Falling rates of marijuana dependence among heavy users

Highlights

• For Daily/Near-Daily (DND) users, DSM-IV dependence fell from 26.5% to 16.1%.

• Decreases held for all DND age and gender subgroups, save for users over 50.

• No significant change in population-level dependence was detected.

• Among symptomatic criterion, all but one (increasing tolerance) fell significantly.

• Potential causal drivers include changing products, policies, or attitudes.

Abstract

Introduction

Marijuana use has become increasingly popular in the United States since the turn of the century, and typical use patterns among past-month marijuana users have intensified, raising concerns for an increase in cannabis use disorders (CUDs). Yet the population prevalence of CUDs has mostly remained flat. We analyzed trends in DSM-IV marijuana dependence among Daily/Near-Daily (DND) users, both overall and by age and gender, and considered potential explanations.

Methods

Using data assembled from the National Survey on Drug Use and Health (2002–2016), rates of self-reported dependence and constituent symptoms are calculated for DND marijuana users; logistic regressions with pre- and post- periods (2002–2004, 2014–2016) and a Cochrane-Armitage trend test are applied to describe temporal changes.

Results

Dependence among DND users fell by 39% (26.5%–16.1%; p < .001), with significant trend. No significant change is detected at the population level. Sub-group analysis shows a steep gradient for age but not for gender. Declines are robust to sub-group analysis, except for users over 50 years old. Among dependence symptoms, most showed significant declines: reducing important activities (p < .001); use despite emotional, mental, or physical problems (p < .001); failing attempts to cutback (p < .001); lots of time getting, using, or getting over marijuana (p < .01); and failing to keep limits set on use (p < .05). Reported tolerance showed no significant change.

Conclusions

Though it is unclear why, the risk of dependence formation among heavy marijuana users appear to have declined since 2002. Further research is warranted regarding explanations related to state marijuana policies, product forms, or social context.

Introduction

Since the turn of the century, the United States has seen dramatic changes in marijuana policy, use, and social context. Eight states and Washington DC have legalized possession, use, production, and sale of marijuana for non-medical use; more than thirty states have legalized some form of medical marijuana, often with dispensaries. Perceptions of marijuana’s risks have been falling (Pacek et al., 2015), and approval for legalization is now nearly two-thirds (Gallup, 2017). Marijuana use has also changed in other ways, e.g., increasing THC potency, an increasing acceptance and availability of marijuana for medical use, and rising popularity of alternative forms of consumption such as edibles, portable vaporizers, and “dabbing” solid concentrates (Davenport and Caulkins, 2016; Kilmer et al., 2014).

Marijuana use has been on an upward trend since roughly the mid-1990s, and intensifying since 2007 (Carliner et al., 2017). The portion of Americans 12 years or older reporting past-month marijuana use rose from 6.2% in 2002 to 8.9% in 2016 (Center for Behavioral Health Statistics and Quality, 2017a). Even among past-month marijuana users, use habits have intensified, for instance as measured by the portion of past-month users reporting daily or near-daily (“DND”) use, i.e., use on more than 20 of the past 30 days, which tripled from one-in-nine in 1994 to one-in-three in 2014. DND users dominate the marijuana market, accounting for 68% of use days and 60% of expenditures in 2012–2013. In total, the number of days of marijuana use reported to NSDUH increased by more than half from 2002–2003 to 2012–2013 (Davenport and Caulkins, 2016).

Use disorders are an important risk of marijuana use. The Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) provides definitions of marijuana abuse and dependence, with diagnostic criteria including spending a great time using or getting marijuana, increasing tolerance, continued use despite physical or mental problems, and reduced time spent on important activities; with the notable exclusion of withdrawal symptoms, due to a lack of evidence at the time of publication (American Psychiatric Association, 2000; Hasin et al., 2013). Questions corresponding to these criteria have been asked by the National Survey on Drug Use and Health (NSDUH) for decades, allowing for the calculation of prevalence of cannabis abuse and/or dependence based on self-reported symptoms.

Though the NSDUH questionnaire remains unchanged, the DSM-V has since introduced changes relating to cannabis use disorders (American Psychiatric Association, 2013; Center for Behavioral Health Statistics and Quality, 2016). First, the DSM-V changed the diagnostic threshold for cannabis use disorder, merging cannabis abuse and dependence and changing some criteria; namely, dropping the criterion relating to legality, citing limited information gained for diagnosis and amid doubts about its use as a clinical indicator, and adding a criterion relating to cravings (Peer et al., 2013). Second, reflecting new research, the DSM-V recognized symptoms for cannabis withdrawal, based on feelings of irritability, anger, anxiety, disturbed sleep, decreased appetite, mood, and/or energy, and at least one physical symptom causing significant discomfort, e.g., abdominal pain, tremors, fever, or headache (Hasin et al., 2008). Comparisons suggest these changes have had little effect on the measured prevalence of cannabis use disorders (Mewton et al., 2013; Peer et al., 2013).

NSDUH data suggest four million Americans suffer from DSM-IV marijuana dependence (Center for Behavioral Health Statistics and Quality, 2017a); and as a percent of the general population, marijuana abuse and dependence have remained relatively stable from 2002 to 2014 at 1.5% (Compton et al., 2016). The National Epidemiologic Survey on Alcohol and Related Conditions records a much larger increase, from 1.5% to 2.9% (Hasin et al., 2015), due largely to changes in NESARC methodology (Grucza et al., 2016).

One might have expected the population prevalence of marijuana dependence to have increased proportionally to the substantial increase in prevalence of marijuana use. That this did not occur presents a puzzle, which this paper seeks to elucidate. Using National Survey of Drug Use and Health (NSDUH) data, we seek to measure levels and changes in marijuana dependence and its symptomatic criteria among the heaviest users, and consider potential causes at hand.