This research is part of my PhD dissertation and I presented some preliminary results at the annual College on Problems of Drug Dependence (CPDD) in San Diego, CA on June 17th. The abstract is provided below along with a PowerPoint file of the poster I presented. A link to download the poster can also be found through my CV.
EVIDENCE FROM A CANNABIS USE DISORDERS MULTIPLE CAUSES MODEL FOR THE CANNABISTOBACCO COMBINATION KNOWN AS ‘BLUNTS’.
Brian Fairman, J C Anthony; Department of Epidemiology, Michigan State University, East Lansing, MI
Aims: Globally, cannabis often is combined with tobacco before smoking. In the US, a prevailing cannabis-tobacco combination involves ‘blunts,’ made by substituting cannabis in a hollowed out tobacco cigar. Cannabis users with a history of blunt smoking seem to be at excess risk for cannabis use disorders (CUD). The aim here is to test a hypothesis that the profile of CUD clinical features is determined, in part, by blunt smoking history among cannabis smokers.
Methods: Data are from the nationally representative 2004 US National Survey on Drug Use and Health (NSDUH), with 7054 recent cannabis smokers age 12+ years, all with past-year cannabis use (6+ times), and with valid CUD and lifetime blunt history assessment. Exploratory factor analysis probed CUD latent structure (e.g., 1 v 2 factor model fit). Then, via a multiple indicators, multiple causes model (MIMIC), the blunt smoking effect was estimated for each CUD clinical feature, holding constant CUD level.
Results: Roughly 3/4s of recent cannabis users had smoked blunts at least once. EFA supported a 1-factor solution, and MIMIC modeling disclosed that blunt smoking was independently associated with “tolerance on” and “spending a great deal of time getting/using” cannabis, with CUD level held contstant. Inverse blunt
associations were found for “giving up important activities”, “failure to fulfill roles”, and “continued use despite social problems” under this model (slope estimates: 0.18, 0.41, -0.25, -0.33, and -0.33, respectively; all p-values<0.02).
Conclusions: We discovered that blunt smoking seems to influence CUD clinical features, over and above CUD level. Whether the observed differences indicate substantive differences in blunt effects and/or disclose measurement bias are open questions. Answers to these questions will be needed if we are to make progress in our understanding of the epidemiology of the cannabis-tobacco combination.
Financial Support: NIDA T32DA021129 (BJF); K05DA015799 (JCA).
