CONTRIBUTIONS TO THE EPIDEMIOLOGY AND MENTAL HEALTH CONSEQUENCES
OF CANNABIS SMOKING
By
Brian John Fairman

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of requirements
for the degree of
Epidemiology - Doctor of Philosophy
2014

ABSTRACT
CONTRIBUTIONS TO THE EPIDEMIOLOGY AND MENTAL HEALTH CONSEQUENCES
OF CANNABIS SMOKING
By
Brian John Fairman

Cannabis smoking might contribute to the incidence and course of major depression, but it is less
clear whether this relationship is particularly pernicious when exposure occurs during adolescence. This
hypothesis is guided by prior evidence that adolescence is a particularly vulnerable period of
neurodevelopment, and early cannabis exposure may have a lasting toxic effect on normal emotional
development. The first study in this dissertation research investigates the degree to which early-onset
cannabis smoking (before age 18 years) might predict the later onset of a sustained spell of depressed
mood in adulthood (herein, a ‘depression spell’), as compared to never and later-onset cannabis users. The
second study complements the findings therein and was motivated to inform on the extent to which
depression spells were clinically relevant, with relevance operationally measured by functional
impairment attributed to depression in an individual's occupational, social, and daily life. Data for these
studies came from the U.S. National Surveys on Drug Use and Health (NSDUH), a program of annual
cross-sectional surveys of large and nationally representative samples of community-dwelling U.S.
residents aged 12 years or older.
The third and fourth studies of this dissertation sought to contribute to our understanding of the
epidemiology of cannabis use disorders via two relatively novel and understudied patterns of cannabis
smoking. In the first, the tobacco-cannabis combination called ‘blunts’, which has become an increasingly
popular method of cannabis consumption in the U.S., was hypothesized to be associated with more
cannabis problems (e.g., when blunt smokers are compared to their non-blunt cannabis smoking peers).
Here also, the NSDUH epidemiological surveys made it possible to estimate the degree to which a history
of blunt smoking is associated with the level of cannabis problems in a representative sample of recent

cannabis users. For the final study in this dissertation research, it was hypothesized that cannabis users
who rapidly transitioned from the first drug opportunity to using cannabis might have a greater risk of
later cannabis problems. Here, for this study, the data are from the collaborative WHO World Mental
Health Surveys (WHO-WMHS), with cross-national data from 14 countries that made it possible to look
beyond the boundaries of the U.S. for novel epidemiological evidence on cannabis smoking.
Results from these studies were informative. Early-onset cannabis smoking was associated with a
later depression spell in adulthood: cases with an adult-onset depression spell were an estimated 70%
more likely to have been exposed to using cannabis in adolescence, as compared to never users. However,
the exposure odds ratio with respect to later-onset cannabis exposure was of similar magnitude,
suggesting that early-onset cannabis exposure per se is less important than was hypothesized, and that the
delay of cannabis onset until adulthood might not greatly affect the risk of a later depression spell. As for
the issue of 'clinical relevance,' cases with recent depression spells suffered noteworthy functional
impairment attributed to this mood disturbance, with greater impairment seen across at higher levels of
cannabis problems. In specific, an estimated 25% (one in four) of the recent depression spell cases
experienced severe to very severe functional impairment attributed to their depression. In the study of
blunt smoking, as hypothesized, the level of cannabis problems was greater for blunt smokers and when
there was more frequent recent blunt smoking, as compared to that experienced by cannabis users with
little or no blunt smoking history. Findings from the cross-national epidemiological surveys were
generally consistent with expectations -- namely, cannabis users who delayed their onset of cannabis
smoking for a year or more after initially being offered the chance to try drugs were at a reduced chance
of experiencing later cannabis-related problem outcomes.

This dissertation is dedicated to my family, friends, and loved ones who believed in and
supported me.

iv

ACKNOWLEDGEMENTS

I wish to acknowledge and thank the numerous people who without their support
and mentorship this dissertation would not have been possible: my mentor,
Professor James C. Anthony; my dissertation committee members, Professors
Naomi Breslau, Siddharth Chandra, and Zhehui Luo; David Barondess and
Mirjana Radovanovic; the T32/D43 pre- and post-doc fellows past and present;
the Faulty and Staff in the Department of Epidemiology and Biostatistics at MSU;
the National Institutes of Health, and specifically, the National Institute on Drug
Abuse (NIDA), and NIDA research and research training awards to support the
dissertation research, mentoring, and cross-national field survey research
(T32DA021129; K05DA015799; R01DA16558).

v

TABLE OF CONTENTS

LIST OF TABLES..........................................................................................................................ix
LIST OF FIGURES........................................................................................................................xi
LIST OF ABBREVIATIONS.......................................................................................................xiv
CHAPTER 1. AIMS AND OBJECTIVES......................................................................................1
1.1. Study 1. Are Early-Onset Cannabis Users At An Increased Risk Of Depression
Spells?..................................................................................................................................1
1.2. Study 2. Does the Level of Cannabis Problems Predict the Level of Functional
Impairment Attributed to Recently Active Depression?......................................................2
1.3. Study 3: Do Cannabis Smokers with a History of ‘Blunt’ Smoking Differ in the
Experience and Degree of Cannabis Problems?..................................................................2
1.4. Study 4: Does Delaying Onset of Cannabis Smoking After Onset of First
Opportunity Account for Variations in Risk for Later Cannabis Problems?.......................3
CHAPTER 2. BACKGROUND AND SIGNIFICANCE................................................................4
2.1. Cultivation and Common Preparations of Cannabis.....................................................4
2.2. Cannabis History and Beginnings of Public Health and Moral Concerns....................6
2.3. The Pharmacology of Cannabinoids...........................................................................10
2.4. The Diagnosis of Cannabis Use Disorders.................................................................13
2.5. The Epidemiology of Cannabis Smoking...................................................................17
2.5.1. Quantity and Location..................................................................................17
2.5.2. Causes..........................................................................................................22
2.5.2.1. Biological Causes and Theories....................................................23
2.5.2.2. Psychological Causes and Theories..............................................25
2.5.2.3. Sociological Causes and Theories................................................27
2.5.3. Mechanisms.................................................................................................28
2.5.3.1. Opportunity to Try Cannabis........................................................29
2.5.3.2. Cannabis Use and Escalation to Frequent and Problematic Use..30
2.5.3.3. Secondary Consequences of Cannabis Smoking..........................31
2.5.4. Prevention and Control................................................................................33
2.6. Cannabis Smoking, Depression, and Functional Impairment.....................................36
2.6.1. The Burden and Epidemiology of Depression.............................................36
2.6.2. Suspected Mechanisms Linking Cannabis Smoking and Depression.........38
2.6.3. Biological Mechanisms................................................................................39
2.6.4. Psychosocial Mechanisms...........................................................................41
2.6.5. Epidemiologic Evidence for Cannabis as a Possible Cause of Depression.42
2.6.6. Functional Impairment, Depression, and Cannabis Use..............................45
2.7. The Blunt Smoking Phenomenon...............................................................................49
CHAPTER 3. MATERIALS AND METHODS...........................................................................53
3.1. The U.S. National Surveys on Drug Use and Health (NSDUH)................................53
vi

3.1.1. Study Background, Design, and Population................................................53
3.1.2. Sampling Approach.....................................................................................54
3.1.3. Methodology, Confidentiality, and Human Subjects Protection.................55
3.1.4. Reliability and Validity Studies...................................................................56
3.1.5. General Strengths and Limitations of the NSDUH Studies.........................57
3.1.6. Measurement of Key Study Variables.........................................................59
3.1.7. Missing Data and Imputation.......................................................................65
3.2. The WHO World Mental Health Surveys Initiative (WMH).....................................65
3.2.1. Study Design and Sampling Procedures......................................................65
3.2.2. Data Collection and Measurement of Psychiatric Disorders.......................67
3.2.3. Protection of Human Subjects.....................................................................68
3.2.4. Measurement of Key Study Variables.........................................................68
3.3. Analysis Plan..............................................................................................................71
3.3.1. Study 1: Early-Onset Cannabis Smoking and Depression Spells……........71
3.3.2. Study 2: Cannabis Problems and Functional Impairment Attributed to
Recently Active Depression Spells........................................................................73
3.3.3. Study 3: Cannabis Problems and History of Blunt Smoking.......................80
3.3.4. Study 4: Time to Cannabis Onset and Cannabis Problems..........................85
CHAPTER 4. RESULTS...............................................................................................................89
4.1. Study 1: Early-Onset Cannabis Smoking and Depression Spells...............................90
4.2. Study 2: Cannabis Problems and Functional Impairment Attributed to
Recently Active Depression Spells....................................................................................96
4.2.1. Sample Characteristics by Depression Status..............................................96
4.2.2. Descriptive Statistics and Distribution of Impairment Outcome
Indicators................................................................................................................97
4.2.3. Factor Analysis..........................................................................................101
4.2.4. Modeling Relationship Between Cannabis Problems and Functional
Impairment Attributed to Depression Spells........................................................105
4.2.5. Post-Estimation Exploratory Analyses......................................................107
4.3. Study 3: Cannabis Problems and History of Blunt Smoking....................................111
4.3.1. Characteristics of Past-Year Cannabis Users.............................................111
4.3.2. Prevalence of Blunt Smoking and Cannabis Problems..............................111
4.3.3. Exploratory and Confirmatory Factor Analysis of Cannabis Problems....114
4.3.4. Recent Blunt Smoking and Level of Cannabis Problems..........................117
4.4. Study 4: Time to Cannabis Onset and Cannabis Problems.......................................122
4.4.1. Frequency and Distribution of Cannabis-Only Users................................122
4.4.2. Cannabis Dependence (CD) and Time to Cannabis Onset (TCO)............126
4.4.3. Association Between Cannabis Problems (CP) and Time to Cannabis Onset
(TCO)...................................................................................................................127
CHAPTER 5. DISCUSSION.......................................................................................................133
5.1. Study 1: Early-Onset Cannabis Smoking and Depression Spells.............................133
5.2. Study 2: Cannabis Problems and Functional Impairment Attributed to
Recently Active Depression Spells..................................................................................141
5.3. Study 3: Cannabis Problems and History of Blunt Smoking....................................145
vii

5.4. Study 4: Cannabis Dependence and Time to Cannabis Onset..................................148
CHAPTER 6. CONCLUSIONS..................................................................................................152
6.1. Cannabis Problems and Functional Impairment Attributed to Depression..............152
6.2. Cannabis Problems and History of Blunt Smoking..................................................153
6.3. Time to Cannabis Onset and Cannabis Problems.....................................................156
APPENDIX..................................................................................................................................158
REFERENCES............................................................................................................................160

viii

LIST OF TABLES

Table 2.4.1. Clinical Features and Criteria for Cannabis Dependence from the Diagnostic and
Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) and the WHO International
Classification of Diseases, 10th Revision (ICD-10).......................................................................15
Table 2.4.2. Features of Socially Maladaptive or Hazardous Cannabis Use.................................16
Table 2.6.1. Major Depressive Episode Diagnostic Criteria Based on the Diagnostic and
Statistical Manual for Mental Disorders, Fourth Edition (DSM-IV)...........................................37
Table 3.1.1. U.S. National Surveys on Drug Use and Health, Overall Samples Sizes and
Response Levels, 2004-2011.........................................................................................................54
Table 3.1.2. Features of cannabis/alcohol problems (CP/AP) based on diagnostic criteria from
the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition.....................................60
Table 3.1.3. Tobacco Dependence (TD) items based on Nicotine Dependence Syndrome Scale
and Fagerstrom Test of Nicotine Dependence...............................................................................62
Table 3.2.1. Sample Sizes and Study Characteristics of the WHO World Mental Health
Surveys...........................................................................................................................................67
Table 3.2.2. Features of cannabis/alcohol problems based on criteria from the Diagnostic and
Statistical Manual of Mental Disorders, 4th Edition and the WHO International Classification of
Diseases, 10th Revision.................................................................................................................70
Table 3.3.1. Frequency and weighted prevalence of cannabis-only use by contributing World
Mental Health country...................................................................................................................85
Table 4.1.1. Sample Characteristics of Depression Spell Cases (n=173,775)...............................91
Table 4.1.2. Prevalence of Depression Spell Cases and Association with Cannabis, Cigarette, and
Alcohol Onset, and Years of Cannabis Involvement.....................................................................92
Table 4.1.3. Exploratory analysis of association between depression spells and fine-grained
cannabis onset exposure.................................................................................................................96
Table 4.2.1. Characteristics of US National Surveys on Drug Use and Health from 2009 to 2011
by Depression Status (n=172,042).................................................................................................97
Table 4.2.2. Descriptive Statistics of Sheehan Disability Scale (SDS) Items Among Sample of
Past-Year Depression Spell Cases (n=13,743)..............................................................................98

ix

Table 4.2.3. Correlation between Sheehan Disability Scale (SDS) Items of Functional
Impairment Due to Depression Spells (n=13,081)......................................................................102
Table 4.3.1. Characteristics of Total Sample Surveyed and Past-Year Cannabis User Analytical
Sample..........................................................................................................................................112
Table 4.3.2. Prevalence of Blunt Smoking History among Past-Year Cannabis Users by
Individual Survey Years..............................................................................................................113
Table 4.3.3. Weighted Prevalence (%) of Cannabis Problems among Past-Year Cannabis
Users............................................................................................................................................113
Table 4.3.4. Correlation Matrix of Cannabis Problem (CP) Items among Past-Year Cannabis
Users (n=3,534)...........................................................................................................................115
Table 4.3.5. Exploratory and Confirmatory Factor Analysis of Cannabis Problems, Standardized
and Unstandardized Factor Loadings Using Random Split Sample of Data from 2004 US
National Surveys on Drug Use and Health..................................................................................116
Table 4.4.1. Frequency of cannabis dependence (CD) among cannabis-only users by World
Mental Health Survey site............................................................................................................122
Table 4.4.2. Distribution of Cannabis-Only Sample by Relevant Characteristics and Cannabis
Dependence..................................................................................................................................123
Table 4.4.3. Unadjusted and covariate adjusted estimates of the association between cannabis
dependence (CD) and time to cannabis onset (TCO) use among cannabis-only users...............126
Table A.1. Spearman rank order correlation matrix between model covariates..........................160

x

LIST OF FIGURES

Figure 2.3.1. Chemical structure of delta-9-tetrahydrocannabinol................................................11
Figure 2.5.1. Annual prevalence of drug use among the population 15-64, 2008-2010. Source:
United Nations Office on Drug and Crime (UNODC), World Drug Report, 2012. Key:
Amphetamine-type stimulants (ATS)............................................................................................18
Figure 2.5.2. Prevalence of cannabis use in 2010 (or latest year available). Source: United
Nations Office on Drug and Crime (UNODC), World Drug Report, 2012. Estimates based on
annual report questionnaire data and other official sources...........................................................19
Figure 2.5.3. Cannabis incidence rates by year. Source: Gfroerer et al., 2002. Note: The
numerator of each rate is the number of persons who first used cannabis in the year, while the
denominator is the person-time exposure measured in thousands of years for persons aged 12 or
older. *Estimated using 2000 data only.........................................................................................20
Figure 2.5.4. Past month use of selected internationally regulated drugs (IRDs) among persons
aged 12 years or older, 2002-2012. Source: U.S. National Survey on Drug Use and Health,
National Findings, 2012.................................................................................................................21
Figure 3.3.1. Hypothesized latent structure of functional impairment attributed to depression
(IMPAIR).......................................................................................................................................75
Figure 3.3.2. Hypothesized latent structure of cannabis problems (CP).......................................76
Figure 3.3.3. Hypothesized latent structure of alcohol problems (AP).........................................77
Figure 3.3.4. Hypothesized latent structure of tobacco dependence (TD).....................................78
Figure 3.3.5. Model of relationships between functional impairment attributed to depression
(IMPAIR) and cannabis problems (CP), alcohol problems (AP), tobacco dependence (TD), and
background characteristics.............................................................................................................79
Figure 3.3.6. Alternative hypothesized two-factor latent structure of cannabis problems (CP),
with one factor of dependence (DEPEND) and a second factor of harmful socially maladaptive
or hazardous use (HARM).............................................................................................................82
Figure 3.3.7. SEM of association between cannabis problems (CP) and blunt smoking
recency...........................................................................................................................................83
Figure 3.3.8. SEM of the association between cannabis problems (CP) and blunt smoking
recency with statistical control for background characteristics and factor correlation with alcohol
problems (AP) and tobacco dependence (TD)...............................................................................84

xi

Figure 3.3.9. Latent measurement model of cannabis problems (CP) in WMH...........................88
Figure 4.1.1. Post-estimation exploratory analysis of the adjusted early-onset cannabis use
(EOCU) estimate with additional statistical control of other covariates using add-one-in (AOI)
and leave-one-out (LOO) approaches............................................................................................95
Figure 4.2.1. Distribution of home management/chore impairment among recently active
depression spell cases....................................................................................................................99
Figure 4.2.2. Distribution of work/school impairment among recently active depression spell
cases...............................................................................................................................................99
Figure 4.2.3. Distribution of close relationship impairment among recently active depression
spell cases.....................................................................................................................................100
Figure 4.2.4. Distribution of social life impairment among recently active depression spell
cases.............................................................................................................................................100
Figure 4.2.5. Post-hoc re-specification of the measurement model of functional impairment
attributed to depression (IMPAIR). Standardized coefficients in parentheses. CFI=0.97;
TLI=0.92; RMSEA=0.04.............................................................................................................102
Figure 4.2.6. Measurement model of cannabis problems (CP). Standardized estimates in
parentheses. CFI=0.995; TLI=0.994; RMSEA=0.02...................................................................103
Figure 4.2.7. Measurement model of alcohol problems (AP). Standardized estimates in
parentheses. CFI=0.992; TLI=0.991; RMSEA=0.02...................................................................104
Figure 4.2.8. Measurement model of tobacco dependence. Standardized estimates in parentheses.
CFI=0.996; TLI=0.996; RMSEA=0.05.......................................................................................105
Figure 4.2.9. between functional impairment attributed to depression (IMPAIR) and cannabis
problems (CP). CFI=0.992; TLI=0.99; RMSEA=0.01................................................................108
Figure 4.2.10. Association between functional impairment attributed to depression (IMPAIR),
cannabis problems (CP), and background characteristics. Reference groups were males, 12-17
year olds, and non-Hispanic Whites. Dashed arrows indicate estimates with p>0.05. CFI=0.97;
TLI=0.96; RMSEA=0.02.............................................................................................................109
Figure 4.2.11. Association between functional impairment attributed to depression (IMPAIR),
cannabis problems (CP), alcohol problems (AP), tobacco dependence (TD), and background
characteristics. Dashed arrows indicate estimates with p>0.05. CFI=0.998; TLI=0.998;
RMSEA=0.01..............................................................................................................................110

xii

Figure 4.3.1. Scree Plot of Eigenvalues from Exploratory Factor Analysis of Cannabis Problems.
Random Half-Split Sample of Data from 2004 National Surveys on Drug Use and Health
(n=3,534)......................................................................................................................................115
Figure 4.3.2. Association between level of cannabis problems (CP) and blunt smoking recency
among past-year cannabis users (n=67,519). CFI=0.95; TLI=0.95; RMSEA=0.02....................118
Figure 4.3.3. Association between cannabis problems (CP) and blunt use recency, sex, age,
race/ethnicity, and population density (n=67,519). Note: Dashed lines indicate estimate with
p>0.05. CFI=0.94; TLI=0.93; RMSEA=0.02..............................................................................119
Figure 4.3.4. Association between cannabis problems (CP) and recent blunt use, background
characteristics, alcohol problems (AP), and tobacco dependence (TD). Dashed lines indicate
estimates with p>0.05. Latent structure and factor loadings not depicted for CP, AP, and TD in
this figure. n=67,519; CFI=0.992; TLI=0.992; RMSEA=0.02....................................................120
Figure 4.3.5. Association between cannabis problems (CP) and days of blunt use in the past
month, background characteristics, alcohol problems (AP), and tobacco dependence (TD).
Dashed lines indicate estimates with p>0.05. Latent structure and factor loadings not depicted for
CP, AP, and TD in this figure. n=67,519; CFI=0.994; TLI=0.994; RMSEA=0.02....................121
Figure 4.4.1. Histogram plot of time to cannabis onset (TCO), all values..................................125
Figure 4.4.2. Histogram plot of time to cannabis onset (TCO) for non-negative values.............125
Figure 4.4.3. Measurement model of cannabis problems (CP). n=6,609; CFI= 0.999; TLI= 0.999;
RMSEA=0.009............................................................................................................................128
Figure 4.4.4. Unadjusted association between cannabis problems (CP) and time to cannabis onset
(TOC). n=6,081; CFI=0.998; TLI=0.998; RMSEA=0.011.........................................................129
Figure 4.4.5. Association between cannabis problems (CP), time to cannabis onset (TOC),
country, sex, and age. Dashed lines indicate p-values greater than 0.05, point estimates not
shown. n=6,030; CFI=0.997; TLI=0.996; RMSEA=0.009.........................................................130
Figure 4.4.6. Association between cannabis problems (CP) and time to cannabis onset (TOC)
adjusting for age of first IRD opportunity, cigarette smoking, alcohol problems (AP), country,
major depressive episode (MDE), and other background influences. Latent structure and factor
loading estimates for AP not depicted in this figure. Dashed lines indicate estimates with pvalues greater than 0.05, with these estimates not depicted. n=6,030; CFI=0.987; TLI=0.986;
RMSEA=0.02..............................................................................................................................131
Figure 4.4.7. Measurement model of alcohol problems (AP) with factor loadings from model
depicted in Figure 4.4.6...............................................................................................................132

xiii

LIST OF ABBREVIATIONS

5HT

Serotonin

ACASI

Audio Computer-Assisted Self-Interviewing

AD

Alcohol Dependence

ADHD

Attention deficit hyperactivity disorder

AP

Alcohol Problems

APA

American Psychiatric Association

ATS

Amphetamine-type stimulants

CAI

Computer Assisted Interviewing

CAPI

Computer-Assisted Personal Interviewing

CBSA

Census core-based statistical areas

CBT

Cognitive Behavioral Treatment

CD

Cannabis Dependence

CFA

Confirmatory Factor Analysis

CFI

Comparative fit index

CI

Confidence Interval

CIDI

Composite International Diagnostic Interview

CM

Contingency Management

CP

Cannabis Problems

CUD

Cannabis Use Disorder

DALY

Disability Adjusted Life Years

DEPEND

Dependence

df

Degrees of Freedom
xiv

DHHS

Department of Health and Human Services

DR

Designated Respondent

DSM

Diagnostic and Statistical Manual

DTC

Delayed Treatment Control

DU

Dwelling Unit

ECA

Epidemiologic Catchment Area

EFA

Exploratory Factor Analysis

EOCU

Early-Onset Cannabis Use

EV

Eigenvalues

FI

Field Interviewer

FTND

Fagerstrom Test of Nicotine Dependence

GWAS

Genome-Wide Association Study

HARM

Harmful Socially Maladaptive or Hazardous

HR

Hazard Ratio

ICD

International Classification of Diseases

IMPAIR

Functional Impairment

IRB

Institutional Review Board

IRD

Internationally Regulated Drug

MAU

Maladaptive Alcohol Use

MCU

Maladaptive Cannabis Use

MDD

Major Depressive Disorder

MDE

Major Depressive Episode

MET

Motivational Enhancement Treatment

xv

MI

Modification Indices

MTF

Monitoring the Future

NA

Noradrenaline

NCS

National Comorbidity Study

NCS-R

National Comorbidity Study – Replicate

NDSS

Nicotine Dependence Syndrome Scale

NHSDA

National Household Survey on Drug Abuse

NIDA

National Institute of Drug Abuse

NLAES

National Longitudinal Alcohol Epidemiologic Survey

NSDUH

National Surveys on Drug Use and Health

OR

Odds Ratio

PSU

Primary Sampling Unit

RMSEA

Root Mean Error of Approximation

RR

Relative Risk

SAMHSA

Substance Abuse and Mental Health Services Administration

SAOC

Subject-As-Own-Control

SDS

Sheehan Disability Scale

SMR

Standardize mortality ratio

SNP

Single nucleotide polymorphisms

SSR

State Sampling Region

SSU

Secondary Sampling Unit

TCO

Time to Cannabis Onset

TD

Tobacco Dependence

xvi

THC

Delta-9-tetrahydrocannabinol

TLI

Tucker-Lewis Index

UNODC

United Nations Office of Drug and Crime

WHO

World Health Organization

WMHS

World Mental Health Survey

Wt

Weighted

YRBS

Youth Risk Behavioral Surveillance

xvii

CHAPTER 1. AIMS AND OBJECTIVES
Hashish has nothing of that ignoble drunkenness about it which the races of the
North obtain from wine and alcohol; it offers an intellectual intoxication.
(Theophile Gautier, 1884, as quoted in Bey & Zug, 2004, p. 417)
This dissertation was conceived around a different sort of ‘intellectual intoxication’ - that of
answering a series of research questions pertaining to the patterns of cannabis smoking that
might lead to problematic use or depressive mood states.

1.1. Study 1. Are Early-Onset Cannabis Users At An Increased Risk Of Depression Spells?
Of important public health concern is whether depression is elevated as a result of youthful
cannabis involvement (C. Y. Chen, Wagner, & Anthony, 2002; Fergusson & Horwood, 1997;
Grant, 1995; B. E. Green & Ritter, 2000). If so, one implication may be that delaying cannabis
onset until adulthood might reduce the risk of later depression. Thus, the hypothesis of this
research focused on whether the first experience of a sustained spell of depressed mood, with
allied psychosomatic features (herein a “depression spell”) might be predicted by a potentially
toxic exposure earlier in life -- i.e., early-onset cannabis smoking. The aim of this research can
be summarized as follows:
AIM 1. Estimate the degree to which early-onset (onset ≤ 17 years) and adult-onset (18+
years) cannabis smoking are associated with a post-exposure first onset of a depression
spell in adulthood, within a conceptual model that accounts for time-invariant
characteristics of sex, age, and race/ethnicity, and variation in tobacco and alcohol use.

1

1.2. Study 2. Does the Level of Cannabis Problems Predict the Level of Functional
Impairment Attributed to Recently Active Depression?
Depression causes not only mental suffering, but can also interfere with the individual’s ability
to carry out everyday tasks, go to work, and engage in meaningful relationships. Little is known
about the degree to which these functional impairments attributed to depression are associated
with cannabis smoking. Further, some readers may question the clinical relevance or significance
of the depression spell concept under study. Therefore, Study 2 proposes to do the following:
AIM 2. Clarify levels of functional impairment attributed to the depression spell under
study, and estimate the degree to which the level of cannabis problems is associated with
the functional impairment attributed to depression in recently active cases, while
accounting for differences in background characteristics, tobacco and alcohol use, and
other relevant covariates.

