Cannabis use disorder in patients prescribed medical cannabis—Should we consider prescription cannabis use disorder as a new diagnostic condition?
Published online on August 02, 2026
Abstract
["Addiction, EarlyView. ", "\nAbstract\n\nBackground and Aims\nThe introduction of prescribed medical cannabis in many parts of the world raises issues regarding how we conceptualise Cannabis Use Disorders (CUD). The Diagnostic Statistical Manual Version 5 (DSM5) provides a framework for ‘Prescription Cannabis Use Disorder’ (Pr‐CUD) when a person is using medical cannabis only‐as‐prescribed, by excluding tolerance and withdrawal items from the diagnosis – with 9 instead of 11 diagnostic criteria. In this study, we compared the 12‐month prevalence rates of CUD in a sample of people prescribed medical cannabis in Australia by a medical or nurse practitioner, according to whether we use conventional DSM5 11‐items for ‘CUD’ versus 9‐items for ‘Pr‐CUD’.\n\n\nDesign\nSecondary analysis of data from CAMS22 Survey – an online anonymous cross‐sectional survey of people self‐reporting use of medical cannabis within the past 12 months.\n\n\nSetting and Participants\n2054 adults using prescribed medical cannabis in the preceding 12 months in Australia for a health condition.\n\n\nMeasurements\nConsumer characteristics, indications and patterns of medical and non‐medical cannabis use, self‐reported DSM5 criteria for CUD.\n\n\nFindings\n2054 participants reported using prescribed medical cannabis in the past 12 months, of which 986 (48%) reported using cannabis only as prescribed, and 1068 (52%) reported using cannabis not as prescribed. By applying 9‐item DSM5 CUD criteria for participants using cannabis as prescribed and 11‐item criteria for participants using cannabis not as prescribed, 640/2054 (31.2%) met Any P‐CUD (>2 criteria) and 241/2054 (11.7%) met Moderate–Severe P‐CUD (>4 criteria) thresholds. This is in contrast to 748/2054 (36.4%) met Any CUD and 273/2054 (13.3%) Moderate–Severe CUD when using 11 DSM5 criteria for all participants. Amongst the 986 participants using cannabis as prescribed, Pr‐CUD (9‐item) prevalence was statistically significantly lower than under the 11‐item framework, both for Any CUD (9% [Pr‐CUD] vs 20% [11‐item]; % difference = 11.8 [95% confidence interval (CI): 8.3, 15.8]; odds ratio (OR) = 1.3 [CI: 1.1, 1.4]) and Moderate–Severe CUD (2% [Pr‐CUD] vs 5% [11‐item]; % difference = 2.9 [CI: 1.4, 5.0]; OR = 2.6 [CI: 1.6, 4.4]).\n\n\nConclusions\nDiagnostic Statistical Manual Version 5 cannabis use disorder (CUD) diagnostic criteria developed for illicit cannabis use may not be appropriate for people using medical cannabis under prescribed conditions. CUD rates are statistically significantly reduced when the ‘Prescription Cannabis Use Disorder’ framework is applied to people only using cannabis products in accordance with the conditions identified on their prescription (including dose, route of administration and indication). This has considerable clinical and policy implications in settings where medical cannabis is legally available under a prescription model.\n\n"]