Prevalence of Cannabis Use Disorder and Reasons for Use Among Adults in a US State Where Recreational Cannabis Use Is Legal

Key Points Question What is the prevalence of cannabis use disorder (CUD) among primary care patients who use cannabis in a state with legal recreational cannabis use, and does prevalence differ by reason for use? Findings In this cross-sectional study, weighted prevalence of any CUD did not vary by reason for cannabis use, whereas the prevalence of moderate to severe CUD did. Prevalence of moderate to severe CUD was higher in those who reported nonmedical use only or both medical and nonmedical use. Meaning In this study, CUD was common among patients who use cannabis in a state with legal recreational cannabis use, with moderate to severe CUD most prevalent among patients with any nonmedical use.


Introduction
1][12][13] Among primary care patients who used cannabis in a state with legal recreational use, where provider recommendation is unnecessary for medical use, the prevalence of patient-reported medical cannabis use was 67%. 13 Greater cannabis use is associated with an increase in risk of cannabis use disorder (CUD), [14][15][16][17] and legalization has contributed to increases, [18][19][20] with 17% of individuals who use cannabis having CUD. 21Among veteran outpatients who used cannabis in a state with legal medical use, the prevalence of CUD varied by reasons for use and was lowest among those who reported medical use only. 12This study assessed whether CUD prevalence varied for patients by self-reported reason for cannabis use in a state with legal recreational use.We estimated the prevalence of CUD based on a diagnostic questionnaire on a confidential cannabis survey among patients who used cannabis, overall and by 3 categories of patient-reported reasons for use: medical use only, nonmedical use only, and both.

Sample
Data for this cross-sectional study were from patients' survey responses and electronic health records.Demographic characteristics (ie, age, sex, race, ethnicity, and insurance) were collected from patients on paper or via patient portal and documented in the electronic health record by the Kaiser Permanente Washington health system before or at the time of screening.The Kaiser Permanente Washington institutional review board approved the study, including waivers of consent (for sample identification) and documentation of consent (survey respondents).The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The study was conducted in Kaiser Permanente Washington, a large health system in Washington State where recreational cannabis use has been legal since 2012.Primary care patients who completed a confidential survey about cannabis use were included in the study.Survey design, sampling, procedures, weighting, and sample characterization, including comparison of respondents to eligible primary care sample and nonrespondents, have been reported previously. 13Briefly, 5000 patients 18 years and older were randomly selected from 108 950 eligible patients with electronic health record documentation of completing a cannabis screen as part of routine primary care from March 28, 2019, to September 12, 2019 (Figure).The single-item screen asks about the frequency of past-year cannabis use (ie, none, less than monthly, monthly, weekly, and daily). 22Sample selection included patients who reported no past-year use as well as stratified oversampling of patients with more frequent cannabis use and patients of minoritized racial and ethnic groups (including American Indian or Alaska Native, Asian, Black, Hispanic, and Native Hawaiian or Other Pacific Islander) in order to obtain representation of subgroups that are often underrepresented in research. 13Race and ethnicity were evaluated as proxies for lived experiences (eg, discrimination) that may influence reasons for or patterns of use.Among invited participants, 1688 (34%) provided informed consent and responded to the survey.Respondents were asked about past-year use and more specific questions about past 30-day use, including reasons, mode, and typical frequency of cannabis use.
Patients who reported past 30-day cannabis use (n = 1463) were included here (Figure ), with results weighted to the primary care sample who used cannabis in the past 30 days (hereafter, patients who used cannabis).

Cannabis Survey Measures
Patients were categorized by their stated reason for using cannabis during the past 30 days: medical use only, nonmedical use only, or both reasons for use.All patients who reported past 30-day use were asked the 15-item Composite International Diagnostic Interview Substance Abuse Module (CIDI-SAM).The CIDI-SAM provides a scaled score (0-11 symptoms) of Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) CUD severity reflecting the number of past-year DSM-5 CUD criteria (ie, 2-3 symptoms = mild, 4-5 symptoms = moderate, 6-11 symptoms = severe). 23e 11 symptoms are reported, as well as any CUD (Ն2 symptoms) and moderate to severe CUD (Ն4 symptoms). 22tients were asked about all modes of cannabis use as well as the primary mode, including inhalation (ie, smoke, vape, dab), ingestion (ie, eat, drink), application (ie, lotion, ointment), or other modes.Questions also included the frequency of past-year use and typical number of days per week and times per day of cannabis use.

