Department of Psychiatry, St Lucas Andreas Hospital, Amsterdam,
1 Amsterdam Institute of Addiction Research, Amsterdam and
2 Department of Clinical Chemistry, St Lucas Andreas Hospital, Amsterdam, The Netherlands
Received 5 July 2000; in revised form 19 April 2001; accepted 7 June 2001
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Diagnostic procedures in this context are part of an administrative legal procedure to evaluate whether the subject has the right to have a driving licence. Under Dutch law, it is demanded that the subject has refrained from alcohol misuse for the last 12 months. In cases where alcoholism is diagnosed, the licence is withdrawn.
The legal context causes two problems in identifying alcoholics. The first problem is the understandably low validity of self-reporting of alcohol problems in DUI subjects (Mischke and Venneri, 1987). Secondly, in many instances, a diagnosis of alcoholism has to be defended in legal procedures. Diagnoses, based on clinical judgement and data with an unreliable correlation with alcoholism, are increasingly challenged in court with questions about the chance of a false positive diagnosis. The accuracy of diagnostic procedures used in diagnosing alcoholism in DUIs is unknown and has, to the best of our knowledge, never been investigated before.
Research suggests a considerable prevalence of AUD in DUI populations. In a review of prevalence reports up to 1986, Vingilis (1989) estimated the prevalence to be between 25 and 50%, depending on the sampling of the population and the criteria used for alcoholism. More recent studies, using DSM-III criteria (American Psychiatric Association, 1980) and biochemical tests show the same prevalence range (Table 1
). However, there is reason for scepticism about the validity of these prevalence values. The number of cases with elevated biochemical markers cannot be equated with true cases of alcoholism, as was done in most of the reviewed studies, mainly because elevated biochemical markers are not in a strict sense markers of alcoholism, but of hazardous use of alcohol. More importantly, research indicates that the sensitivity of biochemical markers drops dramatically in young alcoholics, and also in drinkers with less severe alcoholism (Nystrom et al., 1992
; Allen et al., 1994
; Litten et al., 1995
; Huseby et al., 1997a
; Salaspuro, 1999
). As young drivers represent a relatively high proportion of DUI populations, and severe alcoholics represent only a small minority, the reported prevalence values can be considered as conservative (Vingilis, 1989
; Hasin et al., 2000
).
|
In evaluating diagnostic procedures, one must consider the differences between diagnosing alcoholism in health-care settings and in legal settings. In health care, the main diagnostic aim is to enhance health. Therefore it is important to identify all alcoholic patients. In order to minimize the risk of missed diagnoses, a high sensitivity of diagnostic procedures and tests is important. Usual clinical diagnostic procedures (CDPs) in health care depend on clinical judgement which incorporates all available historical, clinical and laboratory data. In the legal setting of medical examination in a DUI population, the aim is not to enhance health, but to enhance traffic safety. Because diagnosis may be challenged in court, diagnosis is restricted to sure cases. In order to minimize the risk of false positive diagnoses, a high specificity of the diagnostic procedure is important. Therefore more restrictive diagnostic procedures (RDP) are used. Ideally, legal diagnostics must rely on objective, reliable and specific data, such as recent history of alcohol problems, physical signs of alcoholism or specific biochemical tests of hazardous alcohol use. In legal settings, high specificity of diagnostic tests is more important than high sensitivity, because incorrect diagnoses have unacceptable legal consequences.
Understanding the legal dilemma is essential in choosing between the different diagnostic procedures. The dilemma is to find a balance between two opposite aims. On the one side, the requirement is to enhance traffic safety (for the public) each missed diagnosis endangers traffic safety. On the other side, the requirement is to protect the rights of the individual each incorrect diagnosis may have unacceptable consequences (for the individual), such as losing employment after being disqualified from driving.
In the present study, three prevalence estimations, obtained with different diagnostic procedures, were compared with each other and with an unbiased prevalence estimate based on sensitivity and specificity data of biological markers of alcoholism in non-judicial samples; i.e. compared to the estimate of hazardous alcohol use which would be obtained by applying population-based findings to the blood test marker results obtained in the examination.
