1 Addiction Centre Stockholm, Box 125 60, SE-102 29 Stockholm,
2 Karolinska Institutet, Department of Public Health Sciences, Division of Social Medicine, Norrbacka, SE-171 76 Stockholm,
3 Karolinska Institutet, Department of Clinical Neuroscience, Section of Clinical Alcohol and Drug Addiction Research, Karolinska Sjukhuset, SE-171 76 Stockholm and
4 Centre for Social Research on Alcohol and Other Drugs, Stockholm University, SE-106 91 Stockholm, Sweden
Received 5 December 2001; in revised form 30 December 2002; accepted 12 January 2003
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Jonah (1997) has reviewed research on DD, RD and sensation seeking. The vast majority of the 40 studies reviewed showed positive relationships between sensation seeking and RD, particularly for men, with correlation coefficients in the 0.300.40 range. DD and sensation-seeking has been investigated in 18 studies. In all but five, a positive such relationship has been reported.
A substantial number of studies indicate that high alcohol consumption and alcohol misuse increase the risk of mortality and morbidity for several diseases and also of various negative social consequences (Shaper, 1990; Edwards et al., 1994
). The question of whether a fatal car accident might be a disguised suicide has also been addressed and gained some support (Jenkins and Sainsbury, 1980
; Öström and Eriksson, 1993
). Brewer and Morris (1994)
found that arrests for DD substantially increased the risk of eventual death in an alcohol-related accident. In an autopsy study of alcohol-related mortality of all deceased in 1987, aged 1554 years and from Stockholm, Romelsjö et al. (1993)
found a much higher rate of excessive alcohol consumption or alcohol misuse in the medical records and police reports in the male DD group (51.3%), as compared with male non-DD offenders (9.5%). However, according to our knowledge, the association between DD and/or RD and morbidity and mortality has not been previously studied. In the present study, a unique material comprising 8122 male enrolees with a follow-up period of 2123 years was available for investigation. Several longitudinal studies on Swedish military conscripts have been published. One is a 7-year follow-up of the conscripts, with DD and public drunkenness as outcomes (Karlsson and Romelsjö, 1997
). This study showed that risky alcohol use, smoking, sniffing of solvents and certain psychosocial and social variables at conscription were associated with an increased risk for these outcomes.
The aims of the present study were to: (1) investigate social and psychological background characteristics, problem behaviour and alcohol and drug habits of drunk drivers and risky drivers; (2) study the association between DD and/or RD and subsequent hospitalization and mortality prospectively; (3) investigate whether there is a difference in risk between DD and RD in these respects.
![]() |
SUBJECTS AND METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Factors assessed at enrolment
Previous studies of this subject group demonstrated an association on the one hand between various social, behavioural and psychological factors at enrolment, and high alcohol consumption and mortality and hospitalization at follow-up on the other (Andréasson et al., 1988, 1990
; Karlsson and Romelsjö, 1997
). The distribution of such factors among DD and RD (Table 1
) was studied. These factors can be regarded as predictors for the outcomes at follow-up, and are included as potential confounders, partly based on results from several other studies of the conscripts, as the focus is on the relative risk (RR) for drunk drivers and risky drivers. Risky use of alcohol (high alcohol consumption, binge drinking, intoxication, apprehension for public drunkenness, hangovers, relief drinking), drug misuse (sniffing of solvents, illicit drug use) and cigarette smoking were also included as confounders. All these variables were dichotomized except for social class, which had three categories in Sweden since the beginning of the 20th century.
|
Fathers social class. Social class I represents the highest level, and is the reference category for the other two social classes, the middle class (II) and the working class (III). The categorization is based on a method published by Statistiska centralbyrån (SCB, Statistics Sweden) (1956) and is related to the occupational status of the enrolees father. This classification was used in Sweden up to 1980, and used in almost all Swedish studies including the social class concept. The social categorization is comparable to that of other countries (Halldin, 1984
). Social class was one of the strongest predictors of DD in a 7-year follow-up of the same subject group (Karlsson and Romelsjö, 1997
).