1.3. Study 3: Do Cannabis Smokers with a History of ‘Blunt’ Smoking Differ in the
Experience and Degree of Cannabis Problems?
The practice of cannabis ‘blunt’ smoking (i.e., combining cannabis within a tobacco cigar shell)
has gained popularity in the U.S. since the 1990s (Golub, Johnson, & Dunlap, 2005; Timberlake,
2013). Pre-clinical and clinical findings motivate a concern over a functional interaction between
tobacco and cannabis (Agrawal et al., 2009; Castañé et al., 2002; Valjent, Mitchell, Besson,
Caboche, & Maldonado, 2002; Vandrey, Budney, Hughes, & Liguori, 2008). However, it is
unclear whether blunt smokers have a higher level of cannabis problems than their non-blunt
cannabis smoking counterparts (Ream, Johnson, Sifaneck, & Dunlap, 2006; Timberlake, 2013).
The main aim of Study 3 was to produce the following evidence:

2

AIM 3. Estimate the degree to which a history of cannabis ‘blunt’ smoking is associated
with the level of cannabis problems experienced, within a conceptual model adjusting for
the influence of sex, age, race/ethnicity, and other potential time-invariant covariates.

1.4. Study 4: Does Delaying Onset of Cannabis Smoking After Onset of First Opportunity
Account for Variations in Risk for Later Cannabis Problems?
As already mentioned, early-onset cannabis smoking is a marker for later problematic use.
Previous studies have measured this time to cannabis onset in relation to time since birth.
However, not everyone has an equal probability of being offered a chance to try cannabis, and
may be offered this first opportunity at different ages (Van Etten & Anthony, 1999; Van Etten,
Neumark, & Anthony, 1997; Wagner & Anthony, 2002b). Further, individuals may vary to the
degree they use cannabis soon after being presented a chance to try it. This source of variation
has not typically been accounted for or studied in relation to the later risk of experiencing
cannabis problems. Study 4 seeks to understand the relationship between time to cannabis onset,
measured as the number of years from the first chance to try drugs until first cannabis onset, and
the subsequent post-exposure risk of later cannabis problems. The main aim is summarized as
follows:
AIM 4. Estimate the degree to which delaying cannabis onset after the first chance to try
drugs might account for later levels of cannabis problems among a cross-national sample
of adults whose only internationally regulated drug (IRD) use has involved cannabis (i.e.,
'cannabis only' users).

3

CHAPTER 2. BACKGROUND AND SIGNIFICANCE
With hashish, the affective faculties reach the same degree of overexcitement as
do the intellectual faculties. The emotions have the mobility and, at the same time,
the “tyranny of ideas.” The individual uses the power to resist his violent feelings
until finally the mind itself reaches the point of incoherence.
(Moreau, 1845, p. 64)

The purpose of this chapter is to provide the reader with a scholarly and relevant
background on the topics pertaining to this dissertation research. As such, a part of the
background focuses on general related topics of the history, biology, and epidemiology of
cannabis smoking, but in addition, specific attention is paid to depression, functional impairment,
and blunt smoking, since each topic features prominently in this research on cannabis.

2.1. Cultivation and Common Preparations of Cannabis
There are three varieties of cannabis plant, each containing variable amounts of what is
regarded as the primary psychoactive drug compound in this plant, which is known as delta-9tetrahydrocannabinol (THC). The three varieties, Cannabis sativa, Cannabis indica, and
Cannabis ruderalis, are native to Asia, but have now spread throughout the world (Hart, Ksir, &
Ray, 2009). C. sativa is the most common variety associated with use of cannabis as a drug. This
plant can grow up to 20 feet tall, and can thrive under a wide variety of temperature, soil,
sunlight, and moisture conditions, which promotes production of cannabis as local or regional
enterprise rather than requiring international trafficking, such as for cocaine or heroin (Gold,
1991; United Nations Office on Drugs and Crime, 2011). Cannabis is primarily cultivated for its
4

psychoactive properties, but its fiber (called hemp) has historical uses as in manufacture of
clothing and rope (Abel, 1980; Gold, 1991). Cannabis is dioecious, meaning it has separate male
and female plants (Gold, 1991). Female cannabis plants produce a resin, which coats the
flowering tops and leaves of the plant, and this resin protects the plant from moisture loss (Gold,
1991). The psychoactive properties of cannabis is concentrated in the resin, and the parts of the
cannabis plant containing more resin are the most potent (Hart et al., 2009). Humans cultivating
cannabis as a drug breed sub-variations of C. sativa to produce more resin, or often growing it
under environmental conditions thought to maximize potency (Abel, 1980; Hart et al., 2009).
The cannabis plant is processed into different drug preparations that vary in potency
related to the concentration of cannabis resin contained therein. Historically, the most potent
preparation of cannabis has been called hashish, or the extract 'hashish oil' described below,
which is the cannabis resin compressed into blocks or cakes (Gold, 1991). Hashish comes in
various levels of purity, depending on the process of separating the resin from the rest of the
plant matter (Hart et al., 2009). Hashish oil is highly purified resin, dissolved in organic solvents,
and may be added to herbal cannabis in order to increase potency (Gold, 1991; Hart et al., 2009).
Less potent herbal cannabis, commonly referred to the slang term marijuana (or marihuana), is
composed of the dried and crushed remnants of the leaves and flowering tops of the cannabis
plant (Gold, 1991; Hart et al., 2009). In India, which has a long history of cannabis use in its
culture, has three common preparations, bhang, which is a dried powder made from cannabis
leaves and mixed into drinks (least potent), ganja, which is made from the flowers and upper
leaves, and charas, which is pure resin (similar to hashish) prepared from the flowers at the
height of bloom (Abel, 1980; Hart et al., 2009).

5

The most recently introduced cannabis preparation is known as 'dabs,' apparently akin to
'hashish oil,' but not yet well-characterized. The term ‘dabbing’ originates from an American
colloquial phrase of a ‘little dab will do ya’, meaning a small amount will go a long way. The
phrase originated in the 1950s from the marketing campaign for Brylcreem, a men’s hair styling
product (“Brylcreem,” 2014). In the context of a cannabis preparation, a ‘dab’ refers to the
smoking of a small amount of hash oil, particularly butane hash oil (BHO), which has a high
THC concentration (Black, 2012). Butane is used as a chemical solvent to extract THC from
plant matter, with the subsequent step of heating the butane/THC mixture to evaporate the
butane. Since butane is highly combustible, producing BHO can be very dangerous. It is claimed
that the THC content of dabs is greater than the THC content of prior cannabis extract
preparations, and that risk of THC overdose has become a possibility with the 'dabs' formulation
of cannabis products (C. Roberts, 2013). Initial field survey research on users of dabs in
Humboldt County, California, is underway and a NIDA research project proposal currently is
under review (James Anthony, personal communication, 2014). As of April 2014, a
bibliographic search of the scientific and public health literature disclosed no published journal
articles on 'dabs'. For example, the return from a NIH National Library of Medicine PubMed
search for 'dabs & cannabis’ was "No items found" and the PubMed return for 'dabs THC' was
"Your search for dabs THC retrieved no results."

2.2. Cannabis History and Beginnings of Public Health and Moral Concerns
Humans have been cultivating cannabis for thousands of years, and have used cannabis
for a variety of purposes than simply an intoxicant. Cannabis has been variously used to make
rope and clothing, as livestock feed (as seed), as medicine, and for religious purposes (Gold,

6

1991). Across this long time period, humans have debated the physical, psychological, and
ethical consequences of cannabis smoking, which have been shaped by myths, legends,
literature, religion, propaganda, and governments. Covering these historical, social, and legal
issues of cannabis use is important as background for our understanding of the context of the
public health and scientific issues surrounding the study of cannabis.
Humans have cultivated cannabis for the use as hemp dating back at least 10,000 to the
Stone Age (Abel, 1980). It is unclear whether cannabis’ intoxicating properties were known to
pre-history. According to historians, the first references to cannabis being consumed appear in
the Chinese pharmacopeia, the Pen Ts’ao, compiled by Emperor Shen-Nung in 2737 B.C.
Among cannabis’ listed uses were for relief of these ailments: “female weakness (menstrual
fatigue), gout, rheumatism, malaria, beri-beri, constipation, and absentmindedness” (Abel, 1980;
Walton, 1938). Cannabis was well known to other cultures of antiquity, including having
prominent religious significance in the holy Indian Vedas (at least 2000 B.C.), trade as hemp
among ancient Greeks (4th century B.C.), and referenced by the Romans in the influential
materia medica published by Dioscorides in 70 A.D. (Abel, 1980). Among Arabic cultures,
hashish was probably known to them by the tenth century A.D. (Abel, 1980). One of the more
infamous legends about cannabis involved the merchant and traveler Marco Polo. Taken prisoner
in Italy around 1297 A.D., Polo recounted the tale of a ruthless Persian ruler Hasan-ibn-Sabah
“Old Man of the Mountains”, who used his religiously fanatic cult to terrorize and conduct
politically motivated murders (Abel, 1980). While Polo did not directly implicate hashish,
Hasan’s followers became known variously as Heyssessini, Assinini, Hashshashin, or
Hashishiyya, and were associated with the use of hashish. This is reportedly where the word
“assassin” originates (Abel, 1980; Hart et al., 2009; Roffman, Schwartz, & Stephens, 2006).

7

By the 18th and 19th centuries, the properties of cannabis intoxication came to the
attention of Western cultures through colonial expansion and the travels of scientific or literary
figures. In 1798, Napoleon invaded Egypt and soldiers stationed there were introduced to
hashish, which helped infuse French curiosity when they returned with tales of the drug (Abel,
1980). One of the first to study cannabis scientifically was Dr. Jacques-Joseph Moreau (180484). His interest was in the potential for cannabis to produce symptoms similar to insanity
without first suffering mental illness. His study of hashish (in which he recounts his own and
other’s experiences with hashish) was published in Hashish and Mental Illness (1845). It
describes many of the effects of cannabis we know today such as euphoria, distortion of space
and time, and hallucinations or illusions (Abel, 1980; Moreau, 1845). Among the effects of
hashish intoxication Moreau discusses were disturbances of emotions:
One day in the middle of a very intense hashish intoxication, my ears were
suddenly struck by the sound of bells. This was hardly an hallucination but, being
in a sad mood, I interpreted this sound, which I would not normally have been
aware, as the tolling of a funeral bell. (Moreau, 1845, pp. 64–65)
Moreau’s work greatly influenced French Romanticist Pierre Jules Theophile Gautier, who
established the Hashish Club in Paris, and which included literary luminaries such as Alexandre
Dumas, Gerard de Nerval, Victor Hugo, Ferdinand Boissard, and Eugene Delacroix (Abel, 1980;
Gold, 1991; Roffman et al., 2006).
Notwithstanding the fascination for cannabis and hashish among select circles, it
remained an obscure drug to Western cultures, and governments generally showed little interest
in cannabis control before the 20th century, except to encourage its cultivation and sometimes to
raise taxes from its cultivation or sale. Concern over the 'vices' caused by alcohol and opium

8

were more salient, however, this would soon change. Cannabis was widely used in Britishcontrolled India, and in the mid-19th century there was growing concern over the “cannabis
problem”. This led to the formation of a commission to study the cultivation, preparation, and
physiological, psychological, and moral effects of cannabis in the population (Abel, 1980). Ganja
consumption among the lower classes was blamed for everything from ailing health, insanity,
and violent crimes (Abel, 1980). Proponents of cannabis prohibition were weighed against the
strong economic interests of British taxation related to cannabis production. Based upon written
and oral testimony of 1,193 witnesses, most of who had little direct expertise with cannabis or
relied on dogmatic opinions, the Committee concluded that moderate use had little physical
dangers, although excessive use could lead to ‘insanity’ in predisposed individuals, and cannabis
had no significant adverse moral influences (Abel, 1980). Their recommendation was for
continued taxation. Despite methodological flaws noted elsewhere, the Indian Hemp Drugs
Report of 1893-94 was the most comprehensive study on cannabis at the time.
In the United States, public concerns over cannabis consumption grew during the early
20th century due to increased propaganda, sensationalized media reports, and misleading
statements from government officials responsible for drug control. There was a
racial/xenophobic component to the acceptance and perpetuation of these concerns over
cannabis, which was associated with African-American Southern Blacks, Chinese immigrants,
and Mexicans living along the U.S. border (Hart et al., 2009). Though scant empirical evidence
was proffered, public officials like the U.S. Commissioner of Narcotics, Harry Anslinger, hyped
the dangers of cannabis use as being responsible for violence, crime, and sexual perversion. This
paranoia inevitably led to the Marijuana Tax Act of 1937, which taxed every level of cannabis
production, distribution, sale, and use, and along with state laws prohibiting cannabis possession

9

or use, effectively made use of cannabis illegal (Hart et al., 2009). Later government reports in
the U.S. and elsewhere found little empirical support for the myths surrounding the dangers of
cannabis, and despite the Marijuana Tax Act being overturned by the U.S. Supreme Court in
1969, cannabis remains a Schedule I substance (hence illegal) under the Controlled Substances
Act (CSA) of 1970. Nevertheless, the pendulum has been swinging back over the past 20 years
as more U.S. States have allowed cannabis for medicinal uses, and as of January 2014 two states
(Washington and Colorado) had passed laws allowing the legal consumption of cannabis for
recreational purposes.

2.3. The Pharmacology of Cannabinoids
The pharmacology of cannabis is important for understanding the possible biological and
neurological mechanisms as might account for adverse mental health consequences of cannabis
smoking. Cannabis is not a single drug, but is a mixture of over 400 compounds, with at least 60
identified as cannabinoids (Ashton, 2001). Among these, as noted in Section 2.1, the main
psychoactive constituent of cannabis is delta-9-tetrahydrocannabinol (see Figure 2.3.1;
abbreviated ∆-9-THC, or simply THC). THC is found in the plant’s resin and in the leaves,
flowers, and seeds, with higher concentrations of THC in the resin and flowers. Prior reports
have claimed that the THC content of herbal cannabis has increased over the decades since 1960,
with some estimates of a 10-30 fold increase; however, others contend a much more modest twofold change (Ashton, 2001; Hall & Swift, 2007; Mehmedic et al., 2010). In government
confiscated samples, THC content of herbal cannabis ranges between 3-5% (Mehmedic et al.,
2010).

10

CH3

OH

CH3
CH3

O

C5H11

Figure 2.3.1. Chemical structure of delta-9-tetrahydrocannabinol
There have been no reported cases of fatal overdose on cannabis per se; the likely quantify for a
fatal human dose greatly exceeds the amount of even a typical heavy cannabis user (Hall &
Degenhardt, 2009). According to media and online reports, there have been instances of nonlethal 'overdose' (as defined loosely as inadvertent consumption of many times more than the
effective dose of hash oil, dabs, or other high THC concentration products), although the
possibility of fatal overdose remains uncertain (C. Roberts, 2013).
Cannabis can be smoked or ingested, and there are several methods of delivery. The main
route of THC administration is by inhalation of smoke from combusted herbal cannabis. Herbal
cannabis can be smoked by rolling dried plant matter into cigarette paper (the resulting product is
called a ‘joint’), or smoked directly from a pipe or water pipe (a.k.a. a ‘bong’, or a device that
uses water to humidify the smoke for easier and deeper inhalation; Gold, 1991). Tobacco cigars
can also be used as a cannabis delivery device, where some of the tobacco contents are removed
and replaced with cannabis, then smoked. This practice is called ‘blunting’, and the resultant
cigar-cannabis combination is called a ‘blunt’. A more modern high-tech practice is to use a
specialized device to vaporize, rather than burn cannabis, thereby reducing the carcinogenic

11

byproducts and irritating effects of inhaling cannabis smoke. Cannabis is fat and alcohol soluble,
and can be ingested through foods and liquids, as is the case with South Asian 'bhang,' often
prepared by mixing plant material in drinks and candies (Ashton, 2001; Gold, 1991; Hart et al.,
2009).
The mechanisms of THC absorption, metabolism, and excretion greatly affect the
circulating blood concentration of THC. Via inhalation, THC is more rapidly absorbed into the
blood and distributed to the brain (Hart et al., 2009). About 50% of the THC in cannabis smoke
is absorbed by the lungs with inhalation (Ashton, 2001). Physiological and psychological effects
are often felt within minutes, and can last for an average of two hours depending on dose
(Ashton, 2001; Hart et al., 2009). Because THC is lipophilic, it can be stored in the fatty tissues,
and then released slowly in the blood stream (Ashton, 2001; Hart et al., 2009). Cannabis ingested
in food or drink produces around 25-30% lower THC blood concentration than smoking due to
the first pass effect of the liver (Ashton, 2001). The slower absorption delays onset of THC
effects, but also can produce more prolonged and steady intoxication (Ashton, 2001). THC is
metabolized by the liver into at least 45 metabolites, of which 11-hydroxy-delta-9-THC is a
major metabolite (Ashton, 2001; Hart et al., 2009). Some metabolites are psychoactive and may
be more potent than THC itself (Ashton, 2001). THC has a half-life of about 19 hours, and
complete elimination may take up to 30 days (Ashton, 2001; Hart et al., 2009). Excretion occurs
primarily through the gut, which allows for reabsorbed of THC and THC metabolites, but a
smaller portion is eliminated in the urine (Ashton, 2001; Hart et al., 2009). As a result of uptake
into fatty tissue and reabsorption by the gut, there is no clear correlation between blood or urine
THC-metabolite concentration and level of intoxication (Ashton, 2001).

12

Since THC was first isolated by Gaoni and Mechoulam (1964), there have been a number
of advances in the pharmacodynamics of cannabinoids on the human brain. Devane and
colleagues (1988) were the first to identify binding sites for cannabinoids in rat brains, which
today are labeled CB1 receptors. Subsequent work in pig models by Devane and others (1992)
discovered the first endocannabinoid, called anandamide (meaning “bliss” in Sanskrit). A second
type of cannabinoid receptor (CB2) was then identified and appears to play a role in cellular
mechanisms of immune response (Munro, Thomas, & Abu-Shaar, 1993). CB1 receptors have
been found throughout the human body, but within the brain there are greater densities in the
basal ganglia, cerebellum, hippocampus, thalamus, cerebral cortex, and the nucleus accumbens
(Hart et al., 2009; Herkenham et al., 1991). THC, along with other drugs such as opioids,
cocaine, amphetamines, and nicotine, activate the release of dopamine in the nucleus accumbens,
activating reward systems in brain (Ashton, 2001).

2.4. The Diagnosis of Cannabis Use Disorders
Whether termed drug addiction, habituation, or dependence, there comes a point in the
lives of some drug users when their drug-using behavior can be described as 'pathological' or
'compulsive', but some scholars voice objections to this 'medicalizing' of cannabis problems.
Nonetheless, differentiating between pathological and non-pathological patterns of drug use can
have important implications for clinical treatment, public health prevention strategies, and
scientific research. The concept of a syndrome is employed to characterize pathological drug use
where symptoms may encompass physical, mental, and behavioral manifestations, and where no
specific set of symptoms are necessary and sufficient. In this regard, a syndrome is considered to
be “a cluster of symptoms that co-occur in a way that signals the presence of an underlying

13

disorder” (Babor, 2006). Characteristics of a drug dependence syndrome include disturbances of
the mental life, disturbances of behavior, and neuroadaptive changes (Anthony, 2006; Edwards,
Arif, & Hodgson, 1981). Allied with the concept of a drug dependence syndrome has been the
recognition that problematic drug use may be manifested by socially maladaptive or hazardladen consequences, such as interference with important social roles or occurring under
physically hazardous situations (Edwards et al., 1981). This type of maladaptation has been often
termed as drug misuse or abuse. In this dissertation, drug abuse is only referred to as such when
specifically referencing the construct as defined in the DSM. Otherwise, these problems are
collectively referred to as maladaptive cannabis or alcohol use (MCU or MAU), depending on
which drug is being referenced.
The diagnosis of cannabis dependence follows along similar lines as that of other drugs
such as alcohol, opioids, and cocaine. Two similar and converging classification systems of
cannabis dependence come from the World Health Organization (WHO) and the American
Psychiatric Association (APA). The WHO’s nomenclature is detailed in the International
Classification of Diseases, 10th revision (ICD-10), while the APA’s comes from its Diagnostic
and Statistical Manuals (e.g., DSM-IV, DSM-5). Drug dependence clinical criteria from ICD-10
and DSM-IV are listed in Table 2.4.1. Although most criteria for dependence are common across
substances, the DSM-IV excluded cannabis withdrawal as clinically insignificant (American
Psychiatric Association, 1994). This conclusion has been challenged by more recent
epidemiologic studies (Budney, Hughes, Moore, & Vandrey, 2004). Diagnosis of cannabis
dependence under ICD and DSM-IV requires presence of three or more features in the same 12month period. With respect to the clinical features of socially maladaptive or hazardous cannabis
use (DSM “abuse”), these features are listed in Table 2.4.2.

14

During the writing of this dissertation, a new DSM was published (DSM-V; American
Psychiatric Association, 2013), with important changes in diagnostic criteria. First, there is no
longer a distinction between concepts of dependence and abuse; both are subsumed under a
single set of diagnostic criteria called ‘cannabis use disorder’. Second, whereas DSM-IV
cannabis dependence required three or more co-occurring clinical features, DSM-5 requires only
two. Third, a feature of “craving, or a strong desire or urge to use cannabis” was added to the list
of features. Fourth, arrest or legal problems due to or the result of cannabis intoxication is no
longer considered part of the clinical features.
Table 2.4.1. Clinical Features and Criteria for Cannabis Dependence from the Diagnostic and Statistical
Manual of Mental Disorders, 4th Edition (DSM-IV) and the WHO International Classification of Diseases,
10th Revision (ICD-10).
DSM-IV
Dependence or significant impairment or distress, as manifested by 3 or more of the following during a 12
month period:
1. Tolerance, as defined by either a need for markedly increased amounts of cannabis to achieve
intoxication or desired effect, or markedly diminished effect with continued use of the same amount of
cannabis
2. Cannabis is often taken in larger amounts or over a longer period than was intended
3. There is a persistent desire or unsuccessful efforts to cut down or control cannabis use
4. A great deal of time is spent in activities necessary to obtain cannabis (e.g., driving long distances),
use cannabis (e.g., socializing with cannabis using friends), or recover from its effects
5. Important social, occupational, or recreational activities are given up or reduced because of
substance use
6. Cannabis use is continued despite knowledge of having a persistent or recurrent physical or
psychological problem that is likely to have been caused or exacerbated by cannabis (e.g., chronic
cough related to smoking; excessive sedation resulting from repeated use of high doses)

15

Table 2.4.1. (Cont’d)
ICD-10
Three or more of the following manifestations should have occurred together for at least 1 month or, if
persisting for periods of less than 1 month, should have occurred together repeatedly within a 12-month
period:
1. A strong desire or sense of compulsion to use cannabis
2. Difficulties in controlling cannabis-taking behavior in terms of its onset, termination, or levels of use
3. A physiological withdrawal state when cannabis use has ceased or been reduced, as evidenced by: a
characteristic withdrawal syndrome for the substance; or use of the same (or closely related)
substance with the intention of relieving or avoiding withdrawal symptoms
4. Evidence of tolerance, such that increased doses of cannabis are required in order to achieve effects
originally produced by lower doses
5. Progressive neglect of alternative pleasures or interests because of cannabis use, increased amount
of time necessary to obtain or use cannabis or to recover from its effects
6. Persisting with cannabis use despite clear evidence of overtly harmful consequences, such as
depressive mood states consequent to periods of heavy use, or cannabis-related impairment of
cognitive functioning; efforts should be made to determine that the user was actually, or could be
expected to be, aware of the nature and extent of the harm.
Table 2.4.2. Features of Socially Maladaptive or Hazardous Cannabis Use.
Features
1. Cannabis used frequently interfered with work or responsibilities at school, on a job, or at home.
2. Cannabis caused arguments or other serious or repeated problems with your family, friends,
neighbors, or co-workers AND continued to use despite these problems.
3. Used cannabis in situations where one could get hurt, for example when riding a bicycle, driving,
operating a machine, or anything else.
4. Arrested or stopped by the police because of driving under the influence of cannabis or because of
one’s behavior due to being [high/drunk].

16

2.5. The Epidemiology of Cannabis Smoking
The following sections describe the epidemiology of cannabis smoking. This information
is organized around five rubrics that encapsulate five broad questions the field of epidemiology
seeks to hold a lens to concerning the understanding of diseases or adverse health conditions
(Anthony & Van Etten, 1998): 1) Quantity - How many are affected?; 2) Location – Where are
affected cases more likely to be found?; 3) Causes – Why do some people become cases while
others do not?; 4) Mechanisms – What processes link who becomes a case and continues to
remain a case?; 5) Prevention and control – What can be done to prevent or intervene in the
disease process?

2.5.1. Quantity and Location
Cannabis is more widely used around the world than any other internationally regulated
drug (IRD). An estimated 3% to 5% of the global population 15 to 64 years of age use cannabis
annually, representing roughly 119 to 224 million people (Figure 2.5.1) (United Nations Office
on Drugs and Crime, 2012). By comparison, fewer people use cocaine, opiates/opioids,
amphetamine-type stimulants (ATS), and ecstasy combined. Cannabis is more prevalent in
established market economies, such as the United States, the European Union, Canada, and
Australia (Figure 2.5.2). Lower estimates in less economically developed regions such as Africa
(5-13%) and Asia (1-3%) may be due to a lack of high-quality data (United Nations Office on
Drugs and Crime, 2012). Nevertheless, these regions represent a large proportion of the world’s
population, and even low estimates could translate into a substantial number of cannabis users.
Cross-national estimates of cannabis use are also available via the World Health
Organization World Mental Health Survey Initiative (WHO-WMHSI). These surveys estimate

17

the highest lifetime cumulative incidence for cannabis smoking occurs in the U.S. (42%) and
New Zealand (42%), moderate sizes (10-20%) in countries such as Colombia, Belgium, France,
Germany, Netherlands, Spain, and Israel, and the lowest estimates (<10%) in Mexico, Italy,
Ukraine, Lebanon, Nigeria, South Africa, Japan, and the People’s Republic of China
(Degenhardt et al., 2008).
Few studies are available to estimate the prevalence of cannabis smoking in India, despite
its historical tradition of cannabis use in its culture, significant production of cannabis and resin,
and large population. Previously published reports from the United Nations Office of Drug and
Crime (UNODC) estimate the prevalence of cannabis in India to be around 3.2% as of 2000
(United Nations Office on Drugs, 2006). More recent studies report prevalence estimates as high
as 14% among urban adolescent or rural populations, to 4-7% among college students (Baba et
al., 2013; Goel & Chakrabarti, 2010; Gupta, Sarpal, Kumar, Kaur, & Arora, 2013; Ningombam,
Hutin, & Murhekar, 2011).

Figure 2.5.1. Annual prevalence of drug use among the population 15-64, 20082010. Source: United Nations Office on Drug and Crime (UNODC), World Drug
Report, 2012. Key: Amphetamine-type stimulants (ATS).

