The ICADTS Working Group on Guidelines on experimental studies undertaken to determine a medicinal drug`s effect on driving or skills related to driving

June 1999

Medicinal drugs are an integral part of almost all medical treatments. In relieving and curing diseases, they provide the population with a positive contribution to quality of life. Besides the beneficial effects, many commonly used medicinal drugs possess the potential for seriously impairing human performance. Concerning medicinal drugs and traffic safety, DE GIER (1995) states that a very conservative estimate indicates that 10 % of the adult population drives under the influence of impairing drugs with twice the risk of being involved in a traffic accident and that, assuming this figure, those drugs are causing 4,500 deaths, 135,000 injuries and 6,3 billion EURO in property damage and immediate medical care each year in Europe.

Since epidemiological studies on the effects of medicaments are rare, the hazardous potential of medicinal drugs is hard to assess. Thus, findings of experimental studies on drugs and driver fitness are of great importance as the basis for decisions in the field of traffic safety. The empirical studies, as a whole, are suitable as a data base for categorizing the potential hazard of medicinal drugs only when they are based on a sound methodology and when the results of different studies are comparable. A review of the literature leaves the impression that these prerequisites are not met due to, among other things, the considerable variety of elements of the study design, of the sample choices, of the treatment, of the methods of testing driver fitness, and of the statistical evaluation. Due to these methodological differences and to many other reasons -- for example economical ones, reasons of product safety, on the possibility of judging the quality of a study by sponsors of a study or authorities -- a harmonization, optimization, and standardization of the experimental methodology is indispensable.

The aforementioned shortcomings caused the scientific community in the field to select the methodology of experimental studies on medicinal drugs and driver fitness as a central theme in workshops and symposia. Systematic international efforts cosponsored by ICADTS began in 1991. On the basis of a paper and a questionnaire prepared by VERMEEREN, DE GIER and O`HANLON, experts in the field discussed the appropriate questions at an International Workshop in Padova, 1991 (see FERRARA and GIORGETTI, 1992). The results of these discussions were published by DE GIER and LAURELL (1992). In addition, other experts answered the questionnaire, and the results were published by VERMEEREN, DE GIER and O`HANLON (1993, final report containing guidelines with arguments) and VERMEEREN, DE GIER, O`HANLON, LAURELL (1993). The efforts were followed by workshops and presentations of ICADTS Congresses in Cologne (T`92), Adelaide (T`95), and Annecy (T`97) and a Symposium in Nuremberg in 1994 (see KRÜGER, KOHNEN and SCHÖCH, 1995). Due to the relevance and the scope of the endeavour needed, and due to the fact that improvement of methodology is a constant task, the ICADTS established a Working Group on the topic in autumn 1994.

This paper completes the first task of the Working Group, namely to compile the guidelines concerning the adequate methodology of studies on drugs and driver fitness as discussed in the above mentioned papers and events.

This document is specifically addressed to pharmaceutical manufacturers and medicinal drug regulatory authorities who share the responsibility for ensuring the safe use of medicinal drugs by patients who operate motor vehicles. It is hoped that the material presented herein will lead to guidelines for the standardized assessment of each new medicinal drug`s hazard potential for driving as part of the registration process. Programmatic research should lead in all cases to the categorization of the medicinal drug`s hazard potential using a simple scheme that will be understood by the prescriber, dispensers, and ultimate users (see section 6.4).

Even though the topic of illicit drugs may currently be attracting more attention than medicinal drugs, the guidelines at hand are restricted to medicinal drugs. The fundamental reason against the inclusion of illicit drugs is -- in our opinion and probably in the opinion of all doctors prescribing medicinal drugs -- the essential difference between medicinal drugs and illicit drugs. Medicinal drugs are necessary and valuable to patients to relieve and cure diseases. On the contrary, illicit drugs are not necessary and are detrimental to consumers. Therefore, all things should be avoided that combine medicinal drugs with illicit drugs. The important danger in establishing guidelines for medicinal drugs and illicit drugs simultanously would be the trivialization of the problem of illicit drugs.