Emotional control. A psychologist assessed emotional control at the time of enrolment. This variable is a summary assessment of mental stability, emotional maturity, and tolerance to stress and frustration (Stenbacka, 1992). The psychologist rated the subjects on a nine-step scale, which was then collapsed to a five-step scale with a normal distribution. Average control, rather or very good control (categories 35) was used as reference category, compared with poor or rather poor control (categories 12). The inter-rater reliability of the assessments has been reported to be satisfactory (Stenbacka, 1992
). Poor and rather poor emotional control has been shown to be a risk factor for drug use (Stenbacka, 1992
), mortality and other adverse outcomes among these military conscripts.
Psychiatric diagnosis. Enrolees with poor emotional control or who scored poorly on other measures of psychological functioning were evaluated by a psychiatrist with regard to psychosis, neurosis, personality disorder, alcoholism or drug addiction. The primary aim of the evaluation was to decide whether these enrolees were fit to fulfil military service or not, and not primarily whether they were mentally ill at enrolment.
High alcohol consumption. As in other studies of the enrolees, an amount corresponding to consumption of more than 250 g of pure alcohol per week (e.g. one 75 cl bottle of strong liquor) was defined as high alcohol consumption.
Binge drinking. The intake of at least five drinks (approximately 1.5 l of strong beer or 75 cl of wine or 15 cl of spirits, or approximately 60 g of 100% ethanol) on one or more occasions weekly was considered as binge drinking (Karlsson and Romelsjö, 1997).
Risky use of alcohol. This is a composite variable, which includes six items: high alcohol consumption, binge drinking, intoxication drinking, relief drinking, having had hangovers, and having been apprehended for public drunkenness. At least one of the above items had to be present for a risky use categorization.
Factors assessed at follow-up
From Statistics Sweden, with an almost complete coverage of all crimes, we received data on sentences for DD and RD, relating to the Criminal Act of Traffic Violations, up to 1993 which were linked to the enrolment information on an individual basis. Until 1990, the lower limit in Sweden for DD was a blood-alcohol concentration (BAC) of 0.5 pro mille (50 mg/dl); since 1990, it has become 0.2 pro mille (20 mg/dl). DD with a BAC of 1.5 pro mille (150 mg/dl) was considered a serious offence, leading to imprisonment in most cases. The cut-off point was lowered to 1.0 pro mille (100 mg/dl) in 1994. DD offences with a lower BAC (less than 1.5 pro mille) were not included in SCBs register of 19701972. Risky drivers were drivers charged under the Criminal Act of Traffic Violations, with no alcohol involvement. The principal violations were driving without a license, reckless driving and hit-and-run accidents, driving behaviours seemingly in agreement with Jessor et al.s (1977)
characterization. Drivers charged were divided into three different groups, based on the type of traffic offence during the follow-up period: DD, RD, and both drunk and risky driving (DRD) in the same person. The period for DD and RD registrations were 19731993, inclusive, whereas the registrations for hospitalizations and mortality also covered 1994 and 1995. There were altogether 527 subjects in the DD category, 209 in the RD category and 110 in the DRD group.
The computerized data on hospital care for all diagnoses and E-codes in Stockholm County were also linked to the enrolment data set up to 1995. The county includes metropolitan Stockholm and surrounding municipalities, with a population of about 1.5 million in 19691970 and about 1.7 million in 1995. The coverage of the register was about 95% in 1974 and exceeded 99% from 1980 onwards (Andréasson et al., 1990; P. Sjöberg, Stockholm County Council, personal communication, 1998). Mortality data from the Swedish Cause of Death Register at Statistics Sweden were also linked to the enrolment data set. Between 1973 and 1986, the diagnoses for morbidity and mortality were classified according to the International Classification of Deaths and Diseases no. 8 (ICD-8) and from 1987 by ICD-9 (World Health Organisation, 1967
, 1977
). We chose to study diagnoses, which were related to alcohol use in young and middle-aged men, in addition to all hospitalizations and deaths. The following diagnoses according to ICD-9 were included: all cases with an E-code as diagnosis, traffic accidents (E 807849), suicide attempts, suicide and injury whether accidentally or purposely inflicted (E 950959, E 98089), alcohol diagnoses (291, 303, 305A, 980) and narcotic diagnoses (304), and for morbidity, also psychoses (295, 296, 298), except alcohol psychosis (291). Also, hospitalization with any diagnosis (thus including all diagnoses) was studied as an outcome.