18

Figure 2.5.2. Prevalence of cannabis use in 2010 (or latest year available).
Source: United Nations Office on Drug and Crime (UNODC), World Drug
Report, 2012. Estimates based on annual report questionnaire data and other
official sources.

There is a relatively high rate of cannabis consumption in the U.S., and therefore its NIH
finances much of the research on cannabis epidemiology. There are approximately 18.9 million
current cannabis users (i.e., used cannabis within the past month), representing about 7.3% of the
population (U.S. Department of Health and Humans Services, 2013). Peak cannabis incidence
occurred around the mid-1970s, and afterwards there was a sharp decline until a resurgence
during the 1990s (Figure 2.5.3). Since 2002, the prevalence of current cannabis use has been
relatively stable (6.2% to 7.3%), although in recent years there has been a slight statistically
significant increase (Figure 2.5.4). By comparison, the prevalence of other IRDs has been either
flat or slightly declining. Adolescence and young adulthood are peak periods of risk for initiating

19

cannabis smoking (17-18 years), and males are more likely to use than females (Gfroerer, Wu, &
Penne, 2002; Wagner & Anthony, 2002a). With respect to racial or ethnic background, White
youths are more likely to have used cannabis in the past year (15%), than Hispanics (13%),
African-Americans (9%), or Asians (6%) (United States of America, Substance Abuse and
Mental Health Services Administration, & Office of Applied Studies, 2002a). Cannabis smokers
are also more likely to be found among those with lower academic achievement, the
unemployed, and in disadvantaged neighborhoods (United States of America et al., 2002a;
United States of America, Substance Abuse and Mental Health Services Administration, &
Office of Applied Studies, 2002b, 2004).

Figure 2.5.3. Cannabis incidence rates by year. Source: Gfroerer et al., 2002.
Note: The numerator of each rate is the number of persons who first used
cannabis in the year, while the denominator is the person-time exposure measured
in thousands of years for persons aged 12 or older. *Estimated using 2000 data
only.

20

Figure 2.5.4. Past month use of selected internationally regulated drugs (IRDs)
among persons aged 12 years or older, 2002-2012. Source: U.S. National Survey
on Drug Use and Health, National Findings, 2012.

The epidemiology of cannabis use disorders (CUD) has been less studied than use, thus
much of the information comes from studies done in the U.S. and Australia. Based on estimates
from the 2012 U.S. National Survey on Drug Use and Health (NSDUH), there were about 4.3
million persons with CUD in the past year (1.7%), which is a trend that has changed little since
2002 (U.S. Department of Health and Humans Services, 2013). Other large epidemiologic
surveys in the U.S. and Australia report similar findings (Grant & Pickering, 1998; Hall,
Teesson, Lynskey, & Degenhardt, 1999; Kessler et al., 1994; Robins & Regier, 1991; Stinson,
Ruan, Pickering, & Grant, 2006). An estimated 8% of cannabis users experience cannabis
21

dependent within the first ten years, and males are more likely than females to make this
transition (Wagner & Anthony, 2002a, 2007; Wittchen et al., 2008). Peak ages for the first CUD
occur around the same age as starting to use cannabis, which may indicate a relatively short
transition time from first use to disordered use (Stinson et al., 2006; Wagner & Anthony, 2002a).

2.5.2. Causes
Suspected causes and underlying theories on why some people use cannabis and then
continue to a point when use becomes problematic or pathological can be organized into three
broad categories: biological, psychological, and sociological. Most theories of drug use focus on
suspected causes that can be generalized to all psychoactive drugs, rather than cannabis-specific.
Drug epidemiology studies causes at multiple levels of scale. Biological causes and theories
focus mainly on micro-level causes that occur within the individual. Psychological causes and
theories can involve both individual-level causes and interactions with group-level cause.
Sociological causes and theories deal mainly with group-level or macro-level causes. This does
not to say that theories or models exclude aspects of one another. In fact, some integrated
theories of drug use overlap, although emphasis or importance might be placed on a specific
category. The sheer number of different theories on drug use is too numerous to provide a
comprehensive review. However, in this dissertation, I shall attempt to cover a few of the main
relevant theories and evidence in each of the three categories. A more comprehensive review of
theories on drug use are covered elsewhere (Lettieri, Sayers, & Pearson, 1980; Petraitis, Flay, &
Miller, 1995; West, 2001).

22

2.5.2.1. Biological Causes and Theories
Biological theories of drug use tend to focus on physical mechanisms, primarily those
involving genes and drug receptors, as an explanation for why some people initiate cannabis and
exhibit problematic cannabis use phenotypes (W. Martin, 1980; Schuckit, 1980). These
mechanisms have been investigated using experimental studies on animal models, human studies
using twin samples, gene linkage analyses, genome-wide association studies (GWAS), and noninvasive imaging techniques.
Suspected genetic determinants of drug use would be carried from conception, and may
influence the formation of susceptible drug use phenotypes, with possible contributions of
environmental factors. For example, clinical studies of female and male monozygotic twin pairs
estimate that around 40 to 48% of the respective total variation in cannabis initiation can be
explained by shared genetics, and 59 to 51% for cannabis problems (Kendler, Karkowski, Neale,
& Prescott, 2000; Kendler & Prescott, 1998; Rhee et al., 2003; Tsuang et al., 1998; Verweij et
al., 2010). This variation between initiation and disordered use may be due to common genetic
liabilities contributing more to drug initiation in general, and substance-specific genetic factors
possibly playing a larger role in problematic use (Kendler, Jacobson, Prescott, & Neale, 2003;
Xian et al., 2008). Genetic linkage analysis studies, which try to narrow the search for drug userelevant genes to specific chromosomal locations, have found loci for both biologically relevant
and potentially novel genes to cannabis dependence, dependence symptoms, and other use
pattern phenotypes (e.g., initiation, early use, and frequency of use) (Agrawal et al., 2008a,
2008b; Ehlers, Gizer, Vieten, & Wilhelmsen, 2010; Hopfer et al., 2007). Only two known
genome-wide association studies have been currently done, but neither study reported finding

23

single nucleotide polymorphisms (SNPs) relevant to cannabis use that reached acceptable
standards for statistical significance (Agrawal et al., 2011; Verweij et al., 2013).
In addition to genetics, exposure to cannabis can induce pharmacological changes in
receptors, neurological pathways, and gene expressions that may influence the development of
dysfunctional cannabis use patterns. Prolonged exposure to THC produces sex-dependent
downregulation of CB1 brain receptors in both animals and humans, which provides evidence for
both the development of tolerance and sources variation by sex. However, this effect appears
reversible in humans after several weeks of cessation, and may not complete the whole picture
(Hirvonen et al., 2012; B. R. Martin, Sim-Selley, & Selley, 2004). Pre-clinical studies have
further found evidence of common and specific pathways related to THC withdrawal via changes
in gene expression (Le Merrer et al., 2012). Although much more work is needed to better
understand the biological mechanisms underlying cannabis initiation and problematic use, even
proponents of biological theories recognize they explain only part of the observed variation, and
that environmental events contribute to the overall picture (Schuckit, 1980).

2.5.2.2. Psychological Causes and Theories
Psychological causes and theories of drug use emphasize the role of positive and negative
reinforcement or individual personality characteristics in the initiation and continuation of drug
use (Goode, 1989; Jessor & Jessor, 1980; Lindesmith, 1968; McAuliffe & Gordon, 1980). Drug
use behaviors are positively reinforced, and thus repeated, to the extent the experience is
pleasurable, while behaviors are negatively reinforced when they relieve or avoid pain, such as to
alleviate withdrawal symptoms. Pre-clinical work finds THC activates the same mesolimbic
dopamine system hypothesized responsible for the rewarding and reinforcing effects of multiple

24

drug classes (Gardner & Vorel, 1998; Tanda & Goldberg, 2003). In humans, cannabis users
under both naturalistic and laboratory environments commonly report experiencing more
positive feelings, such as euphoria, relaxation, happiness, sociability, creativity, and sexual
pleasure/arousal, than negative feelings of paranoia, anxiety, depressed mood, or irritability (B.
Green, Kavanagh, & Young, 2003). Volunteers under controlled laboratory conditions report
experiencing positive reinforcing properties of both smoked cannabis and oral-THC (Chait &
Zacny, 1992; Mendelson & Mello, 1984). In another study of primarily opioids users, cannabis
was reported as positively reinforcing after the very first use, in contrast to subjectively
described effects of nicotine (Haertzen, Kocher, & Miyasato, 1983).
The degree to which cannabis is reinforcing might manifest as shorter lag times from first
use of the drug to the second use, third, fourth, and so on, prior to other processes (e.g.,
withdrawal) exerting an influence on the process that drives frequency of drug use. This topic
has been informed by early work by Becker (1953) on the experiences of first-time or novice
cannabis users, and most recently by Agrawal and colleagues on the initial reactions towards
cannabis (Agrawal, Madden, Bucholz, Heath, & Lynskey, 2014; Agrawal, Madden, Martin, &
Lynskey, 2013). Notwithstanding this research, time from first cannabis use to more frequent use
(e.g., monthly use) has served as a proxy measure of this process. For example, Crowley and
colleagues (1998) found patients in treatment programs for delinquent or substance use problems
transitioned from first use of tobacco or cannabis to regular (monthly use) within about a year,
while the same transition for alcohol took slightly longer (mean = 2 years). Other studies have
reported either shorter or comparable transition times from first to regular cannabis use as
compared to alcohol, tobacco, and other “illicit” drugs (Ridenour, Lanza, Donny, & Clark, 2006;
Wittchen et al., 2008). Negative reinforcement is commonly associated with the presence of a

25

withdrawal syndrome. While cannabis withdrawal was previously considered to be of little
clinical significance, the current consensus is otherwise (Budney et al., 2004).
Theories emphasizing the role of personality characteristics as determinants of starting to
use drugs have focused on traits of deviance, impulsivity, and risk taking. Here, by ‘deviant’ the
meaning is not pejorative, but refers to behaviors that are outside the norms of the majority.
Jessor and Jessor (1980) hypothesized a social-psychological structure called ‘problem behavior
proneness’, whereby drugs users would be more likely to have “a concern with personal
autonomy, a lack of interest in the goals of conventional institutions like church and school, a
jaundiced view of the larger society, and a more tolerant view of transgression” (Jessor & Jessor,
1980, p. 109). These personality traits shape and are shaped by the perceived social environment
where the compatibility and support of peers and parents exert the most influence. Problem
behavior proneness has some direct empirical support for cannabis use (Jessor, Chase, &
Donovan, 1980; Jessor, Jessor, & Finney, 1973). However, much of the current literature has
focused specifically on conduct problems and other externalizing behaviors as predictors of drug
use (Fergusson, Boden, & Horwood, 2008; Krueger et al., 2002; Miles, van den Bree, & Pickens,
2002; Schubiner et al., 2000).
Impulse control problems may also underlie such deviant behaviors, in which such
individuals are more likely to take risks, and choose more immediate gratification over long-term
negative consequences. Both experimental and observational evidence support findings that
individuals with higher impulsivity or adolescents with attention deficit hyperactivity disorder
(ADHD) are more likely to use cannabis and become cannabis dependent (Day, Metrik, Spillane,
& Kahler, 2013; Kong et al., 2013; C. A. Martin et al., 2002; McDonald, Schleifer, Richards, &

26

de Wit, 2003; Miles et al., 2001; Murphy, Barkley, & Bush, 2002; Rios-Bedoya, Wilcox, Piazza,
& Anthony, 2008).

2.5.2.3. Sociological Causes and Theories
While biological and psychological causes and theories focus on individual-level
differences of scale, sociological causes and theories attempt to explain drug use in terms of
structural factors in society and social relations (Goode, 1989). For example, adolescents who
associate with more deviant and/or drug-using peers are at a greater risk for later cannabis
involvement (Dishion & Loeber, 1985; Fergusson, Horwood, & Swain-Campbell, 2002; Hofler
et al., 1999; Lynskey, Fergusson, & Horwood, 1998; von Sydow, Lieb, Pfister, Hofler, &
Wittchen, 2002). Theories that emphasize peer influence explain drug use as a learned behavior,
where youths may seek out deviant groups that share underlying values, but are also themselves
socialized into the drug subculture that has its own conduct norms (B. D. Johnson, 1980; Kandel,
1980; Oetting & Beauvais, 1987; Oetting & Donnermeyer, 1998; Sutherland, 1939). Aside from
peers, parents are another source of influence on drug use that has been studied. Higher levels of
parental monitoring and poorer parenting practices may increase the chance of trying cannabis
and initiation (Bohnert, Anthony, & Breslau, 2012; J. S. Brook et al., 1998; C. Y. Chen, Storr, &
Anthony, 2005; Chilcoat & Anthony, 1996; Hansen et al., 1987). Self-control theory attributes
inadequate parental socialization and monitoring as leading to children having low self-control
(i.e., more impulsivity), and thus greater likelihood of using drugs (Gottfredson & Hirschi,
1990). In addition to parental influences, cannabis and drug users are less likely to go to church
and be involved in school (J. S. Brook, Adams, Balka, & Johnson, 2002; C. Y. Chen, Dormitzer,
Bejarano, & Anthony, 2004; Cochran, Wood, & Arneklev, 1994; Dornbusch, Erickson, Laird, &

27

Wong, 2001). This has led some sociologists to theorize that drug use may be due to inadequate
social control or weak attachments to parental, religious, and school institutions (Elliott,
Huizinga, & Ageton, 1985; Elliott, Huizinga, & Menard, 1989; Goode, 1989; Hirschi, 1969).
This lack of social control can occur on the neighborhood level, and is referred to as a social
disorganization model, where deviant behaviors like drug use flourish because members of the
neighborhood are unwilling or unable to place controls on such behaviors (Goode, 1989; Petraitis
et al., 1995). Cannabis users are found to cluster within certain neighborhoods, and living in
more disordered or disadvantaged neighborhoods are found to increase the risk of being offered
cannabis and using (Bobashev & Anthony, 1998, 2000; Crum, Lillie-Blanton, & Anthony, 1996;
Wells, Degenhardt, Bohnert, Anthony, & Scott, 2009; Wilson, Syme, Boyce, Battistich, &
Selvin, 2005).

2.5.3. Mechanisms
The mechanisms of cannabis smoking describe the inter-related processes that link the
initial cannabis use experience to repetition of the cannabis use behavior until the development
of a pathological compulsion to use cannabis that may not be entirely volitional in the sense that
there is an attenuation or 'loss of control' over use. This general model of the natural history of
drug use grew from work by Robins (1980), who investigated the predictors of different drugstage transitions among heroin-using Vietnam Veterans (Helzer, Robins, & Davis, 1976). More
recently, Anthony (2010) commented upon this model, where the first step in the process of
becoming a drug user is having an opportunity to use the drug. After the onset of drug use there
is an accumulation of drug use experiences manifested in the frequency of occasions. At some
later point, there begins the first signs or symptoms of a drug problem (e.g., experiences of

28

tolerance or craving), which occur prior to the full expression of a drug use disorder. These drug
problems then can influence or feedback into the process of repetitious drug use, which may
further drive the dependence process. However, transition along this continuum is not certain for
all drug users, and many may stop spontaneously or due to an intervention, while still others may
continue to use drugs without experiencing dysfunction. Nonetheless, heavy or chronic cannabis
smoking (with or without dependence) might lead to adverse health consequences and/or social
disadvantage.

2.5.3.1. Opportunity to Try Cannabis
The earliest stage of cannabis use is marked by the initial opportunity to use cannabis. In
the U.S., as many as half of the population report at least one opportunity, which typically occurs
around age 16 (Van Etten et al., 1997). Among those with an opportunity to try cannabis, around
one-third do so within the first year (Van Etten & Anthony, 1999; Van Etten et al., 1997). Other
investigations, conducted mostly in the U.S. and Latin America, have tried to identify sources of
variation in who does and does not have a chance to use cannabis (C. Y. Chen, Dormitzer,
Gutierrez, et al., 2004; Delva et al., 1999; Neumark, Lopez-Quintero, & Bobashev, 2012;
Rosenberg & Anthony, 2001a; Stenbacka, Allebeck, & Romelsjö, 1993; Storr, Chen, &
Anthony, 2004; Storr, Wagner, Chen, & Anthony, 2011; Wagner & Anthony, 2002b). Males,
those who are less religious or engaged in school, those who are more aggressive or misbehave,
and prior use of tobacco and alcohol are all associated with either being offered cannabis or have
an opportunity at an earlier age (Van Etten, Neumark, & Anthony, 1999; Wells et al., 2011).
Social environment may also play a factor. Youths living in disadvantaged neighborhoods, have
parents who use drugs or are less engaged, and who have cannabis-using friends are more likely

29

to have opportunities (Benjet et al., 2007; Crum et al., 1996; Pinchevsky et al., 2012; Storr,
Chen, et al., 2004). Given a chance to try cannabis, a relatively high proportion end up using
(65%), which is greater than for cocaine or heroin, but similar to hallucinogen use (Van Etten &
Anthony, 1999).

2.5.3.2. Cannabis Use and Escalation to Frequent and Problematic Use
The natural course of cannabis use begins during the period of adolescence, where the
risk of using cannabis starts to increase at 12-13 years, reaches a peak by ages 17-18, and then
has few onsets after 21-22 years (Tucker, Ellickson, Orlando, Martino, & Klein, 2005; von
Sydow et al., 2002; Wagner & Anthony, 2002a; Wittchen et al., 2008). Researchers using
longitudinal growth modeling have tried to identify trajectory patterns of cannabis use from
those who start in adolescence though young adulthood. For example, Schulenberg and
colleagues (2005) identified five main groups: 1) a rare group that had some use, but never
frequent, 2) a fling group characterized by a brief period of frequent use, followed by little or no
later use; 3) increasers, whose frequency of cannabis use escalated over time; 4) decreasers, who
frequently used in the beginning, but then ceased or decreased use over time, and 5) chronic
users who became frequent users and continued to be so throughout follow-up. Other studies
have identified similar groupings characterized by low-level, occasional users, those who
escalate over time, those whose use declines as they get older, and chronic users whose level of
use remain high (J. S. Brook, Lee, Brown, Finch, & Brook, 2011; Caldeira, O’Grady, Vincent, &
Arria, 2012; Tucker et al., 2005; Windle & Wiesner, 2004).
The developmental periods of a cannabis use disorder are also around the same period of
adolescence and young adulthood (15-20 years), with an expected, slightly higher average peak

30

of onset risk around 17-19 years (Stinson et al., 2006; Wagner & Anthony, 2002a; Wittchen et
al., 2008). When a cannabis use disorder does develop, it usually does so within the first few
years after onset, and it much less likely to develop after age 25 (C. Y. Chen, O’Brien, &
Anthony, 2005; Wagner & Anthony, 2002a). Which features of problematic cannabis use emerge
first has been previously studied by Rosenberg and Anthony (2001b), comparing non-dependent
and dependent cannabis users. They found cannabis users experienced subjective loss of control
over use was one of the earliest and most frequent features, and experiences of hazardous use
followed fairly quickly after onset (within the first year). Tolerance emerged relatively later than
previously expected, and withdrawal was one of the least and later features experienced. Among
cases of cannabis dependence, using larger amounts than intended and hazardous use emerged
early in the process, suggesting a heterogeneous course of problematic use.

2.5.3.3. Secondary Consequences of Cannabis Smoking
Part of the natural course of cannabis smoking may include secondary outcomes leading
to increased morbidity or mortality. There have been no known reported cases of fatal cannabis
overdose, and there is some evidence that cannabis users are less likely to die from a drug-related
event (Nyhlén, Fridell, Bäckström, Hesse, & Krantz, 2011). However, similar to persistent
tobacco smoking, the risk of early mortality may accumulate over time due to other causes.
Only a few studies have looked at all-cause mortality and cannabis use. Neumark, Van
Etten, and Anthony (2000) observed that those with a drug dependence diagnosis were 2-3 times
as likely than those without such a diagnosis to die prematurely (average of 22 years), adjusted
for age, race, smoking status, and an alcohol use disorder. Although cannabis dependence was
grouped together with dependence on other drugs, those who only had cannabis dependence

31

made up 65% of all drug dependence cases, and 45% of deceased cases. Three other studies have
examined cannabis all-cause mortality specifically. In a Swedish study of a treatment population,
patients with a primary cannabis use disorder had a standardize mortality ratio (SMR) of 5.3
(Arendt, Munk-Jørgensen, Sher, & Jensen, 2011). This was much lower than for heroin
(SMR=16.9), and cocaine (SMR=7.0), but similar to amphetamines (SMR=5.4). Another study
of Danish males followed over 35-years found cannabis was associated with an elevated
mortality hazard (hazard ratio, HR=4.3) that was similar for stimulants (HR=4.4), but less than
for opioids (HR=2.8) (Davstad, Allebeck, Leifman, Stenbacka, & Romelsjö, 2011). The authors
attributed the lower mortality risk of opioids to the prominence of stimulant use in Denmark.
Finally, a U.S. based study of patients in a large health service organization reported a higher
mortality risk for males of weekly and daily cannabis smokers (relative risk, RR=1.5 and 1.4,
respectively), but not for females (Sidney, Beck, Tekawa, Quesenberry, & Friedman, 1997).
Cannabis may contribute to increased morbidity and mortality through physiological
disease and accidental injury. Acute effects of cannabis smoking may include triggering of
strokes and myocardial infarction episodes, especially in susceptible individuals (Desbois &
Cacoub, 2013; Mittleman, Lewis, Maclure, Sherwood, & Muller, 2001). Intoxication can also
impair reaction times and impair driving, leading to an increased risk of automobile crashes
(Bedard, Dubois, & Weaver, 2007; Blows et al., 2005; Callaghan, Gatley, et al., 2013; Drummer
et al., 2003; Fergusson, Horwood, & Boden, 2008). Long-term effects of cannabis smoking on
the risk of lung or other cancers has been less convincing based on epidemiologic studies
(Hashibe et al., 2005; Mehra, Moore, Crothers, Tetrault, & Fiellin, 2006). This may be due in
part to difficulties measuring lifetime cumulative exposure and sorting out the influence of
tobacco smoking, Two recent case-control studies reported increased cannabis joint-years was

32

associated with a higher odds of lung cancer, even after controlling for tobacco pack-years
(Aldington et al., 2008; Berthiller et al., 2008). These findings were supported by a prospective
study that found “heavy” cannabis smoking (defined as 50 or more uses) was associated with a
two-fold excess risk of lung cancer in a sample followed for 40-years (Callaghan, Allebeck, &
Sidorchuk, 2013). Only one study has purported cannabis use decreases the risk for lung and
other cancers (A. Chen et al., 2008).
Much of the literature on long-term effects of cannabis on mental health has focused on
cognitive deficits and psychosis, with a lesser extent mood disorders and suicide risk. Heavy
cannabis users exhibit cognitive deficits and impaired memory as compared to controls, but it is
unclear whether these effects are long-lasting or continue after cannabis use is discontinued
(Jager, Block, Luijten, & Ramsey, 2010; Pope, Gruber, Hudson, Huestis, & Yurgelun-Todd,
2001; Pope & YurgelunTodd, 1996; Solowij et al., 2002). Among psychiatric disorders, cannabis
use appears to increase the risk of psychosis, but there has been less consistent evidence for
mood disorders (Arseneault, Cannon, Witton, & Murray, 2004; Degenhardt, Hall, & Lynskey,
2003; T. H. M. Moore et al., 2007). Suicide risk is elevated among cannabis users, but it is still
unclear whether if it is caused by cannabis, a consequence, or confounded by other factors
(Beautrais, Joyce, & Mulder, 1999; Lynskey et al., 2004; T. H. M. Moore et al., 2007; Pedersen,
2008; Rasic, Weerasinghe, Asbridge, & Langille, 2012; Wilcox & Anthony, 2004).

2.5.4. Prevention and Control
There is a societal need for prevention and treatment of cannabis problems. In the U.S.,
an estimated 957,000 people received treatment for cannabis use in 2012, more than for cocaine
and heroin (650,000 and 450,000, respectively), but slightly behind treatment for pain reliever

33

use (973,000), and much lower than for alcohol (2.4 million) (U.S. Department of Health and
Humans Services, 2013). Survey samples of U.S. high-school seniors from 1977 to 2005 indicate
that among past-year cannabis users, at least half felt they either should stop or reduce their use
(Terry-McElrath, O’Malley, & Johnston, 2008). Control efforts are directed to developing
interventions and treatments to treat CUDs and reduce their incidence. Cannabis is criminalized
in most countries, and therefore those presenting for treatment may be there due to a court order.
The ethics of legally-coerced cannabis treatment has been debated, yet some evidence points to
better outcomes as compared to non-coerced persons (Caplan, 2006; Copeland & Maxwell,
2007; Miller & Flaherty, 2000).
Primary prevention programs seek to reduce the incidence of CUDs by preventing the
onset of cannabis smoking. Two systematic reviews of school-based interventions tried to
identify characteristics of cannabis prevention programs that were the most effective (Lemstra et
al., 2010; Porath-Waller, Beasley, & Beirness, 2010; Tobler, Lessard, Marshall, Ochshorn, &
Roona, 1999). Increased effectiveness was seen in programs that administered to larger samples,
used multiple prevention models, were longer in duration, used facilitators other than teachers,
targeted high-school students, and stressed interactive social competency development rather
than lecture-based knowledge improvement. However, different conclusions were drawn when
both school-based and non-school based primary prevention programs were studied (Norberg,
Kezelman, & Lim-Howe, 2013). Programs that targeted all-drugs, younger adolescents (10-13
years), multiple modalities (i.e., school, family, and peer-based), and were shorter in duration
with multiple booster sessions had greater effect sizes. These effect sizes ranged from trivial to
large, with most studies reporting trivial to small effects.

34

A few of the more studied treatment programs for CUDs and their conceptual framework
are covered here. Cognitive behavioral treatment (CBT) is based on social learning theory and is
designed to improve the coping skills, and may be appropriate to those already seeking to change
(Stephens, Roffman, Copeland, & Swift, 2006). CBT can be paired with motivational
enhancement treatment (MET), which was designed to reduce ambivalence and increase
motivation to change for those cannabis dependent users who may be uncertain change is
needed. While treatment-outcome studies support CBT/MET over delayed treatment controls
(DTCs), these studies have yet to show CBT/MET to be superior over other treatments (Budney,
Higgins, Radonovich, & Novy, 2000; Copeland, Swift, Roffman, & Stephens, 2001; Dennis et
al., 2004; Stephens, Roffman, & Curtin, 2000; Stephens, Roffman, & Simpson, 1994). For
example, in the Cannabis Youth Treatment (CYT) Study, Dennis and colleagues (2004)
conducted a randomized trial comparing CBT/MET to four alternative treatments on outcomes of
days abstinent and percent in recovery in a sample of adolescents with cannabis related
disorders. While all treatments produced similar results in outcomes, they found CBT/MET was
most cost-effective after controlling for level of severity. An alternative approach frequently
studied is contingency management (CM) interventions, which rely on positive/negative
reinforcement/punishment strategies to induce change. For example, subjects might receive a
voucher of monetary value for clear urine samples, or have the value of the voucher reduced if
incurring relapse. Contingency management, especially voucher-based studies, show a positive
effect on abstinence and may improve effectiveness when combined with CBT and MET
strategies (Budney et al., 2000; Budney, Moore, Rocha, & Higgins, 2006; Carroll et al., 2006;
Kadden, Litt, Kabela-Cormier, & Petry, 2007).

