Research report No. 17

Chapter 5. Risky driving and substance use (PDF 327 KB)

Early adulthood can be a period of considerable risk-taking. The prevalence of substance use reaches a life-time high (Spooner, Hall & Lynskey, 2001), while other forms of risk-taking common at this age include antisocial behaviour, gambling and risky driving. The co-occurrence of risk-taking behaviours can be extremely harmful - even fatal. This chapter focuses particularly on the co-occurrence of risky driving and substance use.

When combined with driving, substance use may have lethal consequences. Substance use can affect the way a driver thinks and feels, the speed with which they react, and their hand-eye coordination. Consequently, drug-taking has the potential to impair a person's ability to drive safely (Drummer, 2008; Shinar, 2006).

The detrimental effects of alcohol on driving ability have been widely documented. Research suggests that alcohol increases a driver's risk of crashing six-fold (Task Force On Drug Abuse, 1996). Furthermore, alcohol is the most commonly detected substance among drivers killed or injured in motor vehicle crashes (Drummer, 2008; Longo, Hunter, Lokan, White, & White, 2000; Mercer & Jeffery, 1995), with alcohol implicated in 25-30% of road fatalities in Australia (Drummer, 2008).

Less is known about the impact of other drugs on driving performance. However, there is increasing evidence that drugs such as marijuana, amphetamines, ecstasy and benzodiazepines impair driving ability (Drummer, 2008; Mallick, Johnston, Goren, & Kennedy, 2007). Further, a combination of drugs, such as alcohol and marijuana or alcohol and benzodiazepines, can have a greater impact on driving ability than when used on their own (Mallick et al., 2007).

Rates of driving when under the influence of alcohol or illegal substances are particularly high among young drivers. A recent Australian study of drivers aged between 16 and 84 years found that 20-29 year olds were the most likely to have driven when under the influence of alcohol in the previous year, while this age group was the second most likely (after 16-19 year olds) to report having driven within three hours of using ecstasy or methamphetamines (Mallick et al., 2007). Additionally, Drummer and colleagues (2003) found that 22-30 year olds were over-represented among alcohol- and drug-affected drivers killed in three Australian states between 1990 and 1999.

There is also evidence that young people who drive when affected by drugs more often engage in unsafe driving practices such as speeding, failing to wear a seatbelt and taking risks in traffic (Donovan, 1993; Fergusson & Horwood, 2001; Steptoe et al., 2002, 2004). Similarly, in previous ATP research, alcohol use, binge drinking, cigarette use and marijuana use were found to be more common among high-level risky drivers aged 19-20 years than among low- or moderate-level risky drivers (Vassallo et al., 2008). Smart & Vassallo (2005) also showed that there is considerable overlap in the risk factors for risky driving, antisocial behaviour and substance use, suggesting that these differing forms of risk-taking shared common childhood and adolescent precursors. Precursors common to all three types of risk-taking included a less persistent temperament style, aggressive and hyperactive behaviour problems, school adjustment difficulties, friendships with antisocial peers and poorer parent-child relationships.

There are many issues that could be explored regarding the co-occurrence of risky driving and substance use. This chapter addresses two main questions. The first focuses particularly on drink-driving and whether young drink-drivers are also prone to engage in other types of risky driving. The second looks at whether young people who are high-level risky drivers also engage in higher levels of substance use, both singly and in combination. The two questions addressed are:

• Do young people who drink and drive engage in other unsafe driving behaviours more often than other young drivers who do not drink and drive?
• Do young people who frequently engage in risky driving use substances more often than other young drivers?

5.1 Co-occurrence of drink-driving and other types of risky driving

Young people who drink and drive were identified on the basis of their responses to two questions: (a) whether they had driven when near or over the alcohol limit during the past month, and (b) on how many of their past ten trips they had driven when affected by alcohol. This combination of questions was used to ensure that all instances of drink-driving were captured. For example, young people who drive very frequently might have undertaken their last 10 trips in a very short period of time (e.g., the past week), and for them it is important to include a question covering a wider time frame. On the other hand, those who drive infrequently might not be captured in the question about drink-driving in the past month, and the question about their last 10 trips might provide a broader sample of their driving behaviour.

Young people who responded affirmatively to either question were classified as drink-drivers (n = 277, 28% drivers), while the remainder were classified as other drivers (n = 703, 72% drivers).

