Although some adults can use marijuana without harm , marijuana users with comorbid psychiatric disorders, such as depression, are at increased risk of experiencing poor symptom and functional outcomes . Yet the degree to which medical or non-medical marijuana use may contribute to adverse clinical outcomes in depression is less clear. Comorbidity of depression and marijuana use has been studied extensively, with evidence showing a high prevalence of depression among marijuana users and vice versa . In the U.S., 7% of adults had major depressive episodes in 2014–2015, and the past year prevalence of non-medical marijuana use among them was 15% . Non-medical medical marijuana use among adults with depression is associated with increased risk of severe psychiatric symptoms high rates of suicidal ideation , and low psychiatric service utilization . A recent study found non-medical users and nonusers were equally prone to develop depression over time , while several others report of significant associations between nonmedical marijuana use and future occurrence of anxiety and depression symptoms and vice versa . Depression and anxiety are associated with more frequent non-medical marijuana use , and in one dispensary-based study,ebb and flow medical use for depression was associated greater problems with use . Since these studies have not concurrently examined non-medical and medicinal use it remains unclear if medicinal users experience some degree of symptom relief or differential impairment in depression.
Several studies report that medical marijuana is currently used as a therapeutic intervention for depression, even among clinical populations, although its safety and efficacy in depression treatment have not been established . Depression is also comorbid with several medical conditions in which cannabinoid drugs have therapeutic value , , suggesting a high prevalence of medical marijuana use among adults with depression, although less remains known about the clinical impact of medicinal use on this population.Clearly, marijuana use has the potential to be clinically problematic in depression, and it is also possible that both non-medical and medical marijuana use could differentially contribute to adverse clinical outcomes and lead to barriers to mental health care in this population. We followed 307 participants in a trial for drug/alcohol use treatment for depression, delivered in a psychiatry setting 1 year post-enrollment to examine: prevalence of non-medical, medical, and non-users of marijuana at baseline; characteristics of non-medical, medical, and non-users of marijuana; and the differential impact of non-medical and medical marijuana use, relative to non-users, over time and 1 year post enrollment clinical and psychiatry service utilization outcomes. Building on our prior work showing that marijuana use has adverse clinical effects on depression , the findings would help to distinguish the impact of non-medical and medical use on clinical outcomes and inform prevention and intervention models. Data were from participants in a randomized controlled trial of motivational interviewing for drug/alcohol use treatment in depression, delivered in an outpatient psychiatry setting. The details of the parent trial have been reported previously . Briefly, a total of 307 participants were recruited from a large outpatient psychiatry clinic from Kaiser Permanente Northern California.
Study clinicians determined eligibility based on inclusion criteria, which required patients to be at least 18, have a Patient Health Questionnaire score ≥5 indicating at least mild depression severity, and either drink at hazardous levels or have used drugs in the past 30 days. All participants provided written informed consent at an in-person appointment in the same clinic where they received usual care. Procedures were approved by the University of California, San Francisco and Kaiser Permanente Northern California Institutional Review Boards. Enrolled participants were randomized to one of two study arms after completing screening procedures, either MI or a non-treatment control. The MI treatment intervention consisted of one 45-minute session followed by two 15-minute telephone “booster” sessions , which were about two weeks apart. Participants in the control arm were given a 2- page brochure, produced by the NIH National Office of Drug Control Policy , on use risks specific to the substances reported by participants at baseline . Participants also continued to receive usual depression care based on current best practices for medication management and empirically supported psychological treatment over the 1 year follow-up. Participants used laptop computers to complete the baseline measures including self-report assessments of past 30-day drug/alcohol use, the PHQ-9 , the Generalized Anxiety Disorder scale , and the Short Form Health Survey . Participants were re-assessed with the same substance use, symptom, and functional assessments three times via telephone interviews bytrained interviewers over the follow-up. Patients were offered $50 gift cards for completing the baseline and 6-month interview, and $100 for completing the 12-month interview. Past 30-day marijuana use was assessed during study interviews via self-report. Patients were asked: “How many days in the past 30 days have use you used marijuana” and “Was the marijuana used in the past 30 days always used for medical purposes, prescribed or recommended by a provider.”