35

2.6. Cannabis Smoking, Depression, and Functional Impairment
2.6.1. The Burden and Epidemiology of Depression
Before discussing the literature on the relationship between cannabis smoking and
depression, some background on the diagnosis, epidemiology, and public health burden of
depression will be helpful. Major depression is characterized by a sustained period of low mood
or markedly diminished interest or pleasure in most activities (i.e., anhedonia), and can be
accompanied by difficulties with sleep, weight or appetite, concentration, energy, or recurrent
thoughts of suicide (see Table 2.5.1 for DSM-IV diagnostic criteria; American Psychiatric
Association, 1994). The World Health Organization World Mental Health Surveys (WHOWMHS) projected as many as one in three people by the age of 75 will experience a mood
disorder in their lifetime, with a greater prevalence among high income countries like the U.S.,
France, the Netherlands, and New Zealand (Kessler et al., 2007). Mood disorders tend to have
onsets in late 20s to early 40s, and there may be evidence of an increased prevalence among
younger cohorts compared to older cohorts (Fombonne, 1994; Kessler et al., 2003, 2007;
Klerman & Weissman, 1989; Weissman et al., 1996). Females are more likely than males to be
depressed (Hankin et al., 1998; Kessler, McGonagle, Swartz, Blazer, & Nelson, 1993;
Lewinsohn, Hops, Roberts, Seeley, & Andrews, 1993; Piccinelli & Wilkinson, 2000). Other
characteristics, such as lower income, lower educational attainment, unemployment, divorced or
separation, other psychiatric disorders, and drug use are associated with depression (Anthony &
Petronis, 1991; Blazer, Kessler, McGonagle, & Swartz, 1994; Hasin, Goodwin, Stinson, &
Grant, 2005; Kessler et al., 2011; Weissman et al., 1996). The public health burden is
considerable. Depression is the leading cause of disability adjusted life years (DALYs) among
mental health disorders, and it is projected to be the second largest overall cause of DALYs by

36

2020 (Lopez, Mathers, Ezzati, Jamison, & Murray, 2006; Murray & Lopez, 1997). Major
depression accounted for 8% of global years lived with disability in 2010, and was the leading
contributor to the burden of suicide and ischemic heart disease (Ferrari et al., 2013). The
economic toll, in terms of treatment costs and lost productivity, exceed well over $80 billion
dollars annually in the United States (Greenberg et al., 2003).
Table 2.6.1. Major Depressive Episode Diagnostic Criteria Based on the Diagnostic and Statistical
Manual for Mental Disorders, Fourth Edition (DSM-IV)
A. Five (or more) of the following symptoms have been present during the same two-week period and
represent a change from previous functioning. At least one of the symptoms is (1) depressed mood or (2)
loss of interest or pleasure.
1) Depressed mood most of the day, nearly every day, as indicated by either subjective report or
observation made by other.
2) Markedly diminished interest or pleasure in all, or almost all, activities most of the day, nearly
every day.
3) Significant weight loss when not dieting or significant gain, or decreased or increased appetite
nearly every day.
4) Insomnia or hypersomnia nearly every day.
5) Psychomotor agitation or retardation nearly every day.
6) Fatigue or loss of energy nearly every day.
7) Feelings of worthlessness or excessive or inappropriate guilt (which may be delusional) nearly
every day (not merely self-reproach or guilt about being sick).
8) Diminished ability to think or concentrate, or indecisiveness, nearly every day (either by
subjective account or as observed by others).
9) Recurrent thoughts of death (not just fear of dying), recurrent suicidal ideation without a specific
plan, or a suicide attempt or specific plan for committing suicide.
B. The symptoms do not meet the criteria for a mixed episode.
C. The symptoms cause clinically significant distress or impairment in social, occupational, or other
important areas of functioning.
D. The symptoms are not due to the direct physiological effects of a substance (for example, a drug of
abuse, a medication), or a general medical condition.
E. The symptoms are not better accounted for by bereavement, i.e., after the loss of a loved one, the
symptoms persist for longer than 2 months or are characterized by marked functional impairment, morbid
preoccupation with worthlessness, suicide ideation, psychotic symptoms or psychomotor retardation.

37

2.6.2. Suspected Mechanisms Linking Cannabis Smoking and Depression
In clinical and representative household samples, depression and cannabis smoking tend
to co-occur more often than one would expect by chance alone (Alpert, Maddocks, Rosenbaum,
& Fava, 1994; C. Y. Chen et al., 2002; Degenhardt, Hall, & Lynskey, 2001; Grant, 1995; Rey,
Sawyer, Raphael, Patton, & Lynskey, 2002; Rowe, Fleming, Barry, Manwell, & Kropp, 1995).
For example, in the nationally representative U.S. National Longitudinal Alcohol Epidemiologic
Survey (NLAES), major depression was associated with a six-fold higher odds of a cannabis use
disorder (Grant, 1995). More modest or null associations have been reported when differentiating
between dependent and non-dependent cannabis use, and after controlling for sex, other
background factors, tobacco, and alcohol use (C. Y. Chen et al., 2002; Degenhardt et al., 2001).
Findings of co-morbidity from cross-sectional studies are suggestive of a causal
relationship, but additional evidence is needed to assure the suspected exposure preceded the
outcome in time. In addition, other considerations aid in making causal inferences, such as the
nine described by Sir Bradford Hill (1965): 1) strength of the association; 2) consistency of the
evidence; 3) specificity of exposure to the disease; 4) temporality; 5) presence of biological
(dose-response) gradient; 6) plausibility with respect to biological mechanisms; 7) coherence
with what is already known; 8) experiment evidence; 9) analogy from similar evidence. In light
of these considerations, experimental and epidemiologic research designs have attempted to
answer two questions, while trying to rule out a third: 1) Does cannabis smoking cause
depression? 2) Does depression cause cannabis smoking? 3) Can the link between cannabis
smoking and depression be explained by some third factor or process? Two main avenues of
research have been applied to these questions. One avenue looks at biological mechanisms
linking the neurological underpinnings of mood dysregulation and the endocannabinoid system.

38

The other proposes indirect psychological or social factors connect cannabis smoking to
depressed mood, or that cannabis is used to alleviate negative affect.

2.6.3. Biological Mechanisms
Pre-clinical studies, mostly using rodent models, have tried to understand the relationship
between cannabinoids, their receptors, and mechanisms involved with mood regulation.
Cannabinoid receptors (CB1) are expressed on neurons throughout the central nervous system,
including the mesolimbic areas of the brain, thought to be involved in stress-response and
emotional regulation through the hypothalamic-pituitary-adrenocortical (HPA) system
(Herkenham et al., 1991; Holsboer, 2000; Weidenfeld, Feldman, & Mechoulam, 1994). Direct
stimulation of CB1 receptors or indirect stimulation by inhibiting the metabolism of
endocannabinoids produces antidepressive-like effects, possibly through stimulation of neurons
that increase transmission of noradrenaline (NA) and serotonin (5HT) (Gobbi et al., 2005; M. N.
Hill & Gorzalka, 2005; Matthew N Hill & Gorzalka, 2005). Inhibiting reuptake of these
neurotransmitters is the basis of many current anti-depressant drugs (Gobbi et al., 2005;
Schildkraut, 1965). Mice lacking CB1 receptors generally exhibit more depressive-like and
anxiety-like behaviors, but results can vary under different environment conditions, especially
when stress is induced (M. Martin, Ledent, Parmentier, Maldonado, & Valverde, 2002; Zanettini
et al., 2011). The biological evidence would appear to support an anti-depression role for
cannabinoid action and receptors.
The role of the endocannabinoid system in the regulation of emotion and mood is
complex and not fully understood. Low doses of cannabinoid agonists can induce anxiolytic-like
responses in mice, but larger doses can produce anxiogenic effects (Kathuria et al., 2003; Patel &

39

Hillard, 2006; Rubino, Realini, et al., 2008). Further, cannabinoid antagonists also have been
shown to produce anxiolytic and antidepressive-like effects (Navarro et al., 1997; Shearman et
al., 2003; Tzavara et al., 2003). Human clinical trials of a CB1 antagonist called rimonabant were
ceased after subjects experienced elevated anxiety and depression (Scheen, Finer, Hollander,
Jensen, & Van Goal, 2006; Van Gaal, Rissanen, Scheen, Ziegler, & Rossner, 2005).
There may also be important and overlooked sex-dependent and maturation-dependent
variation in effects of cannabinoids on mood. During adolescence, the human brain continues to
develop and may be vulnerable to changes induced by exogenous agents like cannabis (Rice &
Barone, 2000; Schneider & Koch, 2003; Schneider, 2008). Following along this line, Rubino and
colleagues (2008) found that administration of THC in adolescent rats and then left to mature
had sex-dependent reduction in CB1 receptor density and G-protein coupling (which is necessary
for the CB1 receptor to be “active”) in the amygdala, ventral tegmental area, and nucleus
accumbens areas of the brain involved in emotional processing and reward. Female rats
administered a THC-agonist exhibited no effects on anxiety-like responses, but they did display
more “behavioral despair” and anhedonia, while males only showed anhedonia. Bambico and
others (2010) similarly found that administration of a CB1 receptor agonist produced anxietylike and depressive-like behaviors in adolescent-treated, but not adult-treated rats. Both low and
high doses led to rats performing more poorly on the forced swim and sucrose preference tests
used as models for behavioral despair and anhedonia, respectively. Anxiety-like differential
responses were produce for high doses only.

40

2.6.4. Psychosocial Mechanisms
An alternative mechanism of cannabis toxicity might involve psychosocial factors,
whereby cannabis involvement initiates or exacerbates a chain of adverse life outcomes, which in
turn increase the risk of depression (Degenhardt et al., 2003). Lower educational attainment,
unemployment, greater financial strain, difficulties maintaining and establishing close
relationships, legal problems, illegal drug involvement, stressful life events, and other difficulties
assuming adult roles have been observed to be more common among heavy cannabis smokers
and predictive of depression (Colman & Ataullahjan, 2010; Fergusson & Boden, 2008;
Fergusson & Horwood, 1997; B. E. Green & Ritter, 2000; Kandel, Davies, Karus, & Yamaguchi,
1986; Marmorstein & Iacono, 2011; Weich & Lewis, 1998). Depression can also contribute to
lower educational attainment (Fletcher, 2008), unemployment (Luo, Cowell, Musuda, Novak, &
Johnson, 2010), and marital disruption (Briscoe & Smith, 1973; Fincham, Beach, Harold, &
Osborne, 1997), suggesting potential feedback loops between depression and poor psychosocial
outcomes.
Another perspective theorizes that depression might increase the incidence of cannabis
smoking via self-medication to alleviate negative affect or disturbances in appetite or sleep that
are among depression's clinical features (Khantzian, 1997). This idea would appear plausible
considering the above mentioned pre-clinical evidence for cannabinoids producing
antidepressive effects, human studies of the positive affective experiences of cannabis
intoxication, and self-reported motives of alcohol and drug use initiation among cases of
dysphoric mood (Bambico, Duranti, Tontini, Tarzia, & Gobbi, 2009; B. Green et al., 2003;
Mueser, Drake, & Wallach, 1998). However, most of the epidemiologic evidence has failed to
show that individuals with depression at baseline are more likely to start using cannabis (Arendt

41

et al., 2007; Bardone et al., 1998; J. S. Brook et al., 1998; Kandel & Chen, 2000; Kandel et al.,
1986; Patton et al., 2002).

2.6.5. Epidemiologic Evidence for Cannabis as a Possible Cause of Depression
A few selected cross-sectional studies are informative as to whether cannabis smoking
predicts later depressed mood. Normally, this is not the case when surveys measure both the
suspected exposure and outcome at a single time point, and there is no other information from
which to guide what event came first. However, chronological assumptions can be made when
participants are asked to self-report the age when specific events occur, such as the onset of drug
use and depression. For example, Chen, Wagner, and Anthony (2002) used age of onset
information to account for the temporal sequence between major depressive episode (MDE) and
cannabis dependence in data from the U.S. National Comorbidity Survey (NCS). They found
cannabis smoking males, with or without daily tobacco smoking, had a modest excess risk of
MDE. Only daily tobacco smoking, non-dependent cannabis-using females were at an increased
risk of MDE over-and-above the excess risk for depression among females.
A similar approach was employed by this dissertation’s author in a paper by de Graaf and
colleagues (2010) using data from the WHO World Mental Health Surveys (WHO-WMHS). We
studied the association between early-onset cannabis use defined any use prior to age 18 years
and the first occurrence of a sustained spell of depressed mood in adulthood (18 years or older)
among all adults. In our analysis, we excluded those with a depression spell onset prior to
adulthood as a violation of the temporal assumptions of the model. A modest association
between early-onset cannabis use and depression spells was observed (odds ratio, OR=1.5),
adjusting for sex, age, tobacco use, and other mental health problems. After further adjustment
for conduct problems, there was no longer a statistically significant effect estimate.
42

Another study, which held constant shared genetic and environmental factors through a
twin study design, found no evidence that early-onset cannabis use predicted major depressive
episode (Lynskey et al., 2004). However, they did find that early-onset cannabis use was
associated with attempted suicide. While caution is warranted when drawing causal inferences
from cross-sectional study designs, these studies do have advantages in that they have no sample
attrition, take less time to complete, and can more easily inform across a broader age span than
typical longitudinal study designs. Nevertheless, prospective studies are considered superior in
terms of making causal inferences with respect to temporally measured exposures and outcomes,
and can be better suited to study time-varying covariates when multiple observations are made
over time. These studies are reviewed below.
The prospective relationship between adolescent cannabis smoking and the risk of
depression in adolescence or young adulthood is unclear. Perhaps the earliest longitudinal study
to test this hypothesis was in a representative sample of public high school students in New York
State (Paton, Kessler, & Kandel, 1977). Cannabis use did not predict depressed mood one year
later, or in a later follow-up at age 24-25 years (Kandel et al., 1986). Past-year cannabis use at
age 15 was associated with depression one year later at age 16 in unadjusted models (odds ratio,
OR=2.7), but not in adjusted models (OR=1.2) in a New Zealand cohort, the Christchurch Health
and Development Study (CHDS;(Fergusson, Lynskey, & Horwood, 1996). Follow-up of this
cohort to age 20-21 did reveal higher rates of depression with increasing cannabis involvement
(Fergusson et al., 2002). Contrasting this trend, weekly cannabis smoking in females, but not
males, was associated with an almost two-fold excess risk of depression (OR=1.9) in an
Australian cohort by age 20-21 years, but later follow-up to 24 years showed evidence of no
effect of either occasional or weekly cannabis use and depression (Degenhardt et al., 2013;

43

Patton et al., 2002). However, the Degenhardt et al. study used a different depression measure
and did not report separate estimates by sex.
The timing of cannabis onset (early vs. late) may modify the relationship with depression.
For instance, cannabis use in childhood (by age 14) and in mid-adolescence (14-16 years)
predicted a modest excess risk of a major depressive disorder (MDD) (OR=1.6 and 1.4,
respectively) by late 20s, but cannabis use in early 20s did not (J. S. Brook et al., 2002). In
another study, both early and late adolescent cannabis onset (before 14 vs. after) was associated
with depression and anxiety in young adulthood (21 years) for frequent cannabis users (ORs=2.3
and 3.0), but no association was observed for early or late onset, occasional users (Hayatbakhsh
et al., 2007).
Prospective studies of adult samples have likewise been inconsistent. For example,
having a cannabis use disorder, but no depression symptoms, at baseline was associated with a
four-fold increased risk of experiencing at least one depression symptom after a 14-16 year
follow up in adults 18 years or older from the Baltimore Epidemiologic Catchment Area (ECA)
Study (Bovasso, 2001). In another study, any cannabis use was associated with major depression
in an adult sample from the Netherlands after a three-year follow-up, but differences were not
seen when separating cannabis users by frequency levels (van Laar, van Dorsselaer,
Monshouwer, & de Graaf, 2007). These positive findings are weighed against others who have
found no evidence of an association, or report null results after accounting for confounding
factors (Harder, Morral, & Arkes, 2006; Manrique-Garcia, Zammit, Dalman, Hemmingsson, &
Allebeck, 2012; Pedersen, 2008).
Variation among prospective studies may be due to a number of factors. Differences in
the age range of exposure studied (adolescence vs. adult), cumulative cannabis exposure or

44

phenotype (any cannabis use, heavy use, dependent use), length and number of follow-up points,
measurement of depression (symptom scale vs. clinical diagnosis), and degree of adjustment for
potential confounding factors could explain some of the variation in estimates. Nevertheless, a
recent meta-analysis of 10 longitudinal studies concluded an overall very modest positive effect
of cannabis smoking on depression (OR=1.2), and a slightly larger effect for heavy cannabis use
(8 studies; OR=1.6; Lev-Ran et al., 2013).

2.6.6. Functional Impairment, Depression, and Cannabis Use
Regardless of the causal relationship between cannabis smoking and depression, it may
still be possible that cannabis smoking exacerbates depressed mood by increasing the severity or
functional impairment. Severity and impairment tend to be positively correlated, but the former
is a function of the “development, spread, and depth of dysfunction of the disease or disorder,
while impairment is an outcome of the underlying disease in a given environment, concerning
what people can do in terms of activities.” (Ustun & Kennedy, 2009). Neither depression
severity or impairment are clearly defined or operationalized in the DSM, even though clinically
significant distress and impairment are part of the qualifying criteria, and an integral to the
clinical judgment of severity (American Psychiatric Association, 1994, 2013). In the DSM, there
are three levels of depression severity (mild, moderate, and severe) with or without psychotic
features, but the concepts of severity and impairment are somewhat conflated:
Mild: Few, if any, symptoms in excess of those required to make the diagnosis,
and the symptoms result in only minor impairment of occupational functioning or
in the usual social activities or relationships with others.
Moderate: Symptoms or functional impairment between “mild” and “severe”.

45

Severe without psychotic features: Several symptoms in excess of those
required to make the diagnosis, and symptoms that markedly interfere with
occupational functioning or the usual social activities or relationships with others.
Severe with psychotic features: Delusions or hallucinations.
(American Psychiatric Association, 1994).
By contrast, the WHO-ICD system avoids this complexity altogether by determining depression
severity (mild, moderate, severe) based on number of symptoms alone, and making no mention
of "impairment" or “disability” (World Health Organization, 1993). The WHO definition of
disability encompasses functional domains such as learning and applying knowledge, general
tasks and demands, communication, mobility, self-care, social functioning, and
school/occupational functioning (Ustun & Kennedy, 2009).
Depression severity has been variously measured as the number of symptoms, intensity
of symptoms, or in reference to suicide co-morbidity (Lux, Aggen, & Kendler, 2010). It may be
an important indicator for the degree of psychiatric co-morbidity with other disorders, affect
treatment outcomes, and predict depression recurrence (Fournier et al., 2010; Judd et al., 2000;
Keller et al., 1992; Kirsch et al., 2008; Nemeroff, 2007; Shapiro et al., 1994). Within the
cannabis-depression literature, depression severity has been studied as the sum of depression
symptoms or a latent dimension using depression symptom scales. These studies of depression as
a continuous outcome tend to be more consistent in finding a positive relationship with cannabis
smoking (Fleming, Mason, Mazza, Abbott, & Catalano, 2008; Griffith-Lendering, Huijbregts,
Mooijaart, Vollebergh, & Swaab, 2011; Groth & Morrison-Beedy, 2010; Horwood et al., 2012;
Otten, Barker, Maughan, Arseneault, & Engels, 2010; Pahl, Brook, & Koppel, 2010; Repetto,
Zimmerman, & Caldwell, 2008).

46

Functional impairment may be just as clinically relevant as severity measures. An
estimated 59% of past 12 month cases of major depressive episode (MDE) rated their
impairment severe or very severe within in a U.S. general population sample (Kessler et al.,
2003). Co-morbidity was common among these MDE cases with severe impairment, with 24%
having a co-morbid substance use disorder. Among adolescents (13-17 years) with major
depression/dysthymia, about a third experienced serious impairment, which was defined as either
moderate impairment in multiple areas (i.e., fears and anxieties that lead to gross avoidance
behavior, episodes of aggression, or antisocial behavior), or severe impairment in one area
(Kessler et al., 2012).
Few studies have reported on the relationship between cannabis smoking, depression, and
functional impairment. In one study, cannabis smoking was part of a scale of risky behaviors that
also included cigarette smoking, alcohol use, sexual intercourse, violence, and suicide (Flisher et
al., 2000). Several correlates of risky behaviors were evaluated, including various stressors,
individual resources (e.g., social competence, verbal intelligence, and parental monitoring),
biological dispositions, and psychopathology (e.g., mood disorder, anxiety disorder, disruptive
disorder, and functional impairment). Higher levels of risky behaviors were associated with
having a mood disorder and greater functional impairment. However, the assessment of
functional impairment was non-specific to any particular cause, and all risky behaviors were
grouped together, hence it is unknown the degree cannabis smoking specifically contributed to
the correlation with depression or functional impairment.
Another study focused specifically on co-morbidity between cannabis use disorders and
mood disorders in the previous 12 months in a household sample of teens from the Houston,
Texas metropolitan area (R. E. Roberts, Roberts, & Xing, 2007). In unadjusted models, teens

47

diagnosed with any cannabis use disorder had three times the odds of having a mood disorder.
The researchers then examined those with a mood disorder and scored below the mean on the
Child Global Assessment Scale (CGAS), a measure of functional impairment with lower scores
indicating greater impairment. Cannabis abuse was associated with a seven-fold excess odds
(95% CI: 2.2, 26.8), but no association was observed for teens with cannabis dependence.
However, models that adjusted for alcohol and other drug use disorders produced either null or
unstable estimates due to too little or no variation in the mood disorder/functional impairment
outcome.
A more recent study of a clinical sample of youths seeking treatment for a primary mood
or anxiety disorder found those with higher scores on a scale of alcohol, tobacco, and other drug
use had higher levels of depression and more impairment (Osuch, Vingilis, Ross, Forster, &
Summerhurst, 2013). However, in these studies it is unclear the degree to which cannabis
smoking might be the causal influence, and whether the impairment experienced was due to
depression or other sources.
One exception was a study of college students whose cannabis use trajectories were
followed for seven years (Caldeira et al., 2012). Students whose cannabis use was characterized
as early-decline, late-increase, college-peak, or chronic had more days of impairment due to
emotions and higher depression scores than non-users or low-stable groups, adjusting for
trajectories of alcohol and tobacco use. Research on this topic could be furthered by studying the
relationship between cannabis smoking, depression, and functional impairment in general
population samples as proposed in this dissertation.

48

2.7. The Blunt Smoking Phenomenon
Studies conducted primarily in the U.S. have reported a gradual increase in the rates of
both cannabis use and dependence beginning in the 1990s and then remaining relatively stable
throughout the 2000s (Compton, Grant, Colliver, Glantz, & Stinson, 2004; Gfroerer et al., 2002;
Golub et al., 2005). Some researchers have suggested this recent resurgence of cannabis smoking
is due to a higher THC content or more early-onset use (Compton et al., 2004). An alternate
viewpoint proposes that youths born since 1970 are eschewing what some might describe as
“hard” drugs, such as cocaine and heroin, in favor of consuming cannabis, particularly in the
form of a ‘blunt’ (Golub & Johnson, 1999). A blunt is constructed from an inexpensive cigar by
replacing the tobacco contents with cannabis. This method of cannabis delivery is contrasted by
an American ‘joint’, in which cannabis is wrapped in cigarette paper. It should be noted that
Europeans commonly mix tobacco and cannabis in a ‘joint’, which is also called a ‘blowtje’ in
the Netherlands (Ream et al., 2006). ‘Chasing’, where smoking a joint or blunt is quickly
followed by smoking a cigar or other tobacco product, may also be similar in terms of the
mechanisms and consequences of tobacco-cannabis co-administration. Blunt smoking originated
and gained popularity among New York City youths, especially African-American males
connected to the emerging hip-hop music scene of early 1990s (Golub, Johnson, Dunlap, &
Sifaneck, 2004).
Blunt smoking has become quite common in the U.S., with evidence of increasing
prevalence over time, and shrinking subgroup differences. A school-based study of youths from
grades 7-12 estimated the overall lifetime prevalence to be as high as 20% (Soldz, Huyser, &
Dorsey, 2003), however, a more nationally representative sample of 12-17 year olds placed
overall blunt smoking prevalence closer to 11% (Golub et al., 2005). Nearly half of all cannabis

49

users have smoked a blunt, but prevalence in some population subgroups may be as high as 8090% with males, African-Americans, older youths, and those living in urban areas being overrepresented (Golub et al., 2005; Substance Abuse and Mental Health Services Administration,
2007; Timberlake, 2013). Between 2000 and 2010, the overall prevalence of blunt smoking
increased from 11% to 14%, with evidence that African-American cohorts persisted in their blunt
use into adulthood, and younger cohorts of other race/ethnic subgroups catching up to AfricanAmerican youths (Golub et al., 2005; Timberlake, 2013).
Public health researchers concerned about consequences of cannabis consumption,
especially among youths and minority groups, should be particularly interested in the blunt
phenomenon. Ethnographers report blunt smokers to be culturally distinct from other cannabis
users, with unique practices, rituals, and terminology, which could have implications for
targeting health messages and interventions (Dunlap, Johnson, Benoit, & Sifaneck, 2005; B. D.
Johnson, Bardhi, Sifaneck, & Dunlap, 2006; Kelly, 2005; Sifaneck, Johnson, & Dunlap, 2005).
A blunt a can hold greater quantity of cannabis than a joint, burn slower and longer, can be more
easily passed around, and is more portable and disposable (Kelly, 2005; Mariani, Brooks, Haney,
& Levin, 2011). Some blunt users (28%) reported adding other drugs to the blunt, including
cocaine, heroin, or psychedelics (e.g. LSD, PCP, mescaline) (Soldz et al., 2003). When properly
constructed, a passing observer might find a blunt to be indistinguishable in both appearance and
smell to a non-cannabis cigar. Youths find this aspect of concealing their cannabis consumption,
especially in more public areas, to be appealing (Sifaneck et al., 2005). In addition, the marketing
of cigar ‘flavors’ (e.g., bubble gum, fruit flavors, vanilla) among brands popular for blunt use,
and ‘blunt wraps’ (tobacco leaf paper), is attractive to youths, novice users, minorities, and
women (Sifaneck et al., 2005). The nicotine content of blunts may be low, yet blunt smokers are

50

more likely to be nicotine dependent and may use tobacco cigarettes a higher rates (Timberlake,
2009). This could differentially expose blunt smokers to greater health risks associated with
tobacco, which are well documented (US Department of Health and Human Services, 2004).
Another potential concern is whether blunt smoking might influence the experience and
development of a cannabis use disorder (CUD). Nicotine and THC might interact
pharmacologically to affect tolerance, withdrawal, and physiological responses (Valjent et al.,
2002). For example, a pre-clinical study found that co-administration of nicotine and THC
produced differential physiological effects (lower body temperature, locomotion, and pain
sensitivity), slowed tolerance to these effects, and increased anxiety-like behaviors greater than
the additive effect of each drug alone (Valjent et al., 2002). Cannabinoid CB1 and nicotinic
receptors may play a role in neuromechanisms underpinning dependence and the reinforcing
effects for both substances (Viveros, Marco, & File, 2006). A qualitative report observed that
blunt smokers believed blunts to be habit-forming due to the nicotine, and not the cannabis
content (Dunlap, Benoit, Sifaneck, & Johnson, 2006). Some researchers have proposed that
alleviating the nicotine withdrawal produced from mixing tobacco and cannabis perpetuates their
co-use (Burns, Ivers, Lindorff, & Clough, 2000; Van Beurden, Zask, Passey, & Kia, 2008).
Social settings, rituals, and normative behaviors associated with blunt smoking might also serve
to reinforce cannabis consumption (Dunlap et al., 2006, 2005, p. 2005; Kelly, 2005; Ream et al.,
2006). While Dunlap et al. (2006) observed that group norms discouraged excessive blunt use or
intoxication, it is not known whether blunt smoking promotes more frequent blunt or non-blunt
cannabis use. Golub, Johnson, and Dunlap (2005) contend that the rise in rates of cannabis use
during the 1990s was due to blunts. An epidemiologic study by Compton et al. (2004) reported

51

that the largest increases of cannabis use disorders during the 1990s occurred for AfricanAmericans, although the authors did not differentiate between cannabis and blunt use.
Only two studies have investigated the relationship between blunt smoking and cannabis
use disorders, or with symptoms of cannabis dependence. In one study, Ream et al. (2008) found
that blunt smoking and ‘chasing’ cannabis with tobacco was uniquely associated with five out of
seven cannabis problems: ‘spent more time using than intended’, ‘neglected usual
responsibilities’, ‘preoccupied with use’, tolerance, and ‘used to relieve negative affect’ as
compared to non-blunt smokers. No relationship was found for ‘wanted to cut down but could
not’ or ‘made psychological problems worse’. In a second study, Timberlake (2009) compared
past month blunt smokers to never blunt smokers and found the former has around twice the
odds of having cannabis use disorder in the past-year, adjusting for tobacco smoking and
background characteristics. Never blunt smokers were compared to groups of other cannabis
users based on how much past month level of blunt smoking exceeded past month level of
cannabis use. Therefore, blunt smokers with relatively little blunt smoking that nevertheless
exceeded cannabis use (e.g., two days of blunt smoking, but one day of cannabis use) may have
been grouped with heavy blunt smokers (e.g., used blunts every day). Not surprisingly, there was
little difference in the odds of CUDs among blunt using groups.

