NDARC Technical Report No. 47 (1997)


Cannabis is the most widely used illicit drug in many Western countries. However, by comparison to tobacco, alcohol and the opiates, little is known about its dependence potential, or the patterns and correlates of symptoms of cannabis dependence. More generally, there has been a lack of research into the characteristics and experiences of long term cannabis users, the group arguably most likely to experience dependence. This study provides detailed information on a sample of long-term cannabis users, recruited and interviewed in Sydney, Australia. It aimed to provide further information on their patterns and experiences of cannabis use, and to compare them to a recently studied population of long-term users in rural NSW. Its particular goal was to study the prevalence and nature of cannabis dependence symptoms among long-term, urban cannabis users, using four different dependence measures.

Two hundred cannabis users, recruited primarily through advertising, were administered a structured questionnaire. Entry to the study required at least weekly use of cannabis for a minimum of five years for males and three years for females. Cross-sectional data were collected on respondents' characteristics and experiences of use and dependence, their health status and other information that may predict behaviour one year later, as assessed at a further interview.

Just over half the sample were male (58%). Respondents had been regularly using cannabis for an average of 11 years. More than half the sample typically used daily (56%), while three quarters (74%) used at least four times a week. The most common route of administration was in a waterpipe or bong, while respondents almost universally smoked the flowering heads of the plant. Cannabis was used in much the same way as people use alcohol, namely, to relax, relieve stress and to feel good. Most (80%) found cannabis easy to obtain, while one in five (19%) currently grew at least some of their own supply. The sample had few cannabis-related criminal convictions, although growing and dealing were common.

Polydrug use was common, with alcohol and tobacco almost universally used on a regular basis. More than half the drinkers in the sample were consuming alcohol at hazardous or harmful levels. Almost half the sample (47%) had experienced problems with drugs other than cannabis, while 20% had sought assistance for these problems. Alcohol and amphetamine had been a problem for one in four respondents (each 27%).

Prevalence of a lifetime DSM-III-R diagnosis of cannabis dependence was 92%, with 40% diagnosed as severely dependent. Three quarters (74%) met an approximation of a lifetime DSM-IV diagnosis. Prevalence of twelve month dependence was 39% on the Severity of Dependence Scale (SDS), 77% on the short UM-CIDI and 72% on a measure of ICD-10 dependence. Only 33% believed they had a problem with cannabis. While most (85%) respondents had moderated their cannabis use at some time, only 16 people had sought help to do so.

Severity of cannabis dependence was correlated with a number of demographic and drug use variables. Older users were less severely dependent than younger users, and the greater the quantity of cannabis consumed, the greater the dependence score. DSM-III-R dependence was more common among females than males, despite their shorter history of cannabis use.

Problem drinking scores were associated with higher DSM-III-R dependence scores (as measured by the CIDI-SAM and the UM-CIDI). Self-reported problematic cannabis use was generally predicted by a combination of current quantity of cannabis used and a cannabis dependence diagnosis.

While there were many similarities between the Sydney and North Coast samples, there were also differences in patterns and contexts of use, and the prevalence and correlates of cannabis depdence. Sydney users were more likely to be diagnosed as dependent and to believe they had a problem with cannabis. Potential social differences between the samples may partly explain these findings. The Sydney sample may have contained a greater range of respondents that were more representative of residents of a major city, whereas the more tolerant attitudes in a rural area in which cannabis use is widely sanctioned may have encouraged potentially heavier cannabis use.

There was general concordance between severity of dependence scores on DSM-III-R and ICD-10 measures, but not between these and the SDS. The SDS was in greatest agreement with the respondent's self-reported belief they had a cannabis problem. Principal components analyses of the dependence measures provided little evidence for a unidimensional dependence syndrome for ICD-10 and DSM-III-R criteria. There was strong support for unidimensionality of the SDS.

Receiver Operating Characteristic (ROC) analyses on the dependence measures revealed adjusted cut-offs may be necessary for diagnosing dependence in samples of long-term cannabis users. A variety of measurement issues need to considered when comparing different ways of measuring dependence, including the selection of appropriate diagnostic cut-offs and the potential for over-diagnosis. The four instruments examined may have utility in different research and treatment contexts.

Most of this sample recognised potential benefits and harms to health from cannabis use. The main benefits identified were reduced stress and general positive psychological effects. These were offset primarily by perceived negative respiratory and psychological effects. This group had higher unadjusted rates of respiratory long-term conditions and wheezy or whistly chests than the Sydney and Australian comparison groups interviewed in the 1995 National Health Survey. However, given the greater proportion of tobacco smokers among the Sydney sample than the general population, it was impossible to separate the effects of tobacco smoking from those of cannabis smoking in producing respiratory symptoms.

Cannabis dependence and polydrug use were related to psychological well-being. The greater the number of dependence symptoms met and the number of drugs currently used, the greater the psychological distress. Nevertheless the pattern of  well-being scores for the sample as a whole did not indicate a clinical profile of psychological distress warranting intervention among this sample of long-term cannabis users.



Wendy Swift, Wayne Hall, Jan Copeland
Date Commenced
29 Aug 1997
Resource Type
Technical Reports