We created a categorical measure to define marijuana use status 30 days prior to each period .Past 30 day alcohol/drug use, other than marijuana use was assessed during study interviews . We created a dichotomous measure to define substance use other than marijuana use, 30 days prior to each period. Patients were coded as using substances if they endorsed ≥ 1 day in the past 30 , providing a dichotomous indicator of substance use, which served as a covariate in multivariable analyses . Frequencies and means were used to describe the sample at baseline. Next, χ2 or analysis of variance tests were used to describe demographic differences between the marijuana use groups, e.g., non-users, medical-users, and non-medical-users. Analyses of interest began with using frequencies and means to describe the baseline rates of psychiatric symptoms, functioning, and psychiatry service use across marijuana use groups . To compare these groups on psychiatric symptoms, functional status, and psychiatry visits at baseline, we used a series of multivariable regression analyses, comparing non-medical users and medical users to non-users . Binary dependent variables were fit with multivariable logistic regression models ,dry racks and fit continuous dependent variables using multivariable linear regression models . Next, 1 year outcomes were examined using similar analytic procedures as described above. We conducted regression models on follow-up psychiatric symptom, functional status, and psychiatric visit data, comparing non-medical users and medical users to non-users on these outcomes. All regression models at baseline and 1 year adjusted for patient characteristics. Control variables were chosen because research suggests that the variable is related to marijuana use or the variable was significant in univariable analyses. All regression models were adjusted for age, sex, marital and employment status, treatment assignment, substance use, past 30-day psychiatry visits , and the number of days of marijuana use in the past 30-days. Differential changes in psychiatric symptoms, functioning, and psychiatry visit trajectories were examined between non-medical users and medical users compared to non-users over the 1 year follow-up, using mixed-effects growth models. This approach to longitudinal data analysis is a form of hierarchical linear/non-linear modeling for repeated measures data, where multiple measurement occasions are nested within individuals . To determine the average rate of change on the outcomes of interest, we computed unconditional growth models, predicting the psychiatric symptom, functional status, and psychiatry visit outcomes from time . Differential changes in these outcomes were then examined between non-medical users and medical users compared to non-users, predicting each outcome of interest from time, and a time varying marijuana use group variable . All conditional models adjusted for age, sex, marital and employment status, marijuana use, and treatment assignment. Additionally, time-varying indicators of psychiatry visits , substance use, and the number of days of marijuana use 30-days prior to each interview, were included as potentially confounding covariates. Analyses were carried out in R version 3.3.1 .
Overall, missing data were modest at <5% over the study. Rather than discard partial study completers and potentially bias the sample analyzed, the expectation maximization method was used to handle missing data during maximum likelihood estimation at the time of analysis. Statistical significance for all tests was defined at p < .05. Over the follow-up, non-medical marijuana use decreased whereas medical marijuana use slightly increased . Patients using non-medical marijuana over 1 year had significantly less improvement in depression symptoms and suicidal ideation . compared to non-users, and a trend was found for less improvement in mental health functioning . Among non-medical marijuana users, a trend was observed of fewer psychiatry visits . Change in medical marijuana use over time was not associated with significant change in symptom, mental and physical functional status, or psychiatry service utilization trajectories; however, there was a trend indicating that those who continued to use medical marijuana over 1 year had less improvement in everyday functioning, , . Given associations between non-medical marijuana use and both fewer psychiatry visits and younger age, we conducted post-hoc analyses in the marijuana use subsample by age. We used Cox proportional hazard survival models to explore differences in the time without psychiatry visits between those aged 18–45 and ≥ 45 who used non-medical marijuana and medical marijuana over 1 year. Results showed a trend where older, but not younger, participants using non-medical marijuana had more psychiatry visits over time than those using medical marijuana . This study examined baseline and longitudinal differences in the characteristics of patients in treatment for depression based on their use of marijuana. Results at baseline revealed that the overall prevalence of marijuana use was slightly higher than prior rates documented among psychiatry samples . This may be explained by the observation that marijuana use is more frequent in states that have permissive marijuana laws , and may reflect more normalized views about marijuana use within California. In addition, only ~28% of patients endorsed using marijuana use for medical purposes, as recommended by a physician. Non-medical marijuana users had higher suicidal ideation, greater depressive symptoms, and poorer mental health functioning compared to non-users at baseline. Non-medical use was also associated with less improvement in each of these domains over 1 year. These findings extend prior work indicating associations among non-medical marijuana use and greater psychopathology , including suicidal ideation , and worse functional outcomes . Future research should develop and test strategies to reduce non-medical marijuana use while improving symptoms and remediating functional impairment in depression. As the parent MI trial found that this intervention was effective in reducing marijuana use , MI may help attenuate the adverse impact of non-medical use on psychological symptoms and functional impairment and warrants further study. Non-medical marijuana users had fewer psychiatry visits at baseline and at the 1 year follow-up, indicating that providers may have fewer opportunities to educate these patients about the adverse clinical effects of marijuana on depression. All participants using nonmedical marijuana underutilized psychiatry services, although post-hoc analyses found this to be less true of older patients. It is possible that older adults using non-medical marijuana have greater psychiatry service use needs due to combined changes in brain plasticity and age-related cognitive decline, which could increase their risk of adverse clinical effects . Future studies of older adults would be valuable in expanding the range of populations examined in marijuana research, which has largely focused on youth and young adults , and is particularly important given changing U.S. demographics. These differences were not observed at 1 year. Because the conditions for which medical marijuana is often used or recommended to alleviate symptoms are also associated with deficits in cognitive and physical capacity , medicinal users could be expected to show less functional improvement. Unfortunately, data on the reason for medicinal use was not collected, and further work will be needed to determine associations between use indications and expected outcomes. Our overall results support our prediction that the degree to which psychiatry patients with depression have adverse clinical outcomes would be influenced by whether marijuana was used for non-medical or medical purposes. Results showed that non-medical use was associated with adverse clinical effects in terms of psychological symptoms and associated functional impairment, and these effects persisted for 1 year. This highlights a need for education efforts in psychiatry treatment contexts around the elevated risks associated with non-medical marijuana use .