Project No: TREN-05-FP6TR-S07.

61320-518404-DRUID

6th Framework Programme

Deliverable (0.1.8)

Final Report:
Work performed, main results
and recommendations
Revision 2.0

Actual submission date: (01.08.2012)

Start date of project: 15.10.2006

Duration: 5 years
Organisation name of lead contractor
for this deliverable: BASt

Project co-funded by the European Commission within the Sixth Framework Programme (2002-2006)

Dissemination Level

Public
PU x
Restricted to other programme participants (including the Commission Services)
PP
Restricted to a group specified by the consortium (including the Commission Services)
RE
CO Confidential, only for members of the consortium (including the Commission Services)
Final Report:
Work performed, main results and recommendations

Status: Public

Project Co-ordinator: Horst Schulze, BASt, Germany

Authors:
Horst Schulze, Markus Schumacher, Raschid Urmeew, Kerstin Auerbach, (all - Federal
Highway Research Institute, (BASt), Germany)

This report is based on contributions provided by DRUID Work Package Leaders:

Anja Knoche (BASt), Inger-Marie Bernhoft (Technical University of Denmark), Marjan
Hagenzieker (SWOV, The Netherlands), Javier Alvarez (University of Valladolid,
Spain), Monika Pilgerstorfer (KfV, Austria), Bojan Zlender (AVP, Slovenia), Han de Gier
(University of Groningen, The Netherlands)

Acknowledgments:
DRUID partner (see the list of partners pp. 6-7)

Project Funded by the European Commission under the Transport RTD Programme of the 6th Framework Program
Index

LIST OF ABBREVIATIONS AND DEFINITIONS 02

PUBLISHABLE EXECUTIVE SUMMARY 03
SECTION 1 – INTRODUCTION 05
SECTION 2 – WORK PERFORMED AND MAIN RESEARCH RESULTS 09
1. Problem situation (WP1 & WP2) 09
1.1. Epidemiology 10
1.2. Experimental studies and Meta-analyses 25
1.3. Synopsis of problem situation based on epidemiology 38
2. Countermeasures 40
2.1. Enforcement 40
2.2. Classification 47
2.3. Rehabilitation 54
2.4. Withdrawal 58
2.5. Dissemination 64
2.6. Legal perspectives 71

SECTION 3 – CONCLUSIONS 77
1. Alcohol 77
2. Illicit drugs 80
3. Psychoactive Medicines 85

SECTION 4 – RECOMMENDATIONS 88
1. Countermeasures to combat alcohol impaired driving 88
2. Countermeasures to combat illicit drug impaired driving 90
3. Countermeasures to combat driving impaired by medicines 92

SECTION 5 – DISSEMINATION AND USE 93

REFERENCES 94

Annex A1 – List of deliverables

Annex A2 – Final plan for using and disseminating the knowledge

1
List of abbreviations and definitions
AC Additional Cost Model
AGL Amendment Guidelines to FP6 contracts
ATC Anatomical Therapeutic Chemical
BAC Blood Alcohol Concentration
CA Consortium Agreement
CC Core Contract (Contract No TREN-05-FP6TR-S07.61320-518404-DRUID)
COOR DRUID Co-ordinator
CPHM Committee for Medicinal Products for Human Use
DG MOVE Directorate General Mobility and Transport
DR Driver Rehabilitation
DRET Driver Rehabilitation Evaluation Tool
DRUID Driving Under Influence of Drugs, Alcohol and Medicines
DUI Driving Under Influence
DUID Driving Under Influence of Drugs
EAB Ethical Advisory Board
EC European Commission
EMA European Medicines Agency
EU European Union
FC Full Cost Model
FCF Full Cost Model with flat rate
GFI Guidelines to Financial Issues
GRPF Guidelines for request for payments form
HCP Health Care Provider
HFI Handbook of Financial Issues
HS Hospital studies
ICADTS International Council on Alcohol, Drugs and Traffic Safety
ICT Information and Communication Technologies
IP Integrated project
OR Odds ratio
ORs Odds ratios
OECD Organisation of Economic Cooperation and Development
OF Oral Fluid
PIL Patient Information Leaflet
PL Package leaflet
PM Project Management
QA Quality Assurance
QAB Quality Assurance Board
QAP Quality Assurance Plan
QM Quality Management
QMS Quality Management System
QRD Quality Review of Documents
PhVWP EMA Pharmacovigilance Working Party
PT Proficiency Test
RH Rehabilitation
RRT Round Robin Test (same as PT)
RSS Road side surveys
SC Steering Committee
SDLP Standard Deviation of Lateral Position
SpC Special Clauses
SmPC Summary of Product Characteristics
TL Task Leader
WG Working Group
WI Work Instruction
WP Work Package
WPL Work Package Leader

2
Publishable Executive Summary

The DRUID Final Report presents main results of the project and evidence based
conclusions relevant for EU and Member States policy makers. The basis for
recommendations included in the Final Report is the content of 50 DRUID deliverables
(available under www.druid-project.eu) in which research outcomes of the seven
project Work Packages (WP) are described in detail.
The Final Report is divided into four main sections: (1) introduction, (2) main research
results, (3) conclusions and (4) recommendations.
Section 1 gives an overview of the project objectives, the DRUID consortium and the
architecture of project activities.
Section 2 contains the main research results of the efforts undertaken within seven
DRUID Work Packages. These results are divided in two groups:

1) A description of the problem situation (results of studies and experiments of the

WP1 and WP2) with regard to DUI/ DUID:
• The data on prevalence of psychoactive substances in the general driving
population that was generated within roadside surveys conducted in 13
European countries according to a uniform study design. For this purpose
samples of body fluids of approximately 50,000 randomly selected drivers have
been analyzed (WP2). The data on prevalence of psychoactive substances in
drivers in accidents that was generated from studies in hospitals in six
European countries and from data on killed drivers in four countries according
to a uniform study design. For this purpose samples of body fluids of
approximately 3,600 seriously injured drivers and 1300 killed drivers have been
analyzed (WP2)
• Risk estimates for driving under influence of psychoactive substances have
been derived from the case-control studies in which data of the roadside
surveys was compared to the data of approximately 4,500 drivers seriously
injured or killed in an accident (WP2)
• Characteristics of drivers tending to drive under the influence of psychoactive
substances (WP2)
• A description of the current state of research on the impact of alcohol, illicit
drugs and medicines on driving is given based on DRUID meta-analyses and
reviews (WP1)
• The results of driving tests conducted to close knowledge gaps on major illicit
drugs and medicines (WP1)

2) Information on countermeasures appropriate to combat driving under influence of

psychoactive substances (results of activities of the WP3 - WP7):
• Results of evaluation of oral fluid screening devices and checklists for
identifying clinical signs of impairment (WP3)
• Results of the cost-benefit analysis of increased drug-driving enforcement
through traffic police (WP3)
• A four level classification and labeling system for medicines regarding their
influence on driving performance (WP4)

3
 • The most comprehensive database on European rehabilitation schemes and
measures as well as on quality assurance measures for rehabilitation programs
(WP5)
• A compilation of practices of driving license withdrawal in European countries
and recommendations concerning best practice for withdrawal/ licensing
strategies (WP6)
• Guidelines for health care professionals on prescribing and dispensing
medicines taking into account their impact on driving performance (WP7)
• Recommendations on how to disseminate the DRUID results to different target
groups, i.a. general public, young drivers, patients, health care professionals,
policy makers (WP7)

Section 3 contains conclusions drawn from the outcomes of DRUID described in the
Section 2, including cross-references between different WPs.
Section 4 presents recommendations based on the scientific outcomes of the project.
As DRUID focuses on the effects of driving under the influence of alcohol, illicit drugs
and psychoactive medicines, recommendations are given with regard to each of these
three groups of substances. By describing a countermeasure, target groups that should
be addressed by this countermeasure are described as well. Further on,
recommendations are given concerning appropriate legal regulations, enforcement
strategies, rehabilitation measures and strategies for driver’s license withdrawal.
Finally, objectives for future research activities are recommended.

4
Section 1 - Introduction

In order to attain the EU target of significantly reducing fatalities in road traffic, it was
necessary to address risks associated with all components of the road transport system:
vehicles, roads, infrastructure, and driver performance.
With regard to the latter, an evidence based consensus exists that consumption of
psychoactive substances is one of the major factors affecting driving performance and
causing accidents. The scientific discussion on this topic was dominated before by the
problem of drunk driving. Whereas public concerns with regard to illegal drugs and medicines
in traffic were growing, our knowledge five years ago was insufficient to address these
concerns.
DRUID was conceived aiming to close this knowledge gap and was implemented as a
comprehensive and integrative effort, absorbing all research results achieved in the past and
conducting pioneering studies addressing new challenges. DRUID embraced all
psychoactive substances, alcohol as well as medicines and drugs. The intention was to
analyse all facets of the problem: consumption, impairing effects, accident risk, enforcement
and licence withdrawal strategies, rehabilitation and prevention. This approach was
substantially new as all prior research studies concentrated on particular problem areas
and/or on particular substance groups.
• Project objectives
The overall objective of DRUID was to provide scientific support to EU road safety policy
makers by making scientific based recommendations concerning combating driving under
the influence of psychoactive substances.
DRUID tried to differentiate between problems common for all psychoactive substances and
consumers and those typical for a group of substances (or for a single one) or for different
target groups. Consequently, measures to combat DUI/DUID that will be developed on the
basis of DRUID recommendations, could be either general or targeted, taking into account
characteristics of consumers and substances. Special attention was paid to patients
consuming medicines.
The first task that DRUID participants decided to implement was a comprehensive state-of-
the-art review of research results in the domain of DUI/DUID. It was foreseen to upgrade this
pre-existing knowledge with the results of DRUID own original studies and experiments.
Project partners planned to establish a brand new theoretical and methodological framework
to be used as a platform on which the pre-existing knowledge and results of manifold DRUID
studies should be integrated and evaluated.
Epidemiological and experimental studies in DRUID were designed aiming to win new
insights in the real scale and danger of consumption of psychoactive substances in traffic
across Europe and to try to assess accident risks caused by consumption of most prevalent
substances.
The project aimed at assessing possibilities to determine blood concentration thresholds for
various drugs analogically to alcohol blood concentration thresholds applied in European
countries to combat driving under the influence of alcohol. A major problem to be solved was
linking the impairment to the accident risk. Calculating accident risks from epidemiological
data is a straightforward approach with face validity. If epidemiological data is sufficient, this
approach enables researchers to calculate accident risks related to substance
concentrations in body fluids. In case the data is poor (as it is common for most illegal
substances and medicines), the impairment approach must be chosen.
One of the DRUID important tasks was practical and scientific evaluation of oral fluid
screening devices for drug detection used for enforcement purposes. The project consortium
intended to evaluate different enforcement strategies and to conduct cost-benefit analysis of

5
these strategies aiming to recommend procedures that would facilitate decision making
concerning drug enforcement policy.
An important objective of DRUID was to develop an empirically based classification and
categorisation system which allows a consistent labeling of medicines with respect to their
impact on driving performance and to enable physicians and pharmacists to use this system.
The approach is characterised by developing so called ‘generic’ manual for training health
care professionals, based on experiences in Belgium, Spain and the Netherlands.
Rehabilitation programmes for consumers of psychoactive substances were systematically
documented and evaluated, taking into account prior experiences. DRUID implemented the
most comprehensive study of drivers’ rehabilitation systems across Europe, evaluated them
and recommended “best practices”.
Accompanying the research on prevalence and risk, special emphasis was given to the legal
situation in the different European countries, their efforts of combating DUI/DUID and their
experience with countermeasures and prevention. It was intended to reveal “best practice”
procedures to be recommended to the Member States.
Health care professionals need to be informed about the potential risk associated with the
use of any kind of psychoactive substance (illicit drugs, medicines, alcohol). DRUID intended
to develop guidelines to make health care professionals aware of their role and to provide
them with relevant information. Further on, it was proposed to develop professional
standards to address the role of physicians and to involve European professional
organisations of physicians in this activity. Adequate training activities were supposed to be
designed and tested.
Information about medicines that are likely to affect driving performance, and thus entail a
higher risk of being involved in an accident, must be communicated to patients in a manner
that ensures the information is fully understood. With this purpose DRUID aimed to evaluate
methods of developing information leaflets, public campaigns and web-sites. An important
objective was to find ways to apply ICT using the DRUID knowledge in computerised
prescribing and dispensing systems.

• DRUID consortium
The DRUID consortium united the following 37 partners from 17 member States and Norway:

## Participant name Short Name Country

1 Bundesanstalt für Straßenwesen BASt DE

2 TÜV Rheinland Consulting GmbH TRC DE
3 Technical University of Denmark DTU DK
4 Centre for research and technology Hellas CERTH-HIT EL
5 National Institute for Transport and Safety Research IFSTTAR FR
6 Université de Caen – Basse Normandie UNICAEN FR
7 Motor Transport Institute ITS PL
8 Institute of Forensic Research IES PL
9 Universiteit Gent UGent BE
10 SWOV Institute for Road Safety Research SWOV NL
11 KLPD KLPD NL
12 Maastricht University, Faculty of Psychology UMaas NL

6
 ## Participant name Short Name Country

13 University of Groningen, Pharmacy RUGPha NL

14 University of Groningen, Psychology RUGPsy NL
15 Universidad de Valladolid UVa ES
Netherlands Organisation for Applied Scientific Research
16 TNO NL
TNO
17 Statens Väg- och Transportforskningsinstitut VTI SE
18 Centrum dopravního výzkumu CDV CZ
19 Centre Regional de Pharmacovigilance UGren FR
20 Bayerische Julius-Maximilians-Universitaet Wuerzburg UWUERZ DE
21 Kuratorium für Verkehrssicherheit KfV AT
22 Jefatura central de trafico DGT ES
23 Società Italiana di Psicologia della Sicurezza Viaria SIPSiVi IT
24 Institute of Transport Economics TOI NO
25 University of Turku U. Turku FI
26 Norwegian Institute of Public Health FHI NO
1
27 Direkcija Republike Slovenije za ceste DRSC SI
28 National Institute for Health and Welfare THL FI
29 Institut Belge pour la Sécurité Routière asbl IBSR BE
30 Ludwig-Maximilians-Universitaet Muenchen LMU DE
31 Universtaetsklinikum Heidelberg UKL-HD DE
32 University of Copenhagen UKBH DK
33 Institut für Therapieforschung IFT DE
34 University of Szeged USZ HU
U.O.C. Tossicologia Forense e Antidoping – Azienda
35 TFA-UNPD IT
Ospedaliera-Universita di Padova
Centre of Post-Graduated Studies in Legal Medicine of the
36 CPS-NILM PT
National Institute of Legal Medicine of Portugal
37 Institute of Forensic Medicine, Mykolas Romeris University TMI LT
38 Javna agencija Republike Slovenije za varnost prometa AVP SI

1
As of 31.08.2010 Javna agencija Republike Slovenije za varnost prometa (AVP) took over rights and
responsibilities of the DRSC. For administrative reasons, DRSC stayed an official partner following Commission’s
request. Thus, AVP became the 38s DRUID partner.

7
• DRUID implementation structure
DRUID was structured objective oriented and aiming to address the following requests
re of the
European Commission:

Work Work package content

EC requirements
package description
To enable policy makers to refer to a substance
Methodology and Experimental
blood concentration threshold defined for driving a WP 1
Research
power-driven vehicle
Too deliver reference studies of the impact on fitness Epidemiological Studies,
WP 2
to drive for alcohol, illicit drugs and medicines Relative Risk Calculation
To evaluate mobile drug detection devices and to Enforcement: Methods and
implement cost-benefit
benefit analysis of enforcement WP 3 Devices, Enforceable
strategies Legislation
To introduce
roduce classification and label
labeling system for
Developing a Classification
medicines with regard to their influence on driving WP 4
System for Medicinal Drugs
performance
To provide
rovide authorities with recommendations
concerning effective drivers rehabilitation schemes, WP 5 Rehabilitation – Good Practice
adapted to individual driver’s situation
To recommend strategies of driving bans, which are Withdrawal – Existing
compatible with the road safety objectives
o and at the WP 6 Practices and
same time respect the need for mobility Recommendations
To define responsibility of health care professionals
vis-à-vis
vis dangerous patients consuming psychoactive
Dissemination and Guidelines,
substances and the role they can
ca play with regard to WP 7
Training Measures
road safety. To develop information and
dissemination instruments for different target groups

The diagram below shows the input-output

input output and cooperation links between Work Packages:

8
Section 2 – Work performed and main research results

The DRUID project was an integrative effort to reduce the danger of alcohol, illicit drugs and
medicines in traffic. It investigated which problems and countermeasures are common for
psychoactive substances as a whole and which problems must be tackled differently with
respect to the characteristics of consumers and substances.
In the following section the work performed within five years is summed up. The main results
are described. Section 2 is divided into two parts: Part 1 describes the problem situation,
which mainly refers to the results of the epidemiological (WP 2) and experimental research
(WP 1); part 2 presents countermeasures with respect to enforcement (WP 3), classification
of medicines (WP 4), rehabilitation of offenders (WP 5), withdrawal of driving license (WP 6),
and dissemination of information to different target groups (WP 7).

1. Problem situation (WP1 & WP2)

The consumption of alcohol, illicit drugs and medicines influences cognitive and motor skills
relevant for safe driving and thereby can alter fitness to drive. It was the objective of DRUID
to collect all available data on the effects of the different substances on fitness to drive. Data
were gathered in epidemiological and experimental studies.
WP1 and WP2 aimed at assessing the situation regarding the prevalence and accident risk
of the use of alcohol, illicit drugs, and (psychoactive) medicines of drivers in Europe.
Therefore in WP2, two of the major attempts to obtain these insights were the DRUID
roadside surveys (D2.2.3) and the DRUID hospital studies on seriously injured and killed
drivers (D2.2.5). Of both data on prevalence are available. In order to get risk estimations,
both data sources were brought together in the DRUID case-control study (D2.3.5). Based
on the prevalence of psychoactive substances in the driving population and of the prevalence
of these substances in the ed killed and injured accident victims, risk estimates “odds ratios”
of getting seriously injured or killed by driving with psychoactive substances were derived.
Results of psychoactive substances already available from published studies were
summarized in meta-analyses. This was done for alcohol (D1.1.2a), for major illicit drugs and
medicines (D1.1.2b; D1.1.2c) and for opioids used in substitution treatment (D1.1.2c). In
addition driving tests were conducted on several illicit drugs (D 1.2.1) and medicines (D1.2.2)
in real traffic, in driving simulators and on closed circuits.
Figure 1 gives an overview about the different sources of data used in DRUID in order to
derive risk estimations: epidemiological data, experimental data, and meta-analyses.

Figure 1: Available sources of data in DRUID (D1.3.1, p. 24).

9
A methodological framework was developed in DRUID that allowed integration of these
different sources of data. It is described in detail in D1.1.1. Its main features are:

1) Impairment by alcohol is used as reference. Therefore the current state of research on

the effects of alcohol on driving and skills related to driving was summarized in a meta-
analysis. In addition in all driving tests an alcohol comparison was done in order to be
able to compare all substance effects to those caused by 0.5g/L BAC.
2) Odds-ratios are used as common risk measures. Epidemiological studies provide odds
ratios as risk measures. All results of experimental studies as well as the results of the
meta-analyses were transferred to ORs, too (see D1.1.1 for more details). By doing so
ORs could be used as common reference for the comparison of the results of different
types of studies.
3) Standardization of procedures. In order to be able to integrate the results of studies
conducted by different partners, they have to be conducted as similar as possible.
Therefore mandatory protocols were developed for the experimental studies (D1.1.1) as
well as for the epidemiological studies (D2.1.2). All meta-analyses were also conducted
according to a comparable methodology to allow the integration of results.

1.1. Epidemiology
Before DRUID only a few surveys have been carried out in Europe as well as in Australia
regarding the prevalence of psychoactive substances in the driving population. These studies
were based on saliva samples and indicated similar results for drivers of passenger cars.
About 1 % took illicit drugs, primarily cannabis/stimulants. About 4-6 % took licit drugs,
primarily stimulants, hypnotics or anxiolytics, or medicines without impairing effects. Recent
studies have been carried out in Denmark, Norway, the Netherlands and the United
Kingdom, the latter three studies were part of the project IMMORTAL of the 5th Framework
Program.

1.1.1. Objectives
The major aim of WP 2 was to assess the increased risk for drivers being involved in a traffic
accident after consumption of various psychoactive substances including alcohol. These
results were mainly obtained by means of case-control studies. Thus, the results reflect both
the most common psychoactive substances in the driving population (prevalence studies)
and the accident risks (case-control studies) while impaired by alcohol and other
psychoactive substances and/or various combinations. The following objectives were
defined:
• Prevalence of alcohol and other psychoactive substances
o in the general population
o in drivers in traffic in general (road side surveys; RSS), including differences in the
patterns of psychoactive substance use between EU countries.
o in drivers who have been injured/killed in traffic accidents (hospital studies; HS)
• The relative risk
o for a car driver of getting seriously injured or killed in a road accident while
positive for alcohol and other psychoactive substances
o for a responsible car driver of getting involved in a fatal car accident while positive
for alcohol and other psychoactive substances
o for patients using psychotropic medicines of getting involved in an accident

10
• Characteristics
o of drivers impaired by alcohol and other psychoactive substances in the general
driving population
o of accident involved drivers
o of general drug users
o of drink and drug drivers

1.1.2. Methodology
In total 25 partners from 15 countries took part in the epidemiological work package, as shown
in Figure 2.

Figure 2: Countries participating

pating in the epidemiological work package.
package

1.1.2.1. Methodology of the studies on prevalence of alcohol and/or

other psycho-active
psycho substances
Prevalence of alcohol and/or other psycho-active
psycho substances was surveyed in the following
populations:
a) The general population
Information on the prevalence of consumption of some frequently used medicines with
effects on the central nervous system (i.e. opioids, antipsychotics, anxiolytics, hypnotics and
sedatives, antidepressants, drugs used in addictive disorders, and antihistamines
a for
systemic use) in a non-hospitalised
hospitalised EU population was collected over the years 2000 to
2005. In the same way, information on the use of illicit drugs in Europe (i.e. cannabis,
amphetamines, ecstasy, LSD, cocaine and crack cocaine, and opioids) opioids) was collected of
retrospective data from the years 1994 until
u 2006 (D2.1.1).

11
b) The general population of car drivers
The prevalence of alcohol and other drugs in the driving population was assessed in thirteen
European countries (Belgium, Hungary, Poland, Czech Republic, Italy, Portugal, Denmark,
Lithuania, Sweden, Spain, the Netherlands, Finland, Norway) based on road side surveys
(D2.2.3). The roadside surveys (RSS) were conducted according to a general design between
January 2007 and July 2009.
Participants, i.e. drivers of passenger cars and vans, were randomly selected using a stratified
multistage sampling design. In the first stage, one or more regions per country were selected.
These regions were meant to be representative for the country with regard to substance use and
traffic distribution. Within the selected regions smaller research areas were selected, and within
these areas, survey locations were selected, where subjects were stopped at random, and were
requested to participate in the study.
Due to voluntary participation non-response and non-response bias are common problems in
epidemiological studies. Non-response bias occurs in the case that not responding people differ
from those who do respond with regard to drug and/or alcohol use. Since only in Italy the drug
tests were mandatory researchers in other countries had to cope with non-response rates. If
drivers under influence would be more likely to refuse participation, results of the roadside
surveys would underestimate the prevalence rate of psychoactive substances. Underestimation
of the prevalence rate among controls will result in overestimation of the risk associated with
psychoactive substance use. In order to exclude a selective non-response bias the response and
the non-response group were compared for other variables in order to show their comparability.
With regard to days of the week and times of the day, the study population sample was stratified
into eight time periods over the week, for each of the survey areas. The time periods did not
overlap and covered all the days of the week and all times of the day. The distribution of the
study population sample by time periods was not proportionate to the distribution of the general
driving population over these periods. This was unavoidable since in many of the thirteen
countries the researchers had to take into account the preferences of the police. Police was
needed to stop the drivers from moving traffic. Weight factors were applied to correct for this
disproportion based on the ratio by time period between the distribution of traffic and the
distribution of the participants.
All countries have used a StatSure Saliva Sampler device for saliva collection, except for the
Netherlands, where saliva was collected by means of ordinary spit cups. Blood samples were
collected in Belgium, Italy, the Netherlands and Lithuania. All four countries used glass tubes for
the collection containing sodium fluoride and potassium oxalate.
Extraction of the substances was based on liquid-liquid (LLE) or solid phase (SPE),
chromatographic separation was performed by gas chromatography (GC) or Liquid
chromatography (LC), detection was done by mass spectrometry.
In total 23 substances were included in the “core substance list”. For each substance an
analytical cut-off was selected based on the lowest limit of quantitation (LOQ) that could be
measured by all toxicological laboratories that were involved in the analysis of the substances.
LOQ's reflect the lowest concentrations for substances at which quantitative results can be
reported with a high degree of confidence. For the final results presented in this report,
equivalent cut-offs, and not the LOQ's, have been used for analysis of the core substances to
correct for differences in concentrations of substances in blood and in saliva. Blood was collected
in some of the countries (Italy and Lithuania), both blood and/or saliva in two countries (Belgium
and The Netherlands), and saliva in the remaining nine countries. Equivalent concentrations for
blood and saliva were developed within the DRUID project to be used for the decision on
whether a sample was positive for a substance or considered negative (D1.4.2). This means that
concentrations of both blood and saliva could be included into the calculations. These equivalent
concentrations have not been known before and therefore were developed in DRUID WP2
(Table 1). Equivalent concentrations are an important finding, that solved the problem of two
different specimen being collected in the road side surveys.