The two groups were then compared on (a) whether or not they had engaged in other risky driving behaviours (speeding, seatbelt or helmet use, driving when fatigued, driving when affected by illegal drugs and mobile phone use when driving); and (b) how often they engaged in other types of risky driving.⁵⁰

Engagement of young drink-drivers in other types of risky driving

We look first at whether or not young drink-drivers engaged in other types of risky driving by comparison with other young drivers.

Drink-driving was strongly associated with speeding, as shown in Figure 22. Significantly more drink-drivers had exceeded the speed limit (by all margins) than those who did not drink-drive.⁵¹ These group differences were particularly evident for the highest speeding category: close to 30% of drink-drivers had exceeded the speed limit by over 25 km/h on a recent trip, compared with only 8% of other drivers.

Figure 22. Percentage who exceeded the speed limit at least once during their past 10 trips, drink-drivers and other drivers, at 23-24 years

Odds ratios can be used to estimate the likelihood of an outcome occurring (e.g., a crash) if a certain factor is present (e.g., speeding). Odds ratios can range from 0 to infinity. Values close to 1.0 indicate no relationship between the factor and an outcome. Values greater than 1.0 suggest that as levels of the factor increase (e.g., speeding), so does the likelihood of an outcome (e.g., a crash), while values smaller than 1.0 suggest that an outcome (e.g., a crash) is less likely as levels of the factor increase (e.g., completion of a defensive driving course).

Odds ratios comparing drink-drivers and other drivers indicated that drink-drivers were more than four times more likely to engage in low-level speeding (up to 10 km/h over limit), 3.4 times more likely to report moderate-level speeding (11-25 km/h over limit) and five times more likely to engage in high-level speeding (more than 25 km/h over limit).⁵²

Young drink-drivers were significantly more likely to have not worn a seatbelt or helmet for all or part of a recent driving trip (Figure 23).⁵³ More than a quarter had driven without a seatbelt/helmet for part of a recent trip, compared with 8% of other drivers (odds ratio of 4.2), while one in ten drink-drivers had driven without a seatbelt/helmet for all of a trip, compared with only 3% of other drivers (odds ratio of 3.0).

Figure 23. Percentage who had not worn a seatbelt for part or all of at least one recent trip, drink-drivers and other drivers, at 23-24 years

Driving when fatigued was significantly related to drink-driving, with a higher proportion of drink-drivers reporting that they had recently driven when very tired, and had nearly fallen asleep or fallen asleep when driving (Figure 24).⁵⁴ Three-quarters of drink-drivers reported that they had driven when very tired on at least one of their past ten trips, compared with 58% of other drivers (odds ratio of 2.1), and one in six had fallen asleep or come close to falling asleep on a recent trip, compared to 9% of other drivers (odds ratio of 2.1).

Figure 24. Percentage who had driven when fatigued at least once during their past 10 trips, drink-drivers and other drivers, at 23-24 years

Young drink-drivers were also significantly more likely to have driven when affected by illegal drugs. As Figure 25 shows, while rates of driving when under the influence of illegal drugs were generally very low, drink-drivers were considerably more likely than other drivers to have driven when affected by marijuana (odds ratio of 5.8), ecstasy (odds ratio of 9.7) and/or amphetamines (odds ratio of 4.5).⁵⁵

Figure 25. Percentage who had driven when affected by an illegal drug at least once during their past 10 trips, drink-drivers and other drivers, at 23-24 years

Finally, young drink-drivers were significantly more likely than other young drivers to have used a handheld mobile phone when driving (Figure 26). Three-quarters had talked on a handheld mobile on at least one of their ten most recent trips, compared with 46% of other drivers (odds ratio of 3.7), while over 80% had used a mobile phone function (e.g., received or sent an SMS) when driving, compared to 59% of other drivers (odds ratio of 3.0).⁵⁶ However, the two groups did not significantly differ in their use of hands-free mobile phones when driving.

Figure 26. Percentage who had used a mobile phone when driving at least once during their past 10 trips, drink-drivers and other drivers, at 23-24 years

Engagement of young drink-drivers in other types of risky driving

A second way of looking at the co-occurrence of drink-driving and other types of risky driving is to explore whether young drink-drivers engage in other risky driving practices more frequently. Table 18 shows that young drink-drivers had engaged in almost all other types of risky driving more often than other young drivers.

