Annaheim, Beatrice

Weeding out : problematic cannabis use : an analysis of instruments to identify "problematic" patterns of use at the population level, in particular the cannabis use disorders identification test (CUDIT) and its implementation / [publication] : / Beatrice Annaheim - [Fribourg Switzerland] : [s.n.], 2013 - 129 p.

[Abstract] BACKGROUND: Cannabis use is widespread in most developed societies despite the illicit status of the substance. While most users remain socially integrated, for a minority their cannabis use is associated with a multitude of health and social problems. The high prevalence rates of cannabis use, coupled with evidence of a correlation with different problems and rising treatment demand, make cannabis use a public health-relevant issue. Different actors have stressed the need for adequate screening instruments (i.e. short questionnaires) to estimate the extent of "problematic" cannabis use at the population level. OBJECTIVES: The aims of the present thesis are to provide an overview of the instruments available to screen for cannabis-related problems, and to validate and further improve one of the instruments (the Cannabis Use Disorders Identification Test, CUDIT) that is best suited to surveys at the population level. In addition, the present thesis seeks to implement this instrument for research on correlates of "problematic" cannabis use (readiness for reducing or quitting cannabis use and sources of cannabis supply) in general population samples. MATERIAL AND METHODS: The thesis comprises six papers, all of which have been published in scientific and professional journals. Five papers are based on data from a comprehensive general population survey conducted in 2004 and 2007 as part of the Swiss Cannabis Monitoring Study. One paper - the overview of available screening instruments - is based on a literature and data base research. A first validation study of the CUDIT applies classical test theory (importance of the single items, internal consistency by Cronbach's alpha, Receiver Operating Characteristics (ROC) analyses), while a second validation is based on a probabilistic statistical approach, Item Response Theory (Samejima's graded response model is estimated to plot Category Response, Item Information, and Test Information Curves). New items (i.e. questions) to supplement the CUDIT are constructed on the basis of ten qualitative interviews with cannabis users and experts. RESULTS: The systematic review of available screening instruments showed that the CUDIT as well as two other instruments (CAST and CUPIT) seem best suited to screen for cannabis-related problems in general population surveys. The first validation of the CUDIT, by means of classical test theory, indicated the poor performance of two items. The second validation, by means of Item Response Theory, showed an improvement in psychometric performance when three original items were replaced. When the CUDIT was implemented for research among cannabis users in the general population, it was shown that experiencing problems, as measured by the CUDIT, motivates users to start thinking about changing their behaviour (i.e. quitting or reducing cannabis use). it was also shown that cannabis users who buy cannabis either from friends or dealers have an increased risk of experiencing cannabis-related problems, while users who procure cannabis from friends are at lower risk of experiencing problems, according to the CUD1T. DISCUSSION: To date, a number of instruments which screen for cannabis-related problems are on their way to becoming standard use. The CUDIT is one of those. In parallel to the present thesis, and on an international level, a few other validation studies have been realised regarding the CUDIT, mainly in a clinical setting. Although the CUDIT has been proven as a viable screening instrument, there is still room for improvement. Concretely, as shown in the present thesis and as regards general population research, the CUDIT needs to be supplemented with additional "easy" items. The endeavour of the present thesis to weed out "problematic" (or "more problematic") from "non-problematic" (or "Iess problematic") patterns of cannabis use implies as a logical consequence that cannabis users are a heterogeneous group. This means that public health measures must be tailored according to the cannabis users' individual level of risk. The present thesis, therefore, supports a Harm Reduction approach towards cannabis.