52

CHAPTER 3. MATERIALS AND METHODS
Chapter 3 details the materials and methods for each of the four studies comprising this
dissertation research project. Studies 1, 2, and 3 relied upon data gathered in the United States
(US) National Surveys on Drug Use and Health (NSDUH). Therefore, the first section of the
chapter (3.1) describes the study methodology these research projects have in common. Study 4
relied upon data gathered in the World Health Organization World Mental Health Survey
Initiative (WHO-WMH). Details of this study are covered in section 3.2. The final section of this
chapter (3.3) explains the analysis approach for each study individually.

3.1. The U.S. National Surveys on Drug Use and Health (NSDUH)
3.1.1. Study Background, Design, and Population
The NSDUH are a series of annual, cross-sectional surveys on the use of tobacco,
alcohol, cannabis, and other drugs, along with mental health correlates and health care utilization
(United States Department of Health and Humans Services, 2012). This series has been ongoing
since the 1970s, with annual surveys since 1990. Previously titled the National Household
Survey on Drug Abuse (NHSDA) and run by NIDA, the NSDUH is currently administered by
the Substance Abuse and Mental Health Services Administration (SAMHSA), which is a
department of the U.S. Department of Health and Human Services (DHHS). The NSDUH study
population now is designated to include the non-institutionalized, civilian population of the U.S.
aged 12 years and older. Approximately 65,000-69,000 persons were surveyed each year. Due to
confidentiality concerns, only a subset of the total sample surveyed was made publically
available by SAMHSA. This dissertation research used the publicly available data collected
between 2004 and 2011 (Table 3.1.1). Overall response levels between 2004 and 2011 ranged
53

from 65% to 70%, as determined by the joint probability of completing the screening interview
to roster eligible household participants, and completing the survey interview conditional on
being selected to participate.
Table 3.1.1. U.S. National Surveys on Drug Use and Health, Overall Samples Sizes and Response
Levels, 2004-2011.
Survey
Year

Sample
Surveyed

Public Use
Sample

Weighted
Screening
Response Rate

Weighted
Interview
Response Rate

Overall
Response Levela

2004
2005
2006
2007
2008
2009

67,760
68,308
67,802
67,870
68,736
68,700

55,602
55,905
55,279
55,435
55,739
55,772

91%
91%
91%
89%
89%
89%

77%
76%
74%
74%
74%
76%

70%
69%
67%
66%
66%
67%

2010
68,487
57,873
89%
75%
66%
2011
70,109
58,397
87%
74%
65%
a
Response level = Screening response rate x Interview response rate
Source: U.S. Department of Health and Human Services. U.S. National Surveys on Drug Use and Health,
Codebooks 2004-2011.

3.1.2. Sampling Approach
The sampling approach involved a multi-stage area probability sample design. The
NSDUH is a nationally-representative household sample of U.S. residents, with sampling frames
that included civilians living on military bases, residents of non-institutionalized group quarters
(e.g., college dormitories and group homes), and dwellings for persons without a permanent
residence (e.g., homeless shelters and hotel residents). Thus, the sampling frame excluded
members of the active-duty military and persons living in institutionalized group quarters (e.g.,
hospitals, prisons, nursing homes, and treatment centers), which represented less than two
percent of the population (United States Department of Health and Humans Services, 2012). The
primary sampling unit (PSU) consisted of geographical State sampling regions (SSRs), and was

54

based on U.S. Census tracts aggregated to include a minimum number of dwelling units (DU)
per SSR. Samples were drawn from each of the 50 U.S. States in addition to the District of
Colombia. A smaller secondary sampling unit (SSU), composed of adjacent census blocks, was
selected within the PSU in cases where the PSU greatly exceeded DU requirements. There was a
50% overlap in use of SSUs between successive survey years. Adolescents and young adults
were oversampled in order to collect approximately the same number of respondents across three
age groups: 12-17 years, 18-25 years, and 26 years or older.

3.1.3. Methodology, Confidentiality, and Human Subjects Protection
Trained field interviewers (FIs) conducted screening and survey interviews with
designated respondents (DRs) using computer assisted interviewing (CAI) methods. Screening
interviews with an adult resident identified all eligible persons within the DU (12 years or older),
including their age, sex, race/ethnicity, and military status. FIs read the study description and
received informed consent from both the screening respondent and DR(s). For DRs 12 to 17
years of age, FIs received verbal consent from a parent or guardian before approaching the minor
for informed consent. Survey interviews were conducted using a combination of computerassisted personal interviewing (CAPI) and audio computer-assisted self-interviewing (ACASI)
methods. CAPI methods collected data on non-sensitive topics, while ACASI methods collected
data on sensitive topics, such as drug use. This method benefited from increased confidentiality
of responses and an increased level of honest reporting. Respondents received $30 after the
interview in appreciation for their time.
All survey materials and procedures were approved by the cognizant institutional review
board (IRB) for the protection of human subjects. To protect respondent confidentiality,

55

identification information elements (e.g., name and addresses) were not linked to respondent’s
interview data. Census tract, state, and other geographical information were also removed from
the public use data. Post-interview statistical disclosure limitation methods identified records
where the confluence of variables such as sex, age, race/ethnicity, and other responses might
compromise confidentiality. In some cases, substitution of variables from a similar donor record
was used. Random sub-sampling, in which random records were removed, also decreased the
probability of identifying a record based upon responses. Values for some core demographic
information such as sex, age, race/ethnicity, and income were imputed either through logical
editing or statistical imputation. Unless noted otherwise in this dissertation, missing values on
other variables were due to bad data, respondent non-response (i.e., left blank, refused to answer,
or did not know the answer), or were not asked the question(s) due to logical skip patterns.

3.1.4. Reliability and Validity Studies
A reliability study using a subsample of respondents from the 2006 NSDUH sample was
conducted five to 15 days after the initial survey by SAMHSA (Substance Abuse and Mental
Health Services Administration, 2010). Cohen’s kappa was one statistic used to measure
consistency between initial and follow-up responses. Kappa values can be interpreted based on
recommendations by Landis and Koch (1977): poor agreement for kappa less than 0.00; slight
agreement for kappa of 0.00 to 0.20; fair agreement for kappa of 0.21 to 0.40; moderate
agreement for kappa of 0.41 to 0.60; substantial agreement for kappa of 0.61 to 0.80; and almost
perfect agreement for kappa of 0.81 to 1.00. Kappa estimates were 0.70 or better for lifetime
drug use variables, 0.47 to 0.85 for age of first use (agreement within 1 year), and 0.60 or better
for drug abuse and dependence measures among all respondents. Reliability for blunt smoking

56

measures (lifetime and past-year use) were 0.66 or greater. For major depressive episode (MDE)
measures among adults 18 or older, kappa values for lifetime, past-year, and age of first
depression were 0.67, 0.52, and 0.65, respectively. Corresponding values among adolescents 1217 years were 0.66, 0.72, and 0.41.
A NHSDA validity study was performed in 2000-2001 among past month drug users
(Harrison, Martin, Enev, & Harrington, 2007). Although not directly related to data used for this
study, results may nevertheless be informative. A maximum of one person per household was
selected and the study excluded residents of Alaska and Hawaii, Spanish-only speakers, and
respondents older than 25 years of age (n = 4,465; participation and biological specimen
response level: 74% and 89%, respectively). Urine and hair specimens were analyzed for the
following drugs or metabolites: cotinine, (for tobacco), marijuana metabolite (delta-9tetrahydrocannabinol carboxylic acid, carboxy-THC), cocaine metabolite (benzoylecgonine or
BZE), amphetamines (amphetamine and methamphetamine), and opioids (codeine and
morphine). Percent agreement between recent self-report measures and urinalysis showed strong
agreement for tobacco (89%), cannabis (90%), and cocaine (98%). Sample sizes and positive
urine tests were too small to draw conclusions for opiates and stimulants. Technical and
statistical issues prevented the use of hair specimens in the study.

3.1.5. General Strengths and Limitations of the NSDUH Studies
Use of NSDUH data had a number of strengths compared to recent and prior
epidemiologic studies of drug use behaviors. First, it is perhaps the largest survey of drug use
behaviors to date. By comparison, sample sizes for the Epidemiologic Catchment Area (ECA)
study, and the U.S. National Comorbidity Study – Replicate (NCS-R) had sample sizes of about

57

21,000 and 10,000, respectively. Large sample sizes helped produce statistically precise
estimates, especially for behaviors or conditions that would be normally rare in community
samples. Second, the NSDUH is ongoing and conducted annually. This allows researchers to
produce timely and relevant estimates of drug behavior, and replicate their results relatively
quickly using independent samples collected using nearly identical methods. Third, estimates
from the NSDUH are more representative of the drug-using population compared to other similar
ongoing surveys, such as the Monitoring the Future (MTF) and Youth Risk Behavioral
Surveillance (YRBS) surveys. For example, the MTF study annually surveys only U.S. students
in grades 8, 10, and 12, while the YRBS study samples U.S. students in grades 9-12 every other
year (Centers for Disease Control and Prevention, 2012; Johnston, O’Malley, Bachman, &
Schulenberg, 2012).
Despite these strengths, some limitations of the NSDUH must be considered. First,
responses to questions about drug use behaviors were based on self-report, which may not be
completely free of recall bias, reporting error, or dishonest response. The use of CAPI and
ACASI methods mitigated some of these limitations, and the reliability study mentioned above
attempted to quantify it. Second, these surveys were cross-sectional rather than longitudinal, and
therefore this research was limited in its ability to make causal inferences when the temporal
sequence between variables was ambiguous. Nevertheless, this research endeavored to take into
account temporality by using age-of-onset information when available. Third, as mentioned
above, these surveys do not target children (less than 12 years), active-duty military, and persons
in institutionalized group quarters (such as prisons), and therefore results may not generalize to
these populations.

58

3.1.6. Measurement of Key Study Variables
Several aspects of drug use behavior were measured for cannabis, blunts, tobacco
cigarettes, alcohol, and other internationally regulated drugs (IRD). Relevant to this dissertation
research were lifetime use, past-year use, past-month use, past-month and year frequency,
features of problematic use, age of first drug use, and age of last drug use. Frequency of drug use
in the past month or year was measured in relation to the number of drug-using days, rather than
quantity (e.g., number of drinks or cannabis joints smoked). Frequency of past-month tobacco
and blunt use was assessed, but not past-year frequency as with alcohol, cannabis, or other drugs.
Questions pertaining to the cannabis and blunt use were asked in separate points in the interview,
however, all questions about problematic cannabis use did not reference blunts. The variable
“blunt smoking recency” was constructed from responses to lifetime, past-year, and past-month
blunt use to have the following categories: a) never blunt smokers; b) used blunts at some point,
but not within the past 12 months (i.e., prior use); c) used blunts sometime within the past-year,
but not within the past month (i.e., past year); d) used blunts within the past month (i.e., past
month). Variables referencing early-onset drug use were based on self-report age of first drug
use and defined as having initiated the drug prior to age 18 years. Similarly, adult-onset was
defined as started to use the drug at age 18 or older. Years of cannabis involvement was defined
as the number of years from first to last use of cannabis.
Except for tobacco, items that measured problematic alcohol or cannabis use features
were based on dependence and abuse criteria in the Diagnostic and Statistical Manual of Mental
Disorders, Fourth Edition (DSM-IV; American Psychiatric Association, 1994). Questions were
drug-specific, covered the respondent’s past 12 month experience, and were asked only if the
respondent had used alcohol or cannabis on more than five days in the past year (see Table 3.1.2

59

for list of features). Participants who used less than this amount were imputed to not have
experienced the feature. Cannabis withdrawal was not measured, as it was not recognized as a
clinical feature under the DSM-IV. Responses to each feature were coded as dichotomous.
Table 3.1.2. Features of cannabis/alcohol problems (CP/AP) based on diagnostic criteria from
the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition.
CP
Variable

AP
Variable

CP1

AP1

CP2

AP2

CP3

AP3

CP4

AP4

CP5

AP5

CP6

AP6

CP7

AP7

CP8

AP8

CP9

AP9

CP10

AP10

NA

AP11

Description

Spent a great deal of time over a period of a month getting, using, or
getting over the effects of [cannabis/alcohol].
Used [cannabis/alcohol] more often than intended or was unable to keep
set limits.
Needed to use [cannabis/alcohol] more than before to get desired effects
or noticed that same amount of cannabis use had less effect than before.
Inability to cut down or stop using [cannabis/alcohol] every time tried or
wanted to.
Continued to use [cannabis/alcohol] even though it was causing problems
with emotions, nerves, mental health, or physical problems
Gave up or reduced involvement or participation in important activities
due to [cannabis/alcohol].
Serious problems at home, work, or school caused by using
[cannabis/alcohol].
Used [cannabis/alcohol] regularly and then did something that might have
put you in physical danger.
[Cannabis/Alcohol] caused you to do things that repeatedly got you in
trouble with the law.
Problems with family or friends caused by using [cannabis/alcohol] and
continued to use cannabis even though you thought using
[cannabis/alcohol] caused these problems.
Reported experiencing two or more alcohol withdrawal symptoms at the
same time that lasted longer than a day after alcohol use was cut back or
stopped.

NA, Not Applicable.

Tobacco dependence (TD) was measured in relation to tobacco cigarette smoking and
only among past month smokers. Table 3.1.3 shows TD assessed using 19 items from the
Nicotine Dependence Syndrome Scale (NDSS) and one item from the Fagerstrom Test of
60

Nicotine Dependence (FTND; Heatherton, Kozlowski, Frecker, & Fagerstrom, 1991; Shiffman,
Waters, & Hickcox, 2004). The NDSS is a scale designed to measure five aspects of dependence:
(1) Smoking drive; (2) Nicotine tolerance; (3) Continuous smoking; (4) Behavioral priority; (5)
Stereotypy. With one exception, responses were coded as 1 (not at all true), 2 (sometimes true), 3
(moderately true), 4 (very true), and 5 (extremely true). One item screened respondents about
whether they had friends who smoked (TD6; yes vs. no) before being asked about whether they
choose to be around friends who smoke (TD7). Responses from TD6 were incorporated into
TD7, such that respondents who said they did not have friends who smoked were given the
lowest value for TD7. Non-smokers or those who did not smoke in the past month were imputed
to have either the lowest value (0) or highest value (6) depending on whether higher or lower
values indicated more or less of the TD trait since some items were reverse coded. One item
from the FTND asked about the length of time elapsed from waking in the morning until the first
cigarette (within 5 minutes, 6-30 minutes, 31-60 minutes, and more than 60 minutes).

61

Table 3.1.3. Tobacco Dependence (TD) items based on Nicotine Dependence Syndrome Scale and
Fagerstrom Test of Nicotine Dependence.
Variable

Description

TD1

Need to smoke to feel less irritable

TD2

Start to crave cigarettes when don't smoke for few hours

TD3

Craving of cigarettes like strong force can't control

TD4

Feel a sense of control over your smoking

TD5

Tend to avoid places that don't allow smoking

TD6

Have friends who do not smoke cigarettes (yes/no)

TD7

Choose not to be around friends who don't smoke

TD8

Rather not travel by airplane because no smoking

TD9

Sometimes worry that you will run out of cigarettes

TD10

Smoke cigarettes fairly regularly throughout the day

TD11

Smoke same amount on weekends as on weekdays

TD12

Smoke same number of cigarettes from day to day

TD13

Number of cigarettes smoke per day often changes

TD14

Have many cigarettes in an hour, then no cigarettes until hours later

TD15

# of cigarettes smoke per day influenced by other things

TD16

Smoking not affected by other things

TD17

Amount of smoking has increased since started smoking

TD18

Need to smoke a lot more to be satisfied

TD19

Smoke much more now before feel anything

TD20

How soon after waking do you have your first cigarette?

Depression was measured via a modified version of the World Mental Health Composite
International Diagnostic Interview (WMH-CIDI), adapted from the NCS-R (Center for
Behavioral Health and Statistics and Quality, 2012). A lifetime occurrence of a major depressive
episode (MDE) was based on DSM-IV diagnostic criteria and required at least five or more
clinical features that occurred every day, or nearly every day over a two-week or more time
period, and where at least one feature was either (1) depressed mood, or (2) loss of interest or
pleasure in most activities (i.e., anhedonia). Additional features included (3) significant changes
in appetite or weight not accounted for by sickness, growth, pregnancy, or dieting, (4) insomnia
or hypersomnia, (5) psychomotor agitation or retardation, (6) fatigue or loss of energy, (7)
62

feelings of worthlessness, (8) difficulty concentrating, thinking, or making decisions, and (9)
recurrent thoughts of death or suicide (American Psychiatric Association, 1994). Respondents
were asked these nine clinical features in relation to a period of time in their life when their
mood was the worst, or if they could not recall the worst time, then their most recent episode. No
exclusions were made due to bereavement, drug-induced depression, or diagnostic hierarchies.
Therefore, there is a slight distinction between MDE defined within this research and that in the
DSM.
This dissertation makes a further distinction between a MDE as just described and a
depression spell. Those who qualified as a case of a MDE were also asked about when they first
experienced a sustained spell of depressed mood or anhedonia that occurred most of the day,
nearly every day for at least two-weeks or longer and had some of the same problems just
described. They were further asked whether such a depression spell occurred in the past 12
months. Therefore, depression outcomes referring to the first or past-year occurrence of
depression will be defined as depression spells, rather than a MDE. Using this information,
respondents could be grouped into three categories based on lifetime depression status: 1) Never
experienced a lifetime MDE; 2) “Prior depression” made up of those who had a lifetime MDE
experience, but did not have a depression spell in the past 12 months; 3) “Past-year depression
spell” consisting of those who experienced depression in the past 12 months.
Functional impairment attributed to depression was measured using the Sheehan
Disability Scale (SDS), a short measure of global functional impairment related to mental health
(Leon, Olfson, Portera, Farber, & Sheehan, 1997; Sheehan, Harnett-Sheehan, & Raj, 1996). The
SDS was developed to measure impairment across three inter-related life domains: work/school,
family/home life, and social life. The NSDUH used a modified version of the SDS, measuring

63

four domains for adults: (1) home management; (2) work; (3) close relationships; (4) social life.
Due to difference in life roles, adolescents were asked about their degree of impairment with (1)
chores, (2) school, (3) family, and (4) social life. In the analysis, categories of home management
and chores were considered the same category when pooling data from adolescents and adults, as
was work and school. SDS items were measured in relation to the time in the past 12 months
when their mood was most severe, and rate the degree to which their depressed mood interfered
in each life domain. Responses were measured via an analog scale from 0-10: 0 (no impairment),
1-3 (mild), 4-6 (moderate), 7-9 (severe), and 10 (very severe). For those who answered one or
more SDS item of this series, an additional item asked about the number of days in the past year
they were unable to carry out daily activities due to depressed mood, (range 0-365).
Conceptual models considered the influence of time-invariant background covariates
such as sex, age, and race or ethnicity. Race or ethnic categories were defined with regards to
standard U.S. census groups: White, Black or African American, Native American or Alaskan
Native, Pacific Islander or Native Hawaiian, Asian, Hispanic, or more than one race. In most
cases it was not possible to produce statistically precise estimates for ethnic/racial subgroups
with relatively small sample sizes (e.g., Native American or Alaskan Native), nor was producing
such estimates an aim of this dissertation. Thus, the main categories of White, Black/AfricanAmerican, and Hispanic were retained, while all other groups were categorized as “Other”.
Additional background characteristics of interest were total family income (<$20K, $20K$49,999, $50K-$74,999, and ≥$75K), and either county metropolitan type (large, small, nonmetro) or population density. Population density was measured in relation to U.S. Census corebased statistical areas (CBSA).

64

3.1.7. Missing Data and Imputation
Respondents were excluded from analysis when missing on the outcome or variables
used to define the analytical sample. Additional specific exclusion criteria were detailed in the
analysis plan for each study below. Missing data on exposure variables were either logically
imputed, dropped from analysis, or carried along in analyses under a ‘missing’ category as
indicated in the description of the analyses. The public use data files provided by SAMHDA
imputed missing data for core background variables (e.g., sex and race), income, lifetime drug
use, and age-of-onset. These missing values that could not be logically imputed were imputed
using a procedure called predictive mean neighborhood (PMN) imputation, which was a
combination of model-assisted and nearest neighbor hot-deck imputation methods (United States
Department of Health and Humans Services, 2012).

3.2. The WHO World Mental Health Surveys Initiative
3.2.1. Study Design and Sampling Procedures
Data for Study 4 come from the World Health Organization World Mental Health
Surveys Initiative (WHO-WMHSI), the largest collaboration of coordinated, cross-national
epidemiologic surveys of mental health disorders and treatment to date. Fourteen countries were
included in this study, which collected data on cannabis use and the opportunity to use drugs
(Table 3.2.1). Data from the Ukraine and South Africa could not be included because questions
about drug opportunity were not asked. Additionally, data collected as part of the European
Study Of The Epidemiology Of Mental Disorders (ESEMeD) had to be excluded due to a
programming error that thwarted collection of data on problematic drug use features. Most
studies were designed to include persons 18 years or older in the study population, with

65

exception of New Zealand (age 16+), Japan (20+), and Israel (21+). The sampling approach
involved a multi-stage area probability sampling design to produce nationally or regionally
representative samples. The first stage of sampling consisted of units analogous to counties or
municipalities. The second stage sampling procedure targeted neighborhoods, towns, or other
equivalent geographical clustering. Response levels varied from 51% (Japan) to 95% (Iraq). To
reduce respondent burden, some sites administered core survey components to the entire sample
(Part 1 sample), and then selected a subsample for interview of non-core components (Part 2
sample). Selection into the Part 2 sample was determined by screening criteria for any mental
disorder and further enriched with a random 25% of remaining respondents. Only respondents
interviewed about drug use provided information for this dissertation study. Table 3.2.1 reports
the resulting sample size, and which part of the sample was administered the drug use interview
module (Part 1 or Part 2).

66

Table 3.2.1. Sample Sizes and Study Characteristics of the WHO World Mental Health Surveys.

Survey

Survey
Years

Response
Level %

Ages

Sample
Size

Sample
Part Used

Americas
Brazil

SPM

2004-6

81

18+

5,037

Part 1

Colombia

NSMH

2003

88

18-65

4,426

Part 1

Mexico

M-NCS

2001-2

77

18-65

5,782

Part 1

United States

NCS-R

2002-3

71

18+

5,692

Part 2

Bulgaria

NSHS

2003-6

72

18+

2,233

Part 2

Northern Ireland

NIMHS

2004-8

68

18+

1,986

Part 2

Romania

RMHS

2005-6

71

18+

2,357

Part 1

Iraq

IMHS

2006-7

95

18+

4,332

Part 1

Israel

NHS

2003-4

73

21+

4,859

Part 1

LEBANON

2002-3

70

18+

1,031

Part 2

NSMHW

2002-4

79

18+

2,143

Part 2

Japan

WMHJ

2002-3

51

20+

1,305

Part 2

China

B-WMH, SWMH

2002-3

75

18-70

1,628

Part 2

Region/Country

a

Europe

Africa and Middle East

Lebanon
Nigeria
Western Pacific

New Zealand
NZMHS
2003-4
73
16+
12,992
Part 1
a
SPM (Sao Paulo Megacity); NSMH (The Colombian National Study of Mental Health); M-NCS (The
Mexico National Comorbidity Survey); NCS-R (The U.S. National Comorbidity Survey Replication); NSHS
(Bulgaria National Survey of Health and Stress); NIMHS (Northern Ireland Mental Health Survey); RMHS
(Romania Mental health Survey); IMHS (Iraq Mental Health Survey); NHS (Israel National Health Survey);
LEBANON (Lebanese Evaluation of the Burden of Ailments and Needs Of the Nation); NSMHW (The
Nigerian Survey of Mental Health and Wellbeing); WMHJ 2002-2003 (World Mental Health Japan
Survey); B-WMH (The Beijing World Mental Health survey); S-WMH (The Shanghai World Mental Health
Survey); NSMHWB (National Survey of Mental Health and Wellbeing); NZMHS (New Zealand Mental
Health Survey)

3.2.2. Data Collection and Measurement of Psychiatric Disorders
Trained interviewers conducted face-to-face assessments using the World Mental Health
Composite International Diagnostic Interview (WMH-CIDI), a fully structured diagnostic
interview based on the WHO-CIDI (Kessler & Ustun, 2008). Consistent interviewer training,
study materials, quality control, and standardized translation procedures were used across
67

surveys to minimize between site variations. Diagnostic measurements of mental, drug, and
behavioral disorders were based on definitions and criteria from the Diagnostic and Statistical
Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association,
1994) and the International Classification of Mental and Behavioral Disorders (ICD; World
Health Organization, 1993).