It goes without saying that the established international and national guidelines relating to clinical, therapeutic experiments that can be adopted for research on drugs and driver fitness should be the basis of the recommended guidelines. For example: guidelines for good clinical research practice (GCP) for trials on pharmaceutical products (WHO, EC), guidelines for the format and content of the clinical and statistical section of an application (FDA), biostatistical methodology in clinical trials in application for marketing authorization for medical products (EC) or structure and content of clinical study report (ICH).

Hence, in the following we concentrate in essence on methodological aspects specific to empirical studies on medicinal drugs and driver fitness. Only those issues of the general guidelines are addressed that often are neglected and that have been found to be partially incomplete in the literature, but that are considered to be essential criteria for the quality of evaluation.

The following guidelines are based fundamentally on the papers of DE GIER and LAURELL (1992) and VERMEEREN, DE GIER and O`HANLON (1993). A completion was done by a paper of BERGHAUS and FRIEDEL (1992a, 1992b), the results of the discussion of the Workshop in Cologne, the Symposium in Nuremberg, and the ICADTS Working Group Meetings in Adelaide and Annecy.

Besides this brief summary of guidelines, a more detailed one with comments to and references of the particular propositions has been established (ICADTS).

We would like to point out that the guidelines recommended are not intended to restrict research on a few established methods but to provide the investigator with a minimum set of quality assured, acceptable guidelines. The researcher is encouraged to include new elements (such as new tests, variables or measurement methods) in order to further develop methodological and psychological knowledge and to obtain a complete differential profile of the medicinal drug effects. The aim, objectives, and methods of the study should be described in advance in a detailed study protocol. Herein, the investigator should give arguments for the use of methods that differ from the suggestions of these guidelines.

Investigators should state their hypothesis concerning a particular medicinal drug`s effect before starting the study. They should distinguish clearly between the following hypotheses:

1. Testing for `difference`:
The drug is assumed to have an effect on performance different from the effect of placebo or some other drug.

2. Testing for `equivalence`:
The drug is assumed, à priori, to have an effect on performance equivalent to that of placebo or some other drug.

It is quite common to consider driver unfitness as an adverse drug effect. In this sense the decision problem is an equivalence rather than a difference testing problem. However, in both cases it is common sense to choose the probability of an erroneous rejection of the hypothesis (type one error probability, denoted by a ) at 5 % and to make sure that the probability of erroneously accepting the hypothesis (type two error probability, denoted by ß) does not exceed a moderate level (10-20 %). Hereby, the test procedure is usually based on two-sided hypotheses. Note that if the intention is to test for difference and it results in a non-significant result, then it is not correct to decide for equivalence. There are various methods to test for equivalence, such as test procedures and confidence interval inclusion rules developed in the area of bioequivalence (see WESTLAKE, 1979). For the equivalence testing problem, one also has to specify the tolerance region or equivalence region which corresponds to the (minimal) relevant difference. The tolerance region may be bounded by the maximal tolerable degree of driving hazard potential.

The sample size should be based on power calculations. In the study protocol, reasons following statistical arguments with respect to the design chosen, the analysis procedure (e.g. multiple comparison procedures, equivalence tests etc.), and the predetermined type one and type two error probabilities should be given to insure the feasibility of the sample size to test the hypotheses under investigation. Usually type one and type two probabilities are fixed in advance, and the sample size is selected subject to the relevant difference to be tested.

The subject sample should be representative (at least with respect to age and gender) of the driving population and the target population of the drug. Subjects should especially have adequate driver experience measured, for example, by minimum annual mileage or minimum overall mileage.

The individuals participating as volunteers in a study may be either healthy humans or patients. A program of drug assessment should generally start with healthy volunteers. A patient study should follow if different effects concerning performance between healthy volunteers and patients are assumed.

