You are on page 1of 11

Journal of Substance Use, August 2007; 12(4): 233242

ORIGINAL ARTICLE

Validation of the CAST, a general population Cannabis


Abuse Screening Test

S. LEGLEYE1, L. KARILA2, F. BECK3, & M. REYNAUD2


1

French Monitoring Center for Drug and Drug Addiction (OFDT), & National Institute for Medical
Research (Inserm) U669, and University Paris XI, 2Paul-Brousse Hospital, 12 avenue Paul VaillantCouturier, 94804 Villejuif, and 3National Institute for Health Prevention and Education (INSERM)
& Research Center on Drugs, Mental health & Society (CESAMES), CNRS, UMR8136; Inserm,
U611; University Paris V Descartes, Paris, F-75006, France

Abstract
Background: Although the use of cannabis among young people has become a major public health
concern, there is no such test in France for general population surveys.
Objective: This article aims to present a short screening test for cannabis abuse among adolescents and
young adults in general population surveys, the CAST (Cannabis Abuse Screening Test) designed at
the French Monitoring Center for Drug and Drug addictions (OFDT).
Methods: The survey was conducted in France among 1728 scholars aged 1422, in secondary
schools, high schools, and university. After presenting internal consistency and factorial structure, the
authors compared the CAST with the part of the POSIT (Problem-Oriented Screening Instrument
for Teenagers) dealing with alcohol and drug abuse among cannabis users in order to validate the
optimal thresholds of the CAST for the high risk of abuse defined for the POSIT.
Results: The CAST appears to be unidimensional and have high internal consistency (Cronbachs
alpha equals 0.81). Among cannabis users whoa are low alcohol consumers, CAST presents very high
sensitivity and specificity compared with the POSIT (respectively, 93 and 81%).
Conclusions: The CAST seems to be an efficient tool in order to screen for cannabis use disorders
among adolescents and young adults. Further research is needed to confirm this result.

Keywords: CAST, cannabis, abuse, general population, France.

Introduction
For several years, surveys carried out on young people in the general population in France
have clearly shown the increasing usage rates of the most common drugs, including
cannabis (Beck, Legleye, & Spilka 2004; Beck, Legleye, & Spilka, 2006a; Beck, Legleye,
Spilka, Briffault, Gautier, Lamboy, Leon, & Wilquin, 2006b). However, these surveys do
Correspondence: S. Legleye, OFDT, 3 Avenue du Stade de France, 93218 La Plaine Saint-Denis, 93218, France. Tel: + 33-1 41
62 77 37. Fax: + 33-41 62 16 00. E-mail: stleg@ofdt.fr
ISSN 1465-9891 print/ISSN 1475-9942 online # 2007 Informa UK Ltd.
DOI: 10.1080/14659890701476532

234

S. Legleye et al.

not yet contain data on harmful use (in line with the classification set out in ICD-10the
International Statistical Classification of Diseases and Related Health Problems) or on
abuse (in line with the criteria set out by the 4th edition of the Diagnostic and Statistical
Manual of Mental Disorders) of these substances. The European Monitoring Centre for
Drugs and Drug Addiction (EMCDDA) has, for several years, been putting forward the
notion of problematic drug use, meaning the injection or long term or regular use of
opiates, cocaine and / or amphetamines. This notion has become one of the five key
indicators adopted by the European Union (EMCDDA, 2002). The EMCDDA is
currently seeking to widen this notion to include cannabis, and is particularly basing its
work on clinical and epidemiological validation work carried out in France (Karila, Legleye,
Donnadieu, Beck, Corruble, & Reynaud, 2004) and abroad (Heishman, Singleton, &
Liguori, 2001; Adamson & Sellman, 2003; Dale, 2003; Okulicz-Kozaryn & Sieroslawski,
2004), as well as making a critical revision of international experiments relating to this
question (Beck & Legleye, 2007). The drawing-up of a European definition is ongoing, but
until the process is completed, the following definition is being used: use that is likely to
cause major health and social problems for the individual or for others. This notion is still
being discussed, and the criteria and indicators that would make it operational are not yet
available. However, this kind of data is proving vital in the field of public health. Being able
to pick up harmful, abusive and problematic substance use at an early stage is crucial. It
enables both information and adapted care to be provided if necessary, in order to prevent
the individual becoming dependent on other drugs.
A full study on problematic drug use would require either retrospective surveys of the
general population or of current drug users who are in difficulty, or longitudinal surveys,
but cross-sectional studies of the general population are also helpful. Work carried out in
English-speaking countries has shown that short tests are effective tools for screening and
that direct questions are often more effective than more complex forms of questioning
(Rost, Burnam, & Smith, 1993; Svanum & McGrew, 1995). The section of the
questionnaire devoted to this is called the CAST (Cannabis Abuse Screening Test) and
it has been drawn up based on a review of existing subject material (Reynaud, 2002).
Several English-language tools enable harmful drug use amongst adolescents and young
adults to be screened for and assessed, but to our knowledge there are no such tools
available in French, apart from the Canadian Depado questionnaire (Bergeron, Landry,
Ishak, Vaugeois, & Trepanier, 1992), which deals with all illegal substances. In order to
progress in this field, the Adolescent, techniques devaluation de lusage nocif (Adotecno)
survey has been implemented in partnership with the Paul-Brousse hospital and the
OFDT. The objective of this survey is to provide an epidemiological overview of harmful
drug use with the help of different screeningCRAFFT (Car Relax Alone Forget Family
or Friends Troubles; Knight, Shrier, Bravender, Farell, Vander Bilt, & Shaffer, 1999),
ALAC (Alcohol Advisory Council of New Zealand; McLaney, Del Boca, & Barbor, 1994),
and CAST and assessment tools (POSITProblem Oriented Screening Instrument for
Teenagers) amongst a population group of 1728 secondary school and higher education
students. CRAFFT deals with drug use; ALAC is a self-evaluation questionnaire on
cannabis use. In the long term, clinical validation of these tests will enable their use to be
validated.
These short questionnaires have not been validated in a clinical situation in France.
However, they are used in drug addiction treatment centres set up by the Interministerial
Mission for the Fight Against Drugs and Drug Addiction (MILDT), as a simple and
efficient way to describe patterns of use and to collect data on this specific public

Validation of the CAST

235

(Obradovic, 2006). The results presented in this paper only refer to the CAST, which aims
to screen for harmful cannabis use.

Method
The survey
A total of 1728 questionnaires were collected in 2003 from the 2000 questionnaires
originally sent out. The non-response rate was 15%, although there was no parental refusal.
The educational establishments surveyed belong to the secondary sector (24 secondary
schools and 19 high schools in the Franche Comte region) and to the university sector,
mainly first age (a school of pharmacy, a school of medicine, three technical schools, two
law schools, one optics technical certificate class and one biology class from the Ile de
France region). The sample is not probabilistic: these establishments were chosen due to
the presence of partner doctors from voluntary health networks who were willing to
participate. Seventeen questionnaires that did not include the sex of the individual or their
year of birth were excluded from the sample. The sample used for analysis was therefore
made up of 1711 subjects aged from 14 to 22 divided up into three population groups:
30.3% of the individuals (n5520) were in the fourth year at secondary school (mean
age515.4, SD50.63, min515, max518), 27.0% (n5462) were in the lower sixth form,
upper sixth form or in the final year of a technical certificate at sixth form college (mean
age517.7, SD50.98, min516, max522) and 42.6% (n5729) were in higher education
(mean age519.8, SD51.08, min517, max522).
The questionnaires were completed in class under the supervision of the school doctor or
nurse during medical visits. A 1-page letter introducing the project including instructions
for completing the questionnaire was distributed to the participants and read by the person
in charge of the survey. Some points were specifically clarified, namely confidentiality,
anonymity, freedom to answer. The participants were asked to respond to each question
without taking into account their responses to previous similar questions. The opportunity
was offered to consult a health professional if necessary. It took about 35 min to complete
the entire questionnaire. Completed questionnaires were put in an envelope by the
participants and sent to the body in charge of inputting the questionnaire data. Data
inputting was supervised and checked by the authors. Checks were carried out on a random
sample of 50 questionnaires: they did not reveal any mistakes.
Ethical and legislative aspects
The study was under the supervision of school doctors, in agreement with the studys
designers; an explanatory letter was sent by the consulting doctor in each region to the
heads and deans of the educational establishments selected to take part. Parents of
secondary school children received an information letter about the survey, its aims and
confidential nature. Parents of pupils aged under 18 could refuse their child(ren) to
participate: in this case, they had to fill-in and to sign the letter and to give it to their
children. Major scholars or students could refuse to participate, but finally, no parents or
adolescents refused to.
Data collecting and statistical analysis were carried out by a different teams. An e-mail
address and telephone number were made available to participants who wished to have
more information on drugs or on the survey, or to find out the results of the study.

236

S. Legleye et al.

Measurement tools
The first part of the questionnaire assesses the frequency of use and amounts taken of
alcohol, cannabis, tobacco, ecstasy, and psychotropic drugs. The questionnaire takes 5 min
to complete; it has been used every year since 2000 as part of the national Escapad survey
(Beck et al., 2004) on 16,000 young people; it follows recommendations laid out by the
European Monitoring Centre For Drugs And Drug Addiction (EMCDDA, 2002). The
CAST questionnaire, developed by the OFDT, screens for different aspects of harmful
cannabis use by assessing the frequency of the following events throughout an individuals
life (never, rarely, sometimes, quite often, and very often): seemingly non-recreational use
(smoking alone or before midday), memory disorders, being encouraged to reduce or stop
using cannabis, unsuccessful attempts to quit, problems linked to cannabis consumption.
The questions are as follows:

N
N
N
N
N
N

Have you ever smoked cannabis before midday?


Have you ever smoked cannabis when you were alone?
Have you ever suffered from memory disorders due to your consumption of cannabis?
Have friends or family members ever told you that you should reduce your cannabis
consumption?
Have you ever unsuccessfully tried to reduce or stop consuming cannabis?
Have you ever had problems due to your consumption of cannabis (arguments, fights,
accidents, anxiety attacks, poor results at school)?

The reference period is the individuals entire life as the questionnaire is basically designed
for adolescents or young adults. Positive response thresholds vary from question to
question. The threshold is set at sometimes for the first two questions as they do not
screen problems but frequencies of use in some contexts, and at rarely for the others. An
individuals final score can therefore be from 0 to 6. The designers of the questionnaire
expect a score of four or above to indicate problematic cannabis use and the present paper
seeks to confirm this statement. This questionnaire has been used since 2002 as part of the
ESCAPAD survey, which is carried out on a yearly basis by the OFDT.
POSIT stands for Problem Oriented Screening Instrument for Teenagers. This 139-item
yes/no screening questionnaire was developed by NIDA (National Institute on Drug
Abuse) in the United States and was validated on adolescents aged 1219 in a school
environment. It is designed to identify psychological, physical, and social health problems,
which may require more in-depth assessment and treatment. It has been chosen because it
was a tool specific to adolescents and young adults, and because a validated French
translation was available. This questionnaire evaluates 10 areas (or factors) that are
classified from A to J, including psychoactive substance abuse (factor A), as well as physical
and mental health, family relations, etc. (McLaney, Del Boca, & Babor, 1994; Hall,
Richardson, Spears, & Rembert, 1998; Knight, Goodman, Pulerwitz, & DuRant, 2001).
Thus, it describes some aspects of the mental health of the respondents. The whole
questionnaire takes around 25 min to complete. The classical two-risk thresholds for the A
factor were used as a reference for defining the optimal thresholds in the CAST.
Statistical analysis
Statistical analysis was carried out at the OFDT. Qualitative variables are described using
percentages and the significance of their difference is evaluated by the x2 Pearson test. The

Validation of the CAST

237

non-response rate for the sample used is low and never exceeded 5%. Principal
components analysis (PCA) followed by a factor analysis (with the varimax rotation) were
carried out for each screening tool and their internal consistency was assessed using
Cronbachs a coefficient. Kappa index were also computed for questions of both CAST
and POSIT questionnaires.
Determination of optimal thresholds for the CAST was calculated using Receiver
Operating Characteristic (ROC) curve and Youdens index. Youdens index states the
percentage of individuals graded positive in the CAST compared with the POSIT test,
according to formula: J5Sensitivity+Specificity2100. Analysis was carried out on SAS V8
and R V2.1. As the sample is a convenient sample, no weighting procedure was used,
neither analytical methods in order to take into account the clustered nature of the data.

Results
Sample description
Regular alcohol intoxication, as well as regular cannabis use is significantly more common
amongst young males than young females. Young females state that they consume more
psychotropic drugs whether they are hypnotic, anxiolytic, anti-depressants, or stimulants,
than young males, which is a classical result. The proportion of users is higher amongst
high school students than amongst the other individuals who were surveyed. This is
particularly the case for regular alcoholic intoxication, and regular or daily cannabis use.
According to the CAST, 9.6% smoked from time to time in the morning, 6.6% smoked
alone from times to times; 11.4% declared having encountered some memory problems
because of their cannabis use and 7.0% heard reproaches from friends and family; 6.1%
said they had had a problem when using cannabis, but only 3.6% said that they had
unsuccessfully tried to reduce or stop their consumption of cannabis. Four-fifths of the
sample obtain a score of zero, 7.1% a score equal to or higher than 3 and 4.9% a score
equal to or higher than 4.
According to the Kappa index, answers are relatively independent: values vary between
0.28 and 0.49 except for the two questions related to context of use (smoking cannabis in
the morning and smoking alone) where the value reaches 0.56. A Principal Components
Analysis (PCA) shows that the first axis (with a value of 3.18) explains 53.0% of the
variance (the second axis explains 15.2%, the third 9.9%). This analysis deals with
continuous variables; however, results are very close to the binary variables used to
calculate the score, even if that is less mathematically correct. The first value would
therefore be 3.14 and the associated axis would carry 52.3% of the total variance, the
second would carry 13.8% and the third 10.3%. It is therefore possible to correctly measure
its internal consistency using Cronbachs alpha coefficient. The result is a50.81.
The A factor of the POSITs test assesses psychoactive substance abuse. The designers
defined the low risk threshold as a score of zero, the moderate risk threshold as a score
between 1 and 6, and the high risk threshold as a score equal to or higher than 7. Almost
two-thirds of the sample (63.5%) obtained a score of zero. The remaining third represented
a moderate risk (score>1) and 4.6% represented a high risk (score>7). No individual
obtained the maximum score of 17.
The A factor in the POSIT test appears multi-dimensional, which is quite natural
according to the number of items: the first principal component analysis axis only includes
30% of total inertia and the second one 8% (variables considered here are all binary). The

238

S. Legleye et al.

factorial analysis suggests that two factors should be retained. Shortly, the first axis shows
drug abuse and dependence, and it resumes most of the information while the second axis
shows non-problematic or non-abusive use. Factor A will, nevertheless, be used in the
standard way following its authors recommendations.
Model calibration
The POSIT questionnaire does not solely deal with cannabis use, like the CAST, but also
covers abuses of all psychoactive drugs. As alcohol was the most widely consumed
substance in the sample group, it was important to carry out this analysis by selecting a subgroup of cannabis and moderate alcohol users so that possible problems screened for by the
POSIT did actually concern cannabis and not another substance. The chosen sub-group is
made up of individuals who had used cannabis during the past 30 days and who said that
they had drunk alcohol fewer than 10 times during that period (which was the case for 90%
of those surveyed, see Table I) and who had experienced a maximum of three alcoholic
intoxications during the past 12 months (84% of those surveyed). This sub-group is
reduced in size (n5123), but this limitation stems from the low levels of cannabis use in the
sample (only 18% of individuals surveyed had consumed cannabis during the past 30 days).
This group is made up of 61% young males (n575) and 39% young females coming in
approximately equal numbers from the three participating types of educational establishment: 24% from secondary schools, 36% from sixth form colleges, and 40% from higher
education. They had all smoked cannabis during the past 30 days: one-third regularly
smoked it (34%), fewer than one in 10 smoked it on a daily basis (9%), and 30% said that
they had smoked at least three joints the last time they used cannabis.
Table I. Psychoactive substance use by sex in the ADOTECNO sample (%)
Sixth form
Secondary School high school
Combined
Combined
(n5520)
(n5462)
Alcohol, lifetime
Alcohol, regular use
(a)
Alcohol, daily use
Drunkenness,
lifetime
Drunkenness, 3
times + (b)
Tobacco, daily use
Cannabis, lifetime
Cannabis in the
past 30 days
Cannabis, regular
use (a)
Cannabis, daily use
Ecstasy, lifetime

University
Combined
(n5729)

Overall sample
Young females Young males Combined
(n5986)
(n5712)
(n51711)

84.5***
6.5***

95.0
13.9

90.0
10.0

87.9*
6.6***

91.3
13.1

89.6
9.9

0.4
30.6***

0.4
64.3

0.8
46.4

0.3
37.0***

0.7
55.4

0.5
46.2

8.9***

25.0

15.1

8.5***

23.0

16.0

19.8***
26.1***
12.3***

32.5
50.3
23.6

26.3
46.9
17.7

24.9
34.8***
25.4***

26.9
47.5
9.8

25.9
41.2
17.7

5.1***

12.3

7.7

3.3***

12.9

8.9

1.6**
2.5***

4.9
5.5

2.2
1.8

1.3***
1.5***

4.2
4.5

2.8
3.0

*, **, ***Chi-squared test used to carry out the comparison between the sexes and between educational
establishments respectively significant at thresholds of 0.05, 0.01 and 0.001. (a) Regular use: used at least 10 times
during the past 30 days. (b) Over the past 12 months. Source: Adotecno, Paul Brousse University Hospital
(HUPB), OFDT, 2003.

Validation of the CAST

239

Table II. CAST optimal thresholds and features relating to POSIT factor A
Last month cannabis
users and light alcohol
users (a): n5123

Threshold

Sensitivity

Specificity

Youden

1
4

76.7
92.9

66.7
81.4

44.4
74.3

Fa>1: n598
Fa>7: n516

(a) Regular cannabis use, no regular alcohol use, maximum of three alcoholic intoxications during the past twelve
months. Source: Adotecno, Paul Brousse Hospital, OFDT, 2003.

In this subgroup, the CAST still appears one-dimensional after analysis of the principal
components (the first value of 2.62 is linked to an axis explaining 43.7% of the inertia and
the second value of 0.95 is linked to an axis explaining 15.9%) and factorial analysis.
Cronbachs alpha coefficient is 0.74.
The CAST proved to be very effective when screening for high risks of abuse (Fa>7):
Youdens index showed that it screens correctly three-quarters of the individuals surveyed
and the area under the ROC curve (in the simple logistic regression High risk in the Factor
A of the POSIT5CAST Score) equals 0.92, which is pretty high. On the other hand, the
CAST is poor when screening for slight risks (Fa>1).
It therefore seems that two risk levels can be put forward for the CAST, namely a low
risk level for young people who obtain between 1 and 3 on Factor A of the POSIT test (and,
therefore, obtain between 1 and 2 on the CAST), a moderate level for those who obtain
scores between 4 and 6 (3 on the CAST), and a high risk level for those who obtain at
least 7 (4 on the CAST). The Table III sums up the features of the CAST at previously
determined optimal thresholds for ordinary frequencies and ways of use. The discrepancy
between these two thresholds highlights the difference between screening for frequency of
use and type of consumption and screening for problematic abusive use.
People who obtain at least 4 with the CAST obtain higher scores to the other
psychopathological dimensions of the POSIT. For example, their mean score for physical

Table III. Features of the CAST screening test for cannabis users
Sensitivity
Specificity
Youdens
Index
n
%
CAST>1

CAST>3

CAST>4

Regularly smoke
cannabis

Smoke cannabis
on a daily basis

122
8.2
92.6
86.9
79.5
63.0
97.3
60.3
48.2
98.5
46.7

41
2.8
97.4
83.1
80.5
79.0
94.7
73.7
68.4
96.7
65.1

Source: Adotecno, Paul Brousse Hospital, OFDT, 2003.

>3 joints smoked the Regularly smoke cannabis


last time cannabis
and >3 joints smoked the
was used
last time cannabis was used
134
7.7
78.2
87.5
65.7
50.4
96.9
47.3
37.0
98.0
35.0

69
4.0
98.4
85.6
84.0
80.3
96.3
76.6
60.7
97.8
58.5

240

S. Legleye et al.

health is worse (3.7 vs. 2.0, p,0.001), as well as their mean score for mental health (4.1 vs.
2.9, p,0.001) and school situation (4.2 vs. 3.2, p,0.001).
Discussion
The Adotecno study, carried out in conjunction with the Paul Brousse University Hospital
Addictions Department (under the leadership of M. Reynaud and L. Karila), and the
OFDT (under the direction of S. Legleye and F. Beck) is part of the implementation of
early screening and assessment strategies for the harmful use of psycho-active substances
amongst adolescents and young adults. In the long term, this research should provide
psychoactive substance abuse screening tools, particularly for cannabis in the general
population but also in clinical contexts.
The goal of the study was to validate the threshold scores recommended by the designers
of the CAST, a screening tool for harmful cannabis use in the general population, using the
drug abuse section of the POSIT questionnaire as a gold standard. Analysis was based on
the answers provided by 1711 young people aged from 14 to 22 who filled in a paper-based,
self-administered questionnaire in class under the surveillance of school medical staff in
their secondary school, sixth form college or university. There are some methodological
limitations to our study. The sample cannot be considered to be representative of the
school or university population in France. Educational establishments were not chosen
randomly, but according to the presence of doctors and heads who volunteered to take part.
Nevertheless, volunteering does not imply that prevention measures were in force in these
establishments nor that they provide specific drugs information, anymore than it means
that pupils or students were specially selected. There must, therefore, be only a slight
impact on the behaviours that were being studied.
Three representative surveys of the general population enable us to compare with the
prevalences measured in Adotecno, for three specific age brackets: 1516-year-olds in the
European School Survey project on alcohol and other drug (Espad: Choquet et al., 2002;
Beck, Legleye, & Peretti-Watel, 2002: n55000), 1718-year-olds (Escaped: Beck, et al.,
2004, n515,000) and 1922-year-olds (Health Barometer, a national phone survey:
Guilbert, Baudier, Gautier, Goubert, Arwidson, & Janvrin, 2001, n52,000). Compared
with all these national surveys, young people who took part in Adotecno appeared to be less
frequent users of cannabis. Escapad 2003 also enabled us to verify that this lower level of
consumption was combined with a smaller proportion of young people who were scanned
as a positive for threshold four of the CAST: thus standardized, the test showed 5.9% of the
1718-year-olds in the Adotecno survey (22% of regular users) were at threshold four as
opposed to 8.9% (5.4% among girls and 12.4% among boys) of those surveyed by Escapad
(53% of regular users). However, this bias does not call into question the validity of the
study insofar as the number of psychoactive substance users is sufficient and the quality of
completed answers satisfactory.
Question repetition in the three screening tools that were used one after the other must
have affected the attention levels of the respondents and must have generated feelings of
weariness. It is difficult to assess these factors, particularly as the session reports filled in by
the interviewers did not show any dispersal of the participants, but the non-response rate
for each CAST question remained less than 4%, which is quite low. In this analysis, we
took into account the non-respondents, assuming that a non-response to an item is a null
response (cotation50). The number of the non-respondents of the A factor of the POSIT
and the CAST is 125: we checked that the exclusion of these non-respondents in this

Validation of the CAST

241

analysis does not affect the results neither the conclusions: those individuals do not belong
to the subsample used for the model calibration. Furthermore, in general population
surveys, the impact of such a non-response rate on prevalence is very low.
The choice of the POSIT test as a reference tool could also be criticized as it has not been
validated in France. However, it is a tool that enables a medical, psychological and social
profile of the subjects to be obtained without being restricted to drug use problems. It was
the POSITs role as a general standard, which rules out the need to have an addiction
consultation using the DSM criteria as a first step, which interested the authors more that
its supposed capacity to screen for young peoples problems in order to then deduce valid
results for the entire population. A forthcoming study focused on 1516-year-olds, based
on ESPAD methodology will soon allow to validate the CAST with the MINI cannabis
among 700 scholars.
Conclusion
The ADOTECNO study, carried out on a non-clinical sample of 1728 subjects in
secondary schools, sixth form colleges, and universities, has enabled the validation of the
CAST threshold scores. The study has also enabled provisional recommendations to be
implemented concerning screening and assessment of harmful drug consumption in this
population group by defining three risk levels for the CAST, which specifically deals with
problematic cannabis use (low and moderate risk for scores under 3, and high risk for
scores of 4 or more). Thus, this research contributes to the definition of a common
European screening tool for cannabis use disorders. After discussion of this results by the
steering committee and approval of the whole members of the project, the CAST has been
recently proposed as an optional module questionnaire for the European School Project on
Alcohol and other Drugs, which will take place in more than 35 countries in 2007 (Hibell,
Andersson, Bjarnasson, Ahlstrom, Balakireva, Kokkevi, & Morgan, 2004).
Nevertheless, other studies are needed to further develop this work, particularly
concerning the use of these tools on patients in hospital or on those who attend specialized
centres. Finally, going beyond the epidemiological and clinical validation of tests, which
screen for problem users, a framework for applying these tests now needs to be drawn up
(targeted population group, location, protocol for administering the test, training for people
in charge of questionnaire completion sessions). This discussion must define the
information with which screened subjects will be provided and / or how to guide them
towards appropriate care.
References
Adamson, S. J., & Sellman, J. D. (2003). A prototype screening instrument for cannabis use disorder, the
Cannabis Use Disorders Identification Test (CUDIT) in an alcohol-dependent clinical sample. Drug and
Alcohol Review, 22, 309315.
Beck, F., Legleye, S., & Peretti-Watel, P. (2002). Alcool, tabac, cannabis et autres drogues illicites parmi les ele`ves de
colle`ge et de lycee, ESPAD France 1999, Volume 2. Available at: http://www.ofdt.fr/BDD/publications/docs/
espad992.pdf (accessed February 2007).
Beck, F., Legleye, S., & Spilka, S. (2004). Drogues a` ladolescence Niveaux et contextes dusage de cannabis,
alcool, tabac et autres drogues a` 1718 ans en FranceESCAPAD 2003, OFDT. Available at: http://
www.ofdt.fr/ofdtdev/live/publi/rapports/rap04/epfxfbka.html (accessed February 2007).
Beck, F., Legleye, S., & Spilka, S. (2006a). Les drogues a` 17 ans, evolutions, prises de risques et contextes
dusageResultats de lenquete nationale Escapad 2005. Tendances, 49, OFDT. Available at: http://
www.ofdt.fr/ofdtdev/live/publi/tend/tend49.html (accessed February 2007).

242

S. Legleye et al.

Beck, F., Legleye, S., Spilka, S., Briffault, X., Gautier, A., Lamboy, B., Leon, C., & Wilquin, J.-L. (2006b). Les
niveaux dusage des drogues en France en 2005, exploitation des donnees du Barome`tre sante 2005. Tendances,
48, 16. Available at: http://www.ofdt.fr/ofdtdev/live/publi/tend/tend48.html (accessed February 2007).
Beck, F., & Legleye, S. (2007). Measuring cannabis related problems and dependence at the population level. In
R. Room, & S. Rodner (Eds.), Cannabis Monograph, European Monitoring Center on Drug and Drug Addiction
(EMCDDA). (In press.).
Bergeron, J., Landry, M., Ishak, I., Vaugeois, P., & Trepanier, M. (1992). Les cahiers de recherche du RISQ,
Montreal. Available at: http://www.risq-cirasst.umontreal.ca (accessed February 2007).
Choquet, M., Beck, F., Hassler, C., Spilka, S., Morin, D., & Legleye, S. (2004). Les substances psychoactives
chez les collegiens et lyceens, consommation en 2003 et evolutions depuis 10 ans. Tendances, 35, 16. Available
at: http ://www.ofdt.fr/ofdtdev/live/publi/tend/tend35.html (accessed February 2007).
Choquet, M., Ledoux, S., & Hassler, C. (2002). Alcool, tabac, cannabis et autres drogues illicites parmi les ele`ves
de colle`ge et de lycee, ESPAD France 1999, Volume 1, OFDT report. Available at: http ://www.ofdt.fr/BDD/
publications/docs/espad991.pdf (accessed February 2007).
Dale, A. (2003). A marijuana screening inventory (experimental version), description and preliminary
psychometric properties. American Journal of Drug and Alcohol Abuse, 29(3), 619646.
EMCDDA (European Monitoring Centre for Drugs and Drug Addiction). (2002) Handbook for surveys on drug
use among the general population, EMCDDA project CT.99.EP.08 B, Lisbon, EMCDDA.
Guilbert, P., Baudier, F., Gautier, A., Goubert, A. C., Arwidson, P., & Janvrin, M. P. (2001). Barome`tre sante
2000. Volume 1, Methode, 144 p. CFES.
Hall, J. A., Richardson, B., Spears, J., et al. (1998). Validation of the POSIT, comparing drug using and abstaining
youth. Journal of child and adolescent substance abuse, 8(2), 2961.
Heishman, S. J., Singleton, E. G., & Liguori, A. (2001). Marijuana craving questionnaire, development and initial
validation of a self-report instrument. Addiction, 96, 10231034.
Hibell, B., Andersson, B., Bjarnasson, T., Ahlstrom, S., Balakireva, O., Kokkevi, A., & Morgan, M. (2004). The
ESPAD Report 2003. Alcohol and Other Drug Use Among Students in 35 European Countries, The Swedish Council
for Information on Alcohol and Other Drugs, Pompidou Group CAN.
Karila, L., Legleye, S., Donnadieu, S., Beck, F., Corruble, E., & Reynaud, M. (2004). Consommations nocives de
produits psychoactifs; resultats preliminaires de letude ADOTECNO. Alcoologie et Addictologie, 26(2), 99109.
Knight, J. R., Shrier, A., Bravender, T. D., Farell, M., Vander Bilt, J., & Shaffer, H. J. (1999). A new brief screen
for adolescent substance abuse. Archives of Pediatrics and Adolescent Medicine, 153(6), 591596.
Knight, J. R., Goodman, E., Pulerwitz, T., et al. (2001). Reliability of the POSIT in adolescent medical practice.
Journal of Adolescent Health, 29, 125130.
McLaney, M. A., Del Boca, F., & Babor, T. A. (1994). Validation study of the Problem Oriented Screening
Instrument for Teenagers (POSIT). J Mental Health, 3, 36376.
Obradovic, I. (2006). Premier bilan des consultations cannabis. Tendances, 50, 14. Available at: http://
www.ofdt.fr/ofdtdev/live/publi/tend/tend50.html (accessed February 2007).
Okulicz-Kozaryn, K., & Sieroslawski, J. (2004). The Development Of Screening Tests For Adolescents Drug- And
Marihuana-Problem Use, paper presented at KBS 2004, Helsinki.
Reynaud, M. (2002). Usage nocif de substances psychoactives, La documentation francaise.
Rost, K., Burnam, M. A., & Smith, G. R. (1993). Development of screeners for depressive disorders and
substance disorder history. Medical Care 1993, 31, 189200.
Svanum, S., & McGrew, J. (1995). Prospective screening of substance dependence, the advantages of directness.
Addictive Behavior, 20, 205213.

You might also like