Table 18. Engagement in risky driving during their past 10 trips, drink-drivers and other drivers, at 23-24 years

	Average number of trips						t-test results
	Drink-drivers			Other drivers			t-value	p
	n	M	SD	n	M	SD
Speeding
Up to 10 km/h over limit	274	5.07	3.22	672	3.09	2.93	-8.82	< .001
11-25 km/h over limit	275	2.20	2.55	670	0.96	1.85	-7.31	< .001
> 25 km/h over limit	273	0.64	1.38	671	0.17	0.86	-5.21	< .001
Failure to wear seatbelt/helmet
For part of trip	275	0.81	1.91	672	0.28	1.27	-4.31	< .001
For all of trip	274	0.33	1.40	673	0.12	0.90	-2.28	[.023]
Fatigued driving
Drove when very tired	275	1.83	1.91	673	1.37	1.74	-3.60	< .001
Nearly fell asleep/fell asleep	275	0.27	0.71	671	0.11	0.40	-3.42	.001
Driving under the influence of illegal drugs
Affected by marijuana	275	0.36	1.36	673	0.10	0.83	-2.90	.004
Affected by ecstasy	275	0.08	0.36	673	0.01	0.10	-3.13	.002
Affected by amphetamines	275	0.11	0.54	673	0.03	0.45	-2.02	[.044]
Mobile phone use when driving
Talked on handheld mobile	275	2.11	2.19	671	1.03	1.59	-7.43	< .001
Used mobile function	274	2.27	2.16	673	1.46	1.87	-5.76	< .001

For example, drink-drivers had engaged in low-level speeding on about half of their ten most recent trips, while other drivers had done so on only about a third of trips. Similarly, while drink-drivers had exceeded the speed limit, used a handheld mobile when driving, or driven when very tired on about 20% of trips, other drivers had engaged in these unsafe driving practices on fewer occasions (between 10% and 15% recent trips). Rates of other forms of risky driving were less frequent (occurring on fewer than 10% of trips), but occurred more often among drink-drivers. There were also trends for differences on failure to wear a seatbelt or helmet for the duration of a trip and driving when affected by amphetamines.

Summary: Co-occurrence of drink-driving and other types of risky driving

Links between drink-driving and other types of risky driving were investigated in two ways. Firstly, drink-drivers and other young drivers were compared to determine whether or not young drink-drivers had engaged in other risky driving practices. A significantly higher percentage of drink-drivers had exceeded the speed limit, driven without a seatbelt or helmet, driven when fatigued, driven when affected by an illegal drug, and used a handheld mobile when driving than other young drivers. Secondly, we examined whether young drink-drivers engaged in risky driving on more occasions, and found that this was the case.

Differences were particularly marked on driving while under the influence of illegal drugs (although it should be noted that rates of driving when affected by an illegal drug were very low overall; see Chapter 3). Odds ratios indicated that young drink-drivers were 9.5 times more likely than their peers to have driven when affected by ecstasy, about 6 times more likely to have driven when affected by marijuana, and 4.5 times more likely to have driven when affected by amphetamines. Rates of speeding, and driving without a seatbelt or helmet for part of a trip were also much higher among young drink-drivers, with drink-drivers being 4 to 5 times more likely than their peers to have engaged in these behaviours on a recent driving trip.

5.2 Comparison of substance use among high-level risky drivers and other drivers

The second question investigated was whether young people who frequently drove in a risky manner more often used legal and illegal substances than other young drivers, both singly and in combination.

To investigate this question, the groups identified at 23-24 years who showed low (n = 630), moderate (n = 261) and high (n = 57) levels of risky driving (as described in Chapter 4, Table 15) were compared. These groups are hereafter referred to as low-, moderate- and high-level risky drivers.

For these analyses, responses to questions about the number of days on which various substances had been used in the past month were examined. The substance use types considered were: alcohol, binge drinking, marijuana, ecstasy and amphetamines. High-level binge drinking was defined as having 5 or more drinks in quick succession for males and 3 for females, while very high-level binge drinking was defined as having 7 or more drinks for males and 5 or more for females.

The use of other types of illegal substances such as heroin, cocaine and hallucinogens was not investigated, as the numbers of young people who reported using these substances was too small for reliable statistical analyses to be undertaken.

Alcohol use and binge drinking

The low-, moderate- and high-level risky driving groups did not differ on the number of days in which they had consumed alcohol during the past month.⁵⁷

However, while the majority of young people in each risky driving group had engaged in binge drinking at least once during the past month, the low-level risky driving group was least likely to have consumed alcohol at high and very high levels (Figure 27).⁵⁸

Odds ratios indicated that the moderate-level risky driving group was almost twice as likely as the low-level group to have engaged in high- and very high-level binge drinking.⁵⁹ While rates of high- and very high-level binge drinking were similar among the high- and moderate-level risky driving groups (Figure 27), the odds ratios comparing the high- and low-level risky driving groups were not significant. It is probable that the size of the high-level risky driving group (n = 57) was too small to detect statistically significant group differences.