Subjects with hearing or visual deficits (esp. colour blindness), alcohol, medicines, drug abuse or addiction, and other factors known to affect performance must be excluded; moreover subjects who are expected (from any genetic or other predisposition) to either not respond or respond differently or adversely to the medicinal drug must not be included. In addition to the normal selection procedures for any subject that involve medical history screening and routine physical examination, extra precaution should be taken for middle aged and elderly subjects in the form of additional medical screening (e.g. blood and urine chemistry, hematological examinations, resting ECGs). Medical screening of female subjects should include a pregnancy test before beginning the trial, and assurance of reliable birth-control should be provided during trial.

The above mentioned selection criteria should be controlled as far as possible by objective tests. For example, a physical examination including the screening of a blood/urine specimen concerning alcoholism indicators should be used instead of only using questionnaires about alcohol addiction.

Cross-over designs reduce the variance of the treatment contrast (if carry-over effects are negligible) and are, therefore, to be preferred for efficiency. In most cases the power is maximized compared to the parallel group design, which leads to a smaller sample size but is often impossible to use in patient studies or in studies with slowly eliminated drugs (carry-over effect). In such cases parallel group designs are necessary.

In addition to the general exclusion criteria for subjects, alcohol and illicit drug use and concomitant medication, except oral contraception, overtiredness and other factors known to influence performance must be checked and/or controlled each test day before starting the test procedures in order to be able to describe the homogenity of the intraindividual trial conditions of each subject.

Apart from the test procedures chosen as operationalization of driver fitness (see chapter 5), additional elements should be included in a test day. Suitable specimens (blood, plasma, urine, saliva) should be taken for checking the compliance with the test conditions and for measuring pharmacokinetic quantities of the test medication. Tests and/or a questionnaire concerning the subjective feelings of the subjects with respect to sedation, experienced performance impairment, impression on driver fitness, and willingness to drive, for examples have to be included. These tests must fulfill the same objective quality criteria as the performance tests (see chapter 5.2).

Studies on investigational drugs generally should include the following control conditions

1. placebo
2. verum.

The standard active drug (verum) serves as a positive control that should demonstrate the sensitivity of the experimental procedure. Studies including the use of a verum should preferably use one of the following: an impairing representative of the same therapeutic class; ethanol sufficient to raise blood alcohol concentration (BAC) to 0.5 or 0.8 mg/ml; or diazepam 10 mg. If the use of a placebo or verum is thought to be not necessary, this should be explained.

Studies undertaken to demonstrate the effect of a new medicinal drug on driver fitness should involve at least the lowest and highest dose to be given therapeutically; or at least 1x and 2x the dose, if there is only one therapeutic dose recommended. Wherever possible, a drug should be tested additionally in healthy volunteers in doses of 4x the standard therapeutic dosage (if the drug is well tolerated in this dosage) to establish a dose-effect curve.

Concerning treatment duration, a study should last at least until a steady-state of the drug plasma concentration has been achieved, or until the desired therapeutic effect has occurred in patients.

In cross-over designs the wash-out phase must be long enough to exclude carry-over effects and should be defined according to the pharmacodynamic properties of the substance (in general, four or five times the drug`s half-life). The rationale of the decision should be made clear.

Pharmaco-kinetic and -dynamic interactions between medicinal drugs commonly used in combination and interactions between alcohol and medicinal drugs should be considered. Assessment of those combinations` effects on performance should always be conducted when it is definitely known that the presence of one interferes with the metabolism of the other; or, when both possess independent mechanisms of action causing impairment so that the possibility of a synergistic (i.e., multiplicitive) effect exists.

Studies for establishing the driving hazard potential of a particular medicinal drug should proceed from conventional laboratory testing to sophisticated driving simulators. Finally, actual driving tests (over-the-road tests in normal traffic conditions or tests on closed circuits under controlled conditions) should be carried out as far as they can safely be applied and will be allowed by the ethic commission. The final evidence that the drug in question would be safe or hazardous to a specified degree should be based on the combined results of all tests in the program. Combining the results from programmatic research should follow scientific guidelines.