Statistical Analysis
All analyses were weighted for sampling strategy and nonresponse to provide estimates reflecting the primary care population who used cannabis. 13We report the prevalence of reasons for cannabis use as well as the characteristics of the sample, overall and stratified by reasons for cannabis use.
Main analyses estimated the prevalence of cannabis measures and CUD symptoms and severity across reported reasons for cannabis use, using multinomial logistic regression for categorical outcomes and logistic regression for binary outcomes, with robust standard errors (SEs) and  adjustment for age, sex, race, ethnicity, and insurance status.Results are presented as the mean predicted probability of outcomes based on recycled predictions; corresponding SEs were estimated using the delta method, which assumes a normal approximation to construct 95% CIs. 24,25Joint Wald tests evaluated differences in CUD across reasons for use.Post hoc sensitivity analyses repeated analyses in a restricted sample excluding patients who reported applied use (topical) as their only mode of cannabis use, due to the expected lower risk of CUD from limited systemic cannabis exposure. 26Analyses were conducted between August and June 2023 using Stata version 17.0 (StataCorp).Two-sided P < .05 was considered statistically significant.

Results
The  a Percentage and standard errors calculated from unadjusted survey data weighted for sampling and nonresponse rates for eligible primary care sample.
b Sample numbers from unweighted survey data.
c Pearson χ 2 test of independence.
d Race and ethnicity data were collected via self-report on paper or via an online patient portal and documented by the health system.These variables were included as proxies for lived experiences (eg, discrimination) that may influence reasons for or patterns of use.
e Patients are provided the option to indicate other, which is undefined, when choosing among 1 or more race categories at appointing or check-in.Patients who indicated more than 1 race are reported as multiracial.
f Indicates data from survey; all other data are from the electronic health record.
g Survey responses for education, employment, marital status, and residence were missing for 9, 5, 6, and 7 patients, respectively.

Table 1 .
Characteristics of Primary Care Patients Who Used Cannabis

Table 2 .
Prevalence of Cannabis Use Disorder by Reason for Cannabis Use Among Primary Care Patients Who Used Cannabis a a Data collected from confidential survey, N = 1463.b P value based on a joint Wald test from logistic regression.c d Any past-year CUD-based DSM-5 criteria.e f Estimates of confidence intervals were not bound at 0; negative values were truncated at zero.

Table 3 .
Prevalence of Cannabis Use Characteristics by Reason for Cannabis Use Among Primary Care Patients Who Used Cannabis Pearson χ 2 test of independence.Estimates of confidence intervals were not bound at 0; negative values were truncated at zero.Patients are provided the option to indicate other, which is undefined, when choosing among 1 or more race categories at appointing or check-in.Patients who indicated more than 1 race are reported as multiracial.29.Matson TE, Hallgren KA, Lapham GT, et al.Psychometric performance of a substance use symptom checklist to help clinicians assess substance use disorder in primary care.JAMA Netw Open.2023;6(5):e2316283.doi:10.1001/jamanetworkopen.2023.1628330.Guo Y, Kopec JA, Cibere J, Li LC, Goldsmith CH.Population survey features and response rates: a randomized experiment.Am J Public Health.2016;106(8):1422-1426.doi:10.2105/AJPH.2016.30319831.Lallukka T, Pietiläinen O, Jäppinen S, Laaksonen M, Lahti J, Rahkonen O. Factors associated with health survey response among young employees: a register-based study using online, mailed and telephone interview data collection methods.BMC Public Health.2020;20(1):184.doi:10.1186/s12889-020-8241-8Sensitivity analysis: Prevalence of cannabis use disorder by reason for cannabis use among primary care patients who used cannabis, with removal of patients (n=22) who reported applied use as their only mode of use eTable 2. Unweighted crosstabulation of survey participants with past 30-day cannabis use who reported cannabis use disorder symptoms, by reason for cannabis use b c d