12
Table 1: Recommended equivalent cut-offs for DRUID core substances.

Equivalent cut-offs

Substance Whole blood (ng/mL) Oral fluid/saliva (ng/mL)

Ethanol 0.1 (g/L) 0.082 (g/L)
6-AM 10 16
Alprazolam 10 3.5
Amphetamine 20 360
Benzoylecgonine 50 95
Clonazepam 10 1.7
Cocaine 10 170
Codeine 10 94
Diazepam 140 5.0
Flunitrazepam 5.3 1.0
Lorazepam 10 1.1
MDA 20 220
MDEA 20 270
MDMA 20 270
Methadone 10 22
Methamphetamine 20 410
Morphine 10 95
Nordiazepam 20 1.1
Oxazepam 50 13
THC 1.0 27
Zolpidem 37 10
Zopiclone 10 25
Tramadol 50 480
7-amino-clonazepam 10 3.1
7-amino-flunitrazepam 8.5 1.0

In total more than 50,000 drivers of passenger cars and vans from the driving population in
the participating countries gave a saliva sample, a blood sample or both samples.
c) Seriously injured and killed car drivers
A cross-sectional survey was conducted to determine the prevalence of alcohol and other
drugs in injured (sampled between October 2007 and May 2010) and killed (sampled
between January 2006 and December 2009) drivers in 9 European countries. Studies in
hospitals of seriously injured car drivers were conducted in six countries (Denmark, Finland,
Lithuania, Italy, Belgium, and The Netherlands); studies of killed car drivers took place in four
countries (Finland, Norway, Sweden, and Portugal). In order to be able to compare the
different studies a uniform design was developed for all participating countries. Obligatory
inclusion criteria were: Driver of a motorized vehicle, injured in an accident on a public road
or in the direct vicinity of a public road, only primary admissions to the hospital (no referrals),
because of traumatological reasons with a time interval between the accident and sampling
of less than 3 hours and an injury severity being MAIS 2 or higher. Each country could
decide upon additional national criteria.

13
Drug concentration was analyzed from blood samples. Extraction was based on liquid-liquid
(LLE) or solid phase (SPE) extraction, chromatographic separation was performed by gas
chromatography (GC) or liquid chromatography (LC): High Performance (HPLC) or Ultra
Performance (UPLC). Detection was done by mass spectrometry (MS) or nitrogen/phosphorus
detection (NPD). Commonly defined DRUID cut-offs were used to define positivity (Table 2).

Table 2: Whole blood cut-offs for DRUID core substances.

Substance Whole blood (ng/mL)

Ethanol 0.1 (g/L)

6-AM 10
Alprazolam 10
Amphetamine 20
Benzoylecgonine 50
Clonazepam 10
Cocaine 10
Codeine 10
Diazepam 20
Flunitrazepam 2
Lorazepam 10
MDA 20
MDEA 20
MDMA 20
Methadone 10
Methamphetamine 20
Morphine 10
Nordiazepam 20
Oxazepam 50
THC 1
THCCOOH 5
Zolpidem 20
Zopiclone 10
Tramadol 50
7-amino-clonazepam 10
7-amino-flunitrazepam 2

Substance concentrations are depending on the pharmacokinetics of the drugs. Therefore

the time lag between e.g. the accident and blood sampling has high impact on substance
concentration in the sample. In the injured drivers, the maximum delay between the accident
and blood sampling was three hours (median = 1.17 hours). The maximum time interval
between accident and taking of the blood sample in killed drivers was 24 hours. Since
substance concentration diminishes over time the measured concentrations will be lower
than the concentration at the time of the accident or at the stop.
Because of the very large interindividual variation in drug metabolism, no back-calculation of
the concentration of the drug for the time of the accident was done. In killed drivers, at the
time of death, metabolism slows down or stops. But post-mortem redistribution can occur. In
order to minimize the risk of post-mortem redistribution, blood was taken from peripheral
sites like the femoral vein. This was done if possible by all the centers that participated in the
killed driver study.

14
Fluoride was added to the tubes in order to slow down the degradation of instable drugs like
cocaine. A total of 3570 seriously injured drivers and 1293 killed drivers were sampled in this
study.2

1.1.2.2. Methodology of the studies on accident risk for driving with

alcohol and/or psychoactive substances
Relative risk is defined as the ratio of two risks, the risk of an event occurring in the group of
exposed subjects and the risk of the event occurring in the group of non-exposed subjects.
The relative risk estimates were approximated to odds ratios, and calculated by means of
logistic regression.
Within DRUID, the accident risk while driving under influence of alcohol and other drugs was
surveyed in the following populations:

a) The relative risk for a car driver of being seriously injured or killed in a road accident while
positive for alcohol and other drugs
The relative risk for a driver of getting seriously injured in an accident while positive for a
given substance was approximated to the odds ratio between the odds for a driver of getting
seriously injured in an accident while positive for a given substance and the odds of getting
seriously injured while negative (D2.3.5). The data from the hospital studies of seriously
injured drivers were used as cases. The data from the road side surveys were used as
controls (D2.3.5).
As the information whether a subject was exposed to a substance or not came from
toxicological analyses of samples from both blood and saliva, it was crucial for this study that
equivalent cut-offs for blood and saliva were developed. The risk estimates have been based
on blood samples of the injured or killed drivers and on blood and saliva samples collected of
the controls in the RSS. Samples were considered positive if the concentration was at or
above the equivalent cut-off either in blood or saliva (Table 1). Six countries contributed to
the study on the relative risk for getting seriously injured: Denmark, Finland, Lithuania, Italy,
Belgium and the Netherlands.
The relative risk as approximated to the odds ratio between the odds for a driver of getting
killed in an accident while positive for a given substance and the odds for a driver of getting
killed in an accident while negative was calculated by means of logistic regression for the
same substance groups as were used in the study of relative risk for seriously injured drivers.
Like for the study on the relative risk for drivers of getting seriously injured, the relative risk
estimates for killed drivers were based on blood samples collected of killed drivers and on
blood and saliva samples collected of the controls in the RSS. Samples were considered
positive if the concentration was at or above the equivalent cut-off either in blood or saliva.
Four countries contributed to the study on the relative risk of getting killed in an accident:
Finland, Norway, Portugal and Sweden.

b) The relative risk of responsibility for a fatal accident while positive for alcohol and other
drugs
Relative risk estimates of impaired car drivers involved in fatal accidents were calculated
based on a responsibility study in France (D2.3.2). Blood samples from car drivers involved
in fatal accidents in the period October 2001 – September 2003, whether killed, injured or
non-injured were confirmation analyzed for alcohol and illicit drugs. In total, 7455 car drivers

2
See D2.2.4 for more information

15
were included in the study. The DRUID cut-offs (Table 2) were used as indication for a
sample of being positive for a drug. The reference group was car drivers with alcohol
concentration below 0.1 g/L. The relative risk estimates were approximated to odds ratios.
The relative risk explains the difference in risk of responsibility between sober car drivers
(relative risk for the reference group=1) and responsible car drivers who were positive for a
substance. The odds ratios were adjusted for age and gender. Similar studies were carried
out in other countries (D2.3.3 and D2.3.4)
c) The relative risk for patients using psychotropic medicines of being involved in an accident
The aim of this study was to assess the association between the risk of a traffic accident and
the exposure to psychotropic medication by means of a case-control study, a so called
pharmaco-epidemiological study. The study was performed in the Netherlands and was
based on the linkage of databases regarding pharmacy prescriptions, police registered traffic
accidents, hospital records and driving licence data (D2.3.1). The case population was
defined as adults, who had a traffic accident between 2000 and 2007 and were driving, and
received medical assistance. The control population was defined as adults, who had a
driving license and had no traffic accident during the study period. In total, 3963 cases and
18828 controls were selected for the case-control analysis.

1.1.2.3. Methodology of the studies on characteristics of drink and

drug impaired drivers
Characteristics of drink and drug impaired drivers were investigated in the following
populations:
a) Characteristics of drivers in the general driving population
Results from the road side surveys (D2.2.3) showed for which age groups and gender of
drivers various substances were most prevalent. In addition to this, the study includes
information on the prevalence by time of the day and week.
b) Characteristics of accident involved drivers
Results from the hospital studies of seriously injured drivers and the studies of killed drivers
(D2.2.5) showed for which age groups, gender and time of the day and week various
substances were most prevalent in injured drivers.
c) Characteristics of general drug users
The main aim of the study that was carried out in Germany (D2.2.2) was to estimate the
prevalence of drugs in drivers in traffic. Instead of detecting drugs in the driving population –
like road-side surveys do – 200 illegal drug users and 100 matched control persons (non-
users of drugs) were queried for four weeks about their driving and drug consumption
behavior by a questionnaire deployed on smart-phones. The questionnaire was filled in daily
for 28 days by the persons involved in the study.
The data about drug use and driving not only assessed the frequency of drug driving and
other information on the trips, so as time, day, distance and passengers, but also data about
situations that led to refraining from driving under the influence of drugs. Finally, the study
included information on person-related characteristics, so as socio-demographic variables,
previous experience and attitudes. Therefore, also individual factors associated with drug
driving could be revealed.
d) Characteristics of drink and drug drivers
Qualitative interviews on motives to drink and drug driving were carried out in Sweden and
Hungary (D2.2.1) based on a uniform interview guide. But due to very limited possibilities of
recruiting drug drivers in Hungary, the results on drug driving are limited to Sweden.

16
1.1.3. Results
The presented results refer to the prevalence, the relative risk and the characteristics of drink
and drug impaired drivers. Only major results are described.

1.1.3.1. Results of the studies on prevalence of alcohol and/or other

psychoactive substances
Prevalence studies were addressed to different populations:
a) The general population
In the general population an increase in the use of medicinal psychotropic drugs and drugs
with central nervous system side-effects could be observed. This increase is in line with the
results of former studies. The major increase was seen in the consumption of
antidepressants and drugs used in addictive disorders. For the other classes of interest
either a slight increase or no increase was noted. Illicit drugs are most prevalent among the
population in the Southern European countries whereas medicines are most prevalent in the
Nordic countries. These data serve as background information for a better understanding of
the European problem of illicit drug use and, subsequently, for comparison with the
prevalence of illicit and medicinal drug use in the driving population.
b) The general population of car drivers
The prevalence of alcohol and other drugs in the driving population was calculated in thirteen
European countries based on roadside surveys. Highest prevalence in general was found for
alcohol, with highest prevalence in the southern countries of Europe (Spain, Italy, Portugal).
There were big differences between the prevalence in the various countries.
The same tendency for the prevalence in the driving population as for the prevalence in the
general population was observed: Higher prevalence of medicinal drugs in the Nordic
countries (Denmark, Norway, Sweden, Finland) and higher prevalence of illicit drugs in the
southern countries of Europe. However, regarding the prevalence of medicine in the
Northern Europe, there were differences in the prevalence between the four countries for the
various types of medicines.
In Eastern Europe (Czech Republic, Hungary, Poland, Lithuania) the prevalence of both
alcohol, illicit as well as medicinal drugs was relatively low compared to the other European
regions whereas drug use in Western Europe (Belgium, the Netherlands) was more or less
on the European average. Combined use of alcohol and drugs and multiple drug use was
more common in the southern countries of Europe.
For illicit drugs cannabis (THC) was the most frequently detected drug in traffic, followed by
cocaine. Amphetamines and illicit opiates were less frequently detected. Illicit drugs were in
general mainly detected among young male drivers, during all times of the day but mainly
during the weekend.
Medicinal drugs were in general mainly detected among older female drivers during daytime
hours. Benzodiazepines were the most prevalent medicinal drug in traffic, z-drugs were less
prevalent. However, considerable differences between countries were observed.
On a European level alcohol was estimated to be found in 3.48% of the drivers, illicit drugs
by 1.90% of the drivers, medicinal drugs in 1.36% of the drivers, drug-drug combinations in
0.39% of the drivers and alcohol-drug combinations in 0.37% of the drivers. However there
were big differences between the means in the four European regions (North, South, West,
East). There were high prevalence rates of alcohol, cocaine, cannabis and combined use in
Southern Europe, partly also in Western Europe, whereas z-drugs and medicinal opioids
were more common, although still low prevalent, in the northern countries.

17
c) Seriously injured and killed car drivers
The prevalence of alcohol alone was between app. 15 and 30% except for Portugal (40%).
Alcohol in combination with other drugs was found in 12% of the samples in Belgium down to
about 2% in Lithuania. Furthermore, the following remarks characterize alcohol use: For
seriously injured drivers, the highest percentage of positive drivers was found in Belgium
(alcohol alone in app. 30% and alcohol combined with other drugs in app. 12.5% of the
drivers). For killed drivers, the highest percentage of positive drivers was found in Portugal
(alcohol alone in app. 40% and alcohol combined with other drugs in app. 6% of the drivers).
Among the positive drivers – both seriously injured and killed, the majority had a blood
alcohol concentration equal to or above 0.5g/L.
The prevalence of illicit drugs varied between the countries with considerable combined use
of various substances. Amphetamine use appeared to be more common in Northern Europe,
both for seriously injured and killed drivers. In Portugal, no killed drivers were positive for
amphetamines. Cocaine use seemed to be more prevalent in Southern Europe, except for
killed drivers in Sweden. In Finland neither any seriously nor killed drivers were positive for
cocaine. The percentages of cannabis (THC and THCCOOH) positive drivers varied, with the
highest percentage in seriously injured drivers in Belgium (app 10% alone and in
combination with other drugs) and the lowest in Lithuania (below 1% alone and in
combination with other drugs). For killed drivers the highest percentage was found in Norway
(app 6% alone and in combination with other drugs) and the lowest in Finland (app 1.3% but
only found combination with other drugs). No illicit opiates were found in the killed drivers.
Regarding medicines a few countries had outstanding high prevalence, that is for
benzodiazepines alone in seriously injured drivers in Finland (app. 10%) and for medicinal
opioids alone in Lithuania in seriously injured drivers (close to 6%). Furthermore, the
following remarks characterize medicinal drug use: Benzodiazepine use appeared to be
more common in Northern Europe, both for seriously injured and killed drivers, with a
maximum for Finland both for seriously injured drivers (app. 10%) and for killed drivers (app.
5%). In the Netherlands no seriously drivers were positive for benzodiazepines. Z-drug use
was mostly found in Northern Europe. No positive findings for Z-drugs were recorded in Italy,
Lithuania and Portugal. Medicinal opioids were found in all countries, with a maximum for
seriously injured drivers in Lithuania (app 6% alone and 2% in combination with other drugs)
and a minimum in the Netherlands (app 0.5%, only found alone). In Lithuania the percentage
of seriously injured drivers who were positive for medicinal opioids was twice as big as in the
other five countries. Similar, in Sweden the percentage of killed drivers who were positive for
medicinal opioids was twice as big as in the other three countries.

1.1.3.2. Results of the studies on accident risk for driving under the
influence of alcohol and/or psychoactive substances
The results of the studies on accident risk for driving under the influence of alcohol and/or
psychoactive substances refer to three different types of risk estimates. The relative risk
estimates were adjusted by age and gender (when there was enough data); if there was not
enough data the crude odds ratio were calculated.

a) The relative risk for a car driver of being seriously injured or killed in a road accident while
positive for alcohol and other drugs (D2.3.5)
The data from the hospital studies of seriously injured/killed drivers were used and as cases.
The data from the road side surveys were used as controls (D.2.3.5). The relative risk for
alcohol varied considerably between countries. The relative risk of getting seriously injured
was not significantly different from 1 for alcohol concentrations between 0.1-0.49g/L. For
killed drivers, the relative risk was already from 0.1g/L significantly above 1. All estimates in

18
this section are based on data from all countries using the results of the analyses with
adjusted odds ratios. The estimate based on data from all countries indicated a risk of getting
seriously injured of 1.18 (CI: 0.81-1.73) and the risk of getting killed in an accident of 8.01
(CI: 5.22-12.29). For alcohol concentrations in the interval 0.5-0.79g/L, the relative risk of
getting seriously injured was significantly increased for some of the countries (Denmark,
Finland and the Netherlands) and for killed drivers in Norway and Portugal. The estimate
based on data from all countries indicated a risk of getting seriously injured of 3.64 (CI: 2.31-
5.72) and a risk of getting killed of 45.93 (CI: 23.02-91.66). For alcohol concentrations at
0.8g/L and above all risk estimates except for Italy were significantly above 1. For alcohol
concentrations between 0.8-1.19g/L, the estimate based on data from all countries indicated
a risk of getting seriously injured of 13.35 (CI: 8.15-21.88) and a risk of getting killed in an
accident of 35.69 (CI: 15.68-81.22). For alcohol concentrations from 1.2g/L and above, the
estimate based on data from all countries indicated a risk of getting seriously injured of 62.79
(CI:44.51-88.58) and a risk of getting killed of 500.04 (CI: 238.07-inf).
Amphetamines alone were too rare in some of the countries to enable the calculation of the
relative risk of getting seriously injured or killed. However the relative risk of death in an road
accident was considerable increased for Finland, Norway and Sweden. The estimates based
on data from all countries indicated a risk of getting seriously injured of 8.35 (CI: 3.91-17.83)
and a risk of killed drivers of 24.09 (CI: 9.72-59.71). The variations in the risk estimates
reflect sparse data with positive concentrations. The results should therefore be handled with
care.
For cocaine alone there were only few positive samples, and the risk estimates for the single
countries varied to a high degree without being significantly above 1. On the contrary, this
was the case when calculating the risk based on all countries with the estimate indicating a
risk of getting seriously injured of 3.30 (CI: 1.40-7.79) and of getting killed of 22.34 (CI: 3.66-
136.53).
The relative risk estimates for cannabis (THC), based on single countries were only
significantly increased for some of the countries but varied between countries to a high
degree. However, based on data from all countries the relative risk of getting seriously
injured and killed while positive for cannabis were not significantly above 1 with a risk of
getting seriously injured of 1.38 (CI: 0.88-2.17) and the risk of getting killed of 1.33 (CI: 0.48-
3.67). But this result should be handled with care because of the very different single country
estimates.
Positive samples were few for illicit opiates, and therefore the risk estimates should be
handled with care. Risk estimates significantly above 1 for the single countries were only
found for fatal injured drivers. The estimate based on data of all countries indicated a
significantly increased risk of killed drivers (crude odds ratio) of 10.04 (CI: 2.04-49.32). The
estimate based on data from all countries of getting seriously injured of 2.47 (CI: 0.50-12.10)
was not significantly increased. The variations in the risk estimates reflect sparse data with
positive concentrations.
The estimates of the relative risk of getting seriously injured or killed while positive for
medicines were not very different between most of the participating countries. However, the
confidence intervals were very large for some of the estimates.
Some of the risk estimates based on calculations of data from single countries were
significantly above 1. However, based on aggregated data of all countries, both the relative
risk of getting seriously injured and the relative risk of getting killed were significantly above 1
for benzodiazepines, Z-drugs and for medicinal opioids. The estimate for benzodiazepines
and Z-drugs of getting seriously injured was 1.99 (CI: 1.36-2.91) and of getting killed 5.40
(CI: 3.90-7.46) based on data from all countries. The estimate for medicinal opioids of getting
seriously injured was 9.06 (CI: 6.40-12.83) and of getting killed 4.82 (CI: 2.60-8.93) based on
data from all countries.

19
The estimates of the relative risk of getting seriously injured or getting killed while positive for
combinations of alcohol and/or drugs were considerably increased in nearly all countries.
Based on aggregated data of all countries, the relative risk of getting seriously injured or
killed while positive for alcohol and drugs was substantially increased. The same was the
case for multiple drug use. The estimate of getting seriously injured while positive for alcohol
and drugs was 28.82 (CI: 18.41-45.11) and of getting killed 31.52 (CI: 16.83-59.05) based on
data from all countries. The aggregated estimate of getting seriously injured while positive for
multiple drugs was 8.01 (CI: 5.34-12.01) and of getting killed was 18.51 (CI: 10.84-31.63)
based on data from all countries.
Based on risk estimates calculated of aggregated data of all countries involved and the risk
estimates calculated for the single countries, an overall assessment of the magnitude of the
relative risk was done. The result by substance group is shown in Table 3.

Table 3: Relative risk level of getting seriously injured or killed in an accident for various
substance groups.

Risk level Risk Substance group

Slightly increased risk 1-3 0.1 g/L ≤ alcohol in blood < 0.5 g/L
Cannabis
Medium increased risk 2-10 0.5 g/L ≤ alcohol in blood < 0.8 g/L
Cocaine
Illicit opiates
Benzodiazepines and Z-drugs
Medicinal opioids
Highly increased risk 5-30 0.8 g/L ≤ alcohol in blood < 1.2 g/L
Amphetamines
Multiple drugs
Extremely increased risk 20-200 Alcohol in blood ≥ 1.2 g/L
Alcohol in combination with drugs

Cannabis and amphetamines: due to very different single country estimates, the risk estimate must be
treated with caution.
Benzoylecgonine, cocaine and illicit opiates: due to few positive cases and controls, the risk estimates
must be treated with caution.

The highest risk was associated with high blood alcohol concentration and alcohol combined
with other psychoactive substances. Other problem groups were medium alcohol
concentrations, multiple drug use and driving with amphetamines. Medium increased risk
was assessed for alcohol concentrations between 0.5 and 0.8g/L, for cocaine and for the
medicines included in the study. The risk associated with cannabis was assessed to be
similar to the risk of driving with a low alcohol concentration. However, it should be noted that
the risk estimates for illicit drugs were based on small numbers of positive samples and
should therefore be handled with care.

20
b) The relative risk of responsibility for a fatal accident while positive for alcohol and other
drugs (D2.3.2)
For amphetamine, cocaine and opiates, adjusted odds ratios of responsibility were not
significantly different from 1. This means that the risk of responsibility for positive drivers was
not significantly different from that of sober drivers.
Among car drivers, positive cannabis detection was found to be associated with increased
risk of responsibility. A significant concentration effect was identified. The effect of cannabis
remained significant after adjustment for age, gender and alcohol: adjusted odds ratio was
1.89 (CI: 1.43-2.51).
For alcohol (≥0.1 g/l), adjusted odds ratios of responsibility were much higher than those
associated to cannabis: adjusted odds ratio for alcohol was 8.39 (CI: 6.95-10.11).
No interaction was statistically significant between alcohol and cannabis. This means that the
odds ratio of responsibility associated to being positive to both cannabis and alcohol is
merely the product of the respective odds ratios of cannabis and alcohol alone.

c) The relative risk for patients using psychotropic medicines of being involved in an accident
(D2.3.1)
The pharmacoepidemiological study that was performed in the Netherlands showed an
increased accident risk for drivers exposed to at least one psychotropic medication of 1.28
(CI: 1.1 - 1.5). The risk was found to be higher for drivers in combination therapy, namely
1.55 (CI: 1.2 - 2.0) and users of modern antidepressants, namely 1.76 (CI: 1.4 - 2.2). The
highest risk groups were new users, intermediate and long half-life benzodiazepine users,
female users, and young/middle-aged users, although only some of the trends in elevated
risk were statistically significant.
The increased relative risks found in this study indicate that psychoactive medications can
constitute a problem in road safety. Therefore, both health care providers and patients
should be properly informed and aware of the potential risks associated with the use of these
medications.

1.1.3.3. Results of the studies on characteristics of drink and drug

impaired drivers
Characteristics of drink and drug impaired drivers were investigated in the four following
populations:

a) Characteristics of drivers in the general driving population

Results from the road side surveys (D2.2.3) showed for which age groups and gender of
drivers various substances were most prevalent. Additionally information on the prevalence
by time of the day and week were given.
As expected, the prevalence of alcohol was significantly higher for male drivers than for
female drivers. In most countries the prevalence of alcohol-positive drivers was highest for
the two oldest age groups (35-49 and ≥ 50 years). This was both the case for male and for
female drivers.
For alcohol concentrations equal to and above 0.1 g/L, the prevalence was significantly
different in different time periods. As expected, the highest prevalence was on weekend
nights whereas the lowest was on weekend days. But surprisingly, there was no difference in
prevalence of concentrations at and above 0.5 g/L over the various time periods.

21
Amphetamines were mainly used by drivers younger than 35 years. In some countries this
drug was more prevalent among male drivers and in other countries more prevalent among
female drivers. The distribution of amphetamines by time period differs between countries.
Cocaine was mainly found in male drivers and in general, the prevalence of cocaine was
very low, but varied significantly by age and country. Cocaine was most prevalent among
male drivers aged 25-34 years, and least prevalent in the age group 50 and above. Cocaine
was detected during all time periods.
Cannabis seemed to be a weekend drug mainly used by young male drivers. There was a
significant difference in the prevalence of cannabis in different time periods, most prevalent
in weekend days and least prevalent in weekend mornings. However, cannabis was found
during all days and hours of the week in most countries.
Illicit opiates were most often used by male drivers aged 35 to 49 years, except for Belgium
where most users were younger than 25. Illicit opiates were not detected among drivers from
northern European countries (Denmark, Finland, Norway and Sweden) and from eastern
European countries (Czech Republic, Lithuania, Poland and Hungary).
Benzodiazepines were in general most prevalent in drivers aged 50 and above and
significantly more prevalent than in the youngest age group (18-24 years) where they were
least prevalent. However, in Italy most benzodiazepines were used by young drivers aged
18-24. In contrast to cannabis, benzodiazepines were drugs that were mainly prevalent in
mature female drivers and during daytime. Thus, prevalence was significantly different over
the time periods; it was most prevalent in the daytime during weekdays and least prevalent in
the evenings of the weekdays.
Z-drugs were not found in Southern Europe. Most drivers positive for Z-drugs were 50 years
and older, except for Hungary where all drivers were between 25 and 34 years old. Z-drugs
were most often detected during daytime hours at weekdays. In none of the countries Z-
drugs were found in weekend nights.
Medicinal opioids were distributed in a similar way as benzodiazepines: their prevalence
differed significantly with age, being most prevalent in the age group 50 and above and least
prevalent in the youngest age group (18-24). Like benzodiazepines, they were significantly
more prevalent among female than male drivers. In general, the highest prevalence was
detected during daytime hours.
Alcohol in combination with drugs had a significantly different prevalence among age groups;
the age group of 25-34 years had the highest prevalence and the age group of 50 and above
the lowest. In general the prevalence for alcohol-drug combinations for male drivers was
higher than for female drivers. There was a significant difference in prevalence over the time
periods; thus the combination of alcohol and drug(s) was most prevalent in daytime of the
weekend and least prevalent in daytime of the weekdays.
Multiple drug prevalence was significantly different among countries; it was most prevalent in
Spain and Italy and least prevalent in Denmark, Poland and Sweden. Drug-drug
combinations were most frequently detected among younger drivers. The distribution over
the four age groups varied largely between countries. In general multi-drug use was more
common among male than among female drivers. There was no significant difference in
prevalence over time periods.

b) Characteristics of accident involved drivers

Results from the hospital studies of seriously injured drivers and the studies of killed drivers
(D2.2.5) showed for which age groups, gender and time of the day and week various
substances were most prevalent.

