Cannabis dependence: predictors, course and treatment seeking

In the Netherlands, it is estimated that approximately 80,000 individuals may be considered as “high risk” cannabis users. Treatment demand for cannabis problems has strongly increased in the past decade. In the public debate, cannabis related problems have been linked with the recent doubling of the concentration of tetrahydrocannabinol (THC) in Dutch domestic grown marijuana. Previous studies have shown that variables involved in initiation of cannabis use differ from those involved in the development of cannabis dependence among users. Although factors involved in initiation of use have been extensively studied, risk factors involved in the transition from risky non-dependent use to cannabis dependence and the course of dependence, including factors related to treatment demand, are largely unknown. The primary objectives of the present study are: 1) To study factors involved in the transition from risky non-dependent cannabis use to cannabis dependence, 2) To study the three-year natural course of cannabis dependence, including factors predicting stability and recovery from cannabis dependence and 3) To study factors related to treatment seeking in cannabis dependent users. Secondary objectives are: 4) To compare characteristics of risky and dependent cannabis users with control subjects from the general population and 5) To validate self-reported measures of cannabis exposure. Putative risk indicators for predicting the transition from risky use to dependence and the course of dependence (persistence and recovery) are selected from the vulnerability-stress model of mental disorders, the affective-motivation model of substance use and from empirical studies. These include demographic variables, cannabis exposure, preference for high potency cannabis, personality (impulsivity, neuroticism), prior or current mental disorders, family history of substance use problems, childhood trauma, life events and functional impairment. Primary outcome is a DSM IV diagnosis of cannabis dependence. Secondary outcome is a change in cannabis use (frequency, monthly dose). Methods: The objectives will be addressed with a combination of prospective and cross-sectional research strategies and with both quantitative and qualitative methods. With targeted sampling techniques, cohorts of 275 high risk non-dependent cannabis users and 275 cannabis dependent subjects will be enrolled in a three years prospective follow-up study (objective 1 & 2). Risky use is defined as cannabis use on 12 or more days per month in the past 12 months, without fulfilling a 12-month diagnosis of DSM IV cannabis dependence. Cannabis dependence is a 12-months diagnosis of DSM-IV cannabis dependence. All subjects will be between 18 and 30 years at baseline. Structured face-to-face interviews will be conducted at baseline (To), first follow up after 18 months (T1) and second follow up after 36 months (T2). Forty cannabis users from the cohort studies will be selected on the basis of transitions in cannabis dependence to be interviewed in-depth twice (shortly after T1 and T2) on the dynamics underlying changes in cannabis dependence status. To better understand reasons for treatment seeking, a sample of 100 cannabis dependent users (18-30 years) applying for treatment at addiction care centres will also be interviewed in a cross-sectional assessment. Reasons for not seeking treatment and ‘unmet need’ are also investigated among non-treatment seekers from the cohort of cannabis dependents users (objective 3). For descriptive analyses, subjects from both cohorts will be compared at baseline with an age and sex matched control group from a representative general population study, NEMESIS 2 (objective 4). Finally, measures of self-reported cannabis consumption are validated by additional toxicological data (cannabinoids in hair and in cannabis samples) and by measuring smoking behaviour in vivo (objective 5). This will be done at T1 among a sub-sample of 50 cannabis users from the two prospective cohorts. Data will be analysed using logistic and negative binomial regression analyses.