3.2.3. Protection of Human Subjects
Each participating country had their study protocols reviewed and approved by either a
human subjects review board or ethics committee for the protection of human subjects (Kessler
& Ustun, 2008). At all WMH sites, interviewers were required to read a statement of voluntary
consent. However, sites varied in the method used to obtain consent depending on the literacy,
culture, traditions, and norms of the country. Some countries required written consent of the
participant, while other countries allowed for oral consent documented by the interviewer.

3.2.4. Measurement of Key Study Variables
Respondents was asked about lifetime history of use and age at which they first used the
following drug sub-types: 1) alcohol; 2) cannabis; 3) cocaine; 4) extra-medical prescription drugs
(e.g., analgesics, stimulants, and sedatives); 5) a catch-all category of ‘other’ drugs including
heroin, LSD, ecstasy, etc. Tobacco cigarette smoking was measured in relation to current status
at time of interview with the following categories: 1) Never smoked or smoked ‘a few times’; 2)
Ex-smoker; 3) Current smoker. Respondents were also asked separately about their first
opportunity to use alcohol or drugs:

68

The next questions are about the first time you had an opportunity to use
[alcohol/drugs], whether or not you used them. By “an opportunity to use” I
mean someone either offered you [alcohol/drugs], or you were present when
others were using and you could have used if you wanted to. Please do not
include times when a health care provider may have offered you free samples.
Thinking back over your entire lifetime, about how old were you the very first
time you had an opportunity to use [alcohol/drugs]?
The variable time to cannabis onset (TCO) was constructed by subtracting the value for age of
first cannabis use from age of first drug opportunity.
Lifetime history of cannabis or alcohol dependence (CD/AD) were diagnosed (present vs.
not present) based on the self-report experiences of clinical features from the DSM-IV and ICD10 (see Table 2.4.1 from Chapter 2). Questions about AD features were alcohol-specific, but
questions concerning dependence features for cannabis were not drug-specific per se, so this
research restricted the analytical sample to respondents who had only used cannabis and no other
IRD in their lifetime (i.e., ‘cannabis-only’). Therefore, the experience of dependence features
could be reasonably attributed to their cannabis use. In order to be asked questions about AD
features, respondents had to have consumed more than once per month during the period of their
life when they drank the most. For a few WMH survey sites, respondents first had to indicate the
presence of at least one maladaptive drug problem in their lifetime prior to being asked questions
about dependence (see Table 2.4.2 from Chapter 2). All problematic cannabis or alcohol drug
features were coded as binary response variables (present vs. not present), and are listed with
labels for each variable below (Table 3.2.2).

69

This research considered a number of covariates that might influence both time to
cannabis onset and the experience of cannabis dependence or problems. These included sex, age,
income (low, low-average, and above average), education (less than secondary vs. secondary
finished), employment (working vs. not working), marital status (never married, previously
married, and married/cohabiting), parent’s highest level of education (primary or less, secondary,
tertiary), and cumulative incidence of a major depressive episode up to the date of assessment
(MDE).

Table 3.2.2. Features of cannabis/alcohol problems based on criteria from the Diagnostic and Statistical
Manual of Mental Disorders, 4th Edition and the WHO International Classification of Diseases, 10th
Revision.
Cannabis

Alcohol

Description

Variables
CP0

AP0

Strong desire to use

CP1

AP1

Spent a great deal of time getting or using

CP2

AP2

Used more often than intended

CP3

AP3

Experienced tolerance

CP4

AP4

Had difficulty cutting down when wanted to

CP5

AP5

Continued to use despite emotional or physical problems

CP6

AP6

Gave up important activities due to use

CP7

AP7

Serious problems home, work, or school

CP8

AP8

Hazardous use

CP9

AP9

Legal problems

CP10

AP10

Continued despite problems with family or friends

CP11

AP11

Experienced withdrawal

70

3.3. Analysis Plan
3.3.1. Study 1: Early-Onset Cannabis Smoking and Depression Spells.
AIM 1. Estimate the degree to which early-onset (onset ≤ 17 years) and adult-onset (18+ years)
cannabis smoking are associated with a later first onset of a depression spell in adulthood, within
a conceptual model that accounts for time-invariant characteristics of sex, age, and
race/ethnicity, and variation in tobacco and alcohol use.

Data for Study 1 came from the 2005-2009 NSDUH (n=278,130). The following groups
were excluded from analysis: 1) Respondents 12-17 years who were not observed for the entire
period of risk for starting to use cannabis (n=90,266); 2) Respondents whose first depression
spell occurred before adulthood (n=11,229); 3) Invalid or otherwise missing data on depression
status or age of first onset (n=2,265); 4) Cases whose cannabis onset occurred after their
depression spell onset, in violation of an assumption that cannabis smoking must precede
depression onset to be considered a viable causal factor (n=595). Thus, the analytical sample
consisted of 173,775 adults. Early-onset cannabis use was defined as a dichotomous variable,
which compared respondents whose age of cannabis onset was equal or less than 17, with those
whose age of cannabis onset was later, or who had not used cannabis by the time of interview.
As described elsewhere (Anthony, 2004; http://www.epi.msu.edu/janthony/, last accessed
28 April 2014), the plan for data analysis was organized in relation to a now-standard "explore,
analyze, explore" cycle. The first cycle involved exploratory steps such as histogram plots and
other exploratory data analyses to shed light on the underlying distributions of the response
variable and each covariate of interest. In this work, study estimates' precision is emphasized; the
focus is on 95% confidence intervals. P-values are presented as an aid to interpretation. In the

71

initial analysis step, the author performed a series of logistic regressions, with occurrence of
depression spell expressed as a function of cannabis onset timing, elapsed time of cannabis
involvement, tobacco cigarette onset, and alcohol onset to produce unadjusted odds ratio (OR)
estimates. All analyses took into account sampling weights and the complex sample structure for
variance estimation purposes. In a subsequent analytical step, the statistical approach involved
constructing a multivariable model that included all previously listed covariates plus adjustment
for sex, age, race/ethnicity, and survey year. Our final exploratory analytical step consisted of
probing the degree to which the early-onset cannabis estimate varied when introducing other
covariates of interest into the model, but whose inclusion might otherwise violate the causal
assumptions of the model. The first approach, called Add-One-In (AOI) analysis, consisted of
estimating the EOCU-depression relationship with one additional covariate added to the previous
multivariable model, removing the added covariate, and then re-estimating using a different
covariate. This process was repeated for each covariate of interest. The second approach
evaluated a statistical model that included all covariates together, and then used a backwards
elimination procedure to reduce the number of covariates based upon having an overall p-value
greater than 0.2. This backwards elimination method of model building has been recommended
by others (Hosmer & Lemeshow, 2000; Royston & Sauerbrei, 2008). A leave-one-out (LOO)
approach was used where by one covariate was left out of the multivariable model in order to see
how sensitive the overall estimates were to the influence of any one particular covariate.
Potential multicollinearity was evaluated using Spearman rank order correlations.

72

3.3.2. Study 2: Cannabis Problems and Functional Impairment Attributed to Recently Active
Depression Spells.
AIM 2. Estimate the degree to which the level of cannabis problems is associated with the
functional impairment attributed to depression in recently active cases, while accounting for
differences in background characteristics, tobacco and alcohol use, and other relevant covariates.

Study 2 is based on NSDUH data collected from 2009-2011 (n=172,042). The analytical
sample was restricted to respondents 12 years and older who experienced a depression spell
within the past 12 months (n=13,743). Contingency tables were used to examine the distribution
of the total sample and lifetime depression status (no depression, prior depression, and recently
active depression spell) by relevant background characteristics (sex, age, race/ethnicity, income,
and county metro). Next, SDS items were summarized using descriptive statistics and histogram
plots were used to characterize their distribution.
Factor analysis was used to fit a hypothesized unidimensional latent trait of functional
impairment (IMPAIR) using the five SDS items (Figure 3.1.1). In these figures, latent factors
were represented by circles or ovals, and observed variables were represented by boxes.
Hypothesized causal relationships between factors and/or observers variables are represented by
directed arrows. Curved double-sided arrows indicated correlations. Prior literature indicated a
single latent factor for the SDS, so no exploratory factor analysis (EFA) was performed
(Arbuckle et al., 2009; Leon et al., 1997; Leon, Shear, Portera, & Klerman, 1992). A latent
measurement model of two or more factors would likely have been under-identified given only
five items in the scale. Factor analysis is based on the following equation:

73

, which is summarized by

Where x is a p x 1 vector of p observed or manifest variables, Λ = [Λij] is a p x m matrix of
factor loadings , f is the m x 1 vector of factors, and ε is the p x 1 vector of error terms assumed
to be uncorrelated among themselves and with factors in f. Factor analysis was similarly used to
summarize hypothesized unidimensional latent traits for cannabis problems (CP), alcohol
problems (AP), and tobacco dependence (TD) (see Figures 3.3.2, 3.3.3, and 3.3.4, respectively).
EFA was not conducted on CP items for this study, but it was performed in Study 3.
Comparative fit index (CFI), Tucker-Lewis Index (TLI), and root mean error of approximation
(RMSEA) values were used to assess model fit. In this research, values for CFI and TFI ≥ 0.95,
and RMSEA ≤ 0.06 were indicative of an excellent fit to the data (Browne, Cudeck, Bollen, &
Long, 1993; Hu & Bentler, 1999). Modification indices (MI) were used to explore how the
measurement models could be improved in case model fit was not adequate. MIs are an
approximate measure of improvement (drop) in the chi-square value that would be gained if
certain model parameters were not assumed to be constrained (Raykov & Marcoulides, 2008).
Model re-specification would push analysis into exploratory mode, and the decision to change
parameter constraints were based on both MIs and plausible substantive theory.
The association between latent traits of cannabis problems and impairment was estimated
in a series of structural equation models (SEMs). The initial SEM estimated the unadjusted
association between CP and IMPAIR. Subsequent SEMs added paths for hypothesized direct and
indirect effects for sex, age, and race/ethnicity variables. Finally, latent traits for AP and TD
were added (Figure 3.3.5). Post-estimation exploratory analyses compared estimates under
74

different measurement model assumptions for impairment, and then tested whether conclusions
might differ after excluding respondents whose cannabis onset did not preceded onset of the first
depression spell (i.e., using age-of-onset data to temporally constrain the hypothesized causal
sequence). All models accounted for sample probability weights and the complex survey design.
Initial analyses were done using Stata 12, but factor analyses and structural equation models
were estimated using Mplus version 7 (Muthén & Muthén, 2012; StataCorp, 2011).

σ2

λ1

SDS1

ε1

SDS2

ε2

SDS3

ε3

SDS4

ε4

SDS5

ε5

λ2
λ3

IMPAIR

λ4
λ5

Figure 3.3.1. Hypothesized latent structure of functional impairment attributed to
depression (IMPAIR).

75

CP1

CP2

CP3

CP4

CP5
CP
CP6

CP7

CP8

CP9

CP10
Figure 3.3.2. Hypothesized latent structure of cannabis problems (CP).

76

AP1

AP2

AP3

AP4

AP5

AP6

AP

AP7

AP8

AP9

AP10

AP11
Figure 3.3.3. Hypothesized latent structure of alcohol problems (AP).

77

TD1

TD2

TD3

TD4

TD5

TD7

TD8

TD9

TD10
TD
TD11

TD12

TD13

TD14

TD20

TD19

TD18

TD17

TD16

TD15

Figure 3.3.4. Hypothesized latent structure of tobacco dependence (TD).

78

Female
ε

AP
18-25

ε
26-34

IMPAIR

35+
ε
Black
CP
Hispanic

Other

TD

ε

Figure 3.3.5. Model of relationships between functional impairment attributed to depression (IMPAIR) and cannabis
problems (CP), alcohol problems (AP), tobacco dependence (TD), and background characteristics.

79

3.3.3. Study 3: Cannabis Problems and History of Blunt Smoking.
AIM 3. Estimate the degree to which a history cannabis ‘blunt’ smoking is associated with the
level of cannabis problems experienced, within a conceptual model adjusting for the influence of
sex, age, race/ethnicity, and other potential time-invariant covariates.

Data for Study 3 came from the NSDUH collected from 2004 to 2011 (n=450,002).
Sample characteristics (i.e., sex, age, race/ethnicity, and population density) were compared
between the total NSDUH sample and the analytical sample (n=77,047). The analytical sample
consisted of respondents 12 years or older, had used cannabis in the past 12 months from the date
of interview, and were assessed for problematic cannabis use. In the first phase of analysis, the
weighted population prevalence of blunt smoking (i.e., ever, past-year, and past-month use) was
estimated across survey years in order to characterize both the extent of blunt use in the
population and its potential variation over time. Next, the prevalence of individual cannabis
problems items were compared between non-blunt users and those who had used in their lifetime
and within the past year.
In the second phase, exploratory factor analysis (EFA) was conducted using the ten
features of cannabis problems (CP) from a random half-split sample of 2004 NSDUH data
(n=3,534). Analysis used a weighted least squares mean variance (WLSMV) estimator and
geomin oblique rotation. A scree plot of eigenvalues (EV) helped to evaluate the number of
factors present. Confirmatory factor analysis (CFA) then was used to compare a single latent
factor model of CP items versus a hypothesized model where CP1-CP6 items loaded on a
cannabis dependence factor (DEPEND) and CP7-CP10 items loaded on a factor of socially
maladaptive or hazardous cannabis use factor (HARM) (see Figures 3.3.2 above for the single
80

factor model, and Figure 3.3.6 below for the two-factor model). The second half-split sample of
2004 NSDUH data was used for this CFA analysis (n=3,520). Results from the EFA and CFA
were used to inform the factor structure of cannabis problems used for later analysis. CFA was
also conducted on alcohol problems (AP) and tobacco dependence (TD), with the assumption of
a single underlying factor for each as depicted in Study 2 (see Figures 3.3.3 and 3.3.4,
respectively).
In the third analysis phase, the association between level of cannabis problems and
recency of blunt smoking was estimated by regressing CP factor scores on dummy coded
indicators of blunt smoking recency, with never blunt smokers as the reference group (Figure
3.3.7). Subsequent SEMs first added observed covariates for sex, age, race/ethnicity, and
population density, and then finally included latent factors for alcohol problems (AP) and
tobacco dependence (TD), allowing for factor correlations between the three latent factors
(Figure 3.6.8). Exploratory analysis investigated whether frequency of past-month blunt smoking
was associated with CP in a model similar to the one depicted in Figure 3.6.7. All analyses took
into account sample probability weights and the complex survey design. Initial analyses were
carried out using Stata 12, but all EFA, CFA, and SEM analyses were conducted in Mplus
version 7 (Muthén & Muthén, 2012; StataCorp, 2011).

81

CP1

CP2

CP3
DEPEND
CP4

CP5

CP6

CP7

CP8

HARM

CP9

CP10
Figure 3.3.6. Alternative hypothesized two-factor latent structure of cannabis
problems (CP), with one factor of dependence (DEPEND) and a second factor of
harmful socially maladaptive or hazardous use (HARM).

82

CP1
λ1

CP2

λ2

CP3

λ3
CP4
λ4

Prior Use

ε
β1

Past Year

CP5
λ5

β2

CP

λ6
λ7

β3

CP6

CP7

λ8

Past Month

CP8
λ9
CP9
λ10
CP10
Figure 3.3.7. SEM of association between cannabis problems (CP) and blunt
smoking recency.

83

Prior Use
β1
Past Year
Blunt
PastMonth
Blunt

β2
AP
β3
ε

Male

β4
ε

18-25

β5

26-34

β6

35+

β7

Black

CP

β8
ε

β9
Hispanic

β10

Other

TD
β11

CBSA
<1 mil

β12

No
CBSA
Figure 3.3.8. SEM of the association between cannabis problems (CP) and blunt
smoking recency with statistical control for background characteristics and factor
correlation with alcohol problems (AP) and tobacco dependence (TD).

84

3.3.4. Study 4: Time to Cannabis Onset and Cannabis Problems.
AIM 4. Estimate the degree to which delaying cannabis onset after the first chance to try drugs
might account for later level of cannabis problems among a cross-national sample of adults
whose only internationally regulated drug (IRD) use was cannabis.

Study 4 focused on a subsample of respondents whose only internationally regulated drug
use (IRD) was cannabis (n=6,609), meaning they had not used cocaine, prescription drugs extramedically, or drugs defined by the ‘other’ catch-all category (i.e., heroin, ecstasy, inhalants, etc.).
Some WMH sites contributed very few cannabis-only users to the analytical sample (see Table
3.3.1), such Iraq (n=1), China (n=3), and Japan (n=6). Prevalence of cannabis-only use was
highest in New Zealand and the United States (29% and 22%, respectively).
Table 3.3.1. Frequency and weighted prevalence of cannabis-only use by contributing World Mental
Health country.
Country

Total

No Use/ Other

Cannabis-Only

Prev. %

Brazil

4,994

4,745

249

5.0

Colombia

4,424

4,164

260

5.9

Mexico

5,766

5,554

212

3.7

United States

5,707

4,329

1,378

24.1

Bulgaria
Northern Ireland

2,211
1,980

2,197
1,770

14
210

0.6
10.6

Romania

2,357

2,341

16

0.7

Iraq

4,288

4,287

1

<0.1

Israel

4,816

4,367

449

9.3

Lebanon

1,032

992

40

3.9

Nigeria

2,132

2,098

34

1.6

Japan

1,293

1,287

6

0.5

China

1,618

1,615

3

0.2

12,988

9,251

3,737

28.8

55,606

48,997

6,609

11.9

New Zealand
Total

a

a

Respondents with missing or bad data on cannabis smoking were excluded (n=68).

85

Initial analyses explored the distribution of cannabis dependence (CD) by WMH country,
cigarette smoking status, alcohol dependence, lifetime MDE, and background characteristics
(i.e., sex, age, income, education, employment, marital status, and parent’s highest level of
education) using contingency tables. Weighted prevalence of CD was estimated across categories
of these relevant covariates. Histogram plots were used to explore the distribution of discrete
continuous variables, such as age of first IRD opportunity, age of cannabis onset, and time to
cannabis onset (TCO). Mean values on these items were compared across CD case status. Time
to cannabis onset logically should have a lower bound of zero, which would indicate opportunity
and use occurred at the same age. However, there were instances of negative values on TCO, and
these respondents were exclude from analyses (n=185; representing 2.8% of the sample).
The next step involved series of logistic regression models, in which the task was to
estimate the association between CD and TCO. Since CD was coded as a binary outcome, a
generalized linear model with a logit link function was used to estimate the change in the log
odds of becoming a case of CD for each unit increase in TCO. The following equation depicts
the logistic regression formula:

Where p is the probability of the outcome occurring, β0 is the baseline log odds of the outcome
occurring, and βi is the log odds of outcome occurrence for every unit change in xi as compared
to xi = 0. The initial logistic regression model included as covariates only TCO and age of first
IRD opportunity. Subsequent models first included statistical control for WMH country, sex, and

86

age, and then for additional background characteristics, cigarette smoking status, and lifetime
MDE.
In the post-estimation exploratory analyses, a single latent trait measurement of cannabis
problems (CP) was modeled using features of cannabis dependence and maladaptive cannabis
use items (Figure 3.3.9). A series of SEMs regressed the factor scores of CP on TCO, age of first
IRD, age, sex, indicator for WMH site, current cigarette smoking status, a latent trait of alcohol
problems (AP), income, education, marital status, and parent’s highest level of education. All
analyses accounted for sample probability weights and the complex survey design. Logistic
regression analyses were done using Stata 12 (StataCorp, 2011). Factor analyses and SEMs were
conducted in Mplus version 7 (Muthén & Muthén, 2012).

87

CP0
CP1
CP2
CP3
CP4
CP5
CP

CP6
CP7
CP8
CP9
CP10
CP11

Figure 3.3.9. Latent measurement model of cannabis problems (CP) in WMH.

88

CHAPTER 4. RESULTS
The initial sections of the results chapter (4.1 and 4.2) build upon the earliest dissertation
project work concerning the link between cannabis smoking and depression spells (as discussed
in Chapter 2). Afterwards, I initiated a dissertation study intended to replicate and expand upon
these findings; for this study, I turned to pertinent data from the U.S. National Surveys on Drug
Use and Health (NSDUH). The fruits of that research I judge to be a reflection of an example of
investigator-initiated research, completed during the course of my doctoral studies, involving no
more than minimal guidance from my primary dissertation committee advisor, J.C. Anthony
(JCA), whose involvement primarily was advisor-appropriate polishing of concepts, approach,
and manuscript copy-editing. This research was subsequently published in a peer-reviewed
journal (see Fairman & Anthony, 2012), but was ‘in press’ at the time the committee members
approved the proposal for this dissertation. With encouragement from JCA, these results have
been included here in section 4.1 of the dissertation. It is my judgment that this prior research
should be woven into the fabric of my dissertation research. Nonetheless, it has not been counted
toward the 'three paper threshold' customarily set for dissertation research programs in our
department. Instead, it qualifies as a fourth separate study as part of the overall dissertation
program of research. This study certainly contributed toward my development as an independent
investigator in epidemiology. Nevertheless, the dissertation committee was asked to judge the
merits of the dissertation’s evidence and significance irrespective of this initial Fairman-Anthony
contribution, based on the remaining three results sections (4.2-4.4).

89

4.1. Study 1: Early-Onset Cannabis Smoking and Depression Spells.
Table 4.1.1 describes the study sample in terms of selected background characteristics
cross-classified by depression spell case and non-case status. About 1 in 10 adults in the sample
had experienced a depression spell at age 18 or in subsequent adult years. Cases were more
likely to be female, white, and older than 25 years of age, have a post-high school education, and
divorced or separated.
From Table 4.1.2, the proportion of cases who had smoked cannabis before age 18 (14%)
was identical to those who had started later (14%). Depression spells were over-represented
among cannabis users with the longest elapsed time of cannabis involvement (11 years or longer,
19%). Varying proportions of depression spells occurred among early-onset users of tobacco
cigarettes (12%) and early onset users of alcohol (12%).
Results from the logistic regression analysis prior to covariate adjustment showed a twofold excess odds of a later depression spell for early-onset cannabis smokers as compared to
never cannabis smokers (estimated odds ratio, OR = 2.2; Table 4.1.2, Unadjusted Model) . While
the estimate for adult-onset cannabis smokers was nearly the same (OR = 2.0), a statistical test of
differences revealed these two estimates to be distinct (p=0.029; not shown in table). When the
estimates were adjusted for elapsed time of cannabis involvement, tobacco onset, alcohol onset,
sex, age, race/ethnicity, and survey year, there was a attenuation towards null (Early-onset: OR =
1.7; Adult-onset: OR = 1.8), but both estimates remained statistically significant at p<0.001. In
contrast to the estimates from models without covariate adjustments, there was no statistically
significant difference between the odds ratio estimates for early-onset and adult-onset cannabis
users (p=0.286; data not shown in table).

90

Table 4.1.1. Sample Characteristics of Depression Spell Cases (n=173,775)
Depression spell ≥ 18
years
Characteristic

Never experienced a depression
spell

n

% (wt)

n

% (wt)

16,108

10%

157,667

90%

18-25
26-34
35-49
50 or older

5,654
2,892
4,978
2,584

7%
11%
12%
10%

77,988
23,133
32,806
23,740

93%
89%
88%
90%

Male

5,536

8%

77,190

92%

Female
Race/Ethnicity

10,572

13%

80,477

87%

White
Black

11,302
1,643

12%
8%

100,370
19,841

88%
92%

Hispanic
Asian

1,995
390

8%
5%

24,731
6,035

92%
95%

778

11%

6,690

89%

2,038
4,736
5,315
4,019

8%
9%
12%
12%

28,465
52,954
44,281
31,967

92%
91%
88%
88%

8,505

10%

83,690

90%

2,773
1,042

11%
12%

28,867
10,267

89%
88%

3,788

10%

34,843

90%

6,252
499
2,783

9%
10%
17%

60,275
3,937
12,863

91%
90%
83%

Total
Age Group

Sex

Othera
Highest Level of Education
Less than High School
High School Graduate
Some College
College Graduate
Employment Status
Full-time
Part-time
Unemployed
Otherb
Marital Status
Married
Widowed
Divorced/Separated

Never married
6,574
9%
80,592
91%
Data: US National Surveys on Drug Use and Health, 2005-2009
Weighted proportion (wt)
a
Included Native Americans, Alaskan Natives, Native Hawaiians, Other Pacific Islanders, and more than
one race/ethnicity.
b
Included, but not in the labor force.

91

Table 4.1.2. Prevalence of Depression Spell Cases and Association with Cannabis, Cigarette, and Alcohol Onset, and Years of Cannabis
Involvement.
Depression Spell
Cases

a

Unadjusted Models

Drug Exposure

Total

n

Prev. %

OR

Cannabis Onset
Never Used
Early-Onset
Adult-Onset
Missing
Years of Cannabis Involvement

90,298
54,891
28,089
497

6,185
6,454
3,439
30

7
14
14
7

1.0
2.2
2.0
1.0

Never Used
< 1 year
1-10 years
11 or more
Missing
Cigarette Onset

90,298
14,688
35,963
11,317
21,509

6,185
1,502
4,217
2,153
2,051

7
12
13
19
12

1.0
1.7
2.0
2.9
1.7

Never Used
Early-Onset
Adult-Onset
Missing

54,945
88,770
29,190
870

3,510
9,697
2,848
53

7
12
10
6

1.0
1.9
1.5
0.9

(95% CI)

(2.0,2.3)
(1.9,2.1)
(0.7,1.5)

(1.6,1.9)
(1.9,2.1)
(2.7,3.2)
(1.6,1.9)

(1.8,2.1)
(1.3,1.6)
(0.6,1.4)

Adjusted Model

P-value

OR

(95% CI)

P-value

<0.001
<0.001
0.986

1.0
1.7
1.8
1.1

(1.5,1.9)
(1.6,1.9)
(0.7,1.7)

<0.001
<0.001
0.745

<0.001
<0.001
<0.001
<0.001

1.0
0.8
1.0
1.4
NA

(0.7,0.9)
(0.9,1.1)
(1.2,1.6)

0.002
0.498
<0.001

<0.001
<0.001
0.689

1.0
1.2
1.1
1.0

(1.1,1.3)
(1.0,1.2)
(0.6,1.5)

<0.001
0.104
0.942

(1.7,2.3)
(1.5,1.9)
(0.7,1.5)

<0.001
<0.001
0.737

Alcohol Onset
22,123
945
5
1.0
1.0
Never
96,531
10,300
12
2.9
(2.6,3.3)
<0.001
2.0
Early-Onset
53,733
4,790
9
2.1
(1.8,2.3)
<0.001
1.7
Adult-Onset
1,388
73
6
1.2
(0.9,
1.8)
0.226
1.1
Missing
Prevalence (Prev), Odds ratio (OR), Confidence interval (CI)
a
Univariable logistic regressions between depression spells and each drug exposure.
b
Multivariable logistic regression model included all variables in the table plus sex, age, race, and survey year.

