Choosing a diagnostic cut-off for cannabis dependence

Area of Science:

• Addiction research
• Psychiatric diagnostics
• Cannabis use disorder

Background:

• Cannabis dependence is increasingly recognized, yet measurement issues in operationalizing the syndrome require further investigation.
• Limited research exists on the diagnostic utility of short dependence measures for cannabis.

Purpose of the Study:

• To investigate the diagnostic utility of three short dependence measures among long-term cannabis users.
• To determine appropriate diagnostic cut-offs for these measures in assessing cannabis dependence.

Main Methods:

• Two hundred long-term, regular cannabis users were recruited in Sydney, Australia.
• Receiver Operating Characteristic (ROC) analyses compared diagnostic performance against DSM-III-R cannabis dependence (gold standard).
• Measures assessed included the short University of Michigan CIDI, ICD-10 dependence criteria, and the Severity of Dependence Scale.

Main Results:

• All three measures demonstrated utility in diagnosing at least moderate DSM-III-R cannabis dependence.
• Optimal diagnostic cut-offs for the short University of Michigan CIDI and ICD-10 measure were unchanged.
• A more liberal cut-off was optimal for the Severity of Dependence Scale, with amended prevalence rates of 77% (UM-CIDI), 72% (ICD-10), and 62% (SDS).

Conclusions:

• The three instruments reliably diagnose cannabis dependence, performing significantly better than chance.
• The measures showed robustness regarding optimal diagnostic cut-offs in long-term cannabis users.
• Guidelines are provided for selecting optimal cut-offs based on specific contexts and desired sensitivity/specificity.