22
As for time periods, in both studies, higher percentage of positives were normally found
among drivers involved in accidents that occurred at night time, either during the week or the
weekend, compared to percentages of positive drivers found among subjects involved in
accidents during daytime. Lithuania was the only country in which the lowest percentage of
positive drivers was found during week nights.
A significant difference was found between gender both for seriously injured drivers and for
killed drivers. In general, the prevalence in male drivers was higher than in female drivers.
Both for male drivers in the group of seriously injured drivers and in the group of killed
drivers, the group aged 25-34 was the one that had the highest percentage of positive
subjects, except for seriously injured drivers in Lithuania and killed drivers in Finland.
Alcohol was mostly found in the younger age groups of males for seriously injured drivers,
whereas in the sample of killed drivers, alcohol was also present in mature drivers.
Cannabis was most prevalent in the younger age groups of male drivers, both for seriously
injured drivers and killed drivers.
Amphetamines, cocaine and illicit opiates were most prevalent in younger age groups of
male seriously injured drivers, whereas the three substance groups were also common in
mature killed drivers.
Benzodiazepines were mostly present in male drivers in the sample of seriously injured
drivers, but were prevalent for all age groups, both for seriously injured drivers and killed
drivers. The diffusion of this substance group among both gender and all age groups may be
explained by the various therapeutic uses, different benzodiazepines being prescribed,
among others, for the treatment of anxiety disorder, sleeping disorders and epilepsy.
Z-drugs appeared to be more common in Northern Europe. They are not used at all in
Southern Europe. Use of these medications was recorded in both genders, and apparently
more frequent in the older age groups starting from 35 years.
Medicinal opioids were also most prevalent in the northern countries. However, there was no
clear picture regarding age and gender.

c) Characteristics of general drug users

The smart-phone study (D2.2.2) revealed that 20.5% of the drug users’ drives were under
the influence of drugs, with cannabis as the most prevalent drug, followed by alcohol. Other
stimulants or multiple drug use only counted for small percentages.
The results showed various differences between drug users and non-users, e.g. the time
they went to bed, was at a later time of the evening for drug users than for non-users, and in
the same way drug users got up later in the morning compared to non-users. In the
evenings, drug users stayed out more often at private locations (i.e. at friends) whereas the
non-users more often visited public locations.
In general, drug users were more mobile at night compared to the non-users, who were more
mobile at usual rush-hour times indicating a daily working routine. But, compared to non-
users that were less mobile during night time, the drug users had less night time trips by car
than the non-users. The drug users seemed to compensate by using alternative modes of
transportation at these hours.
Regarding alcohol consumption, the drug users consumed alcohol more frequently and in
higher doses than the non-users and had the double prevalence of drink driving than the
non-users.

23
High consumption frequency proved to be a striking predictor for frequent drug driving and
highly impaired driving in general, since most substance-positive drives and drives with high
blood concentrations were made by excessive substance users compared to moderate or
heavy drug users.
For alcohol, the subjective feeling of impairment increased with increasing blood alcohol
concentration, whereas for cannabis this relation was only found for moderate to heavy
cannabis users. This indicates that especially moderate substance users are able to judge
their intoxication more realistically. They can be described as being responsible-minded
concerning drugs in traffic. This assumption was derived from a lower number of drives
under the influence, lower blood concentrations on drug drives and a subjective feeling of
impairment depending on the actual intoxication in those users.
Other influencing factors were the perceived risk of detection by the police, the distance to
drive, the availability of alternative transport modes and the presence of companions. The
results also indicated that male drivers less often drive under the influence in case they have
female companions.
Finally, based on German criminal records and self-reported dangerous traffic situations
within the study period, it was demonstrated that the drug users did not seem to be more at
risk than the non-users. So, except from driving under the influence, there was no evidence
that the drug-users showed risky driving behaviour in traffic.

d) Characteristics of drink and drug drivers

The qualitative interviews on motives to drink and drug driving were carried out in Sweden
and Hungary (D2.2.1). The results showed that the interviewees belonged to a very specific
group of people who were addicted to alcohol and or other drugs. Normal sanctions did not
prevent their driving while intoxicated and this applied to participants from both Sweden and
Hungary, partly because they perceived themselves to have very little control over their
behavior, partly because they did not believe that they would be stopped by the police.
Those who had been stopped pointed out that this had happened because the police were
carrying out a routine control, and not because they had driven in an unsuitable way.
The interviewees did not believe that alcohol or drugs would impair their driving and therefore
they did not perceive any real risks of driving. However, one important difference between
drugs and alcohol was that drugs were believed to make them a better driver whereas
alcohol did not make them any worse. Thus, drug driving was not regarded as an offence in
the same way as drink driving.
Following this, respondents who had been caught for drink driving expressed more feelings
of shame than those who had been caught for drug driving. This was partly because of the
offence itself but, as mentioned before, it had more to do with having to admit to others that
they had been drinking and driving and could not control their drinking. Feelings of shame
appeared not to be related to a feeling that the act itself could result in an accident but
somewhat related to if their friends and relatives disapproved. The same feelings were not
expressed by the drug drivers, only later when they were under treatment and they looked
back at their life did the feeling of shame and anguish emerge.
The participants whose drinking and driving was related to problems with alcohol would
argue that losing the license or even to be imprisoned would not have helped them to stop
re-offending. Instead, it was the treatment program which had helped them by providing a
greater insight into their own problems.

24
1.2. Experimental studies and Meta-analyses
A second approach to explore the problem situation of impaired driving was to conduct
experimental studies and meta-analyses. In addition to the epidemiological research which
helps to provide a broad overview of the situation experimental research examines specific
problems in more detail (e.g. effects of a specific psychoactive substance used by a specific
group of subjects in determined driving situations).

1.2.1. Objectives
Experimental studies and meta-analysis were conducted in WP1. The objectives of this Work
Package were3:

1. Providing an overview about the state-of-the-art of research results in the domain of

DUI/DUID with a special focus on impairment caused by alcohol;
2. Closing existing research gaps by conducting driving tests by using a standardized
procedure that allows the comparison of the research results;

1.2.2. Methodology
The experimental work and the meta-analyses were conducted according to a standardized
procedure.

a) Meta-analyses
When a sufficient number of published studies on a certain psychoactive substance is
available, those studies can be summarized by a meta-analysis. In DRUID this was done for
alcohol, major illicit drugs and medicines.
The meta-analysis on alcohol (D1.1.2a) comprises experimental studies published until
2007. A total number of 450 papers in which 5,300 findings concerning alcohol effects on
(driving) performance, social behavior or mood are described were included. For every
finding the presence or absence of a significant alcohol effect is coded. From a meta-analytic
perspective, this procedure belongs to the method of vote counting. This means the
significant findings of effects are summarized for each group of dependent measures for the
same range of BAC. The impairment function derived from the meta-analysis (Figure 3) is
used to determine the concentration of psychoactive substances causing the same
impairment as certain BAC levels. Therefore the percentage of significant findings in those
studies, derived from the meta-analyses of medicines and major illicit drugs, is compared to
that of alcohol (Figure 4). By doing so the substance concentration can be determined at
which the same percentage of significant findings exists as for the different BAC levels.
In the meta-analysis of medicines and illicit drugs (D1.1.2b) 605 publications with a total
number of 13 191 findings were included, that fulfilled certain quality criteria. Only
experimental studies with single oral administration to healthy subjects were included into
this meta-analysis, because studies with either multiple administration to healthy subjects or
with administration to patients are very rare and therefore cannot be analyzed in a meta-
analytic approach. This meta-analysis provides information about the impact of
antipsychotics, anxiolytics, hypnotics, sedatives, antidepressants, antihistamines, and major
illegal drugs (see the detailed list of agents in Table 4) on driving and skills related to driving.

3
The methodological framework is outlined at the beginning of section 2. An integrated summary of all outcomes
of DRUID is given in section 3. Recommendations from a scientific point of view are given in section 4. Therefore
the focus here will be on the experimental studies and the meta-analyses.

25
For opioids mainly used in substitution treatment for opioid addicts (morphine, methadone,
and buprenorphine), narcoanalgetics, and hallucinogens too few published studies were
available. Therefore it was not possible to summarize those results by a meta-analysis.
Instead a review was done for the effects of these agents on driving and skills related to
driving (D1.1.2c).

Figure 3: Using BAC levels (e.g. 0.3g/L, 0.5g/L Figure 4: Application of the impairment
and 0.8g/L) to determine categories of categories for alcohol (no = 0.3g/L,
impairment by fixing the percentage of moderate = 0.5g/L and severe impairment =
significant findings in the meta-analysis. 0.8g/L) to determine the substance
concentrations causing an equivalent
effect as the three BAC levels.

Table 4: Meta-analytically evaluated psychoactive substances in D1.1.2b (sorted by number of

studies per substance group in descending order; D1.3.1 p. 51).

Dose related information is highly relevant for healthcare providers as well as for patients
because they need to know for certain agents which dose might have an impact on driving
ability. Therefore the effects were evaluated with respect to the dosages and the
corresponding impairments found in the studies. An approximation procedure was applied
and the following parameters were calculated:
• dosage that causes the maximum impairment,
• time span after application when the impairment is highest,
• BAC that is equivalent to the maximum impairment,
• duration of the impairment,
• degree of impairment.

26
In order to relate the impairment to the concentration of the substance in blood, a meta-
analysis of pharmacokinetic studies was done. This led to approximated concentration
curves in time course for different substances and different substance concentrations. Using
this approximation the blood plasma concentration of a substance could be estimated by
knowing the dose of the substance administered and the duration between the time the
substance was administered and the time the performance test was done.

b) Driving test (Experimental studies)

13 driving tests were conducted (Table 5) according to a standardized procedure.

Table 5: Overview of psychoactive substances examined in the driving tests (WP1).

No. Substance / Participants

1 Mdma (100 mg) with and without alcohol (0.5.g/kg)

2 Dexamphetamine (10mg) with and without alcohol
3 Mdma (25, 50 and 100mg) before and after a night of sleep deprivation
4 Dexamphetamine (10 and 40 mg) before and after sleep deprivation
5 Zopiclone in patients and controls
6 Medicated vs. unmedicated insomnia patients
7 Alprazolam (0.5 mg) in anxious patients
8 Treated vs. untreated sleep apnea patients (CPAP)
9 Codiliprane and zolpidem in elderly drivers: alone and in combination
10 Codiliprane in healthy volunteers
11 Dronabinol in light and heavy users of THC
12 Opioid analgesics in chronic pain patients
13 Risperidone/paliperidone in patients diagnosed with psychosis

All studies on the impact of illicit drugs were conducted according to placebo controlled,
double-blind, within-subjects study designs. All studies employed representative subject
samples, i.e. recreational users of MDMA, patients who went through a strict medical
screening and selection procedures. Most of the studies employed cross-over designs which
were preferred for their efficiency while providing maximal statistical power with relatively
small sample sizes. Most studies were conducted in real traffic, some in advanced driving
simulators. Due to the national legal and ethical restrictions, one experiment had to be
conducted in a closed circuit.
Driving tests on the effects of illegal drugs were conducted at Tmax, i.e. when drug
concentrations were maximal and, in case of sleep deprivation, also in morning after a night
of sleep loss. More details on study designs, screening, subject characteristics, and in- and
exclusion criteria can be found in the study reports in D1.2.1 and D1.2.2.

27
All studies were adhered to a set of mandatory settings of which the most important were:
• Number of subjects: the minimum number of subjects was 16. The choice for a subjects’
sample-size was always corroborated by a statistical power analysis.
• Subject selection: studies with MDMA and dexamphetamine only included recreational
users of these drugs. Drug-naive, healthy volunteers were excluded.
• Drug screens: subjects were always tested for drugs in urine prior to testing
• Alcohol screens: subjects were always tested for alcohol (breath) prior to testing
• Driving experience: subjects needed to have a driver’s license
• Training sessions: all subjects received training sessions of actual driving tests, simulator
driving tests and/or laboratory performance tests in order minimize learning effects.
Training was performed by all subjects to achieve a stable performance level prior to
study entrance.
• Subjective measures: e.g. alertness, mental effort, etc.
• Ethics: all partners obtained study approval from their local (and national) ethics review
boards and conducted their study according the declaration of Helsinki and Good Clinical
Practice.
• Alcohol calibration: For all parameters alcohol effects on driving were used as a standard
reference to quantify the amount of impairment. Therefore all partners conducted a
placebo-controlled alcohol study in order to calibrate their primary driving parameters for
the effects of BAC 0.5 g/L. Any drug induced performance change > performance change
induced by BAC 0.5 g/L was qualified as a clinical relevant drug effect. Drug effects equal
to those produced by a BAC of 0.5 g/L were also considered to define the “threshold” of
impairment for an individual drug. Drug effects were tested for comparability to BAC 0.5
g/L effects by means of equivalence testing.
• Toxicology: in all studies whole blood, serum and blood spots were collected for
determining concentrations of the investigated active agent.
• Statistics: the statistical analyses consisted of 2 steps: 1) assessment of overall treatment
effect by means of superiority testing; 2) equivalence testing of drug effects relative to the
alcohol criterion.
• Standard set of driving parameters and driving scenarios. These driving parameters
basically covered 3 core levels of driving behaviours: Automated behaviours – well-
learned (over-learned) skills, controlled behaviours – controlled manoeuvres in traffic,
executive, strategic behaviours - interactive functions with ongoing traffic, planning, risk
taking.

Road tracking scenario

The road tracking scenario was based on the Road Tracking Tests that has been used in the
Netherland in over 100 studies for measuring drug effects on driving (O'Hanlon et al., 1982).
Participants were required to drive a 100km course maintaining a constant speed of 95 km/h
and a steady lateral position in traffic lanes. The primary driving measure was the standard
deviation of lateral position or SDLP. SDLP is an index of road tracking error or weaving,
swerving and overcorrecting.
SDLP was measured using an electro-optical device mounted on the rear of the vehicle
which continuously records lateral position relative to the traffic lane. An increase in SDLP,
measured in centimetres, indicated driver impairment, as the driver’s ability to hold the car in
a steady lateral position diminished.

28
Car-Following scenario
The Car Following task was developed to measure attention and perception performance, as
errors in these areas often lead to accident causation. In this task participants were required
to match the speed of a lead vehicle and to maintain a constant distance from the vehicle as
it executes a series of deceleration and acceleration manoeuvres. The primary dependant
variable was reaction time to lead vehicle’s speed decelerations. This test assessed a
driver’s ability to adapt to manoeuvres of other motorists (Brookhuis & de Waard, 1993;
Ramaekers & O'Hanlon, 1994).

Risk taking scenario

Risk taking scenarios were only embedded in studies using a driving simulator. Standard
parameters applied: gap acceptance, number of crashes, number of red light crossings and
number of crashes during sudden event scenarios.

Laboratory tests
In addition, all partners including a number of laboratory tests measuring skills related to
driving. These test included tracking tasks, attention tasks, reaction tasks and cognitive
tasks. Performance parameters associated with these laboratory tests were considered
secondary outcome measures.

1.2.3. Results
In the following the results of the meta-analyses and the experiments are described.

1.2.3.1. Meta-analysis of alcohol

In Figure 5 the percentage of significant findings for all performance categories is shown.
Since this percentage increases to the same degree as the BAC, a linear function is fitted to
the empirical values of the general performance data. The general impairment function
comprises 2914 performance findings. At a BAC of 0.5g/L, 30% of the findings are
significant, while at a BAC of 0. 8g/L about 50% of the findings are significant. More detailed
information about the impairment caused in different categories of performance and mood is
available in the Deliverable 1.1.2a.

29
Figure 5: General impairment caused by alcohol (regarding all performance categories) –
percentage of significant findings for each BAC category, based on the meta-analysis of
alcohol (D1.1.2a, p. 71).

1.2.3.2. Meta-analysis of medicines and illicit drugs

Table 6 shows the impairment caused by certain dosages of medicines based on the meta-
analysis of medicine and illicit drugs. Highly impairing are/is:
• the anxiolytics alprazolam (1 mg), and high dosages of oxazepam (30 mg), diazepam
(20 mg), and lorazepam (2 / 2.5 mg),
• the antidepressants mianserin (10 mg), and amitryptiline (25 / 50 mg),
• the hypnotics/sedatives flunitrazepam (2 mg), triazolam (0.5 mg), zopiclone (7.5 mg),
and zolpidem (20 mg), and
• the antipsychotic promethazine (27 mg), and
• THC (24.5 mg, oral administration).

Neither antihistamines nor illicit drugs cause a comparable high impairment. For more details
see D1.1.2b.

30
Table 6: Degree of impairment sorted in ascending order within the different substance classes
based on the meta-analysis of medicines (D1.3.1, p. 55).

In Table 7 the substances and blood concentrations are listed which cause an impairment
equally to that of 0.5 g/L BAC.
21 studies with 482 effects in total (dose range from 7.5 to 39mg) were included into a meta-
analysis of the effects of oral administration of THC on performance. This reveals that the
impairment caused by 3.7 ng/mL THC (range 3.1 to 4.5) is equal to that caused by 0.5g/L
BAC. An additional meta-analysis on the effects of smoking of THC on performance leads to
a comparable result. 78 studies with a total of 888 effects (doses 1 to 52 mg) were included
into this meta-analysis. Hence THC causes an impairment equivalent to that of 0.5 g/L BAC
at a concentration of 3.7 ng/mL in serum when orally administered and of 3.8 ng/mL in
serum when THC is smoked.

Table 7: Substance concentrations equally impairing as 0.5 g/L BAC (“not calculable” indicates
that not enough data existed to calculate the 0.5 g/nL equivalent dosage; “not reached”,
indicates that a given substances did not cause impairment worse than 0.5 g/L BAC;
“calculable” indicates that the dosage equally impairing as 0.5 g/L BAC could have been
calculated (D1.3.1, p57).

31
Based on the meta-analyses of experimental studies no negative influence of stimulants on
the fitness to drive could be stated. In general, there were more findings of performance
improvements than of performance impairments. D-amphetamine is the agent on which most
studies were available (dosage 1mg to 36 mg). Recent studies focused on the impact of
designer amphetamine MDMA (ecstasy) on performance. In those studies more
improvements than impairments were found, too. Accordingly, there was no performance
decrement during the time of action after consumption of “normal” doses (40mg to 125mg).
Cocaine showed similar effects as amphetamines. From the meta-analysis of experimental
studies no negative influence on the fitness to drive could be stated. Only some case-reports
and non-experimental publications revealed negative effects. But there was a lack of studies
focusing on the impairment during the post acute phase.
The literature review on opioids used in substitution therapy revealed that substitutes may
cause impairments even at low dosages when implicated in a single dose to healthy
subjects. There was no clear evidence if patients treated chronically were able to drive, as
there were big differences in performance decrements between individuals. Many patients in
substitution therapy are using other substances in addition to their medication. Therefore it is
recommended that a screening for other drugs should always be done before a decision is
made if a patient should be allowed to drive.

1.2.3.3. Experimental studies

All studies investigating the effects of stimulants on driving (MDMA and dexamphetamine)
did not reveal impairing effects or increased risk taking caused by the drug consumption itself
(see Table 8 and Table 9 for an overview of the results). In general, low doses of stimulant
drugs produce neutral or even stimulating effects on a range of psychomotor functions.
However there are some studies showing that stimulants may also produce detrimental
effects on specific cognitive functions and increase risk taking behaviours.
Sleep deprivation alone caused impairments equivalent to those observed under the
influence of 0.5 g/L BAC. Only in case of additional alcohol consumption an increased risk
taking behavior could be observed. Stimulant effects of MDMA and amphetamine were not
sufficient to overcome or compensate for driving impairments produced by concomitant
alcohol use or sleep deprivation. No clear relation could be found between drug
concentrations in blood (and plasma) and driving impairment for MDMA and
dexamphetamine.
The pharmacological effects of stimulants and the effects of drug use setting (e.g. poly-drug
use, concomitant alcohol use and sleep deprivation) were intertwined and significantly
contributed to driver impairment. Moreover, users of stimulating drugs were not aware about
post-acute fatigue effects.
A summary of the results of the studies done to investigate the impact of medicinal drugs
on actual and simulated driving is presented in Table 10, Table 11 and Table 12.
Zopiclone (7.5mg) and alprazolam (0.5mg) produced significant driving impairment in
patients as well as in healthy controls. Zolpidem (10mg) produced significant driving
impairment in elderly subjects. Chronic users do not experience subjectively any sedative
effects of zopiclone and alprazolam, whereas infrequent users and healthy users reported
feelings of reduced alertness and sleep. This lack of awareness of (residual) sedative effects
of zopiclone and alprazolam may lead insomnia and anxious patients to belief that car driving
is safe during treatment with these drugs.

32
Results from the insomnia study showed that driving performance and driving related
psychomotor performance did not differ between medicated insomniacs, unmedicated
insomniacs and normal sleepers (i.e controls). These results indicate that driving
performance of insomniac patients does not differ from that of normal sleepers,
even in insomniacs that had been prescribed hypnotic mediation. The lack of driving
impairment in medicated insomniacs could however be predicted from the type of hypnotics
that patients were using. About 2/3 of the patients received short acting hypnotics or low
doses of hypnotics that previously shown not to produce any residual impairment in driving
test ( e.g. zolpidem, temazepam). Sleep apnoea turned out to be strongly correlated to driver
impairment.
Combinations of codeine and paracetamol in general did not produce driving impairment
when administered to healthy volunteers even at higher doses. However, driving impairment
became apparent after the lowest dose when administered to elderly subjects. Thus the
results indicate that the impairing potential of codeine/paracetamol varies with age.
Dronabinol (Marinol®) impaired driving performance in occasional and heavy users in a
dose-dependent way. Equivalence tests demonstrated that dronabinol induced increments in
SDLP were bigger than impairment associated with 0.5 g/L BAC in occasional and heavy
users, although the magnitude of driving impairment was generally less in heavy users.
The results of the driving tests with patients suffering from chronic pain under long-term
treatment with opioid analgesics revealed that driving performance of patients was
comparable to that of healthy controls. Nevertheless neuropsychological tests assessing
skills related to driving revealed that pain patients performed worse compared to healthy
controls on a number of tests.
Patients using risperidone drove with a lateral position that was comparable to that observed
in controls. However the standard deviation of lateral position and reaction time to sudden
events were significantly increased in patients and comparable or bigger than those
observed in controls with 0.5 g/L BAC. The present data thus seems to indicate that patients
under the influence of risperidone do have some impairment that should be considered of
clinical relevance.

33
Table 8: Summary of MDMA and dexamphetamine effects on primary and secondary driving parameters (improvement, no effect or impairment) as
well as subjective measures of arousal and sleep, alone and in combination with alcohol.

Study 1: MDMA-alcohol study Study 2: Dexamphetamine – alcohol study

MDMA Alcohol MDMA + Alcohol Dexam- Alcohol Dexamphetamine + Alcohol

phetamine
Road Decrease Increase Increase SDLP No effects Increased SDLP Increased SDLP
tracking SDLP SDLP Relevance of impairment Relevance of impairment undecided
undecided (95%CI drug effect (95%CI drug effect includes BAC 0.8
includes BAC 0.5 as well as 0) as well as 0)
Car No effect No effect No effect No effect No effect No effect
Following
Risk Taking No effect No effect No effect No effect Shorter gap acceptance; Shorter gap acceptance; increased
increased red light red light crossings and number of
crossings and number of crashes
crashes Relevance of impairment undecided
(95%CI drug effect includes BAC 0.8
as well as 0)
Laboratory Not assessed Not Not assessed No effect Impairment of attention, Impairment of attention, tracking and
measures of assessed tracking and RT RT
skills related
to driving
Subjective Decreased Increased Increased sleepiness Decreased No effect Decreased sleepiness
measures sleepiness sleepiness sleepiness

34
 Table 9: Summary of MDMA and dexamphetamine effects on primary and secondary driving parameters (improvement, no effect or impairment)
as well as subjective measures of arousal and sleep, alone and in combination with alcohol.

Study 3: MDMA – sleep deprivation study Study 4: Dexamphetamine study – sleep deprivation study
MDMA Sleep MDMA + Sleep deprivation Dex- Sleep deprivation Dexamphetamine +
deprivation amphetamine Sleep deprivation
Road No effect Increased Increased SDLP; No effect Increased SDLP Increased SDLP (Relevance of
tracking SDLP; Impairment > BAC 0.8g/L Impairment > BAC 0.5 g/L impairment undecided)
Impairment
> BAC 0.8
g/L
Car No effect No effect No effect Dose related Impairment of phase delay Impairment of phase delay
Following improvement
of phase delay
Risk Taking Not assessed Not Not assessed Improvement Impairment of phase delay Improvement RT to crossing cars
assessed RT to crossing
cars
Laboratory Neutral on Impairment Impairment of attention and Not assessed Not assessed Not assessed
measures of most of attention impulse control
skills related measures, and impulse
to driving Improvement control
on rapid
information
processing
Subjective Increased Decreased Decreased arousal Decreased Increased sleepiness Increased sleepiness
measures arousal arousal sleepiness

35
Table 10: Summary of treatment and sleep disorder effects on primary and secondary driving parameters as well as subjective measures of arousal or
sleep (ZOP=zopiclone; PLA=placebo; BAS=baseline and ALP=alprazolam).