92

b

From Table 4.1.2, one can also see estimates for the unadjusted and covariate-adjusted
depression spell-related odds relative to tobacco cigarette and alcohol onset, and it should be
noted that EOCU is the acronym that refers to early-onset cannabis use, as defined above. Adultonset tobacco cigarette smoking was not associated with depression spells after adjustment for
covariates in the model (unadjusted OR = 1.5; p<0.001; adjusted OR = 1.1; p=0.104).
Nevertheless, there was a very small excess odds that linked adult-onset occurrence of a
depression spell with early-onset tobacco smoking (adjusted OR = 1.2; p<0.001), independent of
the EOCU-depression association. Both early- and adult-onset alcohol drinking were noteworthy
covariates in the covariate-adjusted model (OR = 2.0 and 1.7, respectively; both p<0.001). Postestimation analysis revealed that the difference between these two odds ratio estimates was
statistically significant at p<0.001 (data not shown in table).
Readers might speculate about the degree to which these findings might be affected had
statistical models with covariate adjustment for additional influences, such as education, income,
employment, use of other internationally regulated drugs, etc. (Hereinafter, early-onset cannabis
use sometimes is abbreviated with the acronym, EOCU.)
Given the nature of the data, these covariates might very well be consequences of either
cannabis smoking or depression, and it was decided to evaluate their potential influence on the
EOCU-depression relationship in post-estimation exploratory analyses. Using an add-one-in
(AOI) approach, where only one additional covariate is added to the statistical model at a time,
most of these covariates induced little variation in the EOCU estimate (Figure 4.1.1). Only use of
other internationally regulated drugs besides cannabis shifted the EOCU estimate towards null to
a noteworthy degree, but even then the estimate remained above null (OR=1.2; 95% CI = 1.1,
1.3).

93

An alternative approach estimated a statistical model in which all relevant covariates
were evaluated together in the same model in order to evaluate their potential combined effects
on the EOCU estimate (Figure 4.1.1). A backwards elimination model building procedure was
considered to avoid over adjustment and produce a parsimonious model. However, no single
covariate met the a priori p-value threshold of p>0.20. In this combined model, the EOCUestimate was statistically null (OR=1.0; 95% CI = 0.9, 1.1).
In order to evaluate which covariates might have an overdue influence on the main study
estimate, a leave-one-out (LOO) procedure was used (Figure 4.1.1). This iterative procedure
eliminated one model covariate at a time. A non-null estimate for the EOCU-depression spells
relationship was observed only when excluding years of cannabis involvement (OR=1.2; 95%
CI= 1.1,1 .3) or other illegal drug use (OR=1.3; 95% CI= 1.2, 1.5). Spearman rank order
correlations showed a strong correlation between the cannabis onset variable and years of
cannabis involvement (ρ=0.87), and a weak correlation between cannabis onset and other
internationally regulated drug use (ρ=0.44; a table of the Spearman correlation matrix can be
found in the Appendix).
In the second set of exploratory analyses (Table 4.1.3), onset of cannabis smoking before
age 15 or at 18 years or older was found to be associated with occurrence of later adult-onset
depression spells, even with covariate adjustment, in a comparison with never users of cannabis
as a reference or comparison subgroup (OR = 1.9; OR = 1.7, respectively; Table 4.1.3). In these
analyses there was a probe for possibly evidence that a different age threshold for EOCU might
have produced stronger associations between EOCU and occurrence of the depression spell. No
evidence was found to require reporting of quantitative estimates from this set of exploratory

94

analyses (contrast between ≤15 years and 18 years or older p-value = 0.190; data not shown in a
table).

Figure 4.1.1. Post-estimation exploratory analysis of the adjusted early-onset
cannabis use (EOCU) estimate with additional statistical control of other
covariates using add-one-in (AOI) and leave-one-out (LOO) approaches.

95

Table 4.1.3. Exploratory analysis of association between depression
spells and fine-grained cannabis onset exposure.
Adjusted Modela

Cannabis Onset
Never Used
≤ 15 years
15 years
16 years
17 years

OR

(95% CI)

P-value

1.0
1.9
1.7
1.6
1.6

(1.6, 2.1)
(1.5, 2.0)
(1.4, 1.8)
(1.4, 1.9)

<0.001
<0.001
<0.001
<0.001

1.7
(1.6, 1.9)
<0.001
≥ 18 years
a
Model adjusted for elapsed time of cannabis involvement, tobacco
cigarette onset, alcohol onset, sex, age, race/ethnicity, and survey
year.

4.2. Study 2: Cannabis Problems and Functional Impairment Attributed to Recently Active
Depression Spells.
4.2.1. Sample Characteristics by Depression Status
An estimated 14% of the total NSDUH sample surveyed from 2009-2011 had a major
depressive episode (MDE) at some point in their lives (Table 4.2.1). Among these lifetime cases,
a little more than half were a recently active depression spell cases in the past year (53%;
n=13,743). Differences in background characteristics were apparent by recently active
depression spell status. For example, the population sample tended to more evenly female vs.
male (52% vs. 48%), however, past-year depression spell cases were much more likely to be
female (65% vs. 35%).

96

Table 4.2.1. Characteristics of U.S. National Surveys on Drug Use and Health from 2009 to 2011 by
Depression Status (n=172,042).

Variable
Sex
Age

Race/
Ethnicity

Income

County Metro

Categories

Total

Never MDE/
Unknown

Prior
Depressiona

(n=172,042)

(n=148,291)

(n=10,008)

Recently
Active
Depression
Spellb
(n=13,743)

n

(%)

n

(%)

n

(%)

n

(%)

Male

82,896

(48.5)

75,222

(50.4)

3,423

(37.3)

4,251

(34.8)

Female

89,146

(51.5)

73,069

(49.6)

6,585

(62.7)

9,492

(65.2)

12-17

55,583

(9.7)

48,376

(9.7)

2,642

(7.4)

4,565

(11.3)

18-25

57,503

(13.4)

49,314

(13.2)

3,351

(12.4)

4,838

(16.0)

26-34

17,160

(14.3)

14,509

(14.0)

1,279

(17.0)

1,372

(15.9)

≥35

41,796

(62.6)

36,092

(63.1)

2,736

(63.2)

2,968

(56.9)

White

106,701

(66.7)

90,756

(65.5)

6,875

(77.6)

9,070

(72.4)

Black

22,024

(11.9)

19,712

(12.3)

872

(7.1)

1,440

(10.0)

Hispanic

27,958

(14.7)

24,595

(15.2)

1,364

(10.0)

1,999

(12.2)

Other

15,359

(6.7)

13,228

(6.9)

897

(5.3)

1,234

(5.4)

<$20K

39,845

(18.4)

33,804

(17.9)

2,277

(17.0)

3,764

(25.6)

$20K-$49,999

57,739

(32.8)

49,660

(32.9)

3,323

(30.4)

4,756

(33.4)

$50K-$74,999

28,486

(17.1)

24,631

(17.2)

1,741

(17.1)

2,114

(16.2)

≥$75K

45,972

(31.7)

40,196

(32.0)

2,667

(35.5)

3,109

(24.9)

Large Metro

74,812

(53.2)

64,899

(53.5)

4,229

(51.8)

5,684

(51.0)

Small Metro

60,965

(30.7)

52,192

(30.5)

3,664

(32.6)

5,109

(32.2)

Non-Metro
36,265 (16.0) 31,200 (16.0) 2,115 (15.6) 2,950 (16.8)
Prior depression was defined as individuals with a lifetime occurrence of major depressive episode
(MDE), but not have a depression spell within the past 12 months.
b
Recently active depression spell was defined as a two-week or more spell of depressed mood or
anhedonia with allied symptoms.
a

4.2.2. Descriptive Statistics and Distribution of Impairment Outcome Indicators
On average, recently active depression spell cases experienced a moderate amount of
functional impairment (Table 4.2.2). In a quarter of the cases, impairment was severe or very
severe. Small differences were observed in the mean values for individual SDS items.
Depression spell cases experienced more impairment in social life (mean, M=6.2; 95%CI=6.1,
6.3) and relationships/family life (M=5.8; 95%CI=5.7, 5.9), than for work/school (M=5.0;
97

95%CI=4.9, 5.1) or home management/chores domains (M=5.6; 95%CI=5.5, 5.7). Half were
unable to carry out daily activities due to a depression spell for at least 5 or more days out of the
past year, while a quarter were impaired for a month or longer. Less than 1% of recently active
depression spell cases had missing data any of the first four SDS items, while about 4% were
missing on ‘Days unable to carry out activities’ in the past year (SDS5). The distribution of each
SDS item, along with an overlay of a normal density curve with the mean and standard deviation
of the SDS item is displayed below (Figures 4.2.1-4.2.4). The distribution of impairments with
work/school and home management/chores had the least amount of skew, while impairments in
social life and close relationships were slightly negatively skewed.

Table 4.2.2. Descriptive Statistics of Sheehan Disability Scale (SDS) Items Among Sample of Past-Year
Depression Spell Cases (n=13,743)

SDS1

Home Management/Chores

13663

0

10

5

7

5.6

Std.
Err.
0.04

SDS2

Work/School

13648

0

10

5

7

5.0

0.05

4.9

5.1

SDS3

Relationships/Family Life

13688

0

10

6

8

5.8

0.05

5.7

5.9

SDS4

Social Life
Days unable to carry out
activities

13690

0

10

6

8

6.2

0.05

6.1

6.3

13155

0

365

5

30

60.7

2.01 56.7 64.7

Variable

SDS5

Description

n

Min

Max Q50 Q75 Mean

Data: US National Surveys on Drug Use and Health, 2009-2011
Q = Quartile; Std. Err. = Standard Error; CI = Confidence Interval

98

95% CI
5.5

5.7

2500
2000
1500
1000
0

500

Frequency

0

2
4
6
Interfered with Home Management

8

10

1000
0

500

Frequency

1500

2000

Figure 4.2.1. Distribution of home management/chore impairment among
recently active depression spell cases.

0

2

4
6
Interfered with Work/School

8

10

Figure 4.2.2. Distribution of work/school impairment among recently active
depression spell cases.
99

2000
1500
1000
0

500

Frequency

0

2

4
6
Interfered with Relationships

8

10

1000
0

500

Frequency

1500

2000

Figure 4.2.3. Distribution of close relationship impairment among recently active
depression spell cases.

0

2

4
6
Interfered with Social Life

8

10

Figure 4.2.4. Distribution of social life impairment among recently active
depression spell cases.
100

4.2.3. Factor Analysis
There were moderately strong to very strong positive inter-correlations among SDS items
(Table 4.2.3). For example, the item on impairments with social life and the item on
relationships/family had the strongest correlation (0.72). ‘Days unable to carry out activities’
tended to be less correlated with other items, except for impairment with work/school (0.58).
Factor analysis was conducted using all five SDS items in a hypothesized single factor model
(see Figure 3.3.1 from Chapter 3). This model of overall functional impairment provided a
generally unsatisfactory fit to the data (CFI = 0.926; TFI = 0.770; RMSEA = 0.112).
Examination of modification indices indicated that allowing covariance between the residuals for
impairment items ‘relationships/family’ and ‘social life’ would improve model fit (Figure 4.2.5).
Freeing this parameter resulted in a substantial improvement (drop) in the chi-square value (Δ
2

χ =494.1; df=1). It also made substantive sense for there to be residual correlation between
interpersonal items of functional impairment that might not be captured by the IMPAIR latent
trait. Re-specification of the measurement model improved the fit (CFI=0.97; TLI=0.92;
RMSEA=0.04). In exploratory analyses, this model re-specification did not alter the final
estimates. Model fit to individual, single factor latent models of cannabis problems (CP; Figure
4.2.6), alcohol problems (AP; Figure 4.2.7), and tobacco dependence (TD; Figure 4.2.8) all were
an excellent fit to the data (CFI>0.990; TLI>0.990; RMSEA<0.05 for each model).

101

Table 4.2.3. Correlation Between Sheehan Disability Scale (SDS) Items of Functional Impairment
Due to Depression Spells (n=13,081).
Observed Variable

SDS1

SDS2

SDS3

SDS4

SDS1

Home Management/Chores

1.00

SDS2

Work/School

0.60

1.00

SDS3

Relationships/Family Life

0.52

0.53

1.00

SDS4

Social Life

0.53

0.51

0.72

1.00

SDS5

Days unable to carry out activities

0.49

0.58

0.41

0.44

SDS5

1.00

Data: US National Surveys on Drug Use and Health, 2009-2011

SDS1

2.9 (.44)

SDS2

3.0 (.33)

SDS3

4.3 (.58)

1.0 (.75)
1.3 (.82)
IMPAIR

.92 (.65)

2.2 (.52)

.92 (.66)
SDS4

4.2 (.57)

.35 (.68)
SDS5
Figure 4.2.5. Post-hoc re-specification of the measurement model of functional
impairment attributed to depression (IMPAIR). Standardized coefficients in
parentheses. CFI=0.97; TLI=0.92; RMSEA=0.04

102

CP1
1.0 (.96)

CP2

.92 (.88)
CP3
.96 (.92)
CP4
.88 (.84)
.91 (.88)

CP5

.98 (.94)

CP6

CP

.98 (.94)
CP7
.88 (.85)
CP8
.72 (.70)
.95 (.91)

CP9

CP10
Figure 4.2.6. Measurement model of cannabis problems (CP). Standardized
estimates in parentheses. CFI=0.995; TLI=0.994; RMSEA=0.02.

103

AP1
1.0 (.89)

AP2

.92 (.82)
AP3
.90 (.80)
AP4
.89 (.80)

AP

1.0 (.90)

AP5

1.0 (.93)

AP6

.94 (.84)
AP7
1.0 (.93)
AP8
.97 (.86)
.91 (.82)

AP9

1.0 (.92)

AP10

AP11
Figure 4.2.7. Measurement model of alcohol problems (AP). Standardized
estimates in parentheses. CFI=0.992; TLI=0.991; RMSEA=0.02.

104

TD1

TD2

TD3

TD4

.96 (.94)

-.87 (-.85)

TD5

TD7

.91 (.88)

TD8

.98 (.95)

.99 (.96)
1.0 (.99)
1.0 (.97)

TD9
1.0 (.97)
1.0 (.98)

TD10

.99 (.96)

TD11

TD

.99 (.96)
TD12
-.99 (-.96)

-.95 (-.92)
.98 (.95)

TD13

1.0 (.97)

-.94 (-.92)
1.0 (.97)

TD20

TD19

-.93 (-.90)

TD18

TD17

.97 (.94)

TD16

TD14

TD15

Figure 4.2.8. Measurement model of tobacco dependence. Standardized estimates
in parentheses. CFI=0.996; TLI=0.996; RMSEA=0.05.

4.2.4. Modeling Relationship Between Cannabis Problems and Functional Impairment Attributed
to Depression Spells
As estimated under a model that expresses functional impairment attributed to depression
spells (FIDS) as a function of level of cannabis problems (CPL) and other covariates, functional
impairment attributed to depression spells was at higher levels when level of cannabis problems
was at higher levels. Level of CP was positively associated with higher levels of impairment
attributed to depression in the unadjusted and adjusted models (Figure 4.2.9, Figure 4.2.10, and
105

Figure 4.2.11). Models adjusting for the level of alcohol problems, tobacco dependence, and
background factors such as sex, age, and race/ethnicity partially, but did not completely,
attenuate the hypothesized association linking CP on IMPAIR construct (unadjusted β=0.32 vs.
adjusted β=0.20). Final models (e.g., Fig 4.2.11) also showed increased functional impairment
attributed to depression for those with higher levels of tobacco dependence (β=0.39; p<0.001).
In a comparison of standardized estimates, the hypothesized effect of TD on IMPAIR
was estimated as being greater than for CP (Β=0.23 vs. 0.11; data not shown in figure). These
estimates were statistically distinct, as determined by comparing the final model to one in which
the CP and TD paths to IMPAIR were assumed to be equivalent, which proved to provide a less
satisfactory model fit. By contrast, the level of alcohol problems appeared to have no relationship
with functional impairment attributed to depression.
Other model estimates were of note, but are presented here solely as a source of leads in
future research, whereas the estimates of primary importance in this research involve cannabis
effect estimation. Blacks had higher levels of impairment compared to Whites, while being 1825 year old was associated with lower impairment levels compared to 12-17 year olds. Females
had lower levels of all three latent dimensions of drug problems. For cannabis problems, there
was a positive association for 18-25 year olds, which was inverse for older age groups. This was
contrasted for levels of AP and TD, which were higher for all age groups older than 12-17. As
compared to Whites, Blacks experienced lower levels of AP and TD, while Hispanics and
“Other” racial/ethnic groups had lower levels of TD only.

106

4.2.5. Post-Estimation Exploratory Analyses
Additional post-estimation exploratory analyses to test model assumptions not depicted in
tables or figures are described here, with additional coverage in the Discussion chapter notes on
study limitations. First, it was previously mentioned that the hypothesized measurement model
of functional impairment attributed to depression was re-specified by allowing the residuals for
SDS3 and SDS4 items to correlate, which improved the fit of the model. Re-analysis of the final
model without this modification revealed no appreciable differences in the estimates reported.
Second, the conceptual model assumed CP level determined the level of impairment, however,
since the measurement of both constructs pertained to the same 12 month period prior to
interview, this assumption may not be tenable. Indirect supporting evidence came from an
exclusion of those whose first depression spell occurred prior to or at the same age as their
cannabis onset (n=3,803), which would violate the temporality assumption of the model. Though
based on a smaller sample size (n=9,529), the estimated effect of CP level on impairment was
nearly the same (β=0.19; p=0.035).

107

CP1
1.0

CP2

.92

CP3

.96
CP4
.88
CP5

ε

3.0

SDS2

3.2

SDS3

4.2

1.0

3.6

.91

SDS1

1.3
.98

CP6

CP

.32

IMPAIR

.95

2.2

.91

.98
CP7

SDS4

4.3

.37

.88
CP8

SDS5
.72

CP9
.95
CP10
Figure 4.2.9. Association between functional impairment attributed to depression
(IMPAIR) and cannabis problems (CP). CFI=0.992; TLI=0.99; RMSEA=0.01.

108

ε
3.6

Female
-.28
18-25

.09
-.05

.17

IMPAIR
.33
.46

26-34
-.15

.40
35+

-.53
.19
.14

Black
.09
Hispanic

CP

.04
.01

.14

.93
ε

Other
Figure 4.2.10. Association between functional impairment attributed to
depression (IMPAIR), cannabis problems (CP), and background characteristics.
Reference groups were males, 12-17 year olds, and non-Hispanic Whites. Dashed
arrows indicate estimates with p>0.05. CFI=0.97; TLI=0.96; RMSEA=0.02.

109

Female
-.36
-.11
18-25
.48

.36

.78
-.29

26-34

.12
.12

-.09

35+

.14
-.34
-.06
.14
.30
.18
.25

-.10 .07

.82

.32

.17

.58

-.54
-.16

Black

Other

.04

ε
2.9

IMPAIR
.97 ε
.20

.14

-.14
Hispanic

.78 ε

AP

.54

CP

.39

.40

-.23

.01
-.20

TD

.94 ε

Figure 4.2.11. Association between functional impairment attributed to depression (IMPAIR), cannabis problems (CP),
alcohol problems (AP), tobacco dependence (TD), and background characteristics. Dashed arrows indicate estimates
with p>0.05. CFI=0.998; TLI=0.998; RMSEA=0.01.

110

4.3. Study 3: Cannabis Problems and History of Blunt Smoking.
Here, study aims focus on possible blunt effects that increase levels of cannabis problems.

4.3.1. Characteristics of Past-Year Cannabis Users
Distribution of background characteristics of past-year cannabis users (n=77,047)
differed from the total population sampled by the NSDUH from 2004-2011 (Table 4.3.1). For
example, the weighted proportions of males to females in the total sample were roughly equal,
but past-year cannabis users tended to be male (61%). Youths 18-25 years olds made up over a
third of past-year cannabis users (36%), despite being an estimated 13% of the population.
Differences by race/ethnicity and population density were comparatively small.

4.3.2. Prevalence of Blunt Smoking and Cannabis Problems
An estimated 70% of past-year cannabis users smoked a blunt at least once in their
lifetimes (Table 4.3.2). Slightly under one-half (47%) smoked a blunt in the past year, while
about a quarter smoked one as recently within the past month. Estimates were fairly stable from
2004 to 2011, but there was a statistically significant increase in blunt use during this period. For
example, estimated cumulative incidence increased from 66% to 72% during this period.
Overall, past-year cannabis users commonly experienced ‘spending a great deal of time
getting or using cannabis’ (42%) and tolerance (29%), while few experienced ‘trouble with the
law’ (3%) (Table 4.3.3). Blunt smokers had about twice the prevalence of each cannabis problem
as compared to their non-blunt smoking peers. For example, 13% of past-year cannabis users
who never smoked a blunt had experienced tolerance as compared to 33% of those who had

111

smoked a blunt. A similar pattern emerged when comparing past-year cannabis users who
smoked a blunt in the past year versus those who had not.

Table 4.3.1. Characteristics of Total Sample Surveyed and Past-Year Cannabis User Analytical Sample

Total

Past-Year Cannabis Use
No

Yes

(n=450,002)
n
(wt%)

(n=372,955)
n
(wt%)

(n=77,047)
n
(wt%)

Female

233,672 (51.5)

199,680 (52.9)

33,992 (39.4)

Male

216,330 (48.5)

173,275 (47.1)

43,055 (60.6)

12-17 Years Old

146,438 (10.1)

126,265

(9.8)

20,173 (12.3)

18-25 Years Old

149,844 (13.3)

107,229 (10.6)

42,615 (35.9)

26-34 Years Old

45,269 (14.3)

38,308 (13.6)

6,961 (20.3)

108,451 (62.3)

101,153 (66.0)

7,298 (31.5)

284,551 (67.9)

233,958 (67.7)

50,593 (69.6)

57,056 (11.8)

47,061 (11.6)

9,995 (13.8)

Sex

Age

35 or Older
Race or Ethnicity
White
Black/African-American
Native Am/Alaskan Native

6,723

(0.5)

4,938

(0.5)

1,785

(0.7)

Native HI/Other Pacific Islander

2,044

(0.3)

1,680

(0.3)

364

(0.3)

Asian

15,435

(4.4)

14,153

(4.7)

1,282

(1.8)

More than one race

13,733

(1.2)

10,612

(1.1)

3,121

(1.9)

Hispanic

70,460 (14.0)

60,553 (14.2)

9,907 (11.9)

CBSA ≥ 1 million

184,252 (51.2)

151,947 (50.8)

32,305 (54.1)

CBSA < 1 million

216,376 (40.4)

178,822 (40.5)

37,554 (39.6)

Population Density

Not in CBSA

49,374

(8.4)

Data: US National Surveys on Drug Use and Health, 2004-2011.
Weighted percentage (wt%); Core based statistical area (CBSA)

112

42,186

(8.6)

7,188

(6.3)

Table 4.3.2. Prevalence of Blunt Smoking History Among Past-Year Cannabis Users by Individual
Survey Years.
Total

Smoked Blunts in
Past Year

Ever Smoked Blunts
%

Prev.

95% CI

Prev.

95% CI

Smoked Blunts in
Past Month

Year

n

Prev.

95% CI

2004

9,528 12.4

65.9

(64.1,

67.8)

44.5

(42.9,

46.1)

21.8

(20.6,

23.0)

2005

9,348 12.1

66.4

(64.4,

68.4)

44.2

(42.6,

45.7)

23.1

(21.7,

24.4)

2006

8,971 11.6

69.8

(67.2,

72.4)

47.5

(45.4,

49.7)

24.6

(23.0,

26.1)

2007

9,034 11.7

70.1

(68.0,

72.2)

47.0

(45.1,

49.0)

25.0

(23.6,

26.4)

2008

9,229 12.0

71.6

(69.3,

74.0)

48.3

(46.2,

50.3)

25.4

(24.0,

26.8)

2009

10,022 13.0

72.2

(70.0,

74.3)

46.5

(44.5,

48.6)

25.3

(23.8,

26.9)

2010

10,396 13.5

72.5

(70.6,

74.3)

48.4

(46.2,

50.5)

25.6

(23.7,

27.4)

2011

10,519 13.7

72.1

(70.2,

74.0)

47.8

(45.9,

49.6)

26.8

(25.4,

28.2)

Total

77,047 100.0

70.2

(69.5,

70.8)

46.8

(46.1,

47.5)

24.8

(24.3,

25.2)

Data: US National Surveys on Drug Use and Health, 2004-2011
Weighted prevalence (Prev.); Confidence interval (CI)

Table 4.3.3. Weighted Prevalence (%) of Cannabis Problems Among Past-Year Cannabis Users.

Total

Ever Smoked Blunts

Smoked Blunts in Past
Year

No

Yes

No

Yes

%
SE
41.5 (0.4)

%
SE
18.2 (0.7)

%
SE
48.4 (0.4)

%
SE
23.7 (0.5)

%
SE
55.7 (0.5)

CP2 Unable to set limits

7.3

3.6

8.4

4.4

9.6

CP3 Tolerance

28.5 (0.3)

13.2 (0.5)

33.0 (0.4)

16.3 (0.4)

38.2 (0.4)

CP4 Difficulty cutting down

6.6

(0.2)

4.0

(0.3)

7.3

(0.2)

4.6

(0.2)

8.2

8.5

(0.2)

5.1

(0.4)

9.5

(0.2)

5.8

(0.3)

10.6 (0.3)

CP6 Gave up important activities

11.2 (0.2)

6.2

(0.4)

12.7 (0.2)

7.2

(0.3)

14.5 (0.3)

CP7 Serious role impairment

7.0

(0.2)

3.8

(0.3)

7.9

(0.2)

4.3

(0.2)

9.1

(0.2)

CP8 Hazardous use

7.1

(0.2)

4.3

(0.3)

8.0

(0.2)

4.6

(0.2)

9.2

(0.2)

CP9 Trouble with the law

3.0

(0.1)

1.4

(0.2)

3.5

(0.1)

1.6

(0.1)

4.2

(0.1)

CP10 Problems with family/friends

5.6

(0.1)

3.0

(0.3)

6.3

(0.2)

3.1

(0.2)

7.5

(0.2)

Cannabis Problems (CP)
CP1 Spend great deal of time

CP5

Use despite emotional/
physical problems

(0.2)

(0.3)

Data: US National Surveys on Drug Use and Health, 2004-2011.