A performance test battery as a sound operationalization of the construct `driver fitness` should possess content validity. That means, the test battery as a whole should be representative of the mental and/or behavioral functions relevant to driving and, simultaneously, should be representative of the pharmacological effects of the drug under study.

Concerning the elements of the test battery, studies should include a test to measure divided attention or continous perceptual-motor coordination and a test to measure sedation or drowsiness. Furthermore, the test battery should comprise performance areas like discrete perceptual motor response, speed and accuracy of decisionmaking, sustained attention (vigilance), dynamic visual acuity, short-term spatial memory, risk avoidance and attentional resistance to distraction. This list is not closed.

The duration of the test battery will depend on, among other things, the kind and number of tests included, the medicinal drug, and the question of whether a time course effect should be studied. There should be at least one test in the battery of longer duration (e.g. one hour).

Many representative test procedures exist to test a single mental and/or behavioral function. The choice of an adequate test is left to the investigator who should choose a standard procedure or, if using a new one, the investigator should demonstrate that the procedure is able to discover the impairment addressed by the mental and/or behavioral function under study. In any case, the test selected must meet the requirements outlined in the following.

Performance tests included in a test battery should possess construct validity. That means that the measures of the tests should be simultaneously relevant to driving performance and sensitive to the pharmacological drug effects.

A validation of medicinal drug induced changes in performance with changes in actual driving performance should be made before using laboratory tests in categorization procedures or other legal/regulatory affaires. Laboratory tests should possess either a test-retest reliability coefficient for raw scores measured in the absence of drug effects of r>= .70 or, preferably, a test-retest reliability coefficient for drug-placebo change scores of r>= .50.

A standardization of the measurement procedure of the dependent variables of a test is very important due to reasons of comparability between studies. The training of the subjects on the test should be continued to individual stability of performance (determined by group means, standard deviation, and intertrial correlations).

The term ´simulator´ contains a wide range of constructions from very simple to highly sophisticated devices. Independent of the complexity, they should include tests of reactions on traffic signals, compliance with traffic control devices, passing maneuvers, and turning of intersections.

A similar state of affairs is valid for real driving. Independent of the variety of designing the procedure, closed-course driving should include decision-making, responses to changes in traffic control devices, and interactions with other experimentally controlled road-users. Over-the-road driving tests have to include combined city and highway driving tests. The fundamental aspects of the driver-vehicel-road interaction (e.g. road tracking, speed maintenance, car following, etc.) have to be involved. The ethical implications must be considered with great care.

Standard aspects of the statistical evaluation should be followed. In general, the statistical methods must be described in advance. The description should include remarks on how to proceed with missing values, drop-outs, and extreme values during the analysis. The researcher should employ the most powerful statistical test that can rightfully be applied.

Comparisons among treatments at baseline may reveal the existence of carry-over effects in cross-over studies or differences among groups in parallel group design, thus justifying the use of difference values with respect to the baseline in the statistical analysis.

The performance measures should be recorded on a common scale for all psychometric tests. Most of the statistical procedures available for achieving this depend on population distribution parameters. Hence, investigators should gather these data in a central repository so that all investigators may use the information.

In general, the statistical evaluation is based on several dependent variables which represent different aspects of driver fitness. Therefore, the question is: how to reduce the results (items, endpoints) of the many dependent variables to one decision on driver fitness.

It is obvious that if one is interested in several dependent variables and performs independent tests at a given significance level without significance level adjustment, the probability of an erroneous decision for effect exceeds the given level dramatically. Hence, if several dependent variables are tested seperately the a -inflation has to be taken into account. Investigators should indicate what measures they have taken to correct for a -inflation.