Study 5: Study 6: Study 7: Study 8:

Residual effects of zopiclone 7.5 mg Insomnia patients Alprazolam 0.5 mg in anxious patients Sleep apnoea
patients
Chronic hypnotic Infrequent Healthy controls Medicated insomnia Medicated Unmedicated Healthy Patients with CPAP
users hypnotic users ZOP vs. PLA patients vs. anxious anxious controls vs. patients with no
ZOP vs. PLA ZOP vs. PLA unmedicated patients patients ALP vs. BAS CPAP vs. controls
insomnia patients vs. ALP vs. BAS ALP vs. BAS
healthy controls
Road tracking Increased SDLP; Increased SDLP; Increased SDLP; No difference Increased Increased Increased Increased SDLP in
Impairment > Impairment > Impairment > between groups SDLP; SDLP; SDLP; both patients
BAC 0.5 mg/mL BAC 0.5 mg/mL BAC 0.5 mg/mL Impairment > Impairment > Impairment > groups relative to
BAC 0.5 BAC 0.5 BAC 0.5 controls
mg/mL mg/mL mg/mL No difference
between patient
groups
Car-Following No effect No effect No effect No difference Increased Increased Increased Increased brake
between groups brake reaction brake reaction brake reaction reaction time and
time and time and time and increased time
increased time increased time increased time driven at close
driven at close driven at close driven at close distance to leading
distance to distance to distance to vehicle in both
leading vehicle leading vehicle leading vehicle patient groups
Risk Taking Not assessed Not assessed Not assessed Not assessed Not assessed Not assessed Not assessed Not assessed
Laboratory measures Impairment of Impairment of Impairment of No group differences No effect No effect Decreased No difference
of skills related to memory, memory, memory, in verbal memory, reaction time between groups
driving tracking, divided tracking, divided tracking, divided divided attention,
attention, attention, attention, vigilance and
inhibitory control inhibitory control inhibitory control inhibitory control.
Subjective measures Increased next No effect Decreased next No group differences No effect Decreased Decreased Decreased
(sleepiness/alertness) day alertness day alertness in sleepiness and alertness alertness alertness in patient
alertness groups

36
Table 11: Summary of treatment and pain disorder effects on primary and secondary driving parameters as well as subjective measures of arousal or
sleep (COD= codiliprane; PLA=placebo; ZOL = zolpidem; DRO=dronabinol).

Study 9: Study 10: Study 11: Study 12:

Codiliprane and zolpidem in elderly: alone and Codiliprane in Dronabinol in THC users Opioid patients
in combination healthy
volunteers
COD vs. PLA ZOL vs. PLA COD+ZOL vs. COD vs. PLA Occasional THC Heavy THC users Patients treated with
PLA users DRO vs. PLA opioids vs. controls
DRO vs. PLA
Road tracking Increased Increased No effect No effect Increased SDLP; No overall No difference between
SDLP SDLP Impairment DRO superiority effect groups
10mg > BAC 0.5 of DRO on SDLP
mg/mL Big individual
Impairment DRO variation in
20mg > BAC 0.8 change SDLP
mg/mL after DRO 10 and
20mg : 95% CI
includes zero as
well as alcohol
criterion value
Car-Following No effect No effect No effect Not assessed Decreased Time No effects No difference between
to Speed groups
Adaption
Risk Taking Increased No effect No effect Not assessed Not assessed Not assessed Not assessed
number of
crashes
Laboratory measures of skills Not assessed Not assessed Not assessed No effects No effect on No effect on Patient perform worse than
related to driving Standard Filed Standard Filed controls on some isolated
Sobriety Tests Sobriety Tests tests
Subjective measures Not assessed Not assessed Not assessed Small increase of Increased Increased No difference between
(sleepiness/alertness) sleepiness sedation sedation groups

37
Table 12: Summary of risperidone/paliperidone effects on primary and secondary driving
parameters as well as subjective measures of arousal or sleep.

Study 13:
Risperidone/paliperidone
patients diagnosed with
psychosis and receiving
risperidone vs. controls during
placebo and alcohol
Road tracking Increase in SDLP of patients >
increase in SDLP of controls
during alcohol (BAC=0.5
mg/mL). Alcohol also affected
lateral position. The latter was
not affected in patients
Car-Following Not assessed
Risk Taking RT to sudden events increased
in patients and controls during
alcohol relative to controls during
placebo. Increase in RT in
patients was comparable/bigger
than BAC=0.5 mg/mL
Laboratory measures 9 out 11 patients passed the
of skills related to Vienna driving evaluation test
driving
Subjective measures Not assessed
(sleepiness/alertness)

1.3. Synopsis of problem situation based on epidemiology

Referring to the epidemiological studies of DRUID the main results can be summed up
like this:

Prevalence of alcohol and/or other psychoactive substances in relation to

road safety:

• Alcohol had the highest prevalence in the driving population (up to app. 4%), but
with the majority of the alcohol positive drivers below 0.5g/L as well as in seriously
injured and killed drivers in all countries (up to 15-25%), with the most of the
alcohol positive drivers above 0.5g/L
• There was a higher prevalence of illicit drugs in the southern part of Europe both in
the population and in the driving population.
• There was a higher prevalence of medicines in the northern part of Europe both in
the population and in the driving population.
• Combined use of alcohol an/or other drugs was more common in accident involved
drivers than in the driving population.
• There was no clear picture of the distribution of illicit drugs and medicines among
injured and killed drivers, however, combined use of alcohol and/or other drugs was
much more prevalent in drivers involved in accidents than in the driving population.

38
 Accident risk for driving with alcohol and/or psychoactive substances:

• The risk of getting seriously injured or killed drivers when positive for alcohol of 0.5-
0.8 g/L was medium increased of the magnitude of 2-10 times the risk for sober
drivers. The risk increased exponentially by alcohol concentration, for alcohol
concentrations of 1.2 g/L and above the risk was extremely increased of the
magnitude of 20-200 times the risk for sober drivers.
• The risk of getting seriously injured or killed drivers when positive for most of the
illicit drugs and medicines was medium increased of the magnitude of 2-10 times
the risk for sober drivers.
• The risk of getting seriously injured or killed drivers when positive for multiple drugs
was highly increased and considerably higher than the risk when positive for a
single drug.
• The risk of getting seriously injured or killed drivers when positive for alcohol in
combination with other drugs was extremely increased and comparable to the risk
when positive for high alcohol concentrations.
• Killed drivers, who were responsible for the accident and positive for high alcohol
concentrations, had a highly increased risk compared to the risk of killed
responsible drivers, not positive for alcohol.
• Responsible drivers, positive for alcohol, involved in fatal accidents had a risk of
about 8 times that of responsible drivers, not positive for alcohol. Those positive for
cannabis had a risk of about twice that of drivers not positive for cannabis.

Characteristics of drink and drug impaired drivers:

• Drivers do not think that alcohol impairs their driving, and they think that drugs
improve their driving.
• Drink drivers feel more ashamed than drug drivers. They are mainly worried that
their friends disapprove their behavior.
• Drug drivers do not feel ashamed, but after treatment, they would look back with
shame.
• Rehabilitation reduces recidivism.
• Drink driving with high concentrations and drug driving seems to be associated with
addicted users, whereas moderate users are more responsible.
• Although drug drivers go up later in the morning and go to bed later in the night
than those who do not take drugs, the drug users drive less than other driver
groups in the late hours.

39
2. Countermeasures

2.1. Enforcement

2.1.1. Objectives
DRUID WP 3 dealt with the evaluation of enforcement strategies whereas enforcement
was implemented by traffic police applying oral fluid mobile screening devices for drugs
detection.
The following main objectives were pursued:
a) Large scale practical evaluation of onsite oral fluid screening devices for detecting
of psychoactive substances in drivers.
b) Large scale scientific evaluation of onsite oral fluid screening devices for detecting
of psychoactive substances in drivers and checklist for clinical signs of impairment.
c) A cost-benefit analysis of drug driving enforcement by the police.

2.1.2. Methodology
a) Practical evaluation of oral fluid screening devices
13 manufacturers participated in practical evaluation of their screening devices4. They
provided information on devices and prepared training materials for police teams.
Evaluation was implemented in six European countries (Germany, Belgium, Ireland,
Finland, Spain and the Netherlands) by trained teams of police officers under real
conditions in compliance with a uniform experiment design and applying a uniform
protocol.
The practical evaluation has been focussed on:
• Successful performance
• Duration of collecting a sufficient oral fluid sample
• Duration of sample analysis
• Hygienic aspects
• DPO’s5 impression of reliability of the obtained indication
• DPO’s opinion on the simplicity of the test.
Test persons were asked for their opinion on the tests they performed and the
relevance of enforcement legislation.

b) Scientific evaluation of oral fluid screening devices and a checklist for clinical signs
of impairment
Tests were performed on eight (out of initially 13) devices that have been assessed as
the most promising. Analytical evaluation of oral fluid screening devices and preceding
selection procedures was carried out in the Netherlands, Belgium and Finland from

4
See DRUID Deliverable 3.1.1 for more details.
5
DPO – Dedicated Police Officer

40
October 2007 to December 2009. Tested substance classes were amphetamine(s),
methamphetamine, MDMA, cannabis, cocaine, opiates, benzodiazepines and PCP.
A checklist for clinical signs of impairment (CSI) was evaluated in order to see if visible
signs of impairment can be used as preceding selection criteria for performing an on-
site test. The checklist was based on several existing checklists, e.g. one developed for
the German police and previously used in the European IMMORTAL (Impaired
Motorists, Methods of Roadside Testing and Assessment for Licensing) project.
Study populations consisted of randomly selected drivers from the DRUID roadside
surveys, drivers suspected of driving under the influence of drugs, patients of treatment
centres and rehabilitation clinics and customers of coffee-shops.
Oral fluid was collected as the reference sample. For some cases, in the Netherlands,
whole blood samples were also collected.
The performance of the tests was assessed based on sensitivity, specificity, accuracy,
positive predictive value and negative predictive value for the individual substance tests
of the device. These were assessed based on both DRUID and manufacturer cut-offs.
Sensitivity, specificity and accuracy performance values of 80% or more were set as a
desirable target value.

c) Cost-Benefit Analysis (CBA)

CBA has a foundation in mainstream (neo-classical) economic theory, whereby
economic values are recognised as expressions of individual/household preferences.
Road safety can be regarded as an economic good, something that people have a
demand for, characterised by a mix of private and public aspects.
A benefit-cost ratio was estimated according to formula:

Present value of all safety

benefits
Benefit-cost ratio = --------------------------------------------
Present value of implementation
costs and time use

In addition to CBA and estimates of net benefits and benefit-cost ratios of increased
traffic police enforcement, the study also included cost-effectiveness analysis (CEA).
The CBA should answer two questions: i) To what degree is the enforcement of
legislation against driving under the influence (DUI) of drugs profitable in economic
terms for the society? ii) Which of the existing devices for such enforcement are more
profitable? For these cost-benefit analyses the following data are needed: a) effects of
enforcement, i.e. the reduction of accidents, fatalities, injuries and material damage
due to this kind of enforcement; b) costs of (or positive benefits of preventing)
accidents, fatalities, injuries and material damage; c) costs (negative benefits) of road
user time; d) costs of devices/equipment; e) costs of police time; f) costs of laboratory
analyses; g) costs of the judicial system. The first three elements are handled on the
benefit side, while the latter four enter the cost side of the CBA.
The basic idea of the CBA model is that a particular scenario or group of scenarios will
be compared with the reference situation or baseline, which is a continuation of the
current situation. Thus, CBA basically compares economic benefits and costs from the
implementation of specific policies/projects with a “do-nothing” reference/baseline.

41
DRUID considered three enforcement (increase) scenarios in CBA, for drug/medicine
saliva testing – a low, medium, and high increase (Table 13). Since the CBA was
applied to different countries, with different prevalence of different drugs/medicines
(and alcohol), also the prevalence effect was taken into account. Regarding alcohol
enforcement, one added element in the scenarios was an adjustment of random
alcohol breath testing to maintain current enforcement levels / current resource use,
that is, transfer some share of the alcohol enforcement towards drug enforcement.
Researchers calculated a 10 % reduction in alcohol enforcement, and presented the
CBA of this reduction combined with a tripling of drug enforcement.

Table 13: Three different scenarios for increased drug screening in the DRUID Cost-
Benefit Analysis.
Low enforcement increase level
(50% for drugs/medicines)
Middle enforcement increase level
(tripling for drugs/medicines), with and without a 10%
reduction in alcohol enforcement
High enforcement increase level
(tenfold increase for drugs/medicines)

2.1.3. Results
a) Practical evaluation of oral fluid screening devices
Eight devices were evaluated as “promising”
• Mavand - RapidSTAT
• Securetec - Drugwipe 5+
• Branan - Oratect XP
• Varian - Oralab 6
• Innovacon - OrALert
• Cozart - DDS
• Dräger - Drug Test 5000
• Biosensor – BIOSENS (effective in specific situations in which large number of
persons must be tested in a short time, e.g. discos, large concerts, etc.)

Based on the experience of 137 training sessions and 2960 test a number of Police
User Requirements and Specifications (PURS) has been formulated. These can be
divided in three categories:
• Requirements for training of police officers
• Requirements for operational use of devices
• Requirements for documentation.

42
Police teams involved in this task developed a set of recommendations concerning
combating driving under influence of psychoactive substances from a perspective of
police forces. They propose a zero tolerance approach towards illicit drugs and
impairment approach towards psychoactive medicines. Police officers deem as
necessary to authorize traffic police to implement random tests of drivers on drugs
consumption. From their point of view, a refusal of a driver to undergo an alcohol or
drugs test should be considered as an offence. Breath sample should be used for
detecting alcohol consumption. Oral fluid screening devices should be used for
detecting drugs consumption.

b) Scientific evaluation of oral fluid screening devices and a checklist for clinical signs
of impairment
Tests were performed on eight (out of initially 13) devices that have been assessed as
the most promising. The analytical evaluation of the amphetamine test showed
sensitivity varying from 0% to 87 %. Specificity values were from 91% to 100% and
accuracy values from 84% to 98%.
For cannabis tests, sensitivities ranged from 11% to 59%. Specificities were between
90% and 100% and accuracies from 41% to 82%.
Cocaine tests scored sensitivities of between 13% and 50%, specificities of 99% to
100% and accuracies from 86% to 100%.
Sensitivities of opiate tests ranged from 69% to 90%. Specificities were between 81%
and 100% and accuracies between 75% and 99%.
Benzodiazepine tests had sensitivities from 48% to 67%. Specificities were from 94%
to 100% and accuracies from 77% to 100%.
Not enough positive cases were gathered to successfully evaluate any of the
methamphetamine, MDMA or PCP tests for the devices in which these were included.
None of the tests reached the target value of 80% for sensitivity, specificity and
accuracy for all the separate tests they comprised.
An overall evaluation, wherein any positive drug screening result was viewed as valid
providing that the confirmation sample contained one of the DRUID substances
analysed, was performed as a measure of the usefulness of the devices in police
controls. Three of the devices performed at >80% for sensitivity, specificity and
accuracy in the overall evaluation (Figure 6).
Prevalence of drugs in the study population needed to be considered when assessing
the evaluation results. In addition, the type and prevalence of drugs within the
population for which the device is intended to be used needed to be taken into account
when considering the suitability of the device based on the results presented in this
report.
All countries took their own approach to the evaluation of the checklist for clinical signs
of impairment. The results of the evaluations were not very promising. The indicators
proved to be effective mainly for cases of high concentrations or very recent use. Pupil
reaction test was best predicting parameter, esp. for AMP and THC. The checklist
scored a low sensitivity value (Dutch study), even lower correlation of symptoms and
actual presence of drugs (Belgian study) or there were difficulties in correlating the
symptoms to actual drug use due to the insufficient data collection (Finnish study).
More experience, better training, and selection of time and locations with high
incidence may improve the effectiveness of the CSI checklists.

43
Figure 6: Sensitivity and specificity of the oral fluid screening devices.

c) Cost-Benefits Analysis
Table 14, Table 15 and Table 16 present the results of the economic analyses, for the
three main levels of drug enforcement increase. These are calculations for an average
of the drug screening devices, where the prevalence-weighted average
sensitivity/specificity levels are, respectively, 45.2/96.9% for the Netherlands,
49.8/96.9% for Belgium, and 58.5/96.9% for Finland6. In these calculations no change
in alcohol enforcement level was assumed. The basic requirement for efficiency of
increased drug enforcement is a benefit-cost (BC) ratio of 1.5 or higher (Bickel et al. 2006).

Table 14: 50% enforcement increase - Drugs (average device).

Netherlands Belgium Finland

Annual benefits 5,698,297 13,527,835 2,075,422

Annual costs 291,028 1,683,580 1,636,928

Net benefits 5,407,269 11,844,255 438,494

BC ratio 19.6 8.04 1.27

Simulated BC ratio 18 7.7 1.16

simulated St. dev. 11 4.4 0.45

simulated skewness 5.72 8.28 1.50

simulated kurtosis 10.70 4.39 9.24

Costs per convicted 4,825 4,054 4,147

Costs per prevented fatality 408,481 995,247 5,480,361

Tests per 100,000 inhabitants 9 54 218

6
The average is based on 10 devices (Sun OraLine and the Ultimed Salivascreen VI), yielding only slightly
lower net benefits and BC ratios than averages based on the specified 8 devices.

44
Table 15: 300% enforcement increase (tripling) - Drugs (average device) (100% alcohol).

Netherlands Belgium Finland

Annual benefits 14,258,846 33,886,725 5,170,513

Annual costs 1,038,106 6,654,135 6,505,710

Net benefits 13,220,740 27,232,590 -1,335,197

BC ratio 13.74 5.09 0.79

Simulated BC ratio 12.79 4.90 0.74

simulated St. dev. 4.28 1.70 0.29

simulated skewness 6.73 9.23 7.47

simulated kurtosis 7.23 2.78 0.37

Costs per convicted 4,303 4,006 4,120

Costs per prevented fatality 582,202 1,569,963 8,703,885

Tests per 100,000 inhabitants 18 108 436

Table 16: 1000% enforcement increase (tenfold increase) - Drugs (average device).

Netherlands Belgium Finland

Annual benefits 22,822,067 54,279,481 8,102,222

Annual costs 4,524,467 29,850,057 29,226,691

Net benefits 18,297,599 24,429,424 -21,124,469

BC ratio 5.04 1.82 0.28

Simulated BC ratio 4.70 1.75 0.25

simulated St. dev. 2.67 0.98 0.13

simulated skewness 17.46 9.33 6.15

simulated kurtosis 2.68 0.99 0.12

Costs per convicted 4,168 3,994 4,113

Costs per prevented fatality 1,584,219 4,391,997 24,415,003

Tests per 100,000 inhabitants 61 360 1.455

The indication from Table 14, 15 and 16 is that increased drug control is most profitable
for the Netherlands, and least profitable for Finland. This is logical in terms of baseline
enforcement level, since in Finland the drug enforcement level already is 25 times
higher than in the Netherlands. In the Netherlands even larger increase might be cost
efficient, since the estimated BC ratio is well above 1.5 even for a tenfold increase in
enforcement.

45
When we take into account the different qualities of the drug screening devices, the
indication is that the choice of device does make a difference. The quality differences
have been assessed in DRUID in D3.1.1 and in D3.2.2. If these quality differences, in
addition to affecting the costs (e.g., screening costs and laboratory costs), also are
carried forward to the deterrence effect, affecting benefits, the results of the CBA will
clearly depend on the quality of the selected drug screening device. Figure 7 shows the
difference in BC ratio between the three best rated devices (“above average”), the
three in the middle (“average”), and the two worst rated devices (“below average”).

Figure 7: BC ratios for three quality groups of devices; tripling in enforcement.

The final conclusion is that increased drug driving enforcement based on the roadside
saliva screening is potentially beneficial, particularly for countries which currently have
a low enforcement level. But if the public sector decides to decrease drunk driving
enforcement for the sake of financing increased drug driving enforcement (for a given
budget), the net benefits of police enforcement will decrease (assuming that drunk
driving will increase).

46
2.2. Classification

2.2.1. Objectives
The overall goals of WP4 (Classification) were:
• To review the existing classification and labeling systems regarding medicinal drugs
concerning their impact on driving performance.
• To define the criteria and the methodology on the establishment of European
classification and labeling system of medicinal drugs concerning their impact on
driving performance.
• To develop a methodology to continuously update classification/categorisation
system and labeling system on medicinal drugs concerning their impact on driving
performance.
• To classify the relevant therapeutic groups of medicines available on the European
market according to the classification system developed in the project.
• To ppropose patient‐oriented information on driving for each medicine and propose
appropriate labeling systems regarding medicines and driving

The WP4 activities were planned to be implemented in close collaboration with the
relevant EC Directorates and European agencies (like EMA). WP4 did benefit from
previous experiences at national level (Belgium, France, The Netherlands, Slovenia,
Spain) and from cooperation with international scientific panels (ICADTS, DG MOVE
Working group on alcohol, drugs, medicines and driving, etc.).

WP4 outputs are aimed at physicians/pharmacists and other healthcare professionals,

patients who drive taking medicines, regulatory agencies (European Medicines Agency
(EMA) and national regulatory agencies) and healthcare provider services.

2.2.2. Methodology
To collect the necessary information on already existing classification systems,
questionnaires were distributed to all institutions involved in DRUID and to some
entities outside the DRUID consortium. Additionally internet searches were performed.
To achieve comparability between the systems, the official ATC system maintained by
the WHO Collaborating Centre for Drug Statistics Methodology was used as a base
system.
For the development of input for a European categorisation system it has been decided
to address the Pharmacovigilance Working Party of the Committee for Medicinal
Products for Human Use (CHPM). Together with WP 4 partners three small-scale
invitational workshops have been organised in 2008 in which representatives of
regulatory agencies from nine Member States participated. Based on their discussions
recommendations were derived to develop and implement a European categorisation
system.
During the process of classifying medicines according to the categorization system
developed within DRUID a methodology was used, which included several steps of
evaluation taken into account the conditions of use of a medicine on the European
Union market:

47
1. Pharmacodynamic and pharmacokinetic data
2. Pharmacovigilance data (including prevalence of unwanted effects reported in the
SmPC)
3. Experimental and epidemiological data
4. Additional data derived from the Patient Information Leaflet (PIL) and existing
categorization systems
5. Synthesis

Basically conditions of use of a medicine, pharmacodynamics, pharmacokinetic data,

and pharmacovigilance data (including prevalence of unwanted effects) were derived
from the SmPC, while section 3 was based on a scientific literature search. An
additional step consisted of reviewing section 4.7 of the SmPC “Effects on ability to
drive and use machines” and the PIL section on “driving and using machines” as well
as reviewing the previous categorizations (if available) of a medicine in Belgium,
France, Spain as well as to the ICADTS list. After evaluating all available data, a
provisional category was assigned to each active substance. The provisional category
was proposed and discussed at WP4 meetings where a final and definitive category
was assigned and approved by all WP4 partners.
The existing methodology on DRUID categorization on medicines and driving, allows, if
new evidences emerge, to re-categorise the medicine or confirm the previous
categorisation following the same 5 step process.

2.2.3. Results

2.2.3.1. Review of existing classification efforts

In total, 16 systems were found (Table 17). Some of these systems are no true
categorization systems: Germany directly reproduced ratings from Wolschrijn, and 5
systems have not defined categories (Greece, the Netherlands, Norway, Denmark and
Finland). Only one true categorization system also included warning labels (France II).
Clear relations can be seen between different systems. All categorizations (except
Portugal) are linked to Wolschrijn. With regard to systems structure, the largest
differences concern the number and descriptions of categories. The list by Wolschrijn
included 7 categories. At first, the categories were copied (Belgium), but later on the
categories were summarized and only three categories remained (Spain I). The most
recent and extensive lists (France II and ICADTS) have maintained these three
categories, but have added practical guidelines for patient and physician. One list
(ICADTS) introduced a calibration to BAC levels.
Although different categorisation systems are currently available across Europe, it is
important to point out that none of these classifications have clearly described or
published the criteria for the establishment of a categorisation system for potentially
impairing medications nor were officially adopted at European level.

48
Table 17: Comparison of classification and labeling systems.
Number of
Number of Warning
medicinal Classification Legal
categories label
drugs
Wolschrijn 572 X 7
Germany 406 Scale values
Belgium 182 X 7
Spain I (DGT/UVa) 363 X 3
Spain II (semFYC/UVa) 395 X 4
Spain X
France I (CERMT) 508 X 4
France II (official) 311 X 3 X X
ICADTS 389 X 3
Portugal 241 X 5
Greece I (legal) 89 NA X
Greece II (monographs) 92 NA
The Netherlands 156 NA X
Norway 87 NA X
Denmark 83 NA X
Finland 68 NA X

2.2.3.2. The establishment of criteria for a European

categorization system
After reviewing the most significant of existing categorisation systems in Europe, a
critical discussion has been held to explain the need for such a categorisation system.
It is clear that such system has to address the needs of health care professionals, drug
regulatory agencies, drug manufacturers and patients. Patients need clear warnings
and symbols to use their medicines in the best (and safest) way. Developments in
France show clearly that a multi-level categorisation system is better in showing
difference between the least and most impairing medicine within one therapeutic class
and that warning labels are needed to guide patients in deciding about the use of the
medicine.
Together with WP 4 Partners three small-scale invitational workshops have been
organised in 2008 in which representatives of regulatory agencies from nine Member
States participated. Based on their discussions a first step to harmonize categorisation
systems could be achieved: the adoption of the Guidelines for the Summary of Product
Characteristics in September 2009 (which is valid from 1st of May 2010), in which
categories a) no or negligible influence, b) minor, c) moderate influence, and d) major
influence on driving fitness are specified with some important guidance in special
circumstances.
This adoption is in-line with the classification and labeling system developed by the
DRUID WP4 expert group (Table 18).

