Alcohol and drug use among road users involved in fatal crashes in Norway

Abstract

Objective

To study the prevalence of alcohol and drugs in biological samples from drivers, motorcycle riders, bicyclists, and pedestrians involved in fatal road traffic crashes (RTCs) during 2016-2018 in Norway, both among fatally injured victims and those who survived fatal RTCs.

Methods

Anonymous information was extracted from police data. No personal data were recorded.

Results

There were 330 fatal RTCs with 349 killed road users and 384 survivors during the study period; this included 179 passengers who were excluded from the study. In total, 90% of the studied killed road users and 67% of the survivors were investigated for alcohol or drug use by analyzing biological samples. Alcohol or drugs in concentrations above the legal limits were detected in 21% of the analyzed samples. The proportion impaired by alcohol or drugs (blood alcohol concentrations equal to or greater than 0.05%, or drug concentrations above equivalent limits) was highest among killed bicyclists (43%), higher than among killed pedestrians (24%), car and van drivers (28%) and motorcyclists (20%), and significantly higher than among drivers who survived fatal crashes (4%).

Conclusions

Impairment due to use of alcohol or drugs was often a contributing factor among bicyclists, pedestrians and motor vehicle drivers who died in RTCs. Driving under the influence of alcohol or drugs was more often a contributing factor in cases where the motor vehicle driver was killed than in cases where the driver survived.

Keywords:

• Alcohol
• drugs
• fatal road traffic crashes
• drivers
• bicyclists
• pedestrians

Introduction

Impairment caused by use of alcohol or drugs is among the most common contributing factors for road traffic crashes (RTCs) (WHO 2018; Ringen 2019). Therefore, most countries have implemented legal blood alcohol concentration (BAC) limits (WHO 2018), and an increasing number of countries and states have legal limits for psychoactive drugs (Christophersen et al. 2020) (see also the additional bibliography B1 in the online supplemental material).

Norway implemented a BAC limit in 1936 of 0.05% (g/100 g blood), which was reduced to 0.02% (or 0.1 mg/L expired breath) in 2001. Legislative limits for 20 illicit and medicinal drugs were implemented in 2012 (Vindenes et al. 2012), and eight more drugs were added in 2016 (Ministry of Transport 2016). The legislative limits for drugs correspond to a BAC of 0.02% as well as two graded sanction levels corresponding to BAC of 0.05% and 0.12% for most substances (see online supplemental material, Table S1).

Statistics on alcohol and drug involvement in RTCs may be inaccurate depending on the proportion of involved road users that are investigated by analyzing alcohol in breath or blood samples and drugs in blood samples (Assum and Sørensen 2010). Both the proportion of crashes that are properly investigated and the incidence of driving under the influence (DUI) may change over time. It may therefore be difficult to study trends in the involvement in alcohol and drug use in RTCs, and also to compare official statistics from different countries. In Norway, the investigation of alcohol and drug impairment in RTCs has been performed by analyzing breath or blood samples only if the police suspected alcohol or drug impairment, or wanted to exclude that this had been a contribution cause of the crash. Therefore, only 63-68% of the fatally injured car and van drivers and motorcycle riders in the period 2001-2015 were investigated for alcohol and drug use by analyzing blood samples, in most cases samples taken by legal autopsy (Christophersen and Gjerde 2014, 2015; Valen et al. 2019). Also, about 30% of drivers who were involved in fatal crashes and survived were not subject to analysis of alcohol in blood or breath or drugs in blood samples (Skjønborg et al. 2015). During the last few years, the need for better data on fatal RTCs has been raised. As a result, an increasing proportion of the fatal RTC has been investigated for alcohol and drug involvement after 2015. The Autopsy Act was eventually revised in 2020 requiring legal or medical autopsy of all road users killed in RTCs (MHCS 2020).

The aim of this article is to present detailed statistics on alcohol and drug findings among active road users (drivers, motorcycle riders, bicyclists, and pedestrians) involved in fatal RTCs during 2016-2018, both among fatally injured road users and those who survived fatal RTCs.