Instead of using the method of multiple comparisons, more use should be made of the newer techniques of `multiple endpoints` or `accumulation statistics` in which several dependent variables (that means several aspects of driver fitness) are synopsized. After having compared multiple dependent variables (summation of weighted z-scores; technique of `multiple endpoints`) or computed a score as weighted (weights defined à priori by the scientific community according to the importance of the skills to driver fitness, see chapter 8; technique of `accumulation statistics`), a t-test for equivalence hypothesis is performed (O`BRIAN (1984), POCOCK et al. (1987), LEHMACHER et al. (1991), HILGERS et al. (1998)).

The interpretation of the results should start with an à priori definition of a practically relevant change in performance (see chapter 2). An alteration in performance of at least one population standard deviation should be interpreted as a relevant change in performance. The interpretation should result in categorizing a drug as `presumed to be safe or unlikely to produce an effect`, `likely to produce minor or moderate effects`, or `likely to produce severe adverse effects or presumed to be potentially dangerous` according to the 3-tier warning system for identifying the driving hazard potential of a drug as part of the pan-European registration process accepted by the European Committee for Proprietary and Medicinal Products (CPMP).

7. Publication of Results

An exhaustive description of the material, the methods, and the results is most important due to reasons of scientific precision and due to the growing importance of meta-analyses. Appropriate guidelines are to be found in almost all scientific journals.

With respect to experimental studies on drugs and driver fitness, a precise description of inclusion and exclusion criteria for subjects (i.e., the diseases and addictions checked and the examinations used) and of the methods used to exclude performance influencing factors other than the treatment under study is necessary.

To avoid publishing bias, it is very important to report on all studies even if no cosistent results (i.e. in comparison with the results of other studies) are obtained.

In each study report, the design and the statistical analysis should be described by the a -risk (including the method to correct for alpha-inflation), the ß-risk, the effect size, the sample size, the number of outliers and drop-outs, and the decision criteria to select them, mean values, standard deviations, the analysis of variance table if used or other statistical results, and the statistical package. Reasons for the selection of statistical tests should be given, and it should be demonstrated that the assumptions to use a specific statistical test are fulfilled.

The data exchange should be improved by providing disks with raw data to other researchers on request.

The compendium provides a basis for establishing desirable criteria concerning the experimental methodology of research into the field of medicinal drugs and driver fitness. The guidelines arranged will form the basis for further exploration of how the scientific community looks upon these issues. This intention has many implications.

1. It was not to be expected that all participants of the Workshops and the Working Group would achieve total agreement on the propositions presented in this paper. Differences in opinion are largely attributable to the variety of methods used to investigate the effects of medicinal drugs on driver fitness.

2. The compendium is not to be regarded as complete and all-inclusive. Some investigators will judge that one or other proposition is self-evident whereas other researchers would like to have some propositions more detailed or more scientifically based.

3. Another aspect is the fact that the desirable may not always be achievable. Reasons may include, for example, the prescribed length of a project or the group of subjects investigated, which is limited in quantitative terms. Surely in such cases the investigator should carry out the research, but should indicate why it was impossible to follow the proposals.

4. Further research on experimental methods will be necessary. As a central issue, we would like to point out the development of a sound theory of driver fitness including the establishment of a requirement profile concerning the safe driving. Based on this theory, it should be possible to formulate detailed guidelines for simulator and on-the-road tests and to rank mental and/or behavioral functions according to their importance.

Final practical application of the guidelines, however, will require additional concerted action and development by the parties involved: researchers, psychopharmacological research organizations and institutes, the pharmaceutical industry, and regulatory authorities.

In conclusion, we would like to recall that these guidelines should not restrict research, but they should support a movement towards optimizing, harmonizing, and standardizing experimental procedures for assessing the adverse effects of medicinal drugs on driving performance.