![]() |
SUBJECTS AND METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In accordance with Dutch traffic regulations the following groups were included for referral and examination. (1) DUIs with at least one arrest with a blood-alcohol concentration (BAC) 2.1
(
210 mg/dl) or three DUI arrests with any BAC >0.5
(
50 mg/dl) within 5 years, or refusal to cooperate with breath analysis (examination group). This group is referred for examination by the Dutch Traffic Test Organization, Disqualification Division, which covers the costs. Some basic information on the characteristics of drivers processed under these regulations were obtained from the Dutch Traffic Test Organization, which supplied data on all DUIs examined in The Netherlands in 1997. (2) DUIs who apply for re-granting of the driving licence after previous DUI, medical examination and loss of permanent driving licence for 12 months because of diagnosis of alcoholism (re-examination group). In this group almost all individuals are self-referred, applying for re-licensing, and have to pay for the examination.
Standardized clinical data collection
All DUIs were examined and diagnosed by the same physician. The examination was recorded in a standardized clinical report, which was part of a legal procedure on behalf of the Dutch Traffic Test Organization. The clinical report of each subject consisted of extensive history taking, instruments to assess AUD, namely Structured Clinical Interview (SCID) and the CAGE questionnaire, physical examination, biochemical measurements and a conclusive clinical judgement as to whether it was probable the subject had AUD in the last 3 or 12 months. History taking was focused on clinical symptoms of alcoholism and on possible and probable non-alcoholic causes for elevated biochemical markers. The latter included questions about current and past illness, specifically diabetes, liver diseases, blood transfusions and intravenous drug use [because of the possibility of hepatitis C which can also affect carbohydrate-deficient transferrin (CDT)] (Perret et al., 1997), anaemia, and drugs that could affect biochemical markers (anti-epileptics, folate antagonists, anti-AIDS medication, phenothiazines, some diuretics and thyrostatics).
Alcoholism or AUD is defined as either alcohol abuse or alcohol dependence according to DSM-IV (American Psychiatric Association, 1994).
Blood tests
Venous blood samples for determination of haemoglobin (Hb), haematocrit (Ht), red blood cell count (E), mean cell volume (MCV), carbohydrate-deficient transferrin (CDT), gamma-glutamyltransferase (GGT), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) were taken. Serum samples for CDT were frozen within 4 h after collection and stored at 20°C until use. CDT was analysed in duplicate, using a commercial kit, CDTect, from Pharmacia and Upjohn, Uppsala, Sweden. Measurement of serum GGT, ALT and AST was executed within 4 h with Vitros (Ortho Clinical Diagnostics, Rochester, NJ, USA) at 37°C and re-coded for the value at 30°C. Hb, Ht, E, MCV and samples were kept at room temperature and analysed within 4 h with a Technicon H2 analyser (Bayer Diagnostics, Terrytown, WA, USA). The upper reference limits were: CDTect 20 U/l, GGT
40 U/l, ALT
34 U/l, AST
33 U/l, MCV
100 fl.
Diagnostic procedures
For reasons of comparability with the population-based method, as the diagnostic window of biochemical markers does not exceed 3 months, only current AUD diagnosis (within the last 3 months) was used in the different diagnostic procedures.
Data from clinical reports of every subject were processed in three diagnostic procedures: SCID, RDP and CDP. The three diagnostic procedures are not independent; SCID is incorporated into the RDP and both SCID and RDP are incorporated into CDP. The diagnostic procedures are described below in detail. Essentially SCID identifies alcoholics who are willing to report alcohol problems, RDP identifies those positive with SCID and with elevated biochemical markers which can be seen as proof of current hazardous drinking, whereas CDP identifies those positive on SCID and RDP and subjects with more soft signs' of alcoholism. All resulting diagnoses refer to AUD in the 3 months prior to examination.
Diagnostic procedure 1: SCID. Recent alcohol problems were assessed with the Structured Clinical Interview for DSM-IV Axis I disorders, clinician version, module E (alcohol use disorders) over the last 3 months (SCID-CV; First et al., 1997). The SCID-CV is a semi-structured interview for making the major DSM-IV diagnoses (American Psychiatric Association, 1994
). It was designed for use in clinical settings as a way of ensuring standardized assessments. A study, using earlier versions of SCID, reported testretest kappas for current diagnoses of AUD of 0.75 (Williams et al., 1992
).