Statistical methods
Chi2 tests were performed to analyse the differences in background characteristics between the groups of drunk drivers and risky drivers. In calculating the RRs of hospital admissions and death, we used Poisson regression analysis (Kleinbaum et al., 1988; SAS Institute, 1993
). In Poisson regression, the association between one or more independent variables and an outcome variable, defined as one case per person for all persons included, divided by the total time for all persons (incidence), was analysed. The RR, then, is the ratio between two incidence figures, i.e. for a category of an independent variable and a reference category. Ninety-five percent confidence intervals (95% CI) were also calculated.
Both bivariate and multivariate Poisson regression analyses were performed. The multivariate analyses were performed controlling for the effects of background variables with a statistically significant association to one or both outcomes in bivariate analyses and an association to RD/DD. Ten background variables remained as such confounders in the multivariate analyses (Fathers social class, divorced parents, truancy, ran away from home, criminality, low emotional control, risky use of alcohol, illicit drug use, smoking more than 10 cigarettes a day, sniffing of solvents). Only one of two or more highly correlated variables was included if their pairwise correlation was above 0.4 and if these two variables measured a similar condition.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The RR for persons from the DD, RD and DRD groups, compared with other enrolees, to be admitted into hospital care for certain diagnoses is presented in Table 2. The risk was significantly increased for all diagnoses, and especially for alcohol and drug misuse diagnoses, although with few cases. The association between RD and hospital care for psychosis, alcohol and drug misuse diagnoses was particularly strong. In the DD group, there was a RR of 2.9 for admission to hospital following a traffic accident; in the RD group this RR was 2.8, and in the DRD group 7.2. For all E-codes, the distribution was more even, the RR varying between 2.1 and 4.4. Apart from psychoses, the DRD group was at highest risk overall. The sum on subjects with all diagnoses (any diagnosis) was lower than the sum of the specific diagnoses, as the same subject could be in more than one diagnostic category.
|
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The detection rate for DD is known to be low (Persson, 1980). Thus, there could have been undetected occasions of DD in the RD group, or among other conscripts. On the other hand, a person may be known in the neighbourhood for both reckless driving and DD, which could increase the detection rate for DD, when reported to the police by neighbours. Possible differences in average distances driven by the DD and the RD groups may affect the detection rate, since greater distances would presumably increase the likelihood of detection. Unfortunately, we had no such information.
Factors assessed at enrolment
Regarding the background factors at enrolment, there were significant differences between the DD and the RD groups for several factors, but not with regard to alcohol consumption.
The RD group had a higher frequency of risky alcohol use reported at enrolment, than the DD group, but the figures did not differ significantly. This is consistent with other reports. Scoles et al. (1984) reported that almost 50% of 124 high-risk drivers (with violations such as speeding and reckless driving) not formerly identified as drunk drivers were experiencing significant problems with alcohol. Vingilis (1983)
reported that a substantial proportion of high-risk drivers seemed to be alcoholics or heavy drinkers. Maybe risky drivers are more able to separate alcohol from driving, than individuals in the DD and did not drive as often with a high BAC as the DD group, assuming that the differences cannot be explained by bias in detection by the police. Sniffing of solvents before enrolment was especially common in the RD and DRD groups. Oetting et al.(1985)
described three types of inhalant users: young inhalant users, adolescent polydrug users, and inhalant-dependent adults, and reported that inhalant users are more likely to have emotional problems than non-drug users or marijuana users. Adolescent inhalant users often take drugs and also misuse alcohol. They are also more likely to be deviant and to be involved in crimes. This is consistent with our findings of inhalant misuse as an important risk factor both for DD and RD. Gjerde et al.(1990)
studied driving under the influence of toluene, and found a high recidivism rate among drivers arrested between 1983 and 1987.
