 1–11. Bell, N.S., Amoroso, P.J., et al., 2000. Self-reported risk-taking behaviors and hospitalization for motor vehicle injury among active duty army personnel. Am. J. Prev. Med. 18 (Suppl. 3), 85–95. Centres for Disease Control and Prevention, 1993. From the Centers for Disease Control and Prevention. Risky driving behaviors among teenagers—Gwinnett County, Georgia. JAMA 272 (11), 844–845. Davis, P., McLeod, K., et al., 1999. The New Zealand Socioeconomic Index: develping and validating an occupationally-derived indicator of socio-economic status. Aust. N.Z. J. Public Health 23 (1), 27–33. Evans, L., Wasielewski, P., 1982. Do accident-involved drivers exhibit riskier everyday driving behaviour? Accid. Anal. Prev. 14 (1), 57–64. Evans, L., Wasielewski, P., 1983. Risky driving related to driver and vehicle characteristics. Accid. Anal. Prev. 15 (2), 121–136. Ewing, J.A., 1984. Detecting alcoholism: the CAGE questionnaire. JAMA 252, 1905–1907. Federal Chamber of Automotive Industries (2004). Advertising for motor vehicles: voluntary code of practice. Australia, Advertising Standards Bureau. Fergusson, D.M., Swain-Campbell, N., et al., 2003. Risky driving behaviour in young people: prevalence, personal characteristics and traffic accidents. Aust. N.Z. J. Public Health 27 (3), 337–342. Harrison, W.A., 1997. An exploratory investigation of the crash involvement of disqualified drivers and motorcyclists. J. Saf. Res. 28 (3), 213–219. Henley, N., 2004. Social marketing: ‘selling’ injury prevention. In: McClure, R., Stevenson, M., McEvoy, S. (Eds.), The Scientific Basis of Injury Prevention and Control. IP Communications, Melbourne.
624
S. Blows et al. / Accident Analysis and Prevention 37 (2005) 619–624
Johns, M., 1991. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 14, 54–55. Lam, L.T., 2003. Factors associated with fatal and injurious car crash among learner drivers in New South Wales, Australia. Accid. Anal. Prev. 35 (3), 333–340. Ministry of Health, 1999. Our Health, Our Future: The health of New Zealanders. New Zealand Ministry of Health, Wellington. Norris, F.H., Matthews, B.A., et al., 2000. Characterological, situational, and behavioral risk factors for motor vehicle accidents: a prospective examination. Accid. Anal. Prev. 32 (4), 505– 515. Peck, R.C., 1993. The identification of multiple accident correlates in high risk drivers with specific emphasis on the role of age, experience
and prior traffic violation frequency. Alcohol, Drugs, Driving 9, 145– 166. Preusser, D.F., Williams, A.F., et al., 1991. Characteristics of belted and unbelted drivers. Accid. Anal. Prev. 23 (6), 475–482. Rajalin, S., 1994. The connection between risky driving and involvement in fatal accidents. Accid. Anal. Prev. 26 (5), 555–562. Turner, C., McClure, R., et al., 2004. Injury and risk-taking behaviour—a systematic review. Accid. Anal. Prev. 36 (1), 93–101. World Health Organisation, 2003. Facts About Injuries: Road Traffic Injuries. World Health Organisation, Geneva. Zhang, J., Fraser, S., et al., 1998. Age-specific patterns of factors related to fatal motor vehicle crashes: focus on young and elderly drivers. Public Health 112, 289–295.