Author ORCID Identifier

0000-0002-7263-0701

Date of Award

Summer 8-9-2022

Degree Type

Closed Thesis

Degree Name

Master of Public Health (MPH)

Department

Public Health

First Advisor

Matthew J. Hayat

Second Advisor

Sheryl Strasser

Abstract

Introduction: Despite the high prevalence of mental illnesses, more than half of adults with a mental health illness do not receive treatment, with 24.7% of affected adults reporting an unmet need for treatment in 2019. Thus, the population of U.S. adults may resort to finding alternative ways to deal with symptoms of major depressive episodes, such as recreational marijuana use. Researchers have shown that individuals with mood disorders are more likely to self-medicate with drugs, with marijuana being used more in states with medical marijuana laws. The associations between marijuana use and depression may be attributable to a third common factor that influences both the decision to use marijuana and to depression. There is limited knowledge on the role that access to mental health services plays on the association between marijuana use and the occurrence of major depressive episodes. Therefore, with opposing ideas in the literature regarding whether marijuana use improves symptoms of mental health disorders, or whether it makes conditions worse, it is important to understand what factors may play a role in the use of marijuana in individuals who suffer from mental health disorders so that strategies may be developed, and resources may be allocated to ensure that these individuals have access to appropriate care.

Methods: Publicly available data from the 2019 National Survey of Drug Use and Health (NSDUH) was analyzed for this cross-sectional study (n=42377). Information on demographic characteristics, past year marijuana use, the occurrence of past year major depressive episodes as well as health insurance status were used for this analysis. The outcome variable of interest was marijuana use and was analyzed (1) as a continuous construct, and (2) as a categorical construct with categorization performed in four different ways. Marijuana use as a continuous construct was modeled with a gamma distribution. High occurrence of no use was accounted for with a zero-inflated statistical model. The first part of the two-part model was the zero-model quantifying whether marijuana was used. The second part of the model was modeling the distribution of the number of days of past year marijuana use. The multicategory categorizations were modeled with the multinomial distribution and the dichotomous outcome modeled with the binomial distribution. For assessing past year marijuana use as a categorical variable, unweighted and weighted unadjusted and adjusted analyses were performed.

Results: The majority of individuals (82.4% weighted) reported no past year marijuana use. Of those who reported having used marijuana in the past year (23.5% unweighted, n=9969/42377), the weighted median number of days of past year marijuana use was 59.2 days (Interquartile Range, IQR: [7.5, 258.8]). For the zero model, the odds of past year marijuana use for those who had a past year major depressive episode were 1.83 times the odds of those with no past year major depressive episode (95% CI = [1.70,1.97]), controlling for all other variables. For the continuous gamma model and controlling for all other variables, those with the occurrence of a past year major depressive episode are expected to use marijuana 1.1 times more days than those who do not have the occurrence of a past year major depressive episode (95% CI = [1.03, 1.2]). For past year marijuana use as a dichotomous outcome, the odds of past year marijuana use in those who had a major depressive episode in the past year were 1.86 times the odds for those who did not have a major depressive episode, controlling for all other variables (weighted OR= 1.86, 95% CI= (1.66,2.08)). The odds of past year marijuana use were consistently higher for those who had the occurrence of a past year MDE compared to those who did not, irrespective of categorization. Health insurance was not found to be a statistically significant moderator in any of the associations between past year marijuana use and past year MDE.

Conclusion: These findings support some of the literature regarding the positive association between marijuana use and the occurrence of MDE. However, health insurance, as a proxy to access to mental health care, was not found to be a significant moderator between the association between marijuana use and MDE. It is possible that other factors, or combinations of factors, may serve as a better representation of access to mental health care. Therefore, future studies should focus on identifying other factors that may play a role in the use of marijuana in individuals who suffer from mental health disorders so that strategies may be developed to ensure that these individuals have access to appropriate care and do not have to resort to drug use as a form of self-medication.

File Upload Confirmation

1

Available for download on Sunday, July 28, 2024

Share

COinS