49
Table 18: DRUID Categorization system for medicines and driving.

Information for physicians and pharmacists Warning for patients (with

warning symbols and standard
descriptions per country)

Description of categories with Information on how to advise

levels of impairment their patients

Category 0 Confirm that the medicine will be [no warning needed]

safe for driving, provided that
combinations with alcohol and
Presumed to be safe or unlikely other psychotropic medicines are
to produce an effect on fitness to excluded.
drive.

Category 1 Inform the patient that impairing Warning level 1

side effects may occur especially
during the first days and that
Likely to produce minor adverse have a negative influence on Do not drive without having read
effects on fitness to drive. his/her driving ability. Give the the relevant section on driving
patient the advice not to drive if impairment in the package insert.
these side effects occur.

Category 2 Inform the patient about the Warning level 2

possible impairing side effects
and the negative influence on
Likely to produce moderate his/her driving ability. Advise the Do not drive without advice of a
adverse effect on fitness to drive. patient not to drive during the health care professional. Read
first few days of the treatment. If the relevant sections on driving
possible prescribe a safer impairment in the package insert
medicine, if acceptable by the before consulting the physician
patient. or pharmacist

Category 3 Inform the patient about the Warning level 3

possible impairing side effects
and the negative influence on
Likely to produce severe effects his/her driving ability. Urgently Do not drive. Seek medical
on fitness to drive or presumed advise the patient not to drive. advice after a period of treatment
to be potentially dangerous. Consider prescribing a safer about the conditions to restart
medicine, if acceptable by the driving again.
patient.

* The assigned categories relate to the acute or first time use of the medicine (at the start of treatment)

The DRUID categorisation system should also be used as a tool to motivate health
care professionals to provide patients with clear information, communicate to patients
the risk associated with driving under the influence of medicines, and start HCP-patient
discussion leading to both safer prescriptions and the patient’s conscious decision
whether to drive or not . From the patient point of view, this classification could play an

50
active role in helping them to be involved along the decision-making process, to
understand the hazards of some medications to road safety, and to remind them to use
caution while driving until their individual responses to the therapy have been well
established.
In the last two years (2010 and 2011) the WP4 results were discussed with the
Pharmacovigilance Working Party, resulting in consensus based on the following
compromise. Currently national approaches differ substantially: from France at one end
of the extreme (with three-level pictogram labeling) to Sweden at the other end where
the pictogram was replaced with a generic warning in the patient leaflet. The
consensus was reached that a basic two-level framework would be developed as the
basis for warnings to patients in the Patient Information Leaflet. For medicines without
a potential relevant influence on driving (no or negligible, or minor influence) and for
medicines with a potential relevant influence on driving (moderate influence, or major
influence) the wording has been proposed.
Emphasis is made on improving information related to effects on driving in the Patient
Information Leaflet by simple and patient-centred directions. Therefore collaborative
efforts of DRUID experts, members of the Pharmacovigilance Working Party and other
relevant institutions are recommended, preferably with support of EC (DG SANCO, DG
MOVE). The development of supplementary information for patients (e.g. warning
levels and pictograms) and health care professionals (e.g. prescribing and dispensing
guidelines) should be guided by using DRUID results (D4.2.1, D4.3.1, D7.3.2 and
D7.4.2.) as well as experience in EU Member States.
It is clear that the criteria for a European categorisation system for medicines and
driving should be established based on the consensus among all relevant
stakeholders. It is suggested that European regulatory authorities will be informed
about the DRUID-developed categorisation process, discuss and reach consensus on
the criteria hereby proposed, and carry out special efforts to implement the system at
both international and national level, taking into account country peculiarities. Since the
categorisation needs a constant revision, it is recommended to establish an expert
working group on drugs and driving in order to keep the system functional, up-to-date,
and reliable.

2.2.3.3. Classification/categorization, labeling and patient

oriented information for the relevant therapeutic groups
of medicines available on the European Union market.
The categorisation was first performed according to the active substance (referring to
the ATC classification). However, it is necessary to take into account other factors, e.g.
the excipients (excipient with a notorious effect), the route of administration, the
conditions of use and the general classification for supply (prescription only, OTC or
self-medication).
After evaluating drug’s pharmacodynamic, pharmacokinetic, preclinical and clinical
pharmacological (clinical trials, pharmacovigilance) studies, epidemiological and
additional data, the conclusion on the drug's effect is made integrating three
parameters: likelihood, frequency and intensity.
Fact Sheets were produced for N01-N07 (nervous system) and R06 (respiratory
system - antihistamines) ATC groups of medicines. Each Fact Sheet contains
information on: source of information, presentations, indications, posology and method
of administration, pharmacodynamic and pharmacokinetic properties, possible side-
effects related to driving, Summary of Product Characteristics (SmPC) section 4.7,
studies on psychomotor performance and risk studies, current categorization in some

51
EU countries, proposed DRUID categorization, information for the patient, and place
and date of agreement by the DRUID WP4 members.
In total 3,054 medicines from the above ATC groups were considered by DRUID team
(see Figure 8). Of these 3,054 medicines, 1,513 have not been categorized, because
they are not available on the European Union market. The distribution of the remaining
1,541 categorized medicines was as follows: Category 0 – 50,3%, Category I – 26%,
Category II – 11,2%, Category III – 5,8%, Multiple category – 4,4% and the Depending
on the medicine in combination 2,3%.
The following ATC groups were not categorized: G (Genito urinary system and sex
hormones), H (Systemic hormonal preparations, excluding sex hormones and insulins),
J (Antiinfectives for systemic use), L (Antineoplastic and immunomodulating agents), P
(Antiparasitic products, insecticides and repellents) and V (various).

60%
50,3%
50%

40%

30% 26,0%

20%
11,2%
10% 5,8% 4,4%
2,3%
0%
Category 0 Category I Category II Category III Multiple Depending on
Category the medicine
in
combination

Figure 8: Percentage of medicines categorized within each DRUID category.

DRUID Categorisation: Four categories were proposed regarding the possible effect
of the medicine on fitness to drive.
Category 0: Presumed to be safe on fitness to drive.
Category I: Likely to produce minor adverse effects on fitness to drive.
Category II: Likely to produce moderate adverse effect on fitness to drive.
Category III: Likely to produce severe effects on fitness to drive or presumed to be
potentially dangerous.

Multiple categories: This appeared when a medicine can be included in more than
one category. There can be several reasons for this: In most cases, the different
categorization depended on the route of administration (topical, oral, parenteral, etc). In
the case of some medicines in special ophthalmological preparations (S01), the
different categorization depended on the presentation form of the medication

52
(aqueous-vehicle, cream, drops or ointment, etc.), which is related with the duration of
its influence. In one case, codeine, categorization was based on the dose of codeine
base administered. For two hypnotics, zolpidem and zaleplon, categorization was
based on the time after the medication was taken.
Depending on medicines in combination: This was stated when the categorization
depended on another medicine combined with the one under evaluation.
Table 19 shows the DRUID categorisation of the medicines in the ATC groups A, B, C,
D, M, N, R and S.

Table 19: Number of medicines categorized by ATC groups.

DRUID Categorisation
Not available on
Not evaluated.

combination

TOTAL
medicine in
Depending
categories
ATC GROUP

Multiple
EU market

on the
III
0

II
I

A - ALIMENTARY 243 234 69 8 1 4 4 563

TRACT AND
METABOLISM

B - BLOOD AND 2
BLOOD FORMING 86 135 1 1 225
ORGANS

C- 246 90 200 11 1 548

CARDIOVASCULAR
SYSTEM

D- 156 192 1 4 353

DERMATOLOGICALS

M - MUSCULO- 88 22 44 28 15 197
SKELETAL SYSTEM

N - NERVOUS 346 9 30 86 53 36 560

SYSTEM

R - RESPIRATORY 195 62 24 32 10 5 14 342

SYSTEM

S - SENSORY 153 31 31 6 11 18 16 266

ORGANS

TOTAL 1513 775 400 172 90 68 36 3054

53
2.3. Rehabilitation

2.3.1. Objectives
DRUID WP 5 dealt with rehabilitation of substance impaired drivers. The overall aim of
WP5 was to increase relevant knowledge and to elaborate common European
standards for rehabilitation measures for drivers under the influence of alcohol (DUI) or
illicit drugs (DUID).
The first task of DRUID WP 5 aimed at identifying different types of DUI/DUID
offenders and options for DUI/DUID assessment including different available
approaches. State of the art review of currently existing rehabilitation programmes in-
and outside Europe was implemented. Rehabilitation programmes and measures were
analysed concerning scientific evidences relevant for road safety.
In the second task of DRUID WP 5 the following main objectives were pursued:
1. In-depth analysis on reasons for recidivism of drivers under the influence of
psychoactive substances who participated in driver rehabilitation programmes.
2. Analysis of existing quality management (QM) systems established along with
driver rehabilitation schemes.
3. Development of an evaluation instrument for best practices.
4. Validation of existing driver rehabilitation schemes.

2.3.2. Methodology
Comprehensive reviews of the international literature were carried out concerning the
following topics:
• Identification of different types of DUI/DUID offenders
• Existing DUI/DUID assessment procedures
• Existing DUI/DUID rehabilitation measures
• Addiction treatment and options for dependent DUI/DUID offenders
The investigation of driver rehabilitation measures implemented and applied in Europe
at present was done by means of a questionnaire survey.
The in-depth analysis on reasons for recidivism was realized applying a case-control
study design whereby recidivists (i.e. drivers with a BAC of 1.6g/L or more, who had
participated in a driver rehabilitation course, yet must take part in another driver
rehabilitation course due to a subsequent DUI offence within a trial period) were
compared with a matched control group of non-recidivists (i.e. drivers with a BAC of
1.6g/L or more, who had participated in a driver rehabilitation course and made no
offence during the following trial period) regarding their traffic psychological driver
assessment data. This analysis was supplemented by an analysis of change processes
and components in driver rehabilitation courses by means of a questionnaire survey
addressed to driver rehabilitation courses participants.
The existing QM systems for rehabilitation measures were displayed using the
information derived from country reports. These were developed in several steps and in
collaboration with country experts. In a first step, literature and internet research served
as the basis to define quality management criteria. In a second step, country experts

54
were asked to fill out a tailored questionnaire. After filing reports for each country, the
experts reviewed the report for their country in order to validate the presented data.
Thereby a decision-tree was developed (D5.2.3).
Aiming to introduce a common integrative evaluation methodology, WP5 partners
developed the Driver Rehabilitation Evaluation Tool (DRET). Existing evaluation tools
were reviewed and discussed within the team and with other experts from several
disciplines relevant for or linked to driver rehabilitation.
The validation of existing driver rehabilitation measures was carried out by a
compatibility assessment study. Applying DRET a standard of a good practice was
developed, against which existing European driver rehabilitation programmes were
compared. The assessments were carried out on a quantitative base by the WP5
research team.

2.3.3. Results
The literature review concerning different offender types revealed that the entire group
of DUI/DUID offenders seems to be heterogeneous. Following characteristics of
DUI/DUID offenders were identified:
• Socio-demographic variables: male gender; young age; lower educational or
professional level; lower socio-economic status; single or separated marital status.
• Traffic related variables: prior traffic offence records.
• Consumption habits: heavy to problematic substance use (major risk factors); first
offenders are often moderate drinkers; co-morbidity of substance use problems
with other clinical disorders.
• Personality traits: e.g. sensation seeking or aggression; general risky life style; low
self-control, poor coping styles.
• Decision making processes: deviant attitudes; poor knowledge; low risk
perceptions; influence of the social surrounding, group norms and expectations.

Identified characteristics of the high risk group of DUI/DUID recidivists were:

• Socio-demographic variables: male gender; young age; lower educational level.
• Traffic related variables: the higher the amount of prior records, the higher the
recidivism risk.
The review of assessment procedures shows that DUI/DUID assessments are carried
out to evaluate fitness to drive and to assign offenders to rehabilitation programs. The
context determines the selection of tools and the whole procedure. In contrast to the
assessment for rehabilitation assignment, the legal context of a fitness to drive
assessment requires particularly a high specificity and thus an integrated and
comprehensive approach. Objective parameters like BAC or prior offences can serve
as assignment criteria for deeper assessments or even directly for specific driver
rehabilitation. In Europe DUI/DUID assessment is primarily carried out in the frame of
the fitness to drive decision. It is mostly a multidisciplinary approach, covering medical,
psychological and social aspects.
Driver rehabilitation programmes for DUI offenders are based on a rather long term
tradition in development and practical application in Europe. They are also the basis for
the later developed programmes for DUID offenders. WP5 found out that there is no
uniformity in Europe regarding the implementation and application of DUI/DUID

55
rehabilitation. In the five selected European countries (Austria, Belgium, France,
Germany and Hungary) national regulations on different aspects of DUI/DUID
rehabilitation are established. Regarding the access to DUI/DUID rehabilitation
programmes, the literature shows that European countries use very different
approaches, ranging from voluntary, over recommended, up to obligatory participation.
The review of DUI/DUID driver rehabilitation effectiveness identifies 61 studies on the
topic. European standard group intervention programmes for DUI offenders show an
average recidivism reduction rate of 45.5% (36 studies and 2 reviews) although a large
variation of recidivism reduction rates was observed (15% - 71%).
Based on the provider questionnaire survey, a comprehensive picture of the actual
situation was drawn: At least 47 providers, mainly non-governmental, private
organisations in 12 European countries (Austria, Belgium, France, Germany, Hungary,
Italy, The Netherlands, Poland, Portugal, Sweden, Switzerland, and United Kingdom)
carry out driver rehabilitation services on a regular basis. In total, 87 driver
rehabilitation programmes are in use, thereby 53 for DUI offenders, 21 for DUID
offenders and 13 for mixed groups (DUI/DUID/other traffic offenders). All above
mentioned countries offer programmes for DUI offenders, in addition four Member
States (Austria, Belgium, Germany, and Portugal) for DUID offenders. The vast
majority of driver rehabilitation providers do not offer treatment programmes for
substance dependent offenders.
The main results of the literature review of addiction treatment and options for
dependent DUI/DUID offenders can be summarized as follows:
• Psychosocial treatments of alcohol and drug dependent patients are well
established interventions to support the maintenance of abstinence and to
decrease the scope and frequency of alcohol and drug consumption.
• No superior general strategy could be identified.
• It is important to consider characteristics of the patient, predominant symptoms of
the dependence, and also motivation aspects while matching patients and
treatment approaches. A combination of different treatment strategies provides the
advantage of simultaneously addressing different factors and levels of influence.
• The addiction-specific approach is a fundamental element within the rehabilitation
of dependent DUI/DUID offenders.

For the in-depth analysis on recidivism reasons a sample of n=303 recidivists and a
matched control-group of n=303 non-recidivists were analysed. Group comparisons on
univariate level reveal 20 significant differences between study and control group. On
multivariate level, six of them show predictive value in a regression analysis
additionally.
Based on the entire results, the following risk profile of DUI offenders who might not
profit from a driver rehabilitation course can be deduced:
• Having high BAC levels at the current offence or refusing the breath test
• Having additional prior drink-driving or already several DUI offences (i.e. not the
first one) and consequently having longer suspension periods of driving licence
• Having a habitual drinking pattern in the past and in spite of past or current
abstinence periods having an increased alcohol tolerance, thus having also felt
less impaired at the actual DUI offence
• Denying or not having any alcohol related health problems, being a smoker and
being less aware of own health status

56
• Showing an unrealistic self-perception and less self-reflection whereby alcohol
related risks in traffic are underestimated
• Not living in a partnership
• Being assessed as having an enhanced re-offence risk by a qualified expert (traffic
psychologist).
The results of the participant questionnaire survey indicate that driver rehabilitation
participants feel that such programmes provide strong support for their cognitive and
behavioural change processes. The findings suggest that participants feel encouraged
to establish new behavioural goals and the commitment to stick to them. At the same
time, the participants’ ratings emphasise the important role of the course leader in
encouraging such changes. About 95% of all European DUI offenders who participated
in this study assess the driver rehabilitation course as good or very good. Only about
2% rate the course as bad or very bad (about 3% gave no answer).
The analysis of the existing QM systems revealed that QM systems in driver
rehabilitation schemes are expedient to create transparency by setting rules and
instructions (standards). The compliance with the standards is a necessary
precondition for winning a trust of all stakeholders: legislators, authorities, individuals
and the public. The decision-tree developed in DRUID may serve as a tool to evaluate
the established QM system on a national, provider and programme level. It could also
be used by international and national legislators and providers while implementing a
QM system in the domain of driver rehabilitation.
The Driver Rehabilitation Evaluation Tool (DRET), developed in WP5, covers the main
technical issues of driver rehabilitation measures. 28 items have to be evaluated in
total whereby 11 items focus on (national) driver rehabilitation system issues and 17
items on single programme level. Evaluation of driver rehabilitation system and single
programme is separated, DRET-L refers to the first and DRET-P to the second one. In
order to assess single DRET items against the DRUID WP5 standards, additional
studies were implemented. The evaluation was carried out by means of a categorical
answering mode with four alternatives (yes, partly yes, no, don’t know) supported by a
colour system. In principle, answering could be done either in an electronic or paper-
pencil mode.
In total, 90 driver rehabilitation programmes from 12 countries (Austria, Belgium,
France, Germany, Hungary, Italy, Netherlands, Poland, Portugal, Sweden, United
Kingdom and Switzerland) were included in the validation study, thereby n=55 for DUI
offenders, n=21 for DUID offenders and n=14 for mixed offender groups (alcohol and
drug offenders or both mixed with general traffic offenders). This set of driver
rehabilitation programmes for substance impaired drivers provides a comprehensive
picture on actually existing driver rehabilitation programmes in Europe, although it does
not claim to be fully representative. Out of the assessed 23 driver rehabilitation items
crucial for successful driver rehabilitation interventions, 9 were completely fulfilled and
further 9 at least partly by the assessed European driver rehabilitation programmes.
Only 5 driver rehabilitation relevant topics showed a low compliance with the compared
DRUID WP5 standard, namely existence of a national QM body, definition of operative
tasks of QM body, multidisciplinary approach in case of prior driver assessment,
objective, valid, reliable tools in driver assessment and evaluation of driver
rehabilitation programmes. An additional comparison of driver rehabilitation
programmes for DUI and DUID offenders revealed some differences in the fulfilment of
the legal frame conditions as well as in driver rehabilitation content related
requirements. Thereby, except of the evaluation requirement, DUID programmes
comply better with the WP5 standard than the programmes for the DUI offenders.

57
2.4. Withdrawal

2.4.1. Objectives
The DRUID Work Package 6 dealt with driving licence withdrawal and re-granting
legislation and practices. The aims of DRUID WP 6 were threefold:
1. To review the state-of-the-art on withdrawal of driving licences by collecting and
evaluating information on existing legislations and practices in various European
countries.
2. To assess effectiveness of these strategies, particularly with a focus on conditional
driving licence withdrawal.
3. To develop corresponding recommendations for road safety policy makers at the
national and EU level.

2.4.2. Methodology
In order to gain information about existing legal regulations and sanctioning practices, a
special DRUID questionnaire was designed. The questionnaire consisted of four parts,
each referring to psychoactive substances and driving. The first part dealt with
legislation in the different countries. The focus of the second part was on detection and
police enforcement. The third part concerned toxicology. The fourth part referred to the
sanctioning practice of administrative bodies and/or public prosecutors/courts.
The questionnaires were sent out to country experts. All 27 Member States and 3 non-
EU countries responded to the questionnaire. The responses were subsequently
entered into a database and finally resulted in a report for each participating country.
To ensure the quality of data, all country reports, except for Bulgaria, were once again
reviewed and approved by the country experts.
The analysis of the survey data revealed further research demands, e.g. due to the
various interpretations of the terms “driving ban” and “licence withdrawal” and a lack of
information on some specific issues, e.g. conditional licensing. Therefore, a second
survey among the European countries was carried out in order to get complete, reliable
and differentiated information about the use of the instrument “withdrawal” in case of
driving under impairment of psychoactive substances. Additional data on specific
withdrawal procedures in the European countries were collected.
The questionnaire for the second survey was developed by the Task 6.2 research
team. In order to provide a high standardization of the survey procedure, to ease and to
fasten answering of the questionnaire, a closed answering format was chosen
whenever possible. Nevertheless, an open answering format was included as well to
include answers other than the predefined ones or to give space for exact and further
detailed information or explanations about the national realities. The survey had an
electronic answering format and was sent out by e-mail to experts. Additionally, a
responsible research team member contacted experts via telephone to assist them by
answering the questionnaire.
In addition, findings from empirical primary studies on the general and special
deterrent/preventive impact of withdrawal were analysed. Results from non-empirical
literature were also considered. The search for criminological empirical primary studies
was carried out in journals and databases in order to find research work which deals
with the sanctioning of impaired driving (especially licence measures). Finally, the

58
reference lists of the examined empirical primary studies were checked for further
appropriate empirical research findings.
In a next step, results from other WPs regarding legal issues were summarised. To
supplement all findings, expert workshops were held in order to discuss and cross-
check the research results. Experts provided input on specific problem groups and
withdrawal concerning topics for which empirical results were insufficient. Two kinds of
expert workshops were conducted:
• Country expert workshops
These workshops aimed at getting feedback and additional input from national experts
regarding withdrawal and accompanying issues. Research outcomes of Task 6.2 were
presented and discussed considered from two sides: the legal realities in the European
countries and the scientific findings on the effectiveness of withdrawal and
accompanying measures. In order to make outputs comparable, national workshops
were conducted in each country according to the standard scheme, prepared and
agreed upon within the Task 6.2 research team. The target groups of national experts
to be involved were specified: Ministry of justice; courts, administrative authorities;
Ministry of health; rehabilitation system; police, Ministry of interior; Ministry of transport.
• International expert workshops
These workshops, conducted in close cooperation with the WP1, focussed on
withdrawal related solutions for specific problem groups, namely drivers in substitution
treatment and drivers in long-term medication treatment. The workshops aimed at
developing recommendations based on expert knowledge.

2.4.3. Results
The WP6 created a comprehensive database of the legal systems as well as practices
in European countries with respect to withdrawal of driving licenses as a consequence
of driving under influence of alcohol, illicit drugs or medicines. Besides, procedures and
practices of re-granting driving licences to drivers, from whom a licence has been
withdrawn due to impaired driving, were analysed.
The focus has been put on the following issues:
• Licensing system in general (withdrawal – renewal, special provisions for novice /
young / professional drivers)
• Alcohol / illicit drugs / medicines and driving legislation (zero tolerance or impaired
approach, specified levels, special laws for novice / young / professional drivers)
• Legal regulations for detection (testing at random or on the basis of suspicion;
regulations regarding testing of alcohol / illicit drugs / medicines)
• Different kinds of sanctions for impaired driving (e.g. criminal or administrative
penalty, fine, withdrawal, imprisonment)
The collected database provides reliable data for administrators, politicians and
researchers.
In most EU countries (24/30 countries or 80 %) driving under the influence of alcohol
leads to withdrawal of the driving licence. Thirteen out of 20 countries state a BAC
threshold of 0.5g/L or above. A considerable number of EU countries have a
withdrawal sanction for driving under the influence of both alcohol and drugs (19/30).
About half of the EU countries (12/30 countries) have withdrawal sanctions for all three
groups of psychoactive substances, i.e., alcohol, drugs and medicines. A typical
approach (2/3 of all countries) is a temporary withdrawal of a licence.

59
In 21 of the European countries, conditional withdrawal is not possible, but at least 10
countries provide this possibility. The conditions are mostly time related (e.g.
withdrawal is only effective during weekends). In a few cases, the conditions are
related to geographical area (distance between home and working
place/school/doctor's practice) or vehicle technology (alcohol interlocks). Other
conditions, such as driver's profession, driving licence categories or health reasons
were indicated as well. In 4 of the 10 countries which have implemented conditional
withdrawal some additional diagnostic information is required for the decision on
conditional withdrawal. This information can be obtained through medical or
psychological assessment.
Regarding conditional re-granting, the conditions are mostly time related, i.e. the
licence is re-granted only for a certain period of time or is limited to special situations,
e.g. time for driving to work, for driving kids to school or for medical consultations.
Conditions can also be related to the geographical area (e.g. distance between home
and working place/school/doctor’s practice) and to vehicle technology (e.g. alcohol
interlocks). Apart from these, other conditions can be imposed like addiction treatment,
regular screening of substances, driver rehabilitation. Some other conditions are: type
of driver (professional or non-professional), restriction to a certain speed limit, medical
assessment, proven abstinence for at least one year, stable change of the
consumption behaviour, attendance in a course for drink and drive offenders. In 11 of
the 13 countries which have implemented conditional re-granting, additional diagnostic
information is required for the decision if the driving licence can be re-granted
conditionally. The additional information is acquired by means of medical and/or
psychological assessment. In some countries, it depends on screening results (blood,
urine, hair) or a driving test carried out by the traffic authority’s physician.
Further on, it was inquired if there are any options to reduce the withdrawal period after
the driving licence has been withdrawn. 22 of the countries do not have any options to
reduce the withdrawal period after the driving licence has been withdrawn, while 9
countries have this possibility. Thereby, in some countries the withdrawal period can be
reduced after participation in a treatment, rehabilitation or ignition interlock program.
The withdrawal period can also be reduced for some other reasons, such as personal,
professional or social circumstances, the character of the applying offender, his
conduct after conviction or the nature of the offence.
Altogether, 19 countries have more severe sanctions for recidivists driving under the
influence of illicit drugs, while 11 countries do not differentiate between first time and
persistent offenders. More severe sanctions for recidivists comprise imprisonment, fine,
driver rehabilitation and licence withdrawal.
Generally, there are 9 different conditions for re-granting a driving licence in Europe
after it has been withdrawn for driving under the influence of alcohol, illicit drugs and
medicines:
a. Medical assessment
b. Psychological assessment
c. Screening for substance markers in blood/urine/hair
d. Driver rehabilitation
e. Treatment programme
f. Theoretical driving lessons
g. Practical driving lessons
h. Theoretical driving test
i. Practical driving test

60
Almost all European countries have police control strategies concerning driving under
the influence of psychoactive substances (26/30), while only 4 (Czech Republic,
Denmark, United Kingdom and Ireland) do not have systematic police controls. All 27
Member States use a systematic approach in relation to specific locations or specific
events, weekdays or daytime.
The outcomes of the second survey showed that 22 out of 31 European countries
distinguish between licence withdrawal and driving ban, while 9 do not.
Regarding withdrawal periods, sizable differences between the European countries can
be observed in case of alcohol. The withdrawal periods in most countries depend on
certain BAC limits which differ considerably between the countries. Accidents under the
influence of alcohol as well as recidivism, i.e. another DUI offence within a certain time
period, entail in some countries an extension of withdrawal period.
Only 22 out of 31 countries informed DRUID researchers on withdrawal periods
regarding DUID. Thereby, 8 countries indicated periods up to six months. Withdrawal
periods between six and twelve months exist in 4 European countries. Longer periods,
namely between twelve and 24 months and above 24 months were quoted in 5
countries.
13 out of 31 European countries indicated requirements for patients in substitution
maintenance treatment who would like to own or regain a driving licence. Thereby, the
requirements are often connected with a fitness to drive examination. In some
countries regular medical checks are necessary. 18 of the countries either do not have
any criteria or do not allow patients in substitution therapy to drive at all. One country
does not apply licence withdrawal or driving ban in this case.
The existence of regulations for patients under long term treatment was confirmed by 6
out of 31 European countries, 25 of the responding countries indicated to have no
special regulations for this group. In all of the countries with regulations, a fitness to
drive examination (medical or medical and psychological) is required.
The analysis of the empirical literature revealed that the general deterrent approach
includes three main factors which often overlap:
• punishment certainty (including risk of detection and probability of sanctioning in
the narrow sense),
• punishment severity (including the legal threat of sanctions and the
judicial/administrative practice of imposition),
• punishment celerity.
The perceived risk of detection (not the real risk) is the main general deterrent factor. In
most cases, the first one is higher than the second one, which is mainly influenced by
the intensity of the media coverage. Laws without discretion for the authorities are
major factors to increase the probability of sanctioning in the narrow sense. But this
element must be accompanied by a high level of detection risk to achieve significant
levels of general deterrence. Empirical primary findings support the statement that an
increase of the sanctioning certainty is much more effective than an enhancement of
the sanctioning severity. The punishment celerity is - besides the punishment certainty
- a further important deterrent factor.
The results of driving licences measures can be summarised as follows (Table 20 and
Table 21):

61
Table 20: Results on general deterrence.

Topics Results on general deterrence

General mode of Driving licence measures show measurable closer correlations between
functioning their imposition and the increase of the general deterrent level than
other sanctions (e. g.: imprisonment or fines)
Three main factors (in accordance to their importance for the
achievement of law-abiding behaviour): sanction certainty (perceived
risk most important), sanction celerity and severity

Administrative vs. Advantages of administrative procedure due to sanction celerity and

criminal procedure sanction certainty (especially in case of per se legislation)
Disadvantages of criminal procedure due to huge differences in the
severity of the imposed sanctions

Duration of driving No significant impact of short-term driving ban (12 – 24 hours)

licence measures
Significant deterrent impact of driving licence measures with a duration
between three and six months

Table 21: Results on special deterrence.

Topics Results on special deterrence

General mode of Driving licence measures show measurable closer correlations between
functioning their imposition and the increase of the general deterrent level than
other sanctions (e. g.: imprisonment or fines)
Three main factors (in accordance to their importance for the
achievement of law-abiding behaviour): sanction certainty (perceived
risk most important), sanction celerity and severity

Administrative vs. In many cases: Significant reduction of recidivism rates after

criminal procedure implementation of administrative driving ban laws for both first-offenders
and recidivists
In many cases: Significant impact of withdrawal imposition in an
administrative procedure combined with per se laws

Re-granting Deterrent effect of withdrawal might be highly determined by the re-

granting procedure (promising results can be achieved if a medical-
psychological examination is included)
Some empirical hints for a lack of increased recidivism rates in case of
reduced withdrawal periods after medical-psychological assessment
followed by an educative/rehabilitative measure

Duration of driving Mostly: Significant reductions of recidivism rates in case of driving

license measures licence measures with a duration between three and twelve months.
From twelve months upward, an increase of the recidivism rates can be
observed in a lot of cases.
Significant increase of offences (driving while suspended) in case of
long duration of withdrawal, worst effect from three years upwards.
Higher compliance rates in case of shorter durations.

62
Conditions, Better special deterrent outcome of restricted licence (e. g.: permission
reductions, to drive to work and treatment related activities) combined with
restrictions treatment/rehabilitation compared to the sole imposition of driving
licence measures
Some empirical studies show the worst effects in case of the sole
imposition of restricted licence, valid for both first- and especially
multiple-offenders
Close monitoring of offenders is important

Rehabilitation In case of addiction or misuse, punitive sanctions alone are generally

not effectively deterrent
Best effects can be achieved by a combination of driving licence
measures with treatment/rehabilitation
Treatment/rehabilitation has to complement, but not to substitute licence
measures
In case of combination of treatment/rehabilitation and licence measures,
withdrawal periods can be shortened or replaced by licence restrictions

Within the international expert workshop on driving licence and substitution

maintenance treatment consensus was achieved regarding the following statements:
• It is not adequate to generate over-regulation for a small group which is of minor
importance for road safety.
• There should be no basic difference made between patients in substitution
treatment and patients in other medicinal treatments.
• In comparison to other patients, patients in substitution treatment show a high
compliance.
• An early integration and the option of conditional licence are important and support
the rehabilitation progress.
• A model of conditional licence with regular follow-up-controls is recommendable.
• The development of a rigid assessment and evaluation model is inadequate. It
should be individually adaptable.
• Abstinence of relevant parallel consumption of other drugs (besides alcohol) is
important. Alcohol-driving-abstinence (separation of drinking and driving) is
necessarily required as well as the ability to separate parallel consumption of other
drugs and driving.
• Alcohol addiction or other substance addictions are exclusion criteria.
• Regular checks for other drugs are indispensable. Hair analysis should be carried
out every six months or urine screenings more frequently. Immunological tests are
sufficient. Confirmation analyses are necessary if patients deny positive results.
• “Psycho-social integration” is not a reasonable criterion, psychiatric diagnostic
preferred instead. Hence, a successfully completed therapy is not a reason for
considering the social integration as restored.
• Substitution substance (Methadone vs. Buprenorphine vs. Morphine) and the
height of the daily dosage in milligrams aren’t criteria to judge on fitness to drive.
The substance and dosage being adequate for each client are crucial issues.

63
• A follow-up-period after tapering the dose (treatment end) should be defined
individually.
• For older long-term patients, specific regulations should be implemented as they
often consume prescribed medicines in parallel.
• In cases of the intake of other disease-related prescribed medicine, tests of the
cognitive performance are recommended to assess the fitness to drive.
• Patients treated with diamorphine (pharmacological heroin) are not fit to drive due
to massive secondary impairing disorders as a result of long-term heroin use and
heavy addiction (which are necessary requirements for being treated with
diamorphine). Additionally, patients in diamorphine treatment need to inject the
substance twice a day. As a consequence, they are left in a slight flush until first
signs of withdrawal symptoms appear.
No consensus was found regarding the statement that patients in substitution treatment
are basically not fit to drive Group 2 vehicles (C1, CE, C1E, D, D1, DE, D1E and
passenger transport).
The experts of the international workshop on driving licensing issues and long-term
medication found consensus on the following statements:
• No approach solely based on substance classes is needed.
• Impairment is the key for sanctioning. Patients should be adequately informed on
possible impairing effects and how to recognize them (leaflets, consultation by
physician or pharmacist).
• A model of conditional licensing is only recommendable after an incident which
was sanctioned by full withdrawal.
• A primary preventive approach with a reporting obligation of physicians is
disproportionate.
• Primary preventive licensing measures to be introduced only due to a prescription
of psychoactive medicine are not necessary.
• A medical expertise / assessment should be ordered case-by-case. In this frame, an
individual solution, e.g. regarding certain conditions, can be developed. The
expertise should include the medication at hand and also personal issues.

2.5. Dissemination

2.5.1. Objectives
The main goals of WP7 were the development of guidelines for healthcare
professionals and on risk communication aimed at different target groups based on
DRUID outcomes.
The following objectives were pursued in DRUID WP7:
• State-of-the-art review of existing campaigns concerning psychoactive substances;
• Reflections on improvements of procedures for assessing fitness to drive;
• Development of prescribing and dispensing guidelines for physicians and
pharmacists to select the least impairing medicine within a therapeutic class and to
inform a patient meeting his/her needs;

64
• Evaluation of practice guidelines and protocols in every day medical and
pharmaceutical practice by focussing on different practice models, with and without
the application of Information and Communication Technologies (ICT), as well as
the evaluation of risk communication to patients;
• Development of information aimed at various target groups (general public, drivers
as patients, younger people, health care professionals and policy makers);
• Development of a strategy for risk communication to young drivers.

2.5.2. Methodology
Due to a variety of tasks performed in WP7, different methodological approaches were
required.

Overview over existing campaigns and their impact on road safety

A review was conducted on the state-of-the-art of existing information campaigns
regarding psychoactive substances, as well as the documented effectiveness of those
campaigns (D7.1.1). A questionnaire was sent out to all 37 DRUID partners to collect
information on public campaigns regarding driving under the influence of psychoactive
substances. The effects that were considered when collecting information on the
impact of the campaigns went from effects on the awareness of the campaign
(minimum effect) to effects on attitude and behaviour (maximum effect). Information
was also collected by using Internet (websites of relevant organisations, Google and
YouTube).

Guidelines and Professional Standards

A questionnaire survey among driving licensing authorities and experts was conducted
in 29 European countries (all EU member states, Switzerland, Norway) in order to
obtain better insights into the current situation in Europe concerning guidelines for
physicians on prescribing medicines with impact on driving performance and on
assessing fitness to drive (D7.2.1). In addition, existing guidelines for pharmacists on
advising patients while dispensing those medicines were considered (D7.2.2).

Evaluation and implementation of new technologies

The effectiveness of the implementation of developed protocols and guidelines on the
attitude, knowledge and reported behaviour of healthcare professionals’ (physicians,
pharmacists, nurses) in clinical practice were evaluated via two different approaches: i)
by using an integrated (ICT) tool (additional software integrated into the ICT software
used by professionals in daily practice; country specific development) and ii) by using a
non-integrated tool for presenting the protocols and guidelines (ICT tool developed
within the framework of the project).
The target populations were health care professionals in the primary care setting: i)
physicians (Belgium, Spain), ii) pharmacists (Belgium, the Netherlands, Spain) and iii)
Nurses (Spain). In addition, a “pure” control group was added to evaluate the
effectiveness of current practices with no DRUID-relevant information.
Participants were introduced to the tools/software(s) through a training scheme. Some
of the participants did not receive training (e.g. the integrated group of physicians
(SoSoeMe)). In addition, participants were informed about the DRUID guidelines

65
regarding driving and medicines intake. The time sequence involved a standard
procedure of recruitment, briefing, and consent. Participants filled in the pre-
questionnaire at the start of their training and a post-questionnaire after six months of
using the DRUID guidelines in their practice). They used the software during their daily
practice for either prescribing or dispensing medicines depending on the professional
groups they belonged to. After the testing period ended they filled in a post-
questionnaire in order to evaluate the effectiveness of the tool and the applied
guidelines (D7.4.1 und D7.4.2).

Main DRUID results to be communicated to different target groups and prototype

documents
The Deliverable D7.3.2 contains a summary of the main DRUID results. The authors of
D7.3.2 provided a describing text and a summarizing overview table per WP. These
texts and summaries were discussed in two rounds, first with the DRUID coordinator
and WP7 partners at various work sessions and secondly with WP leaders and authors
of main deliverables.
An on-line survey on the criteria for the design of prototype documents for information
regarding psychoactive substances and driving was conducted to assess the opinions
of experts in various fields (policy makers, physicians, pharmacists, researchers,
professionals working in the field of illicit drugs, etc.). The documents were developed
for various previously defined target populations (D7.3.1).

Effectiveness of pictograms in communicating risk to patients who drive under

the influence of medicines
Two studies using a 2x3 design were conducted to compare the effectiveness of two
pictograms (rating model and triangle model pictograms) in communicating risk
associated with driving impairing medicines to patients and to assess patients’ level of
understanding and intention to change driving behaviour when looking at various
pictograms. In the first study, the respondents (patients with a driving license visiting a
community pharmacy) were exposed to a pictogram (rating model or homologue
triangle model pictogram) and a category (category 1, 2 or 3). In this study, both
pictograms were accompanied by the same side-text (experiment 1). In the second
study the added value of the side-text was examined. Here, the respondents were
exposed to the rating model pictogram with or without side-text and again one of the
three risk categories (D7.3.2).

DRUID outcomes and risk communication to young drivers

In order to develop recommendations for appropriate media-based concepts for risk
communication on the base of the DRUID outcomes a workshop with media experts,
psychologists, social workers, police bodies and representatives from road safety
agencies and governmental institutions was organized. The recommendations of the
experts served as input for the subsequent formative evaluation. A representative
sample of 15 to 24-years olds in Germany was interviewed about (A) their personal
experiences and attitudes concerning driving under the influence of drugs, alcohol
and/or medicines, (B) the knowledge and motivational base for processing thematic
risk communication messages, and (C) specific issues in media use and preferences
for risk message contents and channels (D7.4.3).

66
2.5.3. Results
Overview over existing campaigns and their impact on road safety
A total of 75 campaigns were found, from 13 different countries. The majority of the
retrieved campaigns concerned driving under the influence of drugs and was aimed at
young people. Other possible target populations include the general public, physicians,
pharmacists, teachers, patients, drug users or other types of populations. Most of the
retrieved information campaigns were conducted through the mass media. Brochures
are the type of medium that is used most frequently, followed by posters, paper press,
websites, booklets, TV commercials, radio spots, leaflets, tutorials or another type of
medium. Most campaigns are organized by governmental organisations and road
safety organisations.
Information on the impact of the campaign was found for only 7 campaigns. All these
evaluations documented a positive outcome of a campaign. As only a few evaluations
were found, and these campaigns and their evaluations were performed in different
ways, it is not possible to draw conclusions concerning the association between the
design of the campaigns and their effectiveness. It can be concluded that more
evaluations should be performed on future campaigns concerning driving under the
influence of drugs and/or medicines, and that these evaluations should be made in a
uniform way and design in order to determine guidelines for developing future
campaigns (D7.1.1).

Guidelines and professional standards

Based on the feedback on the questionnaire, an overview of the current European
regulations and guidelines is presented. Concerning prescribing and dispensing of
psychotropic medicines, which might have an impact on the driving performance, it was
concluded that strict and binding regulations are the exception rather than the rule. The
compiled guidelines are typically recommendations, not regulations. The role,
responsibilities and tasks of physicians and pharmacists are not defined uniformly.
Despite the great diversification of recommendations in the different countries one can
deduct a common denominator. Physicians and pharmacists usually should give their
patients the most comprehensive and adequate advice on medicines and their effect
on driving performance. This includes a recommendation of not leaving the patient
alone with the decision whether to drive or not while using medicines.
In most cases physicians and pharmacists will not be made legally responsible in case
an accident happens to one of their patients under a certain medication. But they are
advised to keep a proper record of the consultation, as they might be sued in civil court
cases (by insurance companies).
The regulations in the different countries dealing with the procedures of assessing
fitness to drive are mainly in line with the Council Directive. Practical implementations
and the assignment of responsibilities differ from country to country. It is very difficult to
derive a “best practice” from the present results.
Several opportunities to improve guidelines and procedures for assessing fitness to
drive are presented based on the progress made within DRUID Work Packages 4 and
7. Several reflections on the existing guidelines and regulations, in particular on the text
of Art 15 of Council Directive 91-439-EEC, resulted in 8 recommendations.
Some of the recommendations point at the vague terms that are used in Article 15
(such as “substance abuse”, “regular use”, both for medicines and illicit drugs, etc.),
whereas more internationally accepted terms exist. It is also recommended to include

67
the underlying cause or reason for taking medicines, as well as all co-morbidity factors,
while assessing fitness to drive. Another recommendation points at the term
“combinations of medicines with central nervous system activity”. It is emphasized that
combinations of psychotropic medicines with other medication that can alter the
metabolism of the psychotropic medicine (with a possible consequence of increased
blood levels of the latter) will always call for an individual judgement by prescribing
physicians and dispensing pharmacists. This is especially of interest for drivers with co-
morbidities and in case of polypharmacy.
It is also recommended to apply the DRUID categorization system for medicines
affecting driving performance in developing national requirements on fitness to drive.
Finally it is recommended that in situations where physicians will advise a patient to
start driving again after a period in which the advice was given not to drive while using
the medicine, specific procedures are needed to structure and document a consultation
and to manage the risk of litigation in case an accident could occur.
It will take special efforts to derive a consensus at a European level for the use of terms
and procedures that allow improvements for assessing fitness to drive. Therefore it is
recommended that working groups and expert rounds should discuss the DRUID
recommendations involving physicians, pharmacists, driving licensing authorities and
policy makers (D 7.2.1 and D 7.2.2).

Evaluation and implementation of new technologies

The country studies showed that almost 74% of participants received no education
regarding medicines and driving during their academic studies and their professional
participation in post-graduate education. The information received during the training
did change their knowledge about the potentially detrimental effects of medicines on
driving fitness for more than half the participants (55%). After the implementation of
DRUID guidelines, a 10% increase in the positive change of reported behaviour was
observed in the overall physicians’ samples across the country studies. Patients visiting
pharmacists in the intervention group (Dutch study) were significantly better informed
about driving impairing effects of their medication, but did not change their driving
behaviour. The majority of patients (83.4%) visiting a health service or pharmacy
(Spanish study) would reduce frequency of driving, if a prescribed medicine has the
warning pictogram on the package.
The application of DRUID guidelines was successful and pinpoints the readiness of
health care professionals to adopt them. The findings support the statement that
guidelines are important and can improve the quality of health care. Physicians and
pharmacists have shown a change in behaviour after the implementation of DRUID
guidelines, therefore these guidelines could be successfully incorporated in existing
decision support systems. These guidelines fill in an important “gap” linking prescribing
and dispensing of medicine with both patient and road safety. Physicians are affected
by the DRUID training. However, this training should not be a short-term endeavour,
but flexible, adaptable, and personalized to local settings.
Based on the comments made by the health professionals within the country reports,
the implementation of computerized guidelines and DRUID categorization was highly
accepted as practical information by both physicians and pharmacists and participants
were willing to continue using the DRUID information if integrated in their prescribing
and dispensing computer systems for easier incorporation in their daily practices.
Participants offered ideas for future developments such as inclusion of other medicines
in the categorization scheme and the adaptation of information to the native language.
Future recommendations should also include specialized and elderly directed advices

68
incorporated in the system and adaptation to other target groups and not only drivers
(e.g. heavy machinery usage and seniors information).
A long term goal would be to evaluate the impact of guidelines on the health care
system, various stakeholder groups and to compare it with other studies’ findings. In
addition, further research could facilitate adaptation and customization of guidelines for
different groups of health care professionals and national settings. A set of DRUID
recommendations has been derived from the main conclusions of both composite cross
comparisons and country studies. The key message is clear about the necessity of
diffusion of DRUID information to physicians, pharmacists, and nurses in all clinical
settings (D7.4.1 und D7.4.2).

Main DRUID results to be communicated to different target groups and prototype

documents
A presentation of DRUID results was made taking into account various target groups in
terms of problem definition (alcohol, illicit drugs and medicines) and countermeasures
(legal regulation, enforcement, classification of medicines, rehabilitation, withdrawal,
guidelines for health care professionals, risk communication).

Eight overview boxes were produced to summarize the most relevant information per
topic:
• Alcohol,
• Illicit drugs,
• Psychoactive medicines,
• Enforcement,
• Classification,
• Rehabilitation,
• Withdrawal (of driving license),
• Guidelines/risk communication.

The most relevant issues were extracted for each of the following target groups: (1)
general public, (2) drivers as patients, (3) young drivers, (4) physicians and
pharmacists and (5) policy makers on EU and national level (D7.3.2).
In a broader perspective the theoretic frame of risk communication was described
(definitions and communication theory, risk communication and sources for patients,
managing risk communication related to driving with impairing substances, risk
management framework and risk acceptability).
A risk management framework was developed for effective risk communication in
DRUID by defining all steps that need to be addressed in building good risk
communication. The framework consists of 7 steps: identifying and consulting
stakeholders (initiation), risk identification, risk estimation, risk evaluation, risk control,
implementation and monitoring (D7.3.2).

Effectiveness of pictograms in communicating risk to patients who drive under

the influence of medicines
In two separate studies in Spain and the Netherlands the effects of pictograms for
communicating risk to patients were investigated, comparing the triangle model (e.g.
French model) with rating model developed by DRUID WP 7 Partners (Figure 9).

69
Risk communication based on a developed DRUID rating model pictogram compared to
an existing triangle model shows that a vast majority of patients (70-80%) consider these
pictograms clear and self-explanatory, with a preference for selecting the DRUID model.
78.5% of the patients declared to change driving behaviour if confronted with the
pictogram on the medicine box (D7.3.2).
The results of these studies show that both pictograms were effective in communicating
risk. Those who participated in the studies were able to recognize and understand the
risk of driving under the influence of medicines and have shown their intention to change
their driving behaviour by driving less frequently. In both studies, the rating pictogram
was preferred over the triangle pictogram.
In the Dutch study, for the rating model pictogram, a clear and direct correlation between
the likelihood of changing driving behaviour and the level of impairment of a medicine
has been observed: the higher the category, the more likely a change of behaviour by
decreasing driving frequency.

Figure 9: Pictograms – Rating Model (DRUID) vs. Triangle Model (France).

DRUID outcomes and risk communication to young drivers

In risk communication to young people, relevant content, necessary preconditions and
promising strategies have been derived from a German approach in social marketing
research.
Findings of the survey indicate that susceptibility to drink-driving and drug-driving
applies to about 25% of young people. Permissive attitudes are more common among
males and among formally low-educated individuals. Prospective drivers (mostly aged
15 to 17) display a slightly more positive attitude towards drink-driving and drug-driving
than active, young drivers. This fact should be considered, when defining the target
group for communication. With regard to talking and learning about DUI topics, strong
preferences for peer communication were observed. In terms of important media
channels, social network sites turned out as a promising pathway to reach out to young
people. Overall, these and further results provide important foundations for adjusting
DUI-related risk communication to low- and high-risk groups among prospective and
active young drivers (D7.4.3).

70
2.6. Legal perspectives
It was the aim of DRUID to generate a solid knowledge basis for harmonized, EU-wide
regulations on driving under the influence of alcohol, drugs and medicines that would
embrace empirical, scientifically proven evidences.
The information presented in this section at a glance can be found in detail in D1.4.1
and D1.4.2.

2.6.1. Objectives
A basic prerequisite for developing recommendations concerning legal measures to
combat DUI/DUID is an evaluation of relevant legal measures which are already
installed. This evaluation has to be done from a legal perspective, i.e. to address the
question, how driving under the influence of psychoactive substances can be
effectively combated by means of legal interventions.
In the attempt to establish cut-off limits for drugs, close interdisciplinary collaboration
between researchers involved in empirical research and in toxicology was necessary.
To give recommendations for cut-off limits the experts had to cope with the following
issues:
• Selection of scientific data necessary to estimate substance related accident risk
• Pros and cons of different research methods
• List of criteria for definition of a cut-off
• Selection of the psychoactive substances for which cut-offs should be determined
• Determination of substances that should be included, based on their prevalence
• How to deal with metabolites
• How to deal with combined consumption
• How to deal with legally prescribed medicine use
• Determination of analytical procedures and definition of the analytical substrate(s)
• Quality assurance of laboratory analysis and determination of measurement errors
• Pros and cons of whole blood and plasma as analytes
• Saliva (oral fluid), blood spots and the conversion factors between the different
body fluids

2.6.2. Methodology
For the evaluation of legal measures to combat driving under influence of psychoactive
substances an extensive literature study was done focusing among others on theories
of sanctioning and empirical research results on DUI/DUID sanctioning (e.g. general
deterrence of DUI policies and law changes, deterrence of jail sanctions and fines).
DRUID experts in the fields of experimental studies, epidemiology and toxicology
collaborated closely to give recommendations on how to determine per se cut-off limits.
A questionnaire concerning legal regulations regarding drugged driving and legally
imposed cut-offs for illegal psychoactive substances was distributed in the European
member states, Norway, Switzerland and Croatia. The reply was complemented with
the official data of the EMCDDA.
Recently, three EU countries determined cut-offs for per se legislation. Information
about the procedures and criteria used was collected by directly contacting national

71
experts involved in that topic. Concentration ratios between oral fluid and blood were
established by collecting paired samples of oral fluid and whole blood from drivers in
the RSS and in some of experimental studies. To evaluate the method of dried blood
spots (DBS), samples of whole blood have been taken and concentrations of active
agents have been measured in whole blood and in the blood spots.

2.6.3. Results

2.6.3.1. Evaluation of legal measures to combat DUI/DUID

Driving under influence of alcohol
There are several ways to diminish the risk of alcohol impaired driving. The most
general of those, and also a suitable one for special risk groups, is the enactment and
implementation of legal BAC thresholds. In this respect, the per se laws must be
considered as the most effective approach to combat DUI. The standard per se value
should not be higher than 0.5 g/L as recommended by EU to all member states
(Commission Recommendation 2001/115/EC of 17 January 2001). The effectiveness
of per se BAC values below 0.5 g/L is very much dependent on the prevailing societal,
legal, political environment and the enforcement activity of the police in the respective
member state, but also on habituation.
Results of studies show that informal non-legal consequences (e. g.: social
disapproval, especially by peers and friends) can be very important to bring people to
law-abiding behaviour. Especially on adolescent offenders, they seem to have much
higher impact than formal legal consequences.
Evaluation studies are demonstrating that jail sanctions and prison sentences are not
effective to combat DUI. Therefore they should be avoided as far as possible, in
particular for first-offenders, at least in cases without aggravating circumstances. In
contrast, probation periods are very suitable for first-offenders as well as for repeat-
offenders. Fines are much more effective among adolescent offenders than among
adults.
The implementation of a zero-tolerance approach seems to be very promising for
young and novice drivers. Lowering the legal threshold for a certain period of time for
convicted DUI offenders is an effective countermeasure to enhance road safety, but the
practical implementation has to be assured. For other risk groups (e. g.: professional
drivers, drivers of large vehicles or drivers of vehicles carrying dangerous goods), the
enactment of lower legal per se BAC levels should be discussed with respect to the
specialties of these driver groups. The implementation of BAC thresholds of at
maximum 0.2 g/L was recommended by the European Commission (Commission
Recommendation 2001/115/EC of 17 January 2001).
The risk of detection is crucial because only apprehended drivers are subjected to
punitive or rehabilitative measures. The implementation of random breath testing laws
is inevitable for securing a vigorous enforcement of zero-tolerance laws in general, but
also for higher risk thresholds. As low amounts of alcohol regularly do not cause any
visible signs of impairment, the otherwise effective suspicion based strategy is not
promising here. In consequence, the implementation of analytical procedures that can
be used as evidences is necessary.

72
Driving under influence of illicit drugs
In the near future for most of the drugs, only zero-tolerance and the impairment
approach are available alternatives in combating DUID. As far as the first approach is
favoured, it must be considered, that limits of detection and limits of quantification
(LOQ) are based on technical limits.
In case an offender is above this limit, this doesn’t necessarily indicate recent
consumption of the psychoactive substance or being under influence. Thus, it is better
to implement lower effect limits. Those limits can prove that a negative impact on
driving existed.
Zero-tolerance is more promising particularly with regard to enforcement procedure.
Especially the implementation of random drug tests is necessary to increase the
currently very low risk of detection. As far as the development of valid screening
devices is possible, the enactment of legally imposed analytical procedures is
recommendable.
Until now the effectiveness of enforced impairment laws was relatively low. For the
effective enforcement of impairment laws, it is necessary that policemen are able to
detect all signs of impairment to register them in official protocols. Later on these
records are the basis for a court conviction. Therefore policemen need regular special
training that would enable them to detect all signs of impairment while implementing
roadside checks.
The enforcement strategies must be targeted at special driver groups. On a general
level, it must be distinguished between the detection-based and the deterrence-based
strategy. The detection-based strategy is more effective for high-BAC drivers and
regular consumers of drugs, because they can hardly be deterred by other strategies.
In contrast, the deterrence-based strategy is more effective for the majority of social
drinkers, young drink drivers and occasional drug users. To achieve this objective,
highly visible testing sessions at places and during times with high traffic density, low
density of offenders (e. g. in the early morning hours) must be conducted. For the
detection strategy, places with inverse characteristics must be chosen and controls
must be conducted unobtrusively. For both strategies it is important to conduct tests
regularly.
Overall young drivers are rarely aware of legal interventions and enforcement
strategies. Therefore information campaigns as well as education campaigns are highly
recommended.
Driving under influence of medicines
Zero-tolerance is not appropriate for psychoactive medicines. Normally medicines are
prescribed to treat diseases and complaints, some of which can impair driving by
themselves. Therefore impairment law is more suitable.
Mixed intoxication through alcohol and other psychoactive substances (including
medicines) is a much greater threat to road safety than the sole consumption of these
substances. Consequently, the per se BAC limit in those cases must be lower than for
the single substance consumption.

73
2.6.3.2. Determination of per se cut-off levels
Current state of legislation concerning drug driving
Most European countries adhere to one of two possible approaches or definitions of
the act of drug driving: 11 countries use the impairment approach, 8 countries use
zero-tolerance or per se limits and 9 countries combine these two approaches into a
two-tier system.
Until now, all countries that have per se legislation use analytical cut-offs, i.e. the
concentrations, lower that can be reliably determined by forensic laboratories. In some
countries, these are the lowest limits of quantisation of the forensic laboratories, in
other countries they have been established by experts. In some countries, even if they
are called analytical cut-offs, some consideration was given to a relationship with
effects, e.g. by measuring only the active cannabis component THC, instead of the
inactive metabolite, and using a cut-off that corresponds to a concentration after a
single dose, when the drug still has effects.

Substance thresholds
There are three classes of substance thresholds. “Risk thresholds”: concentrations in
blood that indicate a certain accident risk or impaired driving. “Lower effect limits”: the
lowest concentration where an effect on driving is observed. “Limit of detection” and
“Limit of quantification”: based on technical limitations in order to guarantee a valid and
reliable analytical result and avoid false positive results.
In establishing thresholds, one must realize that the relationship between the
concentration and the effect is not linear for most drugs, and that a given concentration
could correspond to low effects (e.g. in a tolerant individual) or high effects (e.g. in a
drug-naive subject).
The list of drugs to be included in per se legislation will depend on the situation in each
country, e.g. the drugs that are most often found in the driving population or in drivers
involved in an accident. Most countries have a very limited list of 10 substances or less.
There is a consensus on not including medicinal drugs in the list. It is not reasonable to
define cut-off values for patients in long-term treatment. Even high doses may lead to
fewer effects because of tolerance.
Norway and the Netherlands recently tried to determine safe driving limits and they
arrived to very similar values, e.g. 3ng/mL THC or 48 and 50ng/mL for MDMA in whole
blood. Both countries defined a risk threshold for THC, where the impairment is
equivalent to 0.5 g/L BAC.
In determining “lower effect limits”, stimulant drugs like amphetamines and cocaine
pose a particular challenge. The correlation between drug concentration and risk of
traffic accidents/impairment is variable or insufficiently documented. In experimental
studies, at the (rather low) doses that were given, driving performance increases rather
than decreases.
However, in epidemiological studies indications of increased accident risk could be
found. To define cut-offs the results of experimental studies and of epidemiological
studies should be taken into account. In case there is not enough data available,
another approach has to be chosen: The pharmacokinetics of these substances could
be used together with data on consumption patterns to determine cut-offs. The cut-off
can be set at a certain time after use, e.g. the duration of the effects.

74
Usually, inactive metabolites are not included in the legislation, except when the parent
drug is unstable and is metabolized very rapidly. Some drugs like THC have a very
rapid metabolism, and if the delay between the stop or accident and the blood sampling
is long, the concentration could decrease remarkably (based on a half-life of 1.4 hours,
3ng/mL of THC decrease to 0.68ng/mL after 3 hours). The “lower effect limits” should
be established taking this into account. Another possibility, but less easy to implement,
is that the lowest concentration that can be accurately measured (LOQ, limit of
quantisation) is used instead of the lower effect limit when the sampling delay is longer
than 2 or 3 hours.
The epidemiological studies in DRUID have shown that people very often use more
than one drug. The question has been raised if the per se lower effect limits or the LOQ
should be used when more than one drug (or alcohol) is detected. Some have
recommended using the LOQs. One of the problems of using lower effect or safe
driving limits is a definition of the dose that can be taken still remaining under the limit.
One should realize that establishment of lower effect limits does not mean that one
excuses drug use. In many countries (e.g. Sweden and Finland), people who are
sanctioned for driving under the influence of narcotics will also be sanctioned for drug
use, or sanctioned for drug use even if inactive metabolites are detected. But to
achieve the compliance of the population, a clear legislation should be implemented,
which differentiates drug and traffic policy.
It is not a problem to limit the list of drugs in per se legislation to a few substances, if
the per se law is combined with an impairment law, where all other impairing
substances are covered. In this scenario, a quick and easy to enforce procedure exists
for the most common drugs, and a more elaborated procedure exists for the less
frequent cases, including medicinal drugs and combination of drugs. It is not realistic to
develop cut-offs for all existing medicinal and recreational drugs. Moreover, for new
drugs it might take some time before cut-offs have been established.

Usability of conversion factors between analytes

In DRUID many experimental results were collected and a literature review was
performed on conversion factors between plasma and whole blood and between whole
blood (B) and oral fluid (OF). The drug concentration ratios between oral fluid and
blood (sometimes called oral fluid to blood conversion factors) were studied by
collecting paired samples of oral fluid and whole blood from drivers. Oral fluid to blood
(OF/B) concentration ratios were calculated but large variations were found between
individuals; typically the coefficient of variation (relative standard deviation) was 50 to
100%. Therefore, conversion factors cannot be used to accurately estimate drug
concentrations in blood based on drug concentrations in oral fluid. The estimated
equivalent cut-off concentrations for oral fluid and blood were used for the calculations
of drug prevalence (D2.2.3) and for the odds ratio calculations (D2.3.5).
Drug analysis in samples of oral fluid can be used to estimate the drug prevalence in
blood if using equivalent cut-off concentrations. Three formulas were used for
estimating equivalent cut-off concentrations using the average OF/B ratio, median
OF/B ratio or percentile regression. To determine which formula fits best the original
paired data, the prevalence of samples above selected cut-off concentrations in blood
was estimated using the formula and compared with the actual prevalence in blood.
The accuracies of the three procedures were calculated for the chosen cut-off
concentrations in blood and for concentrations corresponding to 2.5 times and 5 times
the analytical cut-off. The procedure with the least average percent deviation (in
absolute value) from the actual number of subjects with drug concentrations above the
cut-offs in blood was identified as the best one for each substance separately. Based
on this equivalent cut-offs were established for oral fluid.

75
Dried blood spots as alternative method for collecting blood samples
The use of dried blood spots (DBS) has potential as a precise and inexpensive option
for the determination of several analytes in small blood samples. But the small sample
volume requires very sensitive techniques. By using LC-MS/MS, all investigated
analytes could be determined with sufficient lower limits of quantisation (LLOQs).
Evaluation data showed no significant differences in precision as well as lower limits of
detection (LLODs) and LLOQs. Analysis of DBS is feasible with the advent of
increasingly sensitive MS technologies such as LC-MS/MS. The DBS/B ratios were
very close to 1.00, and the relative standard deviations 8.56%. Thus, the use of DBS in
routine analysis will result in simplifying blood sampling, transport and storage as well
as sample processing in the laboratory. DBS drug analysis can be regarded as a
valuable and inexpensive alternative to determination of substances from whole blood.
It should be noted that no evaluations for THC have been performed yet. Therefore
these results should be taken with some care.

76
Section 3 – Conclusions
The following section summarizes main results of the integrated project DRUID and
aims at drawing conclusions relevant for policy makers on EU and Member States
level. 50 deliverables, in which the outcomes obtained in the seven work packages of
the project are described in detail, can be found on the DRUID website (www.druid-
project.eu). Deliverable D7.3.2 “Main DRUID results to be communicated to different
target groups” gives an overview of the project achievements across all work
packages. In the deliverable D1.3.1 “Driving under the influence of alcohol, illicit drugs
and medicines: Risk estimations from different methodological approaches” the results
obtained in epidemiological and experimental studies conducted in DRUID are
integrated. Deliverable D1.4.2 “Per se limits - Methods of defining cut-off values for
zero tolerance” aims at giving pertinent considerations when a nation wants to
determine per se cut-off levels.
As DRUID focuses on the effects of driving under the influence of (1) alcohol, (2) illicit
drugs, and (3) psychoactive medicines, the following conclusions are divided into three
corresponding chapters.

1. Alcohol
The roadside surveys (RSS) that were conducted in DRUID to obtain prevalence rates
of driving under influence of psychoactive substances indicate that alcohol is the most
frequently detected psychoactive substance in the general driving population as well as
in the seriously injured and killed drivers in Europe (hospital studies, HS). On average
in all European countries in which RSS were conducted7 (later on referred to as EU
mean) 3.5% of drivers were driving under influence of alcohol (≥ 0.1g/L). There were
rather big national differences (0.15-8.59%). Nevertheless alcohol was the most
common toxicological finding in most countries. Referring to blood alcohol
concentrations of 0.5g/L or above the EU mean prevalence in the general driving
population reduces to 1.5% (range: 0.07-5,23%) and to 0.4% (range: 0.01-1.47%) for
heavy alcohol consumers (≥1.2g/L).
The prevalence rates of alcohol impaired drivers being involved in an accident are even
much higher in the participating countries8. In seriously injured (≥ 0.1g/L: 17.7-42.5%; ≥
0.5g/L: 16.1-38.2%) and killed drivers (≥ 0.1g/L: 19.0-44.9%; ≥ 0.5g/L: 16.3-35.1%)
alcohol was the most frequently detected substance. This means that drivers involved
in an accident have more often higher alcohol concentrations in blood than drivers in
the general driving population.
Referring to the results of these case-control studies the risk of serious injury or
fatality for alcohol (≥ 0.5g/L) is significantly increased compared to sober drivers (i.e.
drivers who are below the DRUID cut-offs for any substance). When the whole group of
alcohol impaired drivers is divided in different alcohol categories the risk of injury and
fatality continuously increases: 0.1-0.5g/L: 1-3 times higher risk; 0.5-0.8g/L: 2-10 times
higher risk; 0.8-1.2g/L: 5-30 times higher risk; ≥ 1.2g/L: 20-200 times higher risk
(D 2.3.5). Taking into account the responsibility for the accident the DRUID results
reveal that the risk of being responsible for a fatal accident is 5-8 times higher for an
alcohol impaired driver (≥ 0.1g/L) than for a sober driver. Severely intoxicated drivers

7
RSS were conducted in Belgium, Czech Republic, Denmark, Finland, Hungary, Italy,
Lithuania, Netherlands, Norway, Poland, Portugal, Spain and Sweden.
8
HS on injured/killed drivers were conducted in Belgium, Denmark, Finland, Italy, Lithuania,
Netherlands, Norway, Portugal and Sweden.

77
(≥ 1.2g/L) have a 15-21 times higher risk of being responsible for a fatal accident
(D2.3.2). The risk of alcohol impaired driving is clearly confirmed in the experimental
studies: Alcohol has a negative impact on driving performance and on skills related to
driving (D1.1.2B, D1.2.1, D1.2.2).
Thus, alcohol is still a most dangerous psychoactive substance in traffic. This holds
true for all EU Member States in which studies were implemented. Regarding this
problem the following conclusions can be drawn:
• The first priority of countermeasures should always be on alcohol; other
psychoactive substances are of second priority. This implies that any activities,
regarding e.g. policy issues, enforcement, education or campaigns, should never
be carried out at the cost of alcohol countermeasures.
• The concurrent implementation of diverse countermeasure, like drivers’ education,
enforcement, rehabilitation, etc. is necessary to combat driving under the influence
of alcohol.
• The scientific assessment of the “alcohol problem” in the driving population should
be carried out on a regular basis. DRUID established standards how to conduct
national roadside surveys to gain comparable and valid data in Europe. A
continuation of this work offers the possibility to assess the problem in the long run
and to study the development of prevalence rates of alcohol cases. This could also
be a valuable method for the evaluation of national campaigns against drunk
driving. However, the conduction of roadside surveys is cost and time expensive.
Further on the problem of national and/or European representativity is very
complex and hard to solve. As participation in a roadside survey is voluntary, there
is always a risk of bias through non-participation of impaired drivers. Therefore
other ways of monitoring the prevalence of alcohol (and illicit drugs and
psychoactive medicines) should be sought for as well.

In addition to determining the significance of the alcohol problem in the general driving
population and in the population of seriously injured/killed drivers, DRUID tried to
characterize problem groups. In the driving population alcohol was mainly detected
among older male drivers: In most countries the share of alcohol positive male drivers
was the highest for the two oldest age groups (35-49 and 50+).
Within accident involved drivers (i.e. seriously injured or killed drivers) alcohol was
mainly detected among younger (25-35 years) male drivers with high BAC levels. High
BAC level, young age and speed are associated with an increased risk. Further on the
risk of having an accident multiplies with combined use. DRUID results demonstrate
that the combined use of alcohol and drugs is very common and in terms of risk a
serious problem (D2.2.3, D2.2.5, D2.3.5).
• Therefore countermeasures to combat driving under influence of alcohol should
take into account driver characteristics. This means that enforcement strategies,
educational activities or legislative measures should address special target groups
and not the driving population as a whole.
• Countermeasures especially for young male drivers should be promoted. In
Germany in 2007 zero tolerance for young (< 21 years) and novice drivers was
implemented (“Null Promille für Fahranfänger”). The rate of alcohol related
accidents within the target group could be lowered about 15% in the first year.

78
Appropriate countermeasures to combat alcohol impaired driving are rehabilitation
programs (D5.1.1, D5.2.4). An average recidivism reduction rate of 46% (range: 15-
71%) was found in standard group intervention programs for DUI offenders. The
implementation and the quality of such programs in Europe vary from country to
country. DRUID established the most comprehensive database on European
rehabilitation schemes and measures as well as on quality assurance measures for
rehabilitation programs, compiled up-to-date statistics of recidivism, analyzed
recidivism reasons and suggested best practices in the field of rehabilitation.

Referring to alcohol rehabilitation programs the following conclusion can be drawn:

• Driving rehabilitation should be harmonized within Europe, e.g. by using common
European standards and recommendations on rehabilitation good practices
developed within DRUID.
• Driver assessment and rehabilitation should be legally regulated and based on
defined criteria.
• Alcohol offenders should be treated in separate groups, not together with drug
offenders.
• Non-addicts and addicts should be distinguished as they require different
interventions or treatments.
• Multiple offenders and offenders with a BAC ≥ 1.6g/L should undergo an
examination aiming to enable a provider to detect/exclude driver’s addiction.
• An alcohol ignition interlock can be used during the rehabilitation phase, yet it
should be combined with rehabilitation/treatment and close monitoring.

Further on DRUID analyzed practices of driving license withdrawal in European

countries. The study revealed that national strategies are very heterogeneous. It can
be stated that withdrawal is an effective general and special deterrent factor, if sanction
certainty and celerity are assured. Sanction severity is less important. This means that
an immediate withdrawal/suspension of driving license and a high level of perceived
detection risk are decisive. Sanction certainty can be increased by strict enforcement
(e.g. implementation of random alcohol and drug controls).
Additionally it can be stated that the combination of license withdrawal and
treatment/rehabilitation is more effective in terms of deterrence than license withdrawal
alone. This holds true especially for addicted drivers and in cases of medicines misuse.
Conditional withdrawal is a measure to support the reintegration process of drivers.
Conditional withdrawal seems to be effective when it is combined with rehabilitative
measures and close monitoring (D6.2).

DRUID conclusions concerning withdrawal:

• Practices of driving license withdrawal should be harmonized across Europe.
• The withdrawal duration should be between 3 and 12 months.
• Driver rehabilitation should be an integrated part of driving license withdrawal.
• Conditional withdrawal should always be combined with rehabilitative measures
and close monitoring.
• Zero alcohol limits for novice drivers are very effective.

79
2. Illicit drugs
The prevalence of illicit drugs in the general driver population (based on RSS) is
much lower than the prevalence of alcohol. The estimated EU mean for all investigated
illicit drugs is 1.9% (individual countries’ range 0.2 to 8.2%), but there is a high national
variability. Cannabis (THC) (EU mean prevalence: 1.32%; range: 0.0-5.99%) and
cocaine (EU mean prevalence: 0.42%; range: 0.0-1.45%) are the most frequently
detected illicit substances in most countries (D2.2.3).
Within the accident involved drivers the prevalence of the different substances shows
great national variability as well, so no clear picture of the distributions can be
identified. THC seems to be one of the most prevalent illicit drugs, followed by cocaine
and amphetamines. The majority of illicit drugs is used in combination with other
psychoactive substances, mainly with alcohol. Among seriously injured and killed
drivers, those having consumed alcohol together with drugs, are the second most
common group in all countries (except for LT). The group of drug-drug combined users
represents either the third (BE, DK, FI, I) or fourth (LT, NL) most common group
(D2.2.3, D 2.2.5, D2.3.4)9.
The DRUID results based on case-control studies show that the injury risk is
extremely increased with combined use of drugs and alcohol comparable to the risk of
alcohol consumption alone ≥ 1.2g/L. Drug-drug combinations causes the second
highest risk of injury , showing a highly increased risk comparable to the risk of alcohol
consumption alone 0.8-1.2g/L.
The high risk of combined drug-alcohol or drug-drug consumption is confirmed by the
experimental studies. With regard to MDMA it was found that drug use alone does not
impair, but MDMA in combination with alcohol (or sleep deprivation) impairs driving
performance dramatically. These results are validated by on road and simulator
experiments (D1.2.1).
For the different investigated illicit drugs ORs between 2 and 7 were found. This
increase of risk corresponds to the risk caused by BAC of 0.5-0.8g/L, which has an OR
of about 4. Due to the low prevalence of illicit drugs in the RSS and HS it was not
possible to differ between concentrations. Therefore drivers with (very) low and (very)
high substance concentrations of the investigated agent were pooled. Furthermore due
to the low numbers of positive cases and controls the confidence intervals are very
wide. This indicates low accuracy of the ORs. In addition the ORs differ between the
countries involved in DRUID (D2.3.5).
The sole use of illicit drugs is not frequently detected in Europe. There are great
national differences in prevalence. The main problem has to be seen in combined
consumption, i.e. when illicit drugs are consumed in combination with other
psychoactive substances, especially with alcohol. Though the prevalence rates of
combined consumption are not high either, the injury risk is clearly increased in these
cases. The following conclusions can be drawn:
• Due to the great national variability regarding prevalence rates of illicit drug use
countermeasures should be adapted to national requirements.
• Increase of drug enforcement is a countermeasure, which is potentially cost-
beneficial for countries that currently have a low enforcement level. It is not
beneficial when the increase is financed at the costs of drink-driving enforcement.
• Enforcement activities should take into account that especially combined
consumption (drug-alcohol or drug-drug) is most dangerous.

9
In these Deliverables „drugs“ means both illicit drugs and medicines.

80
• Referring to combined drug-alcohol consumption it is supposed that even an
alcohol level of 0.1 or 0.2g/L is enough to increase accident risk. This has to be
taken into account when enforcement or legislative measures are adopted.
• The scientific monitoring of the “illicit drug problem” in the driving population should
be carried out on a regular basis. However, RSS are very expensive and time
consuming and their efficiency is doubtful. Time and amounts to be invested in
RSS will have to be even increased in the future. Otherwise, taking into account
low prevalence rates, it will be impossible to detect enough positive drivers in order
to come to comparable and valid European data. The risk of bias in prevalence
rates due to voluntary participation in roadside surveys aggravates this problem. In
order to prevent drivers positive for illicit drugs from refusing to give a sample of
any body fluid, the legal regulations would have to be changed in a way that would
make mandatory collection of random samples possible. Therefore (even to a
greater degree than for alcohol) other ways of monitoring the prevalence and risk of
illicit drugs (and psychoactive medicines) should be investigated.

To make countermeasures more effective characteristics of drivers tested positive

for illicit drugs were studied. DRUID shows that in the general driving population illicit
drugs are mainly detected among young (< 35 years) male drivers, during all times of
the day but mainly at the weekend. Combined use of alcohol and drugs is more
prevalent among young (< 35 years) male drivers during night time hours. Combined
drug-drug use is in general most common in middle aged male drivers. Referring to
injured and killed drivers the use of illicit drugs is most prevalent in young and middle
aged male drivers as well. Consequently countermeasures should be targeted to this
driver group:
• Countermeasures against drug driving should take into account driver
characteristics and should be target-group-specific.
• Enforcement strategies, educational activities or legislative measures should be
addressed especially to young drivers.

Drug detection, especially in the context of roadside surveys and/or everyday drug-
driving police enforcement, is a great challenge. In contrast to alcohol that can be
easily and reliably detected by breathalyzers, illicit drugs have to be detected by on-
site drug screenings. Police officers tested, within DRUID, the practicability of
available oral fluid drug screening devices. Besides, scientific requirements on these
devices were formulated and devices were evaluated correspondingly. The results
show that most of the investigated systems are not effective enough concerning
specificity and sensitivity. Thus, the roadside assessment of illicit drug consumption
might be deficient, depending on the used device. Further on it has to be taken into
account that large-scale random drug testing (which has the largest general deterrence
effect) is not feasible, because the devices are too expensive and it takes too much
time to collect and analyze samples.

Checklists (Clinical Signs Inventory, CSI) used by the police to preselect suspected
drivers would be a good method to support on-site drug screenings. But DRUID results
were not very encouraging in this regard. More experience and better training of police
officers may improve the results.
Oral drug screenings and checklists have the advantage of being non-invasive
methods, but are lacking high standards of reliability and validity. Blood analyses to the
contrary allow precise detection of drug amounts, but are invasive and therefore not
feasible for on-site screenings. The use of dried blood spot testing (DBS) might open
new perspectives, since it is a minor invasive method and it has potential to be a
precise and inexpensive option for the determination of several analytes in small blood

81
samples (D1.2.1, D1.2.2., D1.4.2). Analyses of DBS are feasible with the advent of
increasingly sensitive technologies such as LC-MS/MS. The DBS-blood ratios are very
close to 1.00, and the relative standard deviations ≤ 9%. The use of DBS will result in
simplified handling during blood sampling, transport and storage as well as sample
processing in the laboratory. Thus the DRUID results regarding dried blood spot
analyses are very promising, though the practicability in the daily work of police has not
been studied yet.

Conclusions with regard to illicit drug detection can be formulated as follows:

• Drug screening devices should be optimized regarding scientific requirements and
practical use. National and regional circumstances should be considered when
selecting screening devices.
• Checklists to preselect suspected drivers should only be used by sufficiently trained
police officers.
• Dried blood spot testing is a very promising method and should be further
developed and tested, especially for practicability regarding enforcement measures
by the police.

Referring to blood analyses to detect illicit drug use DRUID established common
analytical cut-offs and standardized analyzing procedures. These measures are
crucial for comparing study results of different countries. The DRUID-standards already
gained acceptance worldwide.
• An agreement should be established across Europe on the kind of body fluid to be
used for drug analyses. Particularly important is to make a choice between whole
blood and plasma.
• European harmonization of drug analyses should be carried on, i.e. common
analytical cut-offs and standardized analyzing procedures should be introduced.

The development of conversion factors between plasma and whole blood and
between whole blood and saliva became a major objective in the course of DRUID.
Oral fluid to blood conversion factors were calculated by analyzing paired samples of
oral fluid and whole blood from drivers in BE, FI, I, and NO. Large variations were
found between individuals (coefficient of variation: 50-100%). Therefore, conversion
factors cannot be used to accurately estimate drug concentrations in blood based on
drug concentrations in oral fluid for most psychoactive substance. However, drug
analysis in samples of oral fluid can be used to estimate the drug prevalence in blood if
using equivalent cut-off concentrations. Therefore a method developed in DRUID can
be used (D 1.4.2).
Valid detection of impaired drivers and appropriate analyses of body fluid samples are
just one side of the coin. The other side is the establishment of effective sanctions for
offenders. Almost all of the research findings refer to DUI, but of course general
deterrent principles are valid for DUID offenders, too (see section “Alcohol”).
A major problem is – in contrast to alcohol – that no legally defined risk thresholds are
available. The implementation of thresholds (per se laws) increases sanction certainty
and thus general and special deterrence. Clear information about thresholds and the
compulsory consequences of driving under the influence need to be communicated by
public campaigns. Especially young people should be better informed about the risks.

82
Taking these results into account we conclude that:
• Driver assessment (e.g. driving tests, medicinal or psychological examinations)
should always be carried out for the decision on fitness to drive and further
intervention (rehabilitation/treatment).
• Measures to improve sanction certainty and sanction celerity should be undertaken.

As already stated for alcohol, driver rehabilitation helps to prevent people from
impaired driving. It can be concluded:
• Driver rehabilitation should be an integrated part of a comprehensive
countermeasure system. Legal regulation of participating in rehabilitation measures
should be established in order to assure interventions by offenders.
• Driving rehabilitation should be harmonized within Europe, e.g. by using common
European standards and recommendations on rehabilitation good practices
developed within DRUID.
• Different types of DUI/DUID offenders should be treated separately according to
their special needs.
• Non-addicts and addicts should be distinguished as they require different
interventions or treatments.

One major objective of DRUID is to make recommendations on establishing cut-off

levels for drugs in per se legislation with regard to driving under the influence. Three
classes of cut-offs can be used (D1.4.2):
(1) “Risk thresholds”: Concentrations in blood that indicate a certain accident risk or
impaired driving.
(2) “Lower effect limits”: The lowest concentration where an effect on driving is
observed.
(3) “Limit of detection” and “Limit of quantification” (analytical cut-offs): Based on
technical limitations in order to guarantee a valid and reliable analytical result.

Currently all countries that have per se legislation use analytical cut-offs for illicit drugs.
In some countries these are the lowest limits of quantisation of the forensic
laboratories, in other countries they have been established by experts. In some
countries, although if they are called analytical cut-offs, the relations to effects have
been considered, e.g. by measuring only the active cannabis component THC, instead
of the inactive metabolite.
To determine risk thresholds for illicit drugs DRUID partners intended to find, for each
substance, a concentration in blood at which the accident risk is equivalent to the risk
associated with 0.5g/L BAC. This approach starts from the premise that alcohol
impaired driving is tolerated up to a BAC of 0.5g/L in most European countries. This
means that a certain risk is accepted and that this approach should be applied, also
quantitatively, to the use of illicit drugs as well. The list of drugs to be included in per se
legislation could embrace the drugs most frequently found in the driving population
and/or in drivers involved in an accident. To calculate risk equivalents it is necessary to
include impaired drivers with different substance concentrations (in blood).
However, in the epidemiological studies the number of drug impaired drivers was too
small to determine different concentration classes. In addition from the RSS mostly
only saliva samples were available. DRUID affirmed that it is not possible to convert the
saliva concentration of a certain agent into blood concentration (D1.4.2). Therefore it
was not possible to calculate risk thresholds for all substances.

83
Risk thresholds could be formulated only for THC which was the most prevalent illicit
drug in the general driving population and in injured/killed drivers. The prevalence of
THC across all countries that participated in DRUID is 1.37%. This is about one third of
the alcohol prevalence. The epidemiological, the experimental and the meta-analytical
approaches result in rather low risk estimations. Epidemiological case-control studies
assess at maximum a 2.4-fold risk for injury, experimental studies and meta-analysis
rank the risk between 0.5 and 2 times than that of sober driving. So THC seems to be
much less impairing and risky than most of the other examined substances. Although a
relationship between THC concentration and accident risk was found in the
epidemiological studies, it was only possible to set an exact THC cut-off by a meta-
analysis of experimental studies. Thereby it was found that the serum concentration of
3.8ng/mL THC (≈2ng/mL in whole blood) causes the same amount of impairment as
0.5g/L alcohol. This value could be an empirical basis for a threshold discussion. The
meta-analysis could also be used to define limits comparable to lower BAC levels.
Any threshold discussion should address the question if the DRUID approach to
determine risk threshold as equivalents to 0.5g/L alcohol is feasible. From a scientific
point of view it can only be justified to accept the same risk for all psychoactive
substances (including alcohol). From a political point of view the determination of risk
thresholds as equivalents to 0.5g/L alcohol might be questionable, because a BAC of
0.5g/L is not a legal limit in all European countries. Some Member States have lower
alcohol limits and therefore risk threshold calculations for THC would have to be
adapted accordingly. Besides, in European countries in which presently a certain risk is
accepted, a discussion continues concerning alcohol zero tolerance approach.
It has to be taken into account that the results of DRUID epidemiological studies have
to be interpreted cautiously for most substances due to low numbers of drivers tested
positive on illicit drugs in the RSS as well as in the HS. The confidence intervals, which
are indicating the reliability of the calculated risks, are rather wide. This indicates that
that the risk measures are not very precise.
In general the following top-down procedure for cut-off determination is recommended:
• use the epidemiological data on the accident risk of different single substance
concentrations. If this data is not sufficient,
• use the experimental data. If this data is not sufficient,
• let national expert rounds determine cut-offs by using additional information
(e.g. pharmacokinetic drug profiles, consumption behavior). If this information
is not sufficient,
• use the limit of quantification (here the advantage is that new drugs may easily
be implemented into the list of impairing substances).
The (empirical) determination of risk thresholds is much more complex and costly,
especially when they should be based on case-control-studies. The list of drugs in per
se legislation can even be limited to a few substances, if the per se law is combined
with an impairment law, where all other impairing substances are covered. Moreover,
with regard to new drugs it might take some time before the different cut-offs have
been established.
In the course of DRUID two EU countries changed their legislation regarding drug risk
thresholds. In The Netherlands and in Norway efforts were made to define cut-off levels
for drugs in per se legislation. Both countries tried to determine lower effect limits and
risk thresholds. The cut-off levels are based on the expertise of scientists of different
subjects and politicians. Norway and The Netherlands introduced a THC risk threshold
of 3ng/mL in whole blood (D1.4.2).

84
The following conclusions can be drawn:
• It is advised to establish legal risk thresholds for illicit drugs on the basis of solid
empirical data, i.e. experimental and epidemiological results. If the empirical basis
is too weak, pharmacokinetic substance characteristics could help to define a lower
effect limit.
• The risk thresholds for drugs should reflect the impairment equal to that of 0.5g/L
BAC or to any other legally relevant BAC.
• The risk threshold for THC should be set adequate to 0.5 BAC at 3.8 ng/mL serum
plus an added value for measurement error and confidence interval.
• For all other illicit drugs a zero-tolerance should be implemented, as the empirical
database does not allow a definition of risk thresholds yet. In order to determine the
limit of effect, all information regarding the drug effects and its pharmacokinetics
should be considered by national expert teams. Those teams could use the
comprehensive data provided by DRUID. Thereby it could be avoided that further
development of the analyzing methods leads to a constant decline of the limits of
quantification.

3. Psychoactive Medicines
DRUID prevalence studies indicate that medicines (benzodiazepines, medicinal
opiates and opioids, and Z-drugs) are less prevalent in the driving population (EU
mean, all psychoactive medicines: 1.4%; range across countries: 0.17-2.99%) than
alcohol and illicit drugs. The same holds true for the prevalence among accident
involved drivers (D2.2.3, D2.2.5). Due to low prevalence rates risks can be estimated
only for broad substance categories, i.e. benzodiazepines, medicinal opioids, and Z-
drugs. Further on the national results show large variations.
In the DRUID prevalence studies an EU mean of 0.9% (range: 0.14-2.73%) is
assessed for benzodiazepines, which is rank four in the prevalence ranking of all
investigated substances. Among the injured drivers benzodiazepines (0.0-10.2%) were
the third most frequent finding after alcohol and THC. Among the killed drivers the
benzodiazepines (range: 1.8-13.3%) was the second most frequent toxicological
finding after alcohol, followed by amphetamines. The risk of benzodiazepines was also
a subject of the responsibility studies. About 6% of all tested drivers were under the
influence of psychoactive substances, mainly benzodiazepines. Medicinal opioids are
less common in the general driving population. For the prevalence of these substances
an EU mean of 0.35% (range: 0.00-0.79%) is estimated. Medicinal opioids are
relatively often used in combination with other psychoactive substances. The same
holds true for Z-drugs. The estimated EU mean for Z-drugs is 0.09% (range: 0.00-
0.69%).
DRUID results show that psychoactive medicines can be a problem for road safety.
Although medicines are normally prescribed to treat diseases and complaints, some of
these by themselves will impair driving fitness of patients. Patients fitness to drive will
only be improved by medicines under specific conditions: a) use of medication by the
patient according to the prescription, b) patient refrains from alcohol and other
medicines besides those prescribed by the attending physician, c) patients restrain
from driving at the beginning of a treatment or when medication or dose is changed.
The risk estimated by the DRUID case-control studies for benzodiazepines and Z-
drugs is 2-3 for a serious injury and 5-7 for fatality. For medicinal opiates and opioids
the European risk estimation is 5-8 for a serious injury and around 5 for fatality.

85
Experimental driving tests conducted in DRUID show a comparable or even impairment
level for patients taking different hypnotics or opioids than drivers impaired by 0.5g/L
alcohol. Zopiclone was less impairing than the alcohol reference as well. Only
alprazolam proved to be highly impairing showing a 5-fold risk by already medicated
anxiety patients and a 16-fold risk by healthy volunteers, both compared to 0.5g/L
alcohol. This result indicates that the development of tolerance and habituation plays a
major role for resulting impairment.
A meta-analysis conducted in DRUID reveals that impairment strongly depends on the
kind of substance and dosage (concentration). The relevant studies have been
conducted with healthy volunteers with single application. The results deliver important
information regarding driver fitness after single intake and/or the beginning of a
persisting medical treatment. But these results cannot be transferred to the situation of
patients under long-term treatment because habituation leads to decrease of impairing
side effects. In addition the medicinal treatment by itself could result in improving
driving fitness.
In the epidemiologic studies (general driving population) psychoactive medicines were
mainly detected among older (> 35 years) female drivers during daytime hours. The
same holds true for accident involved drivers impaired by psychoactive medicines.
These driver characteristics already imply that the associated target group that should
be focused by countermeasures differs from alcohol and drug offenders (who are
mostly young and male).
Medicines are usually prescribed to people suffering from a disease or complaint in
order to treat or stabilise their condition. In case of impairment of psychomotor
functioning by their disease condition, medicines could reduce impairment caused by
the illness or complaint. Although the treatment effects of medicines are different
comparing patients and health subjects, both groups can suffer from impairing side
effects on driving fitness There are evidences that, while being under long-term
medical treatment, depressed patients and patients using analgetics perform better
than patients without treatment or in comparison to healthy controls. A DRUID expert
workshop was conducted on this topic, resulting in three main statements:
(1) A proper prescription of a medicine includes correct information given to a patient
by a practitioner. Patients under long-term treatment with psychoactive medicines
should not be stigmatized by obligation to carry a special “medication passport”.
Other than by drug users, the responsibility and willingness to follow instructions
regarding medicines use and driving under long-term treatment is usually observed
in patients with insight in their condition and circumstances of impaired fitness to
drive..
(2) It is not reasonable to define cut-off values for patients in long-term treatment. Even
high doses may lead to fewer effects. The correlation between dosage and
impairment is only intra-individual. There is no clear inter-individual correlation.
Dosage effects are mostly investigated and observed with single users or new
users.
(3) Alcohol increases impairment and interacts with many medicines in an unfavorable
way. Hence, a separation of drinking, medicine consumption and driving is
necessary and the respective information should be part of the physician’s
consultation.

The above statements together with the empirical DRUID results implicate:
• In general the preservation or recovery of a patient’s mobility should be the
ultimate ambition. Therefore patients taking psychoactive medicines are
different compared to alcohol/drug offenders, i.e. the use of legally prescribed
medicines should not be controlled by the same legal countermeasures as

86
 provided for alcohol/drug offenders. Risk thresholds for psychoactive
medicines should not be established. Moreover it is nearly impossible to define
those thresholds because the impairment and accordingly the risk are
depending on the type of substance, the dosage and the duration of the
treatment as well as the interaction between disease and medication.
• The adequate countermeasure to combat impaired driving is information. A
comprehensive information system for medical doctors, pharmacists and
patients in order to inform them about the potential risk of the different
psychoactive substances, the maximum impairment, and the duration of intake
after which habituation has taken place, etc. should be implemented.
• Drinking, medicine consumption and driving should be separated. The
respective information should be part of the physicians and pharmacists
consultations.

DRUID suggests the implementation of a four level classification and labeling

system (D4.2.1, D7.3.2) regarding the influence of medicines on driving performance.
Over 3000 medicines were reviewed and over 1500 of them were categorized (D4.3.1,
D4.4.1). It is suggested to integrate the classification and labeling system in existing
computerized prescribing and dispensing systems for physicians and pharmacists
(D7.4.1, D7.4.2). The system was developed in close cooperation with EMA and is in
line with the recently approved SmPC guidelines. DRUID results are compatible with
any existing national classification system and can be integrated in them. Thus the way
for implementing the DRUID system is paved.
• DRUID proposes a four level classification and labeling system regarding the
influence of medicines on driving performance. The European (or even worldwide)
implementation of this system should be supported.
• The categorization of medicines should be continued. This implicates that the
scientific examination of the effects of medicines on driving has to be encouraged.
The same is true for medication only available on prescription as well as over-the
counter medication.

A special problem arises when medicines are misused by patients or by healthy

drivers. In these cases the medicines tend to lose their desired effects and might
reduce fitness to drive. Therefore the corresponding legal procedures and
consequences should be in line with combating DUID policies. The same holds true for
the combined consumption of medicines and alcohol.
• Legal procedures and consequences of misuse of medicines should be in line with
DUID.
• Legal measures should only be taken after an incident in traffic, whereby
impairment is the key for sanctioning.

Another special case is patients in substitution treatment. There should be no basic

difference made between patients in substitution treatment and patients in other
medicinal treatments. However, it has to be considered that consumption of or even
addiction to other psychoactive substances (alcohol, medicines and/or illicit drugs) is
often a problem (D1.1.2c). Addiction to other substances is clearly an exclusion
criterion for driving. In other cases a conditional licensing based on the results of a
fitness to drive examination might be promising.
Each patient in substitution treatment should be assessed individually regarding fitness
to drive, taking into account addictions to and abuse of other substances as well as
kind of substance used for treatment (e.g. patients treated with diamorphine are not fit
to drive).

87
Section 4 – Recommendations

In the following section we formulate recommendations for policy makers on the EU

and Member States levels based on the empirical research evidences that were
generated in DRUID. The recommendations are conceived as a scientific support to
developing countermeasures to combat impaired driving. Recommendations are
divided into three groups referring to three classes of investigated psychoactive
substance - “alcohol”, “illicit drugs” and “medicines”.

1. Countermeasures to combat alcohol impaired driving

Alcohol is still a most prevalent psychoactive substance in traffic, a problem common
for all EU Member States. The number of drivers in the general driving population with
BAC > 0.5g/L is rather low. Accident involved drivers (injured or killed) often show
higher BAC. Further on the combined use of alcohol and illicit drugs or medicines is a
rather seldom but dangerous problem.

Target groups • Young male drivers with high BAC

• Male drivers above 50 years of age
• Drivers addicted to alcohol and alcohol misuse
• Drivers with combined consumption of alcohol and illicit
drugs

Legal regulations • The legal limit of BAC 0.5g/L that is established in most
European countries is reasonable as the injury risk of
drivers being impaired by 0.1-0.5g/L is rather low. There are
no scientific reasons to alter this risk threshold.
• Countries that have established lower legal limits than BAC
0.5g/L, have in general lower prevalence rates of alcohol
impaired drivers in the general driving population.
Nevertheless injured drivers with high alcohol concentration
in blood are still a problem.
• The establishment of lower legal limit for specific target
groups is promising (e.g. BAC 0.0g/L for novice and
inexperienced drivers as proposed in the Commission
Recommendation of 17 January 2001 on the maximum
permitted blood alcohol content (BAC) for drivers of
motorized vehicles (notified under document number
C(2000) 4397)).
• For combined consumption lower legal limits should be
imposed (e.g. BAC 0.0g/L).
• Mandatory alcohol testing for drivers involved in accidents
with injuries should be introduced.

Enforcement • Drink-driving enforcement is cost-beneficial. Previous efforts

strategies should be continued and if necessary (countries with high
prevalence rates for alcohol) extended.
• The first priority of countermeasures should always be the
alcohol; other psychoactive substances are second priority.
• To enhance general deterrence effects random police
checks are appropriate.

88
Rehabilitation • Driver rehabilitation should be harmonized, e.g. by applying
measures common European standards and using recommendations
on good practices for rehabilitation measures developed
within DRUID.
• Driver assessment and rehabilitation should be legally
regulated and based on defined criteria.
• Alcohol offenders should be treated in separate groups, not
together with drug offenders.
• Non-addicts and addicts should be treated in separate
programs as they require different interventions or
treatments.
• Multiple offenders and offenders with a BAC ≥ 1.6g/L should
undergo an examination to preclude addiction.
• An alcohol ignition interlock can be installed during the
rehabilitation phase, yet should be combined with
rehabilitation/treatment and close monitoring.

Withdrawal • Practices of driving license withdrawal should be

measures harmonized across Europe.
• Withdrawal is an effective general and special deterrent
factor. Immediate withdrawal/suspension of driving license
and a high level of perceived risk of detection are decisive.
Sanction certainty can be increased by strict enforcement
(e.g. implementation of random alcohol and drug controls).
• The withdrawal duration should be between 3 and 12
months.
• Driver rehabilitation should be an integrated part of driving
license withdrawal.
• Conditional withdrawal should always be combined with
rehabilitative measures and close monitoring.

Future needs of • The collection of epidemiological data on a regular basis is

scientific useful having implemented new legal alcohol limits or
investigations sanctions.
• As the case-control studies are very time-consuming,
expensive and encounter legal and ethical restrictions, it is
advised to find alternative study methods of collecting
reliable data.

89
 DRUID final report 2012

2. Countermeasures to combat illicit drug impaired driving

The prevalence of illicit drugs in the general driver population is much lower than the
prevalence of alcohol. The estimated EU mean for all investigated illicit drugs is 1.9%.
Compared to alcohol (3.5%) the prevalence of single illicit drugs is very low. THC and
cocaine are the most frequently detected illicit substances in most countries. There are
high national variations in prevalence.

Target groups • Young male drivers

• Drivers with combined consumption of illicit drugs and
alcohol; several illicit drugs
• Drug consumers (stimulants, e.g. MDMA) with sleep
deprivation

Legal regulations • European agreement regarding the body fluid (especially

whole blood versus plasma) to be used for drug detection
• Regulations should be based on scientific findings; if
epidemiological and experimental data are not sufficient, an
expert team should determine cut-offs taking into account
other findings (e.g. pharmacokinetic profiles).
• European harmonization of drug analyses (e.g. cut-offs;
standardized analyzing procedures).
• Based on scientific results a risk threshold for THC
equivalent to 0.5 BAC at 3.8 ng/mL serum plus an added
value for measurement error and confidence interval could
be defined.
• Regarding legal regulations a two-tier system is
recommended. The combination of per se limits with an
impairment approach allows graded sanctions: a less
severe sanction when drugs are present above the per se
limit and a more severe sanction when the driver was
impaired in addition.

Enforcement • Increase of drug enforcement is potentially cost-beneficial,

strategies especially for countries that currently have a low
enforcement level. It is not beneficial from the road safety
point of view (in terms of fatalities) if it happens at costs of
drink-driving enforcement.
• The use of only those screening devices which fulfill
practical and scientific requirements is advised.
• Training of police officers (drug recognition expert
programs) to improve drug detection.
• Drug detection roadside actions should be conceived taking
into account pre-selection by time, place and target group
(e.g. alcohol impaired drivers) and national prevalence data.

90
Rehabilitation • Harmonization of driver rehabilitation (see section on
measures alcohol)
• Driver assessment and rehabilitation should be legally
regulated and based on defined criteria.
• Drug offenders should be treated in separate groups, not
together with alcohol offenders.
• Non-addicts and addicts should be distinguished as they
require different interventions or treatments

Withdrawal • Withdrawal in case of regular consumption of drugs should

measures be combined with adequate rehabilitation programs

Future needs of • Improvement of drug recognition expert programs and

scientific impairment checklists
investigations • Improvement of on-site screening devices which fulfill
practical as well as scientific requirements
• Further development of dried blood spot method
• The collection of epidemiological data on a regular basis is
useful having implemented new legal limits or sanctions.

91
3. Countermeasures to combat driving impaired by medicines
On a European level medicines are taken by 1.4% of drivers, as was estimated based
on the results of 13 countries. The use of medicines varies very much per country. The
risk assessment reveals medium increased accident risk when driving under influence
of psychoactive medicines.

Target groups • Health care providers and patients

• Female drivers above 50 years; especially drivers using
benzodiazepines and medicinal opiates

Legal regulations • No thresholds should be defined for medicines.

• The adequate countermeasure to combat impaired driving
is information about the possible side effects and how to act
and decide on using medicines in a safe manner while
driving. Therefore a comprehensive information system for
physicians, pharmacists and patients should be
implemented.
• Implementation of the four level classification and labeling
system developed in DRUID

Enforcement • Only appropriate if medicines are misused by patients or by

strategies healthy drivers. Legal procedures and consequences of
misuse of medicines should be in line with combating DUID
policies.
• Should focus on combined consumption of medicines and
alcohol

Rehabilitation • In case of misuse comparable to recommendations on

measures combating illicit drugs

Withdrawal • In case of misuse and combined consumption with alcohol

measure comparable to recommendations on combating illicit drugs
(see section on illicit drugs)

Future needs of • Expansion of research on impact of medicines on driving

scientific fitness
investigations • Expansion of research to examine the association between
the use of commonly prescribed psychoactive medicines
and road traffic accidents.
• Expansion of research to determine effective risk
communication strategies to inform the general public,
health care providers and patients on medicines and traffic
safety.
• Development of procedures for the assessment of fitness to
drive

92
Section 5 – Dissemination and use

All deliverables submitted and approved by the European Commission. See the
complete list of outputs and dissemination activities in Annex I.

93
References

Bickel, P., Friedrich, R., Burgess, A., Fagiani, P., Hunt, A., De Jong, G., Laird, J., Lieb,
C., Lindberg, G., Mackie, P., Navrud, S., Odgaard, T., Ricci, A., Shires, J. &
Tavasszy, L. (2006). “Proposal for harmonised guidelines.” Deliverable 5,
Developing Harmonised European Approaches for Transport Costing and Project
Assessment (HEATCO), Project funded by the European Commission under the
Transport RTD Programme of the 6th Framework Programme, Institut für
Energiewirtschaft und Rationelle Energieanwendung (IER), Universität Stuttgart,
Stuttgart, Germany.
Brookhuis KA, de Waard D (1993): The use of psychophysiology to assess driver
status. Ergonomics 36:1099-1110.
Krüger, H.-P., Kohnen, R., Diehl, M. & Hüppe, A. (1990). Auswirkungen geringer
Alkoholmengen auf Fahrverhalten und Verkehrssicherheit (Problemstudie).
(Forschungsberichte der Bundesanstalt für Straßenwesen, V 213).
O'Hanlon JF, Haak TW, Blaauw GJ, Riemersma JB (1982): Diazepam impairs lateral
position control in highway driving. Science 217:79-81.
Ramaekers JG, O'Hanlon JF (1994): Acrivastine, terfenadine and diphenhydramine
effects on driving performance as a function of dose and time after dosing. Eur J
Clin Pharmacol 47:261-266.

94