Thus, like Donovan et al.(1985), Wilson (1992)
and Vingilis (1983)
, we found both similarities and differences between drunk drivers and risky drivers.
Factors assessed at follow-up
The main findings of our study concerned hospitalization. The RRs were most elevated for the DRD group. The risk of hospitalization with an alcohol diagnosis in the multivariate analysis was significantly elevated for both the RD and the DD groups, although the RR tended to be higher for the RD, than for the DD, group. Risky drivers were more prone to be hospitalized for E-codes in general than drunk drivers, and had a much higher risk for hospitalization following a drug misuse diagnosis. One explanation could be that some risky drivers had a more reckless lifestyle generally, even in other situations. For drunk drivers, there were higher risks of suicide attempts and traffic accidents than among risky drivers. Windle and Miller (1990) found that problem drinking was associated with depression among DD offenders. This was expected, as subjects in this category had been sentenced both for DD and RD, and as the RD and the DD groups separately were associated with increased RRs.
Mortality in specific diagnostic groups could not be appropriately analysed, due to too few cases. The RR for mortality is substantially increased, but imprecise, as suggested by the wide CIs due to too few cases in the multivariate analysis, especially in the RD group (Table 4).
Because of missing data, the number of cases in the multivariate analyses was lower than in the bivariate analyses, especially for risky drivers. The comparison between these two kinds of analyses was consequently hampered. The low number of cases is reflected in wide CIs, showing that the risk estimates are not precise. In the DD and RD groups, the RRs were lower in the multivariate analysis, but were still significant.
The RR for mortality was somewhat higher for the DD, than for the RD, group as previously mentioned. However, the number of deaths was quite small, as the highest age was only 45 years at the end of the follow-up period.
Colquitt (1987) found that 1/3 of the 252 patients hospitalized after motor vehicle accidents had a BAC of 1 pro mille (100 mg/dl) or more. Few of these were referred to alcoholism-rehabilitation programmes or to courts for prosecution. The authors suggested that it should be mandatory for physicians to send all injured patients with high BAC to treatment, and also that prosecution for DD should be made easier by simplifying the procedures for providing courts with the BAC measurements. If so, a considerable proportion of drunken drivers would be identified and offered treatment. This would be greatly facilitated if all drivers from traffic accidents were asked about alcohol and drug use, and that laboratory tests were used to obtain further information, when needed. Further such information should also be obtained from subjects who have been taken for RD. There are good reasons to support these suggestions.
![]() |
ACKNOWLEDGEMENTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
FOOTNOTES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Andréasson, S., Allebeck, P. and Romelsjö, A. (1990) Hospital admissions for somatic care among young men: the role of alcohol. British Journal of Addiction 85, 935941.[ISI][Medline]
Brewer, R. D. and Morris, P. D. (1994) The risk of dying in alcohol-related automobile crashes among habitual drunk drivers. New England Journal of Medicine 331, 513517.
Colquitt, M. (1987) Drunk drivers and medical and social injury. New England Journal of Medicine 317, 12621266.[Abstract]
Donovan, D. M., Quesser, H. R., Salzberg, P. and Umlauf, R. (1985) Intoxicated and bad drivers: subgroups within the same population of high-risk men drivers. Journal of Studies on Alcohol 46, 375382.[ISI][Medline]
Edwards, G., Anderson, P., Babor, T., Casswell, S., Ferrence, R., Giesbrecht, N., Godfrey, C., Holder, H., Lemmens, P., Mäkelä, K., Midanik, L., Norström, T., Österberg, E., Romelsjö, A., Room, R., Simpura, J. and Skog, O. J. (1994) Alcohol Policy and the Public Good. Oxford University Press, Oxford.
Gjerde, H., Smith-Kielland, A., Normann, P. T. and Morland, J. (1990) Driving under the influence of toluene. Forensic Science International 44, 7783.[CrossRef][ISI][Medline]
Halldin, J. (1984) Prevalence of Mental Disorder in an Urban Population in Central Sweden with a Follow up of Mortality, pp. 1516. Karolinska Institutet, Sundbyberg, Sweden (Dissertation).
Jenkins, J. and Sainsbury, L. (1980) Single-car road deaths disguised suicides. British Medical Journal 281, 1041.
Jessor, R. M., Turbin, M. and Costa, F. (1997) Predicting developmental change in risky driving: the transition to young adulthood. Applied Developmental Science 1, 416.
Jonah, B. (1997) Sensation seeking and risky driving: a review and synthesis of the literature. Accident Analysis and Prevention 5, 651665.[CrossRef]
Karlsson, G. and Romelsjö, A (1997) A longitudinal study of social, psychological and behavioural factors associated with drunken driving and public drunkenness. Addiction, 92, 447457.[CrossRef][ISI][Medline]
Kleinbaum, D. G., Kuppper, L. L. and Muller, K. E. (1988) Applied Regression Analysis and Other Multivariate Methods. Duxbury Press, Belmont.
Oetting, E., Edwards, R. and Beauvais, F. (1985) Social and psychological factors underlying inhalant abuse. NIDA Research Monograph 85, 172203.
Öström, M. and Eriksson, A. (1993) Single-vehicle crashes and alcohol: a retrospective study of passenger car fatalities in northern Sweden. Accident Analysis and Prevention 25, 171176.[CrossRef][ISI][Medline]
Persson, L. G. W. (1980) Actual Drunken Driving in Sweden I: Hidden Criminality Theoretical and Methodological Problems. Empirical Results. Sociologiska Institutionen, Stockholm (Dissertation).
Romelsjö, A., Henningsohn, L., Jakobsson, S. W. and Karlsson, G. (1993) The prevalence of alcohol-related mortality in both sexes: variation between indicators, Stockholm. American Journal of Public Health 83, 838844.[Abstract]
SAS Institute Inc. (1993) SAS Technical Report P-243, SAS/STAT Software, the Genmod Procedure.
Scoles, P., Fine, E. W. and Steer, R. (1984) Personality characteristics and drunk patterns of high-risk drivers never apprehended for driving while intoxicated. Journal of Studies on Alcohol 45, 411416.[ISI][Medline]
Shaper, A. G. (1990) Alcohol and mortality: a review of prospective studies. British Journal of Addiction 85, 837847.[ISI][Medline]
Statistiska Centralbyrån (Statistics Sweden) (1956) Kodlista för socialgruppsindelning. Statistiska centralbyrån (Statistics Sweden), Stockholm.
Stenbacka, M. (1992) Initiation into Drug Abuse: The Role of Social, Behavioural and Psychological Factors in Initiation of Cannabis Use. Karolinska Institutet, Department of Social Medicine, Sundbyberg (Dissertation).
Wilson, J. (1992) Convicted impaired drivers and high-risk drivers: how similar are they? Journal of Studies on Alcohol 53, 335344.[ISI][Medline]
Windle, M. and Miller, B. (1990) Problem drunk and depression among DWI offenders: a three-wave longitudinal study. Journal of Consulting and Clinical Psychology 58, 166174.[CrossRef][ISI][Medline]
Vingilis, E. (1983) Drunk drivers and alcoholics. Are they from the same population? In Research Advances in Alcohol and Drug Problems, Smart, R., Glaser, F., Israel, Y., Kalant, H., Popham, R. E. and Schmidt, W. eds, pp. 299342. Plenum Press, New York.
World Health Organisation (1967) International Classification of Diseases, 8th revision (ICD-8). World Health Organisation, Geneva.
World Health Organisation (1977) International Classification of Diseases, 9th revision (ICD-9). World Health Organisation, Geneva.