BERGHAUS, G., FRIEDEL, B.: Quality assurance elements and standardization of empirical studies on the subject "drugs and traffic safety". J Traffic Med 20, 71 - 82 1992a

BERGHAUS, G., FRIEDEL, B.: Empirical man-machine interaction studies. Proposed flow-chart. In: FERRARA, S.D., GIORGETTI, R. (Eds): Methodology in man-machine interaction and epidemiology on drugs and traffic safety. Experiences and guidelines from an International Workshop. Addiction Research Foundation of Italy (ARFI)-Research Monograph Series Nr. 6, Padova, 1992b

O`BRIAN, P.C.: Procedures for comparing samples with multiple endpoints. Biometrics 40, 1079-1087, 1984

EC: Good clinical practice: guidelines for essential documents for the conduct of a clinical trial. - In: Note for Guidance. CPMP - Commission of the European Communities (1994). Biostatistical methodology in clinical trials in application for marketing authorizations for medicinal products. - In: Note for Guidance. CPMP - Commission of the European Communities, 1994

FERRARA S.D., GIORGETTI, R. (Eds): Methodology in man-machine interaction and epidemiology on drugs and traffic safety. Experiences and guidelines from an International Workshop. Addiction Research Foundation of Italy (ARFI)-Research Monograph Series Nr. 6, Padova, 1992

FDA: Guidelines for the format and content of the clinical and statistical sections of new drug applications. U.S. Department of Health and Human Services, Public Health Service, Food and Drug Administration, Center for Drug Evaluation and Research, Rockville Md, 1988

DE GIER, J.J., LAURELL, H: Studies on man-machine interaction. In: FERRARA S.D., GIORGETTI, R. (Eds): Methodology in man-machine interaction and epidemiology on drugs and traffic safety. Experiences and guidelines from an International Workshop (pp. 33-50), Addiction Research Foundation of Italy (ARFI)-Research Monograph Series Nr. 6, Padova, 1992

DE GIER, J.J.: Drugs other than alcohol and driving in the European Union. Study conducted with support of the European Road Safety Federation and the Directorate General for Transport of the Commission of the European Communities. Institute for Human Psychopharmacology (IHP), University of Limburg, Maastricht, 1995

HILGERS, R.D., FRIEDEL, B., BERGHAUS, G.: Äquivalenztestung im Rahmen experimenteller Untersuchungen zur Fahrtüchtigkeit. In Kongreßbericht 1997 der Deutschen Gesellschaft für Verkehrsmedizin e.V., pp. 130 to 134, Berichte der Bundesanstalt für Straßenwesen, Mensch und Sicherheit, Heft M92, 1998

International conference on harmonization of technical requirements for registration of phamaceuticals for human use (ICH): Structure and content of clinical study reports. Efficacy Topic E3, Oct. 27, Draft 10, 1994

International Council on Alcohol, Drugs and Traffic Safety (ICADTS): Guidelines on empirical studies undertaken to determine a medicinal drug`s effect on driving or skills related to driving - Detailed report - , 1999

Krüger, H.-P., Kohnen, R., Schöch, H. (Hrsg.): Medikamente im Straßenverkehr. Fischer Verlag, Stuttgart, 1995

Lehmacher, W., Wassmer, G., Reitmeir, P.: Procedures for two sample comparisons with multiple endpoints controlling the experimentwise error rate. Biometrics 47, 511- 521, 1991

Pocock, S.J., Geller, N.L., Tsiatis, A.A.: The analysis of multiple endpoints in clinical trials. Biometrics 43, 487-498, 1987

Vermeeren, A., De Gier, J.J., O`Hanlon, J.F.: Methodical guidelines for experimental research on medicinal drugs affecting driving performance - an international expert survey. Institute for Human Psychopharmacology (IHP), University of Limburg, Maastricht, 1993

Vermeeren, A., de Gier, J.J., O`Hanlon, J.F., Laurell, H.: An expert survey on methodology of experimental research in drugs and driving. In: Utzelmann, H.D., Berghaus, G., Kroj, G. (Eds): Alcohol, drugs and traffic safety - T92. pp. 795- 803, TÜV Rheinland, Cologne, 1993

Westlake, W.J.: Statistical aspects of comparative bioavailability trials. Biometrics 35, 273-280, 1979

Working Group T`92 ICADTS: Quality assurance and standardization concerning empirical trials on drugs and driving. Results of the workshop on the occasion of ICADTS - T`92. In: Utzelmann, H.D., Berghaus, G., Kroj, G. (Eds): Alcohol, drugs and traffic safety - T92. pp. 804-806, TÜV Rheinland, Cologne, 1993

WHO: Guidelines for good clinical research practice (GCP) for trials on pharmaceutical products, current draft. WHO Division of Drug Management and Policies, Geneva.