Methods

Anonymous statistics for fatal RTCs during 2016-2018 were extracted from police data. Involved road users were divided into the following groups: 1) drivers of cars and vans; 2) drivers of busses, minibuses, trams, trains or trucks; 3) motorcycle and moped riders; 4) drivers of other motor vehicles (agriculture tractors, all-terrain vehicles, snowmobiles, and other tracked vehicles); 5) bicyclists; 6) pedestrians; and 7) passengers. Data for passengers were not recorded, firstly because they are seldom directly responsible for the crash, and secondly because less than 25% were investigated for alcohol or drug use by analysis of blood samples. Information about type of road user, crash outcome (killed or survivor), and results of alcohol and drug testing were transferred to a research database. To protect privacy, the involved road user’s age or sex, the time or place of the RTC, and other parameters that might contribute to identifying the involved persons were not recorded. Only aggregated data for groups of five persons or more with the same characteristics (type of road user and crash outcome) are presented in compliance with the definition of anonymity (Langtvedt 2015). The Data Protection Official of the Norwegian National Road Policing Service approved the collection and publication of the anonymous data. Projects handling data anonymously do not need approval from the Regional Committee for Medical Research Ethics, according to the Norwegian Research Ethics Act and the Act on Medical and Health Research.

For survivors, breath samples were analyzed for alcohol by the police using approved breathalyzers in cases where the police did not suspect drug impairment. If this breath test indicated alcohol concentration above the legal limit, drivers were either brought to the police station for evidential breath testing or to a medical doctor for taking a blood sample for quantitative analysis of alcohol. If the police suspected impairment by drugs instead of, or in addition to alcohol, a blood sample was taken.

For fatally injured road users, blood samples were either taken at arrival to hospital (if still alive) or during medical or legal autopsy. Blood samples were analyzed at Oslo University Hospital (Oslo, Norway) or St. Olav University Hospital (Trondheim, Norway). Alcohol was analyzed by head-space gas chromatography and drugs by liquid chromatography – mass spectrometry (see bibliography B2 in the online supplement for references). Both laboratories participate in international inter-laboratory comparisons and proficiency testing programs and are accredited by the Norwegian body for accreditation of laboratories (Norwegian Accreditation, Lillestrøm, Norway; http://www.akkreditert.no/en). All samples were analyzed for alcohol; drugs assigned with legislative limits (see online supplemental material, Table S1) were analyzed if requested by the police. Results for GHB were disregarded as this substance may be formed postmortem. Samples were also analyzed for some psychoactive drugs that had not been assigned with legislative limits: codeine, tramadol, pregabalin, and some New Psychoactive Substances (“designer drugs”). Results for medicines given as medical emergency treatment were disregarded.

Road users who died within 30 days after the crash were defined as fatally injured. However, if alcohol and drug testing was performed in autopsy samples for road users who died in hospital a day or more after the crash, the test results were not used in this study.

Test result for alcohol or drugs that had been assigned with legal limits were regarded as positive if the concentration was equal to or larger than the legislative limits for motor vehicle drivers listed in the Norwegian road traffic act corresponding to a blood alcohol concentration (BAC) of 0.02%, and regarded as likely impaired if equal to or larger than the concentrations corresponding to the graded sanction limits corresponding to a BAC of 0.05%; see online supplemental material, Table S1.

Results

There were 330 fatal RTCs in Norway during 2016-2018 with 349 killed road users and 384 survivors. The crash investigation teams of the Norwegian Public Roads Administration concluded that 25% of those crashes were related to alcohol or drug use (Ringen 2019). When excluding passengers, a total of 554 active road users were involved in those crashes. Of those ones, 56% (n = 308) were fatally injured.

In total, 90% of the killed road users were investigated for alcohol or drug use (Table 1), a much higher proportion than in previous studies of killed drivers and riders during 2001-2015, where only 63-68% had been investigated by analyzing blood samples (Christophersen and Gjerde 2014, 2015; Valen et al. 2019). Both alcohol and drugs were investigated for 89% of the killed road users; and one percent was investigated only for alcohol.

Table 1. Proportions of active road users involved in fatal road traffic crashes during 2016-2018 who were investigated for alcohol or drug use by analyzing samples of blood or breath.

	Survivors	Killed	Total
Drivers of cars and vans	n = 139 T: 65.5%; A: 19.4%	n = 158 T: 92.4%; A: 1.3%	n = 297 T: 79.8%; A: 9.8%
Drivers of trucks, buses, minibuses, trams, trains	n = 95 T: 68.4%; A: 35.8%	n = 12 T: 91.7%; A: 0%	n = 107 T: 71.0%; A: 30.8%
Riders of motorcycles and mopeds	n = 2 T: 50.0%; A: 0 %	n = 57 T: 87.7%; A: 0%	n = 59 T: 86.4%; A: 0 %
Drivers of other vehicles	n = 8 T: 100%; A: 62.5%	n = 13 T: 92.3%; A: 7.7%	n = 21 T: 95.2%; A: 28.6%
Bicyclists	n = 1 T: 0%; A: 0%	n = 27 T: 77.8%; A: 3.7%	n = 28 T: 75.0%; A: 3.6%
Pedestrians	n = 1 T: 0%; A:0%	n = 41 T: 92.7%; A: 0%	n = 42 T: 90.5%; A: 0%
Total	n = 246 T: 67.1%; A: 26.8%	n = 308 T: 90.3%; A: 1.3%	n = 554 T: 80.0%; A: 12.6%

n = total number of persons, T = total proportion investigated for alcohol or drug use, A = proportion investigated for alcohol use only.

Among the surviving active road users (n = 246), 67% were investigated for alcohol or drug use (Table 1); 40% were investigated for both alcohol and drugs, whereas 27% were tested for alcohol only, mostly by breath testing (19%).

Alcohol or drugs were detected in concentrations above the legal limits for motor vehicle drivers in 20% of the analyzed samples (see details in online supplemental material, Table S1). Alcohol was the most commonly found substance, followed by sedatives and hypnotics, THC (tetrahydrocannabinol; cannabis), stimulants, and opioids.

Among the fatally injured road users who were investigated for alcohol and drug use, 19% tested positive for one substance and 10% tested positive for two or more, whereas among survivors, 4% tested positive for one substance and only one person tested positive for two substances. The most common combinations were alcohol + THC (n = 10), amphetamines + benzodiazepines (n = 7) and THC + benzodiazepines (n = 7). About 90% of the killed road users who tested positive had concentration in blood corresponding to the graded sanction limit corresponding to a BAC of 0.05% or higher, which in most cases is associated with clinical impairment and increased crash risk (Borkenstein et al. 1974; Blomberg et al. 2009).

In 85% of the cases where alcohol was detected, the BAC was above the highest graded sanction limit (0.12%), indicating significant impairment. For sedatives, hypnotics and opioids, also the majority of the cases had concentrations above the highest graded sanction limits; whereas for cannabis, about half of the cases had THC concentrations above the highest graded sanction limits. For stimulants, legislative graded sanction limits have not been established; however, most findings indicated problem use of very high doses which is not compatible with safe behavior in road traffic.

Alcohol or drugs with legislative limits were detected in blood samples from about 30% of the killed car and van drivers (Table 2), whereas among killed motorcyclists, the proportion that tested positive for alcohol or drugs was somewhat lower (26%). Among killed pedestrians and bicyclists, 26% and 48%, respectively, tested positive for alcohol or drugs.

Table 2. Alcohol and drug findings among fatally injured road users. Results are presented as percent of the investigated persons.

	At or above legal limits^a	Alcohol or drug concentrations equivalent to BAC of 0.05% or higher
	% (n)	Total % (n)	Alcohol % (n)	Sedatives and hypnotics % (n)	THC % (n)	Stimulants^b % (n)	Opioids % (n)
Drivers of cars and vans (n = 146)	29.5 (43)	28.1 (41)	15.1 (22)	11.6 (17)	7.5 (11)	3.4 (5)	2.1 (3)
Drivers of trucks... trains (n = 11)	0	0	0	0	0	0	0
Riders of motorcycles and mopeds (n = 50)	26.0 (13)	20.0 (10)	8.0 (4)	10.0 (5)	8.0 (4)	6.0 (3)	2.0 (1)
Drivers of other vehicles (n = 12)	41.7 (5)	41.7 (5)	41.7 (5)	0	0	0	0
Bicyclists (n = 21)	47.6 (10)	42.9 (9)	38.1 (8)	4.8 (1)	0	4.8 (1)	0
Pedestrians (n = 38)	26.3 (10)	23.7 (9)	15.8 (6)	10.5 (4)	10.5 (4)	2.6 (1)	5.3 (2)
Total (n = 278)	29.1 (81)	26.6 (74)	16.2 (45)	9.7 (27)	6.8 (19)	3.6	2.2 (6)

^a Concentrations of alcohol or drugs equivalent to BAC of 0.02% or higher. ^bFive times the legislative limit for amphetamines, cocaine, MDMA, and methylphenidate.

The proportion that tested positive for alcohol or drugs was much lower among those who survived fatal crashes (Table 3); only about eight percent of the involved car and van drivers had alcohol or drugs above the legal limits, and only one of the 65 drivers of heavier vehicles, including buses, minibuses, trams, trains, or trucks tested positive; that driver had a low concentration of a medicinal drug indicating low-dose therapeutic use not expected to affect the ability to driver safely. In total, about 4% of drivers of all categories who survived fatal RTCs were impaired by alcohol or drugs.

Table 3. Alcohol and drug findings among road users surviving fatal crashes. Results are presented as percent of the investigated persons.

	At or above legal limits^a	Alcohol or drug concentrations equivalent to BAC of 0.05% or higher
	% (n)	Total % (n)	Alcohol % (n)	Sedatives and hypnotics % (n)	THC % (n)	Stimulants^b % (n)	Opioids % (n))
Drivers of cars and vans (n = 91)	7.7 (7)	6.6 (6)	3.3 (3)	2.2 (2)	0	2.2 (2)	0
Drivers of trucks... trains (n = 65)	1.5 (1)	0	0	0	0	0	0
Riders of motorcycles and mopeds (n = 1)^c	–	–	–	–	–	–	–
Drivers of other vehicles (n = 8)	0	0	0	0	0	0	0
Total (n = 165)	4.8 (8)	3.6 (6)	1.8 (3)	1.2 (2)	0	1.2 (2)	0

^a Concentrations of alcohol or drugs equivalent to BAC of 0.02% or higher. ^bFive times the legislative limit for amphetamines, cocaine, MDMA, and methylphenidate.

^c Data was not regarded as anonymous as the number was too low.

In addition to the results presented above, one percent of the samples tested positive for psychoactive drugs that had not been assigned with legislative limits. This included the illegal sedative drug diclazepam in one blood sample, tramadol in nine, codeine in two, and pregabaline, which is used for the treatment of anxiety, epilepsy, and nerve pain, in three samples. Those findings are not included in Tables 2, 3, and supplemental Table S1.

Discussion

The key finding in this study was that impairment by alcohol or drugs was more common among killed road users than among survivors involved in fatal crashes. In total, about one quarter of the killed road users were impaired by either alcohol or drugs. Impairment was more common among certain road users groups, such as killed bicyclists and killed drivers of agriculture tractors, all-terrain vehicles, and tracked vehicles, where about 40% were impaired. None of the killed drivers of trucks, buses, minibuses, trams, and trains were impaired, while 20-30% of killed motorcycle riders and car drivers were impaired.

Previous studies of alcohol and drug use among fatally injured road users in Norway included only car and van drivers and motorcycle riders. In those studies, only 63% of the victims were investigated for alcohol and drug use by analyzing blood samples (Christophersen and Gjerde 2014, 2015; Valen et al. 2019), in contrast to 88-92% in our study. The selection bias in those studies might have been higher than in our study, as a larger proportion of drivers who were not suspected for alcohol or drug use might not have been subjected to collection of blood samples for alcohol or drug testing. This may have caused an inflated prevalence of alcohol or drugs as compared with the actual proportion among all killed drivers.

A previous study found that 40% of the tested fatally injured car and van drivers during 2001-2010 had concentrations of alcohol or drugs above the legal limits corresponding to a BAC of 0.02% (Christophersen and Gjerde 2014). Another study found that 34% of those killed during 2005-2010, and 26% of those killed during 2011-2015, had alcohol or drug concentrations above the graded sanction limits corresponding to BAC of 0.05% (Valen et al. 2019), the same limits for impairment as we used in our study. Our findings for the period 2016-2018 shows that 28% had concentrations above the limits corresponding to BAC of 0.05%, and are thus similar to the findings for 2011-2015. However, our results are based on the analysis of blood samples from 92% of the killed drivers, whereas the former studies were based on the analysis of blood samples from only 63%.

A previous study of motorcycle and moped riders killed in road traffic crashes during 2001-2010 found that 27% of those tested had alcohol or drugs in concentrations above the legal limits corresponding to BAC of 0.02% (Christophersen and Gjerde 2015), whereas Valen et al. (2019) found that 23% of those killed during 2005-10 and 15% of those killed during 2011-2015 had alcohol or drug concentrations above the graded sanction limits corresponding to BAC of 0.05%, which is similar to our finding for 2016-2018 (20%). Also for the study of motorcycle riders, we included a larger proportion of the victims (88%) than the former studies (63%). However, the numbers of riders included in our study and in previous studies were fairly low; therefore the findings had low statistical power, making comparisons difficult.

The highest prevalence of alcohol or drug impairment was found among killed bicyclists (43%). Also other studies have found that many injured or killed cyclists were impaired by alcohol or drugs, although with lower prevalences than in our study. A previous Norwegian study of injured road users admitted to hospital found that 14% of bicyclists tested positive for alcohol or drugs (Bogstrand et al. 2011), whereas a later study found that 19% of adult injured bicyclists admitted to hospital were impaired by alcohol or drugs (Mjåland et al. 2019). A Finnish study found that 31% of injured bicyclists tested positive for alcohol (Airaksinen et al. 2018). A British study found that 22% of cyclists who were killed in RTCs during 2000-2006 and subjected to autopsy tested positive for alcohol, in most cases with higher concentrations than the legal limit for motor vehicle drivers in the UK (Elliott et al. 2009). A study based on data from the American Fatality Analysis Reporting System (FARS) found that the percentage of fatally injured cyclists with high BACs (≥0.08%) declined from 28% to 21% from 1982 to 2014 (Eichelberger et al. 2018). The high prevalence of alcohol and drugs in our study may be caused by a selection bias as only 78% of the killed bicyclists were analyzed for alcohol and drug use. Also, the low number of cases might have given poor accuracy. Differences between studies may partly be related to cultural differences that affect the incidence of riding a bicycle under the influence of alcohol, and there may also be differences between killed and injured bicyclists regarding alcohol and drug impairment.

We found a fairly high prevalence of alcohol or drug impairment among killed pedestrians (24% had BAC ≥ 0.05% or equivalent drug concentrations). Bogstrand et al. (2011) found that 41% of pedestrian injured in RTC and admitted to a Norwegian hospital tested positive for alcohol or drugs, although using lower concentration limits than in our study. In the UK, 47% of autopsied pedestrians killed in RTCs during 2000-2006 tested positive for alcohol (Elliott et al. 2009). Data from FARS indicated that the percentage of fatally injured pedestrians with high BACs (≥0.08%) declined from 45% to 35% from 1982 to 2014 in the USA (Eichelberger et al. 2018). A Polish study of autopsied pedestrians killed in RTCs found that about 52% tested positive for alcohol, with a mean BAC of 0.2% (Pawłowski et al. 2019). A study of autopsied pedestrian fatalities (286 victims) in northern Sweden from 1977 to 1995 found that 19% had been drinking alcohol with a median BAC of 0.16% (Öström and Eriksson 2001). Thus, walking under the influence of alcohol or drugs also increases the risk for RTC involvement.

To conclude, we were able to include a larger proportion of fatally injured drivers than previous Norwegian studies; our findings may therefore represent more accurately the actual involvement of alcohol and drugs, although a fairly low number of cases gave low statistical power for some road user groups. We were also able to include surviving drivers involved in fatal RTC, showing that the prevalence of impairment by alcohol or drugs was lower than among killed drivers, in spite of the fact that a lower proportion was tested for alcohol or drug use. Both the results of our study and previous studies show that impairment by alcohol or drugs is an important contributing factor for road traffic injury and death also among bicyclists and pedestrians, not only for motor vehicle drivers and riders. That problem has previously often been overlooked when discussing the effect of alcohol and drug impairment on road traffic safety, as the focus has been on the involved motor vehicle drivers.

The local police departments decide whether road users involved in fatal crashes are investigated for alcohol or drug impairment. According to Norwegian legislation, such investigation could until 2020 be done if a violation of the road traffic act was suspected, or if there was a need for excluding this possibility. The police has to cover the costs for legal autopsies, and for alcohol and drug testing. Budget restraints may therefore have affected their decisions. In an evaluation performed by the Norwegian National Road Policing Service, it was concluded that a large proportion of serious RTCs had not been investigated properly, including lack of alcohol and drug testing (Skjønborg et al. 2015). It was recommended that the routines for RTC investigations should be improved.

Future studies in Norway may obtain more accurate data on alcohol and drug use as the Norwegian Prosecuting Authority ordered the police in 2016 to perform alcohol and drug testing of all drivers involved in serious RTC. In addition, legal or medical autopsy of all killed RTC victims became compulsory after May 2020 by revising the Norwegian Autopsy Act (MHCS 2020).

Limitations

We assessed drug concentrations only for those substances that have been assigned with legislative limits. It is possible that some of the involved road users had been impaired by other psychoactive substances. It is also possible that detection of opioids or diazepam in a few cases might have been a result of acute medical treatment, but not recorded in police data, and should therefore have been disregarded. Postmortem changes in drug concentrations might in some cases have caused misclassification of likely impairment.

Alcohol testing had not been performed for 10% of the killed road users involved in fatal RTCs, and drug testing had not been performed for 11%. Also, alcohol testing had not been performed for 33% of the survivors, and drug testing had not been performed for 60%. It is likely that the prevalence of alcohol or drug use among those who were not subject to alcohol and drug testing was lower than among those who were tested. It is therefore likely that the alcohol and drug findings may not accurately reflect alcohol and drug use among road users. We also expect that the collected blood samples did not always reflect the alcohol and drug concentrations at the time of the crash accurately due to metabolism between crash and collection of blood samples, and postmortem re-distribution when the data were based on autopsy samples.

Data availability statement

We do not have permission to share research data.

Additional information

Funding

This study was sponsored by the Norwegian Public Roads Administration.