Figure 27. Engagement in high- and very high-level binge drinking in the past month, low-, moderate- and high-level risky drivers, at 23-24 years

Marijuana, ecstasy and amphetamine use

Significant connections were found between risky driving and marijuana use.⁶⁰ Considerably more high-level risky drivers (30%) had used marijuana in the past month than moderate- (15%) and low-level risky drivers (10%). Odds ratios indicated that the high-level group was almost 4 times more likely than the low-level group to have used marijuana in the past month.⁶¹ Young people in the low- and moderate-level groups did not significantly differ in their likelihood of using marijuana.

The risky driving groups also significantly differed in their use of ecstasy. Fourteen per cent of the high-level group, and 12% of the moderate-level group had used ecstasy in the past month, compared with 7% of the low-level group. Odds ratios showed that moderate-level risky drivers were more than twice as likely as low-level risky drivers to have used ecstasy. As found earlier in relation to binge drinking, while ecstasy use was a little more common among the high-level group than the moderate-level group, the high-level group did not significantly differ from the low-level group in their likelihood of ecstasy use. Once again, the small size of the high group may have been responsible for this finding.

Finally, amphetamine use was significantly related to risky driving, with those in the low-level group the least likely to have used this type of drug in the past month.⁶² Fourteen per cent of the high-level group and 10% of the moderate-level group had recently used amphetamines, compared with 6% of the low-level group. As levels of risky driving increased so did amphetamine use, with moderate-level risky drivers almost twice as likely as low-level risky drivers to have recently used amphetamines, and high-level risky drivers 2.8 times more likely to have used amphetamines within the past month.

Multi-substance use

To investigate connections between risky driving and multi-substance use, the three risky driving groups were compared on (a) the number and frequency of use of differing substances in the past month, and (b) rates of combined marijuana use and binge drinking.

Number and frequency of differing substances used

As noted earlier, young people provided information about the number of days during the past month they had used various substances (alcohol, marijuana, ecstasy, amphetamines, methamphetamines, heroin, cocaine, inhalants, hallucinogens, or other illegal drugs). A total multi-substance use score was calculated for each individual by summing the number of different substances he or she reported using during the past month. The maximum possible score on this index was 10.

Figure 28 shows that two-thirds of ATP study members had used one type of substance in the past month, about 10% had used two different substances, and a similar proportion had used three or more substances. The highest number of different substances reported was 7, while 12% had not used alcohol or other drugs in the past month.

Figure 28. Number of substances used in the past month, ATP sample, at 23-24 years

Comparisons of the three risky driving groups revealed that individuals in the high- and moderate-level groups had used significantly more substances than those in the low-level group.⁶³ On average, the high-level group had used 1.6 different types of substances in the past month (SD = 1.3), whereas those in the moderate-level group had used 1.4 types (SD = 1.1) and those in the low-level group had used 1.2 different substances (SD = 0.9).

We next sought to determine whether high-level risky drivers used multiple substances more often than other drivers. Two steps were used to identify frequent multi-substance users as follows:

• those in the highest 20% of the ATP sample on binge drinking, marijuana, ecstasy or amphetamine use were classified as frequent users of the particular drug type, and the remainder of the sample was classified as not frequent users of the drug; and
• a total frequent multi-substance use score was calculated for each individual by summing the number of different substances on which he or she had been classified as a frequent user. The maximum possible score was 4.

Figure 29 shows the ATP sample's distribution of scores on the frequent multi-substance use index. Approximately two-thirds did not frequently use any substance. Just over one in five were frequent users of one substance, and about 7% were frequent users of two different substance types. Very few (about 3%) were frequent users of more than two substances.

Figure 29. Frequent multi-substance use in the past month, ATP sample, at 23-24 years

Comparison of the three risky driving groups revealed that the high-level group was significantly more likely than the low-level group to have frequently used several substances (Figure 30).⁶⁴ Odds ratios indicated that high-level risky drivers were almost five times more likely than low-level risky drivers to have frequently used two or more different substances within the past month, and were also 2.4 times more likely to have been a frequent user of one substance.⁶⁵ In addition, moderate-level risky drivers were more than twice as likely as low-level risky drivers to have frequently used two or more substances, and one-and-a-half times more likely to have frequently used one substance.⁶⁶

Figure 30. Frequent multi-substance use in the past month, high-, moderate- and low-level risky drivers, at 23-24 years

Risky driving and combined binge drinking and marijuana use

Four groups with differing patterns of frequent binge drinking and marijuana use were formed, using the criteria described in the previous sub-section to identify frequent multi-substance use. The four groups were: a neither group, a binge drinking only group, a marijuana use only group, and a both group. The proportions in each group are shown in Table 19.

Table 19. Definitions and sizes of multi-substance use groups, at 23-24 years

Groups	Description	N	%
Neither	Not frequent on both binge drinking and marijuana use	683	68.3
Binge drinking only	Frequent binge drinking, not frequent on marijuana use	180	18.0
Marijuana use only	Frequent marijuana use, not frequent binge drinking	73	7.3
Both	Frequent on both binge drinking and marijuana use	64	6.4

High- and moderate-level risky drivers significantly differed from low-level risky drivers in their combined use of alcohol and marijuana (Figure 31).⁶⁷ Only one-quarter of the low-level group were frequent binge drinkers, marijuana users or both, by comparison with half the high-level group and over a third of the moderate-level group.

Differences were particularly marked for combined alcohol and marijuana use, with 18% of the high-level group found to be frequent users of both substances, compared with only 5% of the low-level group and 6% of the moderate-level group. Odds ratios indicated that high-level risky drivers were almost 6 times more likely than low-level risky drivers to both binge drink and use marijuana frequently, and 3.3 times more likely to frequently use marijuana alone.⁶⁸ Moderate-level risky drivers were about twice as likely to frequently use marijuana, and about one and a half times more likely to frequently binge drink than low-level risky drivers.⁶⁹

Figure 31. Frequent binge drinking and marijuana use in the past month, high-, moderate- and low-level risky drivers, at 23-24 years

Summary: Comparison of substance use among high-level risky drivers and other drivers

Young people who engaged in high- or moderate-level risky driving tended to have higher rates of substance use at 23-24 years of age than young people of the same age who rarely engaged in unsafe driving.

Binge drinking, marijuana use, ecstasy use and amphetamine use were more common among the high- and moderate-level groups than the low-level group. Differences were particularly evident for extreme forms of binge drinking (7 drinks in quick succession for males, 5 for females) and marijuana use. Over 70% of high- and moderate-level risky drivers had engaged in binge drinking at this level during the past month, compared with 57% of low-level risky drivers. Close to a third of high-level risky drivers had used marijuana within the past 30 days, compared to 15% of moderate- and 10% of low-level risky drivers.

The high- and moderate-level groups did not differ from the low-level group in their general inclination to drink alcohol; rather, differences were evident in their tendency to consume alcohol at harmful levels.

Strong links were found between risky driving and multi-substance use. While almost all 23-24 year-olds had recently used at least one of the 10 different legal and illegal substances examined, young people who frequently drove in a risky manner had used a larger number of different substances during the past month than their less risky counterparts.

Additionally, high-level risky drivers (and to a lesser extent, moderate-level risky drivers) were more likely than low-level risky drivers to have engaged in frequent multi-substance use. The high group were about five times more likely than those in the latter group to be frequent users of two or more of the following: binge drinking, marijuana, ecstasy and amphetamines. The moderate-level group was twice as likely as the low-level group to have frequently used more than one of these substances.

An examination of rates of combined marijuana use and binge drinking among the three risky driving groups found that high-level risky drivers were significantly more likely than other drivers to be frequent users of both substances. Almost one in five high-level risky drivers had frequently used both drugs during the past month, compared with only 5-6% of moderate- and low-level risky drivers.

5.3 Discussion and implications

This chapter examined connections between risky driving and substance use in two ways. Firstly, it looked at whether young drink-drivers were more likely to engage in other forms of risky driving than other young drivers. Speeding, driving without a seatbelt, when fatigued, under the influence of an illegal drug or when using a mobile phone were all more common among young drink-drivers than other young drivers. Differences were particularly evident on speeding, driving while under the influence of illegal drugs, or driving without a seatbelt.

Secondly, low-, moderate- and high-level risky drivers were compared to determine whether high- and moderate-level risky drivers were more likely to use substances than low-level risky drivers, both singly and in combination. Binge drinking and marijuana, ecstasy and amphetamine use were significantly higher among the high- and moderate-level groups than the low-level group, with the strongest differences found on binge drinking and marijuana use. Further, high- and moderate-level risky drivers were more likely to use multiple substances, and did so more frequently.

Some implications that can be drawn from these findings and earlier findings from this report are discussed next.

Most young people planned ahead to avoid drink-driving

As noted earlier (Table 8), 23% of 23-24 year olds had driven when under the influence of alcohol on at least one occasion in their past ten trips. As alcohol has been shown to increase a driver's risk of crashing six-fold (Task Force on Drug Abuse, 1996) and is implicated in 25-30% of road fatalities in Australia (Drummer, 2008), this rate is concerning. Clearly, road safety efforts targeting drink-driving among young people continue to be necessary.

Current initiatives encourage young people to plan ahead to avoid drink-driving and Section 3.1 showed that young people used a variety of strategies to achieve this, such as forward planning, reducing consumption, finding another way home or staying overnight. These differed across males and females, those living in metropolitan and non-metropolitan areas, and those with differing levels of educational attainment. In all, two-thirds of the ATP cohort always planned ahead, while a further 17% made plans most of the time. However, almost one in ten (9%) rarely or never made plans to avoid drink-driving. Further efforts may be needed to reach this group. Of the 91% who made plans to avoid drink-driving, three-quarters subsequently never engaged in drink-driving. However, about one in five of this group went on to drink and drive on rare occasions, while 1% did so most of the time or always.

These findings suggest that current initiatives are hitting their mark, since the majority of young people successfully devised and used strategies to avoid drink-driving. Nevertheless, a small number did not make plans, and of those who made plans, almost a quarter ended up drink-driving at least occasionally. Further work is needed to understand why some young people do not plan ahead or are unable to carry out their plans. For example, is this a result of their personal style (such as impulsiveness, low self-control), peer group influences, lack of availability of alternative forms of transport, other circumstances, or a mix of these factors? The intervention implications arising from these alternative explanations would differ considerably. Identification of the types of drink-driving avoidance strategies that have the highest rates of success could also provide valuable information to inform road safety efforts and would enable better targeting of sub-groups who use differing strategies (e.g., metropolitan and non-metropolitan youth, males and females).

Drink-driving co-occurred with other forms of risky driving

The findings provide insight into the broader driving behaviour of young people who drink and drive. They suggest that drink-driving is not an isolated phenomenon and tends to occur in tandem with other forms of risky driving. Thus, odds ratios showed that young drink-drivers were 5 times more likely than other young drivers to have driven in excess of 25 km/h above the limit, not worn a seatbelt during a trip, or to have used a mobile phone when driving; between 4 and 9 times more likely to have driven when under the influence of an illegal drug (depending on the type of substance used); and twice as likely to have driven when very tired. Hence, there was considerable overlap between drink-driving and other types of risky driving.

These findings are in keeping with other research showing inter-connections between differing types of risky driving practices. Laapotti and Keskinen (2008), when comparing the fatal crashes of young and middle-aged Finnish male drivers, concluded that drink-driving is an aspect of a more pervasive risky driving style for young male drink-drivers. Donovan (1993) showed that drink-driving co-occurred with driving under the influence of illegal substances and other forms of risky driving (e.g., speeding, unsafe passing, lane changing), with inter-correlations between the three problem driving outcomes of between .24 and .46. Other research has shown that young people who drink and drive are more likely to drive above the speed limit and be apprehended for driving-related offences (Engström et al., 2003; Palamara & Stevenson, 2003; Steptoe et al., 2004).

Taken together, the findings of the present study and other research suggest that drink-driving is not an inadvertent or aberrant occurrence, but is often part of a more general risk-taking approach to driving. Hence, road safety efforts that emphasise conscious decision-making and pre-planning, while successful for less risky drivers, may not be as effective with individuals who engage in high-level risky driving practices.

The attractions of drink-driving need to be considered and tackled. Fry and Holden (2007), in a qualitative study of 53 Australian 17-25 year olds, found that the sub-group who regularly drove after binge drinking saw this as part of their personal "mobility" and "independence", found drink-driving exciting, and had fatalistic attitudes towards the possibility of a crash or apprehension by police. They also tended to believe that they remained skilled drivers when drunk, with heightened perceptions. In contrast, their more law-abiding counterparts focused on moral and societal obligations, feelings of responsibility towards others, and the potential negative impacts they might experience as a result of drink-driving (e.g., on their mobility, employment and career prospects, interpersonal relationships).