• Prof. Dr. F. J. Alvarez
  Dpto. Farmacologia y Terapeutica
  Facultad de Medicina
  47005 Valladolid
  SPAIN

• Prof. Dr. G. Berghaus
  Institut für Rechtsmedizin der Universität Köln
  Melatengürtel 60
  50823 Köln
  GERMANY

• Dr. M. Burns
  Department of Psychology and Department of Psychiatry and Behavioral Sciences
  University of California
  Los Angeles
  4138 Royal Crest Place
  Encino, CA 91436
  USA

• Prof. Dr. S.D. Ferrara
  Istituto Medicina Legale
  Via Falloppio 50
  35100 Padova
  ITALY

• Prof. Dr. B. Friedel
  Bundesanstalt für Straßenwesen
  Brüderstr. 53
  51427 Bergisch Gladbach
  GERMANY

• Dr. J.J. de Gier
  Vinkenslag 32
  4901 AP Oosterhout
  THE NETHERLANDS

• Dr. R. Giorgetti
  Istituto Medicina Legale
  Via Falloppio 50
  35100 Padova
  ITALY

• Dr. R.-D. Hilgers
  Institut für Medizinische Statistik, Informatik und Epidemiologie der Universität Köln
  Joseph-Stelzmann-Str. 9
  50931 Köln
  GERMANY

• Dr. R.S. Kennedy
  1040 Woodcock Road
  Suite 227
  Orlando, Florida 32803
  USA

• Prof. Dr. H.-P. Krüger
  Psychologisches Institut III der Universität Würzburg
  Röntgenring 11
  97070 Würzburg
  GERMANY

• Dr. H. Laurell
  Swedish Road Safety Office TSV
  78186 Borlänge
  SWEDEN

• Dr. Ch. Mercier-Guyon
  Traffic Medicine Research Center CERMT
  BP 132
  74004 Annecy, Cédéx
  FRANCE

• Prof. Dr. H. Moskowitz
  Department of Psychology and Department of Psychiatry and Behavioral Sciences
  University of California
  Los Angeles
  4138 Royal Crest Place
  Encino, CA 91436
  USA

• Dr. J. Perl
  Clinical Forensic Medicine Unit
  New South Wales Police Service
  151-241 Goulburn Street
  Surry Hills, NSW 2010
  AUSTRALIA

• Dr. G. Starmer
  Department of Pharmacology
  University of Sydney
  Sydney NSW 2006
  AUSTRALIA

• Dr. A.C. Stein
  Safety Research Associates, Inc.
  4739 La Canada Boulevard
  La Canada, CA 91011
  USA

• Drs. A. Vermeeren
  Institute Brain and Behavior
  Maastricht University
  P.O. Box 616
  6229 ET Maastricht
  THE NETHERLANDS

• Drs. E.R. Volkerts
  Netherlands Institute for Drugs and Doping Research (NIDDR)
  University of Utrecht
  P.O Box 80.082
  3508 TB Utrecht
  THE NETHERLANDS

• Dr. M. Vollrath
  Psychologisches Institut III der Universität Würzburg
  Röntgenring 11
  97070 Würzburg
  GERMANY

• Dr. P.F. Waller
  The University of Michigan Transportation Research Institute (UMTRI)
  2901 Baxter Road
  Ann Arbor, Michigan 48109-2150
  USA

We would like to indicate that the detailed version of the guidelines can be obtained from the ICADTS on request. Thanks to Kathleen Bartlett of Robert Kennedy`s office who corrected for language flaws and grammar.