113

(0.2)

(0.3)

(0.2)
(0.2)

4.3.3. Exploratory and Confirmatory Factor Analysis of Cannabis Problems
A correlation matrix of CP items did not reveal a clear pattern of factor structure (Table
4.3.4). An exploratory factor analysis was then conducted using the first half of a random split
sample of data from the 2004 NSDUH (n=3,534). Figure 4.3.1 presents a scree plot of
eigenvalues (EV). There were two EV larger than one, however, almost 60% of the total variance
could be explained by the first factor (EV=5.7), while a second factor would contribute only an
additional 13% (EV=1.3). Examining factor loadings from one and two factor solutions revealed
additional insights (Table 4.3.5). All items loaded strongly on the single factor, with loadings
ranging from 0.56 for ‘difficulty cutting down’ to 0.85 for ‘giving up important activities’. Item
loading in the two factor solution did not conform to DSM-IV hypothesized constructs of
dependence and abuse. Most items loaded strongly on one factor dominated by socially
maladaptive or hazardous problems (CP5-CP10), while a second factor appeared to tap problems
related to repetitive use behavior (CP2 and CP4). As noted earlier, ‘spend a great deal of time’
(CP1) and tolerance (CP3) were common problems, and these items had moderate loadings on
both factors.
Confirmatory factor analysis (CFA) was then conducted on the second half of sample
data from the 2004 NSDUH (n=3,520). A single latent factor model was compared to a two
factor model with CP1-CP6 items loading on a latent trait of dependence (DEPEND), and with
CP7-CP10 items loading on a latent trait of socially maladaptive or otherwise harmful cannabis
use (HARM). Both models were a good fit to the data (Table 4.3.5: CFI>0.95; TLI>0.95;
RMSEA=0.01). However, latent factors were strongly correlated (0.86). Thus, the totality of the
evidence favored a single factor, which was used for subsequent analyses.

114

Table 4.3.4. Correlation Matrix of Cannabis Problem (CP) Items Among Past-Year Cannabis Users
(n=3,534).

1

2

3

4

5

6

7

8

9

10

CP1 Spend great deal of time

1.00

CP2 Unable to set limits

0.56 1.00

CP3 Tolerance

0.67 0.47 1.00

CP4 Difficulty cutting down
Use despite emotional/ physical
CP5
problems
CP6 Gave up important activities

0.50 0.71 0.39 1.00

CP7 Serious role impairment

0.47 0.44 0.50 0.34 0.66 0.79 1.00

CP8 Hazardous use

0.59 0.40 0.45 0.29 0.55 0.57 0.59 1.00

CP9 Trouble with the law

0.39 0.27 0.50 0.18 0.47 0.48 0.68 0.52 1.00

CP10 Problems with family/friends

0.57 0.41 0.52 0.33 0.66 0.72 0.73 0.68 0.63 1.00

0.56 0.48 0.57 0.46 1.00
0.61 0.50 0.56 0.36 0.68 1.00

Data: Random half-split sample of data from US National Survey on Drug Use and Health, 2004

7

6

Eigenvalues

5

4

3

2

1

0

0

1

2

3

4

5

6

7

8

9

10

11

Factor

Figure 4.3.1. Scree Plot of Eigenvalues from Exploratory Factor Analysis of
Cannabis Problems. Random Half-Split Sample of Data from 2004 National
Surveys on Drug Use and Health (n=3,534).

115

Table 4.3.5. Exploratory and Confirmatory Factor Analysis of Cannabis Problems, Standardized and Unstandardized Factor Loadings Using
Random Split Sample of Data from 2004 US National Surveys on Drug Use and Health.

Cannabis Problems

Exploratory

Confirmatory

(n =3,534)

(n =3,520)

1 Factor

2 Factors

1 Factor

%

1

1

2

Latent

2 Factor

β

B

SE

Latent

β

B

SE

CP1

Spend great deal of time

42

0.79

0.36

0.57

CP

0.70

1.00

DEPEND

0.73

1.00

CP2

Unable to set limits

8

0.65

0.05

0.78

CP

0.59

0.84

0.06 DEPEND

0.60

0.83

0.06

CP3

Tolerance

28

0.71

0.38

0.46

CP

0.68

0.96

0.06 DEPEND

0.69

0.95

0.06

CP4

Difficulty cutting down

8

0.56

0.10

0.84

CP

0.59

0.84

0.06 DEPEND

0.60

0.83

0.06

CP5

Use despite emotional/
physical problems

10

0.78

0.61

0.26

CP

0.72

1.02

0.07 DEPEND

0.74

1.02

0.07

CP6

Gave up important activities

12

0.85

0.78

0.12

CP

0.86

1.22

0.07 DEPEND

0.89

1.23

0.07

CP7

Serious role impairment

8

0.84

0.94

0.09

CP

0.86

1.22

0.06

HARM

0.92

1.00

CP8

Hazardous use

9

0.71

0.63

0.15

CP

0.59

0.85

0.07

HARM

0.63

0.68

0.04

CP9

Trouble with the law

4

0.65

0.78

0.13

CP

0.71

1.01

0.08

HARM

0.75

0.81

0.04

CP10

Problems with family/friends

7

0.82

0.84

0.02

CP

0.75

1.06

0.07

HARM

0.78

0.85

0.04

Model Fit Indices
CFI

0.96

0.99

0.95

0.96

TLI

0.96

0.99

0.95

0.96

RMSEA

0.01

0.05

0.01

0.01

Factor Correlation

NA

0.55

NA

0.86

Data: US National Survey on Drug Use and Health, 2004
SE, standard error; β, standardized estimates; B, unstandardized estimates; CFI, comparative fit index; TFI, Tucker-Lewis index; RMSEA, root
mean square error of approximation

116

4.3.4. Recent Blunt Smoking and Level of Cannabis Problems
Factor scores for the latent trait of cannabis problems (CP) were regressed on a
categorical variable indicating time since recent blunt smoking experience (Figure 4.3.2). Never
blunt smokers were compared to mutually exclusive categories of those who smoked blunts
within the past month (‘past month’), smoked blunts more than 30 days ago but within the past
year (‘past year’), or more than 12 months ago (‘prior use’).
In an unadjusted model, any history of blunt smoking was associated with higher levels
of CP, with larger estimates for more recent blunt use. For example, those who had smoked
blunts within the past month had higher levels of CP than ‘past-year’ blunt smokers (β=0.98 vs.
0.62, respectively). Likewise, ‘past-year’ blunt smokers had higher levels of CP than ‘prior
users’ (β=0.28). When the model adjusted for the influence of sex, age, race/ethnicity, or
population density, estimates for blunt use recency did not appreciably change (Figure 4.3.3).
Males, younger age groups, Hispanics, “Other” racial/ethnic groups, and residents of rural areas
were associated with higher levels of CP. African-Americans/Blacks were no different than
Whites on level of CP, nor were there differences observed between those living in more urban
areas (CBSA ≥1 million vs. CBSA<1 million). The final model incorporated the influence of
alcohol problems (AP) and tobacco dependence (TD) within the model. Estimates for blunt
smoking recency did not change appreciably (Figure 4.3.4).
As a probe into the possibility that the relationship between CP and blunt smoking might
differ by frequency of blunt smoking, the model was then re-estimated using days of blunt use in
the past month, with a response range from zero to 30 days. There was a positive association
such that for every added day of blunt smoking, there was a 0.03 increase along the dimension of
CP (Figure 4.3.5). This estimate cannot be interpreted as an effect because temporal sequencing

117

and feedback loops have not been taken into account, notwithstanding a fundamental difference
in the scales between count and categorical variables. In exploratory analyses, past-month blunt
frequency was categorized (0 days; 1-7 days; 8-14 days; 15-21 days; 22-30 days), and results
showed an estimated increase in effect size for each category as compared to those who did not
use blunts in the past month: β=0.45 (1-7 days); β=0.59 (8-14 days); β=0.73 (15-21 days);
β=0.84 (22-30 days) (data not shown in table or figure).
CP1
1.0

CP2

.90

CP3

.94
CP4
ε
.60

Prior Use
.28
Past Year

.86
CP5
.99

.62

CP

1.1
1.1

.98
Past Month

CP6

CP7

.87
CP8
.84
CP9
1.0
CP10

118

Figure 4.3.2. Association between level of cannabis problems (CP) and blunt
smoking recency among past-year cannabis users (n=67,519). CFI=0.95;
TLI=0.95; RMSEA=0.02.

Prior Use
.29
Past Year
Blunt

CP1
.57

PastMonth
Blunt

1.0
.95

Male
.15
18-25

.91
ε
.60

-.20

.93

26-34

-.31

35+

-.32

1.1

<.01

.87

.07

.84

.06

1.0

Black

CP3
CP4

.87
.99

CP

CP2

1.1

CP5
CP6
CP7
CP8

Hispanic
Other
CBSA
<1 mil

CP9

.03
CP10
.10

No
CBSA
Figure 4.3.3. Association between cannabis problems (CP) and blunt use recency,
sex, age, race/ethnicity, and population density (n=67,519). Note: Dashed lines
indicate estimate with p>0.05. CFI=0.94; TLI=0.93; RMSEA=0.02.

119

Prior Use
.29
Past Year
Blunt
.57
PastMonth
Blunt

AP

.95

Male

ε
.60

.15
18-25

.60
σ2
.21

-.20
.14

26-34

-.31

35+

-.32

Black

CP

σ2

<.01

.13

.60

.07
Hispanic
.06
Other
CBSA
<1 mil

TD

.03
.10

No
CBSA
Figure 4.3.4. Association between cannabis problems (CP) and recent blunt use,
background characteristics, alcohol problems (AP), and tobacco dependence
(TD). Dashed lines indicate estimates with p>0.05. Latent structure and factor
loadings not depicted for CP, AP, and TD in this figure. n=67,519; CFI=0.992;
TLI=0.992; RMSEA=0.02.

120

Days of
Blunt
Use

AP

.03

Male

ε
.64

.17
18-25

.57
σ2
.23

-.20
.16

26-34

-.36

35+

-.53

Black

CP

σ2

<.01

.16

.87

.08
Hispanic
.06
Other
CBSA
<1 mil

TD

.03
.08

No
CBSA
Figure 4.3.5. Association between cannabis problems (CP) and days of blunt use
in the past month, background characteristics, alcohol problems (AP), and
tobacco dependence (TD). Dashed lines indicate estimates with p>0.05. Latent
structure and factor loadings not depicted for CP, AP, and TD in this figure.
n=67,519; CFI=0.994; TLI=0.994; RMSEA=0.02.

121

4.4. Study 4: Time to Cannabis Onset and Cannabis Problems.
This study's aim was to estimate the degree to which early-onset cannabis smoking might affect
risk of developing cannabis problems. The data are from the World Mental Health Surveys.

4.4.1. Frequency and Distribution of Cannabis-Only Users
Among 6,609 cannabis-only users in the sample, 80 (1.2%) had qualified as cases of
cannabis dependence (CD) when assessed at WMHS baseline (Table 4.4.1). Eight WMH
countries contributed no cases of CD among cannabis-only users, either because there were too
few cannabis-only users, and/or CD was rare. Cannabis dependence was based on qualifying
under either DSM-IV or ICD-10 criteria, with most CD cases qualified under both criteria
(n=73). In addition, all but two respondents with CD had one or more features of socially
maladaptive or hazardous cannabis use experiences (data not shown in table).
Table 4.4.1. Frequency of cannabis dependence (CD) among cannabis-only users by World Mental
Health Survey site.
Country

No

Yes

Total

Brazil

245

4

249

Colombia

257

3

260

Mexico

210

2

212

1,365

13

1,378

14

0

14

208

2

210

16

0

16

1

0

1

449

0

449

Lebanon

40

0

40

Nigeria

34

0

34

Japan

6

0

6

China

3

0

3

New Zealand

3,681

56

3,737

Total

6,529

80

6,609

United States
Bulgaria
Northern Ireland
Romania
Iraq
Israel

122

Distribution of the cannabis-only sample by CD and other variables of interest are shown
in Table 4.4.2. Mean age of first IRD opportunity was three years younger for those with CD
compared to those without (15 vs. 18), and age of first cannabis use also differed by three years
(17 vs. 20). Two-thirds of cannabis-only users have smoked tobacco cigarettes (66%) and 5.8%
had alcohol dependence (AD).
Table 4.4.2. Distribution of Cannabis-Only Sample by Relevant Characteristics and Cannabis
Dependence
CannabisOnly

Cannabis Dependence

(n=6,609)
Variable
Age of 1st IRD
Opportunity
Age of 1st Cannabis
Use
Cigarette Smoking
Status

Category

Total

%Yes

18

-

18

15

-

(Mean)

20

-

20

17

-

Never/Only a few

2,246

34.0

2,230

16

0.7

Ex-Smoker

1,752

26.5

1,728

24

1.4

Current Smoker

2,610

39.5

2,570

40

1.5

1

0.0

1

0

0.0

6,223

94.2

6,187

36

0.6

386

5.8

346

40

10.4

Female

3,246

49.1

3,213

33

1.0

Male

3,363

50.9

3,316

47

1.4

37

-

37

32

-

No Dependence
Dependence

Sex

Yes
(n=80)

(Mean)

Missing
Alcohol Dependence

No
(n=6,529)

%

Age

(Mean)

Income

Low

1,371

20.7

1,331

40

2.9

Low-Average

1,780

26.9

1,758

22

1.2

Above Average

3,175

48.0

3,158

17

0.5

283

4.3

282

1

0.4

< Secondary

2,240

33.9

2,193

47

2.1

Secondary

4,369

66.1

4,336

33

0.8

Not-Working

1,508

22.8

1,478

30

2.0

Working

5,101

77.2

5,051

50

1.0

Never married

3,814

57.7

3,773

41

1.1

Previously married

897

13.6

887

10

1.1

Married-cohabiting

1,897

28.7

1,868

29

1.5

1

0.0

1

0

0.0

Missing
Education
Employment

Marital Status

a

Missing

123

Table 4.2.2. (cont’d)
Parent's Highest Level
of Education

Primary or less

989

15.0

969

20

2.0

Secondary

3,098

46.9

3,060

38

1.2

Tertiary

1,978

29.9

1,960

18

0.9

Missing
Major Depressive
Episode
a

544

8.2

540

4

0.7

No Depression

5,049

76.4

4,994

55

1.1

Depression

1,560

23.6

1,535

25

1.6

Not-working category: students, homemakers, retired, and other non-employed.

The distribution of the time to cannabis onset (TCO) variable, operationally defined as
the number of years between age of first IRD opportunity and age of first cannabis use, was
examined using histogram plots (Figure 4.4.1). Nearly half of cannabis-only users (47%) had an
opportunity to try drugs and used cannabis for the first time at the same age. Another third (32%)
used cannabis within the three years after their first drug opportunity. The histogram plot also
revealed that some respondents had negative values due to reporting a younger age of cannabis
onset than IRD opportunity (n=185; 2.8%). Figure 4.4.2 shows the same distribution of TCO
when observations with negative values were removed.

124

3000
2000
0

1000

Frequency

-60

-40

-20
0
20
Years from IRD opportunity to MJ onset

40

Figure 4.4.1. Histogram plot of time to cannabis onset (TCO), all values.

Figure 4.4.2. Histogram plot of time to cannabis onset (TCO) for non-negative
values.
125

4.4.2. Cannabis Dependence (CD) and Time to Cannabis Onset (TCO)
We use SLT as an abbreviation for a short lag time from first chance to try cannabis to
actual use of cannabis, and TCO to refer to time elapsed from cannabis onset until onset of
cannabis use. In the initial logistic regression model, SLT was inversely associated with odds of
developing cannabis dependence after adjusting for age of first IRD opportunity (β=-0.2; 95% CI
= -0.3, -0.1; p<0.001; Table 4.4.3). This means that for every year increase in TCO, there was
about a 20% decrease in the odds of experiencing cannabis dependence. Additional models
controlled for the influence of country, sex, age, income, education, employment, marital status,
parent’s education, and major depression. No difference in the estimate between CD and TCO
was found in these models. A smaller sample size in Models 2 and 3 versus Model 1 was due to
some countries having too few cannabis-only subjects for multivariable analysis. Thus, Models 2
and 3 were based on data from Colombia, Mexico, New Zealand, Brazil, Northern Ireland, and
the United States, which had sufficient numbers of cannabis-only users. Estimates were no
different in Model 1 if restricted to the same sample used for Model 3 (data not shown in table).

Table 4.4.3. Unadjusted and covariate adjusted estimates of the association between cannabis
dependence (CD) and time to cannabis onset (TCO) use among cannabis-only users.

Model
Model 1

n

β

6,074

-0.2

-0.3

-0.1

<0.001

-0.2

-0.3

-0.07

0.001

-0.3

-0.06

0.004

95% CI

P-value

Adjusted for age of first IRD opportunity.
Model 2

5,512

Adjusted for Model 1 covariate, country, sex, and age.
Model 3

5,512

-0.2

Adjusted for Model 2 covariates, smoking, income, education, employment, marital status, parent's
education, major depression.

126

4.4.3. Association Between Cannabis Problems (CP) and Time to Cannabis Onset (TCO)
Factor analysis was conducted using the eight manifest items for cannabis dependence
(CP0-CP7) and four items characterized as socially maladaptive or hazardous (CP8-CP11).
A single latent factor was observed (Figure 4.4.3). This model fit the data well (CFI= 0.999;
TLI= 0.999; RMSEA=0.009). An inverse association was observed (β=-0.05; p<0.001) when
level of CP was regressed on TCO alone (Figure 4.4.4). Adding statistical control for age of first
IRD opportunity, country, sex did not appreciably change the point estimate for TCO (Figure
4.4.5; β=-0.08; p<0.001). This model also revealed that males had a higher level of CP (β=0.31),
while age and age of first IRD opportunity were inversely related to CP (β=-0.01 and -0.07,
respectively). The final model depicted (Figure 4.4.6) controlled in the model for additional drug
influences (i.e., cigarette smoking and a latent factor of alcohol problems), background
characteristics (i.e., income, education, employment, marital status, and parent’s education), and
presence of lifetime occurrence of major depressive episode (MDE). The latent factor structure
of alcohol problems (AP) used in the model is shown in Figure 4.4.7. Despite possible model
misspecification concerning some covariates (that is, the level of CP could potentially be a
determinant of income, educational attainment, or marital status), the relationship between CP
and TOC remained relatively unchanged (β=-0.07).
In addition, the data exploration disclosed positive relationships between level of CP and
level of AP (β=0.53), being a current or ex-cigarette smoker versus never smoking (β=0.48 and
0.56, respectively), having been previously married versus never married (β=0.27), and lifetime
MDE (β=0.38). Finishing secondary school (compared to only finishing primary school) and
having a parent with at least a secondary school education (versus primary) were associated with
lower levels of CP. Having a higher income, being currently married, and having parents with a

127

higher level of education appeared to have no influence over level of CP. There was no
consistent relationship between a specific WMH country and level of cannabis problems across
models.
CP0
1.0

CP1

.99
CP2
1.0
CP3
.99
.85

CP

CP4

.99

CP5

.99

CP6

.95

CP7

.91

CP8

.80
CP9
.96
.97

CP10
CP11

Figure 4.4.3. Measurement model of cannabis problems (CP). n=6,609; CFI=
0.999; TLI= 0.999; RMSEA=0.009.

128

CP0
1.0

CP1

.99
CP2
1.0
CP3
.99
ε
.90

TOC

-.05

.85

CP

CP4

.99

CP5

.99

CP6

.95

CP7

.91
CP8
.80
.96
.97

CP9
CP10
CP11

Figure 4.4.4. Unadjusted association between cannabis problems (CP) and time
to cannabis onset (TOC). n=6,081; CFI=0.998; TLI=0.998; RMSEA=0.011.

129

Brazil

Israel

Mex.

Leb.
.24

Col.

CP0

Japan
China

Nig.

N.Z.

1.2

Bul.

1.0

CP1

.99
CP2
1.0
CP3

N.I.

.99

Rom.

.85

CP4

Iraq
TOC

-.08

CP

CP5

.99

CP6

.95

.31
.90
ε

Male

.99

CP7

.91
CP8

-.01

.80

Age

.96

-.07

.97

Age IRD
Opp.

CP9
CP10
CP11

Figure 4.4.5. Association between cannabis problems (CP), time to cannabis
onset (TOC), country, sex, and age. Dashed lines indicate p-values greater than
0.05, point estimates not shown. n=6,030; CFI=0.997; TLI=0.996;
RMSEA=0.009.

130

Above
Ave. Inc.

MDE

LowAve. Inc.
Current
Smoker
ExSmoker
TOC
Male
Age
Age IRD
Opp.

.38

Mex.

Brazil

Israel

Japan

Col.

-.38

Leb.

China

Bul.

Nig.

N.Z.

N.I.

Rom.

1.0
.99

Iraq

.48

CP0

1.0
.56

.85

.47
-.02

CP

-.06
.58
.59

Secondary -.23
Finished
.27
Previously
Married

Parent Edu:
Secondary

CP2
CP3

.99

-.07

Married/
Cohabit.

CP1

ε

CP4

.99

CP5

.99

CP6

.95

CP7

.91
CP8
AP

.80
.96
.97

CP9
CP10

-.24
CP11

Parent Edu:
Tertiary

Figure 4.4.6. Association between cannabis problems (CP) and time to cannabis
onset (TOC) adjusting for age of first IRD opportunity, cigarette smoking, alcohol
problems (AP), country, major depressive episode (MDE), and other background
influences. Latent structure and factor loading estimates for AP not depicted in
this figure. Dashed lines indicate estimates with p-values greater than 0.05, with
these estimates not depicted. n=6,030; CFI=0.987; TLI=0.986; RMSEA=0.02.

131

AP0
1.0

AP1

1.0
AP2
1.1
AP3
1.0
1.0

AP

AP4

1.0

AP5

.99

AP6

.97

AP7

.99

AP8

.83
AP9
.96
.97

AP10
AP11

Figure 4.4.7. Measurement model of alcohol problems (AP) with factor loadings
from model depicted in Figure 4.4.6.

132

CHAPTER 5. DISCUSSION
This final chapter presents a discussion of the main findings and results of each study
herein. In this discussion, one of my goals is to interpret the results within the context of existing
knowledge, including the prior literature and underlying theoretical concepts, as introduced in
the background chapter. In addition, I wish to be open to each study’s limitations, a malady
which no scientific research is immune, but also highlight what I think to be the research’s many
strengths. In the final chapter (Chapter 6), I discuss what can be concluded from the research,
particularly any implications for future research and possibly for public health practice.

5.1. Study 1: Early-Onset Cannabis Smoking and Depression Spells
As has been previously mentioned, the research for Study 1 has already been published at
the time of completing this dissertation. Thus, the discussion that follows is largely reproduced
here with appropriate updated information since publication. For the most part, the material in
Section 5.1 qualifies as an extended quotation of the previously cited source material (Fairman
& Anthony, 2012).
In this study, the primary research aim was to estimate the degree to which early-onset
cannabis smoking initiated prior to age 18 years might signal an increased risk of a later adultonset incident depression spell. No more than a modest excess odds for this relationship was
found in the comparison of early-onset users with individuals who had never smoked cannabis.
This association remained statistically robust after covariate adjustment for elapsed time of
cannabis involvement, onset of tobacco cigarettes and alcohol, sex, age, and race/ethnicity,
consistent with prior studies predicting a modest association between various forms of adolescent
cannabis use and later depression in adolescence or young adulthood (D. W. Brook, Brook,
133

Zhang, Cohen, & Whiteman, 2002; Fergusson et al., 2002; B. E. Green & Ritter, 2000; Groth &
Morrison-Beedy, 2010; Hayatbakhsh et al., 2007; Pahl et al., 2010; Patton et al., 2002; Rey et al.,
2002). Our study’s results add to the literature on the association between cannabis and
depression by extending the timeline to include the first occurrence of a sustained spell of
depression in age groups past adolescence. This refinement was made possible via a constraint
on the temporal sequencing of first cannabis use and depression spell, made possible by using
age-of-onset data. However, for completeness, several studies with null results should be noted,
in which original unadjusted associations were rendered null once potential confounders were
taken into account, much as the associations under study here were much-attenuated with
statistical adjustments for covariates of interest (de Graaf et al., 2010; Degenhardt et al., 2010;
Fergusson & Horwood, 1997; Griffith-Lendering et al., 2011; Lynskey et al., 2004; Pedersen,
2008; Windle & Wiesner, 2004).
As mentioned elsewhere (Fairman & Anthony, 2012), readers may interpret the overall
population estimate as too modest to sustain a causal inference due to uncontrolled sources of
potential confounding, such as a diathesis that might determine both EOCU and adult-onset
depression spells. For example, de Graaf and colleagues (2010) suggested that early behavioral
dysregulation or conduct problems might be one such indication of a shared diathesis. If
controlled, these now-omitted background characteristics might well render the association null,
as in de Graaf et al. (2010). Whereas the present study included no measures of early behavioral
or conduct problems, Greenland and O’Rouke (2008) provide an approach that can be used to
gauge whether (and by how much) an association might shift toward the null if an unmeasured
covariate had been controlled. Their method is based upon the ratio of adjusted and unadjusted
estimates as derived from prior research. To illustrate, according to estimates reported by de

134

Graaf et al. (2010), the ratio of adjusted to unadjusted OR estimates is 0.9. As forecast by the
Greenland-O’Rourke method, multiplication of 0.9 times the present study’s point estimate of
1.7 yields an odds ratio estimate of 1.5 (i.e., closer to the null). Accordingly, this study's
conclusion continues to be that the link from early-onset cannabis smoking to later adult-onset of
a depression spell is quite modest (but is non-null).
This study's secondary research question was whether the estimated associations with
depression spell might be different when studied in association with adult-onset cannabis
smoking. In this study’s estimates, little difference between the early-onset and adult-onset
subgroups of cannabis smoking onset was found with respect to occurrence of adult-onset
depression spells in models with covariate adjustment for potential confounders and background
characteristics. One might expect that the delay of cannabis onset until adulthood might have a
diminished effect on the incidence of depression spells, in congruence with theories of either a
direct influence of exogenous cannabinoids on the maturing adolescent brain (Rice & Barone,
2000; Schneider, 2008), or an indirect influence through difficulty in assuming adult roles that
lead to depression (B. E. Green & Ritter, 2000; Kandel et al., 1986; Marmorstein & