Assessing problematic cannabis use

Abstract

Many cannabis users do not report experiencing harm from use, suggesting that not all use is cause for concern. The objective of this research is to establish a threshold beyond which cannabis use becomes problematic. Thresholds from the cannabis portion of the Alcohol, Smoking, Substance Involvement Screening Test (ASSIST) and the Cannabis Use Disorders Identification Test (CUDIT) were compared against a simple frequency of use question (daily use) for their ability to identify problematic users. Data are drawn from the 2008 Canadian Alcohol and Drug Use Monitoring Survey (n = 16,674). Effectiveness of the thresholds for predicting signs of problematic use was assessed in terms of sensitivity and specificity and in terms of positive and negative predictive values (NPV). The ASSIST (threshold of 8) and the CUDIT (thresholds of 6 and 8) outperformed the frequency measure with regards to sensitivity and specificity; the CUDIT with a threshold of 6 struck the best balance. With regards to positive predictive values, the CUDIT with a threshold of 8 revealed the highest value, whereas the daily use threshold, the CUDIT (threshold of 6 and 8) and the ASSIST (threshold of 8) revealed approximately the same NPV. Rates of harm were low for those using less than daily. Findings suggest that the CUDIT with a threshold of 6 may be best for use in population surveys when there is typically no reason to favor either sensitivity or specificity. Health care professionals may achieve greater precision using a threshold of 8 on the CUDIT.