AUD diagnosis was made if the subject scored positively on one of the SCID criteria of alcohol abuse or scored positively on three criteria of alcohol dependence, in the 3 months prior to interview.
Diagnostic procedure 2: RDP. We devised a RDP for detection of alcoholism, with the aim to maximize reliability and specificity of diagnosis. From the standardized recorded history, only data from the SCID interview, the four CAGE questions (Mayfield et al., 1974) and data from history and medication were used to check for possible non-alcohol causes for raised test results. From physical examination, only liver palpation was used. All biochemical measurements were used. The RDP is described in Fig. 1
.
|
In order to diminish the small chance (in this population) of incorrectly diagnosing subjects with non-alcoholic liver disease, without increasing the much greater chance (in this population) of missing diagnosis in non-abstinent subjects with alcoholic liver disease, we used a higher cut-off value of MCV in the combinations of elevated MCV, with elevated ALT, AST or GGT, as an extra safeguard against incorrect diagnosis.
The algorithm for RDP was made before analysis. Two items (subject states that he feels no effect of drinking 4 AU' and blood pressure
160/95) were deleted post-hoc, as these items did not provide any additional differentiating value.
The choice of the different combinations of elevated biochemical measurements in RDP was motivated by our aim to achieve a specificity of ~95%. In non-clinical settings, reported specificities are: GGT 8090%, ALT >80%, AST >90%, elevated MCV in men >90% and elevated CDT >80% (Sillanaukee et al., 1992; Conigrave et al., 1995
; Salaspuro, 1999
). As hazardous use of alcohol elevates CDT, MCV and the enzymes GGT, ALT and AST through partly independent biological pathways, these markers can be considered as partly independent tests (Bean and Daniel, 1996
). Demanding that a pair of diagnostic tests are simultaneously elevated before making a positive diagnosis of alcoholism maximizes specificity and minimizes false positive labelling of innocent patients, at the price of many missed diagnoses (Sackett et al., 1991
). If CDT, GGT and MCV are independent tests, simultaneous elevation of CDT and GGT, or CDT and MCV, or MCV and GGT will result in specificity values between 96 and 99%. This assumption has been partly confirmed in a study that measured specificity of simultaneously elevated CDT and GGT in non-alcoholic controls (van Pelt, 1997
).
Diagnostic procedure 3: CDP. In this diagnostic procedure, a diagnosis was reached through clinical judgement after evaluation of all available data, according to usual clinical practice. Besides biochemical measurements, historical data, clinical signs, and instruments to assess alcohol problems were used.
Histories included time and circumstances of arrest. A police report of BAC, data of earlier DUI and reports on earlier medical examinations after DUI were available. No information from the applicants' general practitioner was requested (which is the procedure in the UK DUI system, see Morgan and Major, 1996).
Recent alcohol intake was assessed by means of a structured interview. This included questions about the exact amount of alcohol units (AU) in the week prior to the examination, an estimate of the average AU per week during the last year, and questions about changes in quantity and frequency of drinking in the last year.
Hazardous drinking is defined as the level of persistent alcohol consumption being likely to result in adverse health effects: >280 g ethanol/week (Saunders and Lee, 2000). As 1 AU is defined as a standard drink of ~10 g alcohol, hazardous use signifies an average of more than 28 AU weekly.
Lifetime and current alcohol problems were assessed using the CAGE questions, SCID, and questions about any past treatment for alcohol problems. A subject was considered to have had a lifetime alcohol problem if CAGE score of 2 was obtained, or if a subject was ever treated for alcohol problems, or received a SCID diagnosis in the 12 months prior to the interview.
As in RDP, history also included questions about different diseases and drugs, to control for possible confounders in regard to non-alcoholic causes for elevated biochemical markers or liver enlargement.
Physical examination included breath smelling of alcohol during examination (but no alcohol breath test), blood pressure, liver palpation and observation of skin abnormalities indicative of liver dysfunction and neurological dysfunction indicative of polyneuropathy or withdrawal symptoms.
Diagnosis of recent AUD in the CDP procedure was based on clinical reasoning. All data and clinical signs were assessed as either diminishing the chance of recent AUD, increasing the chance of recent AUD, or confirming AUD diagnosis. A positive diagnosis of recent AUD was made if the above-described SCID and RDP procedures resulted in an AUD diagnosis, or if several AUD chance-increasing data were present without the presence of confounding effects of illness or drugs.
