Director, Epidemiology Indivior, Inc. North Chesterfield, VA, USA, United States
Background: Despite increasing cannabis use in the United States, the prevalence of cannabis use disorder (CUD) has remained relatively stable in national survey data, though with some increases in specific symptoms (i.e., tolerance) and increases in some specific sub-populations, such as patients treated at the Veterans Health Administration.
Objectives: To use a data-driven approach to shed light on patterns of CUD symptoms among past-year cannabis users, which could be obscured by current diagnostic approaches.
Methods: A sample of respondents (n=61,708) reporting past year cannabis use was drawn from the 2015-2020 National Survey on Drug Use and Health. Latent class analysis was utilized to identify heterogenous subgroups, based on patterns of nine DSM-5 CUD symptoms (excluding craving and withdrawal, which were not included in 2015-2019 surveys). The association of subgroups with demographic and clinical characteristics, including past-year treatment utilization, was explored.
Results: A 5-class model provided the best fit: No Symptoms (77.3% of the sample), reporting no symptoms; Time & Tolerance (15.1%), reporting time spent using/getting/getting over effects (‘time’) and tolerance; Uncontrolled Use (2.8%), characterized by time, tolerance, using more than intended, and inability to cut down/stop; Neglecting Activities (3.1%), characterized by time, tolerance, and neglecting activities; and Severe (1.8%), with a high likelihood of most symptoms. All individuals in the symptomatic classes met criteria for at least mild CUD (2+ symptoms); moderate/severe CUD criteria (4+ symptoms) was met by 2.1% in Time & Tolerance, 78.0% in Uncontrolled Use, 43.0% in Neglecting Activities, and 100% in Severe. Past-year illicit drug treatment was uncommon, reported by 2.0% in No Symptoms, 4.9% in Time & Tolerance, 6.0% in Uncontrolled Use, 7.0% in Neglecting Activities and 17.6% individuals in the Severe class.
Conclusions: Five distinct classes were identified representing CUD symptom profiles that varied both quantitively and qualitatively. Though most respondents reporting past-year cannabis use reported no CUD symptoms, all individuals in the 4 symptomatic classes met criteria for at least mild CUD and few received any illicit drug treatment, even among those with nearly all CUD symptoms. These results suggest a heterogeneity of CUD that can inform development of individualized treatment programs.