These findings reveal two quite different mindsets, suggesting that these groups may need to be targeted differently. For instance, these findings suggest that drink-drivers are more individualistic, and self-focused in their attitudes than other drivers; hence it is possible that road safety efforts that focus on the negative consequences of drink-driving on others will not resonate as strongly with this group as with other drivers. Rather, initiatives that focus on adverse consequences of drink-driving for the drivers themselves might have greater impact. Hence, interventions focusing on the negative personal impacts of drink-driving (e.g., loss of friends, reputation, etc.) might be one means of reaching this group. However, further research is needed to confirm the efficacy of such an approach. Additionally, this group's misconceptions of having enhanced driving skills and heightened perceptions when driving under the influence of alcohol are clear-cut targets for road safety efforts.

A second implication of these findings is that road safety campaigns aimed at decreasing risky driving practices in general are likely to have a flow-on effect on drink-driving. As Shope and colleagues have suggested (Shope, Waller, & Lang, 1996; Shope, Waller, Raghunathan, & Patil, 2001) interventions to prevent risky driving will also reduce drink-driving and driving under the influence of illegal drugs, since these risky driving practices are substantially inter-linked. Thus, specifically targeted strategies (e.g., Victoria's booze bus drink-driving program) and more general road safety initiatives (e.g., population-targeted publicity campaigns) can both play a role in decreasing drink-driving among young people.

High-level risky drivers used alcohol and other drugs more often

A further issue addressed in this chapter was whether high-level risky drivers engaged in alcohol and other drug use more often than low-level risky drivers. This was clearly the case; for example, high-level risky drivers were almost 4 times more likely to have used marijuana in the past month and were almost 3 times as likely to have used amphetamines. Additionally, they were more likely to be multi-substance users, and to use multiple substances more frequently.

These findings are consistent with other research (e.g., Caspi et al., 1997). For instance, Beirness and Simpson (1988) found that risky driving - as measured by behaviours such as deliberate risk-taking when driving, failure to wear seatbelts, and driving while under the influence of alcohol or drugs - was associated with alcohol and other drug use in a sample of Canadian high school students. Similarly, Shope and Bingham (2002) found that young US adults who drove when affected by alcohol were more likely than other young people to report problem drinking, drug use and delinquency.

These findings also extend earlier ATP research (Vassallo et al., 2008) showing that risky driving co-occurred with several externalising problems at 19-20 years (alcohol, cigarette and marijuana use; binge drinking; antisocial behaviour), with co-occurrence being most noticeable for antisocial behaviour and marijuana use at this age. Similar links between substance use and risky driving were evident here at 23-24 years of age. As at 19-20 years, marijuana use was particularly associated with risky driving at 23-24 years. Additionally, just over half of the high-level group had frequently used alcohol, marijuana or both, a rate that was almost double that of the low-level group.

One conclusion that could be drawn from these findings is that risky driving is one element of a risk-taking lifestyle for a number of young people. As such, young risky drivers are likely to benefit from interventions that not only target their behaviour on the road, but also other aspects of their lives, such as their use of legal and illegal substances. Thus, while specific programs targeting risky driving will remain essential, more broad-based "common solutions" approaches may also play a role in decreasing risky driving.

Interestingly, 23-24 year old moderate-level risky drivers also significantly more often engaged in binge drinking, ecstasy and amphetamine use than low-level risky drivers, and were twice as likely to be frequent multi-substance users. This finding demonstrates a co-occurrence between risky driving and substance use even at moderate levels of risky driving. It is possible that the moderate-level group was also attracted to risk-taking, albeit at lower levels than the high-level group. If so, similar broad-based interventions might have success in reducing this group's engagement in problem behaviours, including risky driving. As the numbers who engage in moderate-level risky driving tend to be much larger than the number who are high-level risky drivers (261 vs 54 in this particular study), the benefits to road safety could be quite large.

Risky driving may reflect an underlying propensity for problem behaviours

The co-occurrence of risky driving and substance use found here could be seen as supportive of Problem Behaviour Theory (Jessor, 1987; Jessor et al., 1997), which proposes that problem behaviours are interrelated and reflect a basic underlying trait or propensity. This theory implies that both targeted and more generalist intervention and prevention approaches will be effective in decreasing a range of problem behaviours, including risky driving. For example, programs addressing risky driving among young traffic offenders, broader initiatives that target multiple problems (e.g., drink-driving, driving under the influence of other drugs), and population-wide initiatives that aim to increase community awareness (e.g., media campaigns addressing speeding) could be expected to affect risky driving and other problems behaviours as well.

However, the overlap between risky driving and substance use was partial, not complete. For example, only approximately a third of 23-24 year old high-level risky drivers had used marijuana within the past 30 days and just over two-thirds had not. These findings support Willoughby, Chalmers, & Busseri's (2004) conclusion that while problem behaviours share common elements, they remain distinct phenomena. The findings also demonstrated that the high-level risky driving group was heterogeneous, containing young people showing differing profiles of problem outcomes that range from involvement in a single problem to involvement in several types of problem behaviours. Given this heterogeneity, no single approach is likely to be solely effective in reducing risky driving. Rather, a range of intervention and prevention efforts would appear to be needed to reduce risky driving among young people.

Conclusions

This chapter looked at connections between risky driving and substance use. Firstly, young people who drove when under the influence of alcohol more often engaged in other types of risky driving (e.g., speeding, not wearing a seatbelt) than their peers who did not drink and drive. Secondly, high- and moderate-level risky drivers more often used legal and illegal substances than low-level risky drivers. Thus, risky driving and substance use frequently co-occurred, and may be one element of a risk-taking lifestyle for a number of young people.

• ⁵⁰ Due to the number of statistical tests that were performed in relation to each topic, an adjusted significance level of p < .01 is used to reduce the risk of Type 1 error (a finding of significant differences when there is no such difference). For reader interest, results that are significant at the conventional p < .05 level are reported as trends, but are not interpreted.
• ⁵¹ Up to 10 km/h over limit: χ²(1) = 32.24, p < .001; 11-25 km/h over limit: χ²(1) = 68.10, p < .001; more than 25 km/h over limit: χ²(1) = 78.04, p < .001.
• ⁵² Up to 10 km/h over limit: odds ratio = 4.28, confidence intervals (2.50-7.33); 11-25 km/h over limit: odds ratio = 3.36, confidence intervals (2.50-4.51); > 25 km/h over limit: odds ratio = 5.15, confidence intervals (3.49-7.59).
• ⁵³ Drove without a seatbelt or helmet at all: χ²(1) = 14.63, p < .001; drove without a seatbelt for part of a trip: χ²(1) = 59.46, p < .001.
• ⁵⁴ Drove when very tired: χ²(1) = 22.75, p < .001; fell asleep or nearly fell asleep when driving: χ²(1) = 11.98, p = .001.
• ⁵⁵ Drove when affected by marijuana: χ²(1) = 38.97, p < .001; drove when affected by ecstasy: χ²(1) = 23.48, p < .001; drove when affected by amphetamines: χ²(1) = 13.11, p < .001.
• ⁵⁶ Talked on handheld mobile phone while driving: χ²(1) = 69.56, p < .001; used a mobile phone function while driving: χ²(1) = 41.25, p < .001.
• ⁵⁷ χ²(1) = 4.98, p = .083.
• ⁵⁸ High-level binge drinking: χ²(1) = 12.30, p = .002; very high-level binge drinking: χ²(1) = 16.84, p < .001.
• ⁵⁹ Comparison of low- and moderate-level risky driving groups—high-level binge drinking: odds ratio = 1.83, confidence intervals (1.27-2.64); very high-level binge drinking: odds ratio = 1.75, confidence intervals (1.29-2.38).
• ⁶⁰ χ²(1) = 20.28, p < .001.
• ⁶¹ Odds ratio = 3.76, confidence intervals (2.02-7.01).
• ⁶² χ²(1) = 10.11, p = .006.
• ⁶³ F(2, 945) = 9.47, p < .001.
• ⁶⁴ χ²(4) = 27.88, p < .001.
• ⁶⁵ Comparison of the high- and low-level risky driving groups on frequent use of—two or more different substances: odds ratio = 4.74, confidence intervals (2.25-10.03); one substance: odds ratio = 2.39, confidence intervals (1.28-4.48).
• ⁶⁶ Comparison of the moderate- and low-level risky driving groups on frequent use of—two or more different substances: odds ratio = 2.14, confidence intervals (1.31-3.49); one substance: odds ratio = 1.55, confidence intervals (1.10-2.18).
• ⁶⁷ χ²(6) = 31.81, p < .001.
• ⁶⁸ Comparison of high- and low-level risky driving groups—frequent binge drinking and marijuana use: odds ratio = 5.71, confidence intervals (2.53-12.87); frequent marijuana use only: odds ratio = 3.31, confidence intervals (1.35-8.11).
• ⁶⁹ Comparison of moderate- and low-level risky driving groups—frequent marijuana use only: odds ratio = 2.02, confidence intervals (1.17-3.47); frequent binge drinking only: odds ratio = 1.56, confidence intervals (1.08-2.26).