Population-based prevalence estimate of hazardous use
Studies have shown that sensitivity and specificity of markers of hazardous alcohol use depend on the distribution of severe and mild cases of alcoholism in the studied cohort. A high ratio of severe/mild cases heightens, while a low ratio lowers, sensitivity. Because we assumed that our population consists of a high-risk population of hazardous users, alcoholics and social drinkers without AUD, we used sensitivity and specificity values found in studies with two high-risk populations (Sillanaukee et al., 1993; Huseby et al., 1997b
). Sillanaukee et al. (1993) compared hazardous drinkers, with some signs of AUD, to social drinkers and found a sensitivity of 57% and a specificity of 79%. Huseby et al. (1997b) compared alcohol-dependent patients to non-dependent patients from a population of men admitted to a surgical ward and found a sensitivity of 55% and a specificity of 85%. Sensitivity and specificity values refer to the relation of AUD and elevated CDT or GGT.
The estimated prevalence of AUD was computed with the following formula:
![]() |
Statistical analysis
Statistics Package for Social Sciences (SPSS for Windows 9.0, 1999) was used for computation of frequencies. Comparison of groups was performed with the t-test. Comparisons of multiple groups were conducted with analysis of variance.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
In all, 107 out of 212 DUIs reported lifetime alcohol problems according to our definition (if the CAGE score is 2, or if a subject was ever treated for alcohol problems, or received a SCID diagnosis in the 12 months prior to the interview). As expected, the re-examination group reported more lifetime alcohol problems (61.3%) than the examination group (36.6%). The last percentage is probably the result of under-reporting. According to an epidemiological study performed in 1996, the 1-month, 12-month and lifetime prevalence of AUD in the Dutch male population was respectively 8.5, 13.4 and 28.3% (Bijl et al., 1998
).
Prevalence according to different diagnostic procedures
SCID. Applying SCID over the last 3 months as a diagnostic procedure identified seven DUIs with AUD in the examination group and only one in the re-examination group. According to SCID, the estimated prevalence of AUD over the last 3 months in the examination group was 7.5%, but only 0.8% in the re-examination group (Table 3). Applying SCID over the last 12 months identified 21 DUIs with AUD in the examination group (22.6%) and four in the re-examination group (3.4%).
RDP. The RDP resulted in an AUD diagnosis in 32 DUIs from the examination group (prevalence according to RDP 34.4%) and 18 from the re-examination group (prevalence of AUD in the re-examination group 15.1%).
CDP. Using all data, this resulted in an AUD diagnosis in 54 DUIs in the examination group (prevalence according to CDP 58.1%) and 43 DUIs in the re-examination group (prevalence according to CDP 36.1%).
Population-based prevalence computation
The total number of subjects with elevated CDT or GGT was 101 (51 from the examination group and 50 from the re-examination group). The proportion of DUIs with elevated biochemical markers was 101/212. Using the sensitivity and specificity values found by Sillanaukee et al. (1993) resulted in an estimated prevalence of AUD for all DUIs in our study of 74%. Using the sensitivity and specificity values found by Huseby et al. (1997a) resulted in an estimated prevalence of 82%.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
On the one hand, one has to consider the possibility that the low sensitivity of biochemical markers, used in the population-based method, inflates the estimated prevalence of AUD beyond results of earlier research and beyond face validity. Another possible explanation is that the estimated prevalence found with the population-based method, either 82 or 74%, can be considered as maximal prevalence only. As hazardous drinking encompasses a larger group than the group with AUD, the criterion can be only used as maximal reference level. On the other hand, one has to consider the possibility that the diagnostic tools to detect alcoholism in DUIs result in considerable under-diagnosing. SCID identifies maximally 5% of all AUD found with the unbiased estimate. This performance was not unanticipated; SCID identifies only those alcoholics who are aware of, and willing to be open about, their problems. For obvious reasons, most DUIs will not be open about their alcohol consumption (which was reported as three times lower than the mean in the Dutch population) or about the frequency of experiencing alcohol problems (which was also reported as lower than in the Dutch population).
RDP identified six times as many as the SCID procedure only, and at least 28% and maximally 31% of the unbiased AUD estimate. This is a significant gain, compared to SCID. At the same time, it is evident that the sensitivity of the RDP is low. This result is also according to expectation: RDP will result in under-diagnosis, because physical signs of alcoholism are late symptoms of alcoholic disease, because ~520% of alcohol-dependent patients and 4060% of alcohol abusers show no elevations of biochemical tests (Sillanaukee et al., 1993; Litten et al., 1995
; Hillman et al., 1998
) and because 31% of our population consisted of subjects younger than 35 years. In young subjects, biochemical markers have a low sensitivity for detection of alcoholism. Another reason for under-diagnosis by RDP is that any possible non-alcoholic cause had also automatically led to exclusion of the diagnosis.
CDP identified at least 56% of the unbiased AUD estimate and up to 60% of the examination group. Even if one assumes that the prevalence of AUD found with CDP is rather high, one has to consider that the prevalence values in this study refer to prevalence of AUD found several months after the DUI arrest. It seems reasonable to assume that prevalence of AUD at the time of arrest would be much higher.
The above-mentioned prevalence is dependent on the administrative selection of DUIs for examination, which varies in different countries. The issue here is to provide the clinician, working within a legal situation, with a method to calculate the PPV and NPV for different diagnostic procedures. The diagnostic gain of CDP above RDP has significant legal disadvantages that can be illustrated by the consequences for the PPV of this procedure. If we use this procedure in a population with a 40% prevalence of AUD, under the optimistic assumption that CDP has a specificity of 80% and a sensitivity between 60 and 95%, the PPV of CDP will vary between 66 and 75%. This may be quite acceptable in health-care settings, but is evidently not acceptable in legal settings. The high chance of false positive diagnosis makes CDP unacceptable in the legal context of AUD diagnosis in DUI populations. Until better markers are available, we advise physicians who participate in diagnosing AUD in DUI populations to use RDP enhanced with secondary data such as circumstances of arrest.
It remains to be researched if RDP (enhanced or not) has a high enough PPV and an acceptable NPV. However, it is too optimistic to hope that such research will be able to replace clinical reasoning completely (Gilg, 1999). As different sub-groups of DUIs have different a priori prevalences (Table 3), and test parameters of biochemical markers are dependent on age and gender, different norms must be used in diagnostic procedures. Even if precise knowledge of the PPV of different diagnostic procedures in different groups becomes available, one has still to answer a social, as well as the legal, question: how sure one has to be of diagnosis in diagnosing alcoholism in DUI populations?
|
![]() |
ACKNOWLEDGEMENTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
FOOTNOTES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
American Psychiatric Association (1980) Diagnostic and Statistical Manual of Mental Disorders, 3rd edn. American Psychiatric Association Press, Washington, DC.
American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders, 4th edn. American Psychiatric Association Press, Washington, DC.
Bean, P. and Daniel, P. (1996) Carbohydrate-deficient transferrin; current facts and future projections for the insurance industry. On the Risk 12, 4348.
Bijl, R. V., Zessen, G. van and Ravelli, A. (1998) Psychiatric morbidity among adults in the Netherlands: the NEMESIS study. II. Prevalence of psychiatric disorders. Nederlandse Tijdschrift voor Geneeskunde 141, 24532460.
Conigrave, K. M., Saunders, J. B. and Whitfield, J. B. (1995) Diagnostic tests for alcohol consumption. Alcohol and Alcoholism 30, 1326.[Abstract]
de Zwart, W. M. (1998) Ontwikkelingen in het gebruik van alcohol en drugs. In Verslaving, Mensink, C. and Spruit, I. P. eds, pp. 1642. Bohn, Stafleu en van Loghum, Houten.
Dunbar, J. A., Ogston, S. A., Ritchie, A., Devgun, M. S., Hagart, J. and Martin, B. T. (1985) Are problem drinkers dangerous drivers? British Medical Journal 290, 827829.[ISI][Medline]
Dutch regulations on driving ability (1994) Regeling Maatregelen rijvaardigheid en geschiktheid. Regeling van de Minister van Verkeer en Waterstaat, houdende nadere regels met betrekking tot maatregelen rijvaardigheid en geschiktheid, bedoeld in Hoofdstuk VI, paragraaf 9, van de Wegenverkeerswet 1994. RV 216650.
First, M. B., Spitzer, R. L., Gibbons, M. and Williams, J. B. W. (1997) Structured Clinical Interview for DSM IV Axis I Disorders (SCID I) Clinical Version. American Psychiatric Press, Washington.
Gilg, T. (1999) Einsatzmöglichkeiten von CDT in der rechts- und verkehrmedizin. In Klinische Alcoholismusdiagnostik, Soyka, M. ed., pp. 118129. Steinkopf, Darmstadt.
Gjerde, H. and Mørland, J. (1987) Concentrations of carbohydrate-deficient transferrin in dialysed plasma from drunken drivers. Alcohol and Alcoholism 22, 271276.[ISI][Medline]
Gjerde, H., Sakshaus, J. and Mørland, J. (1986) Heavy drinking among Norwegian male drunken drivers: a study of gamma glutamyltransferase. Alcoholism: Clinical and Experimental Research 10, 209212.[ISI][Medline]
Hasin, D., Paykin, A., Endicott, J. and Grant, B. (2000) The validity of DSM-IV alcohol abuse: drunk drivers versus all others. Journal of Studies on Alcohol 60, 746755.[ISI]
Hillman, A., Sykes, R. A. D. and McConnell, A. A. (1998) Limitation in the use of gamma glutamyltransferase estimations in alcohol-dependent subjects. Alcohol and Alcoholism 33, 626630.[Abstract]
Huseby, N. E., Nillssen, O., Erfurth, A., Wetterling, T. and Kanitz, R. D. (1997a) Carbohydrate-deficient transferrin and alcohol dependency: variation in response to alcohol intake among different groups of patients. Alcoholism: Clinical and Experimental Research 21, 201205.[ISI][Medline]
Huseby, N. E., Nillssen, O. and Kanitz, R. D. (1997b) Evaluation of two biological markers combined as a parameter of alcohol dependency. Alcohol and Alcoholism 32, 731737.[Abstract]
Iffland, R. and Grassnack, F. (1995) Epidemiologische untersuchung zum CDT und andere indikatoren für alcoholprobleme in blut alcoholauffalliger Deutsche PKW fahrer. Blutalcohol 32, 2741.
Jaster, R. and Wegener, R. (1993) Diagnosis of chronic alcohol abuse in intoxicated drivers: CDT in combination with other parameters. Blutalcohol 30, 257265.
Litten, R. Z., Allen, J. P. and Ferrite, J. B. (1995) Gamma glutamyl transpeptidase and carbohydrate deficient transferrin: alternative measures of excessive alcohol consumption. Alcoholism: Clinical and Experimental Research 19, 15411546.[ISI][Medline]
Lutz, F. U. and Bausach, S. (1992) Leberenzymwerte (Gamma-GT, GOT, GPT) bei trunkenheitstätern zum tatzeitpunkt. Blutalcohol 29, 211215.
Mayfield, D., McLeod, G. and Hall, P. (1974) The CAGE questionnaire: validation of a new alcoholism screening instrument. American Journal of Psychiatry 131, 11211123.[ISI][Medline]
Michiels, W. and La Harpe, R. (1996) Les ivresses dans le trafic a Genève: abus occasionels ou alcoolisme. Soz Präventivmed 41, 2835.
Miller, B. A., Whitney, R. and Washousky, R. (1986) Alcoholism diagnoses for convicted drinking drivers referred for alcoholism evaluation. Alcoholism: Clinical and Experimental Research 10, 651656.[ISI][Medline]
Mischke, H. D. and Venneri, R. L. (1987) Reliability and validity of the MAST, Mortimer Fuilkins Questionnaire and CAGE in DWI Assessment. Journal of Studies on Alcohol 48, 492501.[ISI][Medline]
Morgan, M. Y. and Major, H. E. (1996) The use of serum carbohydrate deficient transferrin in the assessment of high risk offenders in Great Britain. Alcohol and Alcoholism 31, 625628.[Abstract]
Nickel, W. R. (1995) Regranting of Driving Licenses Report of the Working Group on Regranting of Licenses. International Council on Alcohol Drugs and Traffic Safety, December 1995.
Nystrom, M., Peraslo, J. and Salaspuro, M. (1992) Carbohydrate-deficient transferrin (CDT) in serum as a possible indicator of heavy drinking in young students. Alcoholism: Clinical and Experimental Research 16, 9397.[ISI][Medline]
Östrom, 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.[ISI][Medline]
Papoz, L., Weill, J., lHoste, J., Chich, Y., Got, C. and Goehrs, Y. (1986) Biological markers of alcohol intake among 4796 subjects injured in accidents. British Medical Journal 292, 12341237.[ISI][Medline]
Perret, R., Froehlich, F., Levanchy, D., Henry, H., Bachman, C., Pécoud, A., Bianchi, L. and Gonvers, J. J. (1997) Is carbohydrate deficient transferrin a specific marker for alcohol abuse? A study in patients with chronic viral hepatitis. Alcoholism: Clinical and Experimental Research 21, 13371342.[ISI][Medline]
Pikkarainen, J. and Pentilla, A. (1989) Frequency of heavy and problem drinkers among Finish drivers. In High Alcohol Consumers and Traffic: Proceedings of the International Workshop; High Alcohol Consumers and Traffic, November 2830, 1988, pp. 3748. Inrets, Institut National de Recherche sur les Transports et leur Sécurité, Arcueil.
Poole, K. W., Flynn, P. M., Rao, A. V. and Cooley, P. C. (1996) Estimating the prevalence of drug use from self-reports in a cohort for which biological data are available for a subsample. American Journal of Epidemiology 144, 413420.[Abstract]
Ruud, J., Gjerde, H. and Mørland, J. (1993) Alcohol consumption among convicted drunken drivers in Norway. Journal of Traffic Medicine 21, 121125.
Sackett, D. L., Haynes, R. B., Guyatt, G. H. and Tugwell, P. (1991) Clinical Epidemiology. Little Brown and Company, Boston.
Salaspuro, M. (1999) Carbohydrate-deficient transferrin as compared to other markers of alcoholism: a systematic review. Alcohol 19, 261271.[ISI][Medline]
Saunders, J. B. and Lee, N. K. (2000) Hazardous alcohol use: its delineation as a subthreshold disorder, and approaches to its diagnosis and management. Comprehensive Psychiatry 41, 95103.[ISI][Medline]
Scoles, E. A., Fine, E. W. and Steer, R. A. (1986) DUI offenders presenting with positive blood alcohol levels at pre-sentencing evaluation. Journal of Studies on Alcohol 47, 500502.[ISI][Medline]
Sillanaukee, P., Seppa, K. and Koivula, T. (1992) Acetaldehyde-modified hemoglobin as a marker of alcohol consumption: comparison of two new methods. Journal of Laboratory and Clinical Medicine 120, 4247.[ISI][Medline]
Sillanaukee, P., Seppa, K. and Koivula, T. (1993) CDT by anion-exchange chromatography by RIA as a marker of heavy drinking among men. Alcoholism: Clinical and Experimental Research 17, 230233.[ISI][Medline]
van Pelt, J. (1997) Carbohydrate deficient transferrin: a new biochemical marker for chronic excessive alcohol consumption. Nederlands Tijdschrift voor Geneeskunde 141, 773777.[Medline]
Vingilis, E. (1983) Drinking drivers and alcoholics: are they from the same population? In Research Advances in Alcohol and Drug Problems, Smart, R. G. et al. eds, pp. 299342. Plenum Press, New York.
Vingilis, E. (1989) Are drinking-drivers alcoholics? A review of the literature. In High Alcohol Consumers and Traffic: Proceedings of the International Workshop: High Alcohol Consumers and Traffic, November 2830, 1988, pp. 165182. Inrets, Institut National de Recherche sur les Transports et leur Sécurité, Arcueil.
Waller, J. A. and Turkel, H. W. (1966) Alcoholism and traffic death. New England Journal of Medicine 275, 532536.[ISI][Medline]
Williams, J. B., Gibbon, M., First, M. B. et al. (1992) The structured clinical interview for DSM-III-R (SCID) multisite test retest reliability. Archives of General Psychiatry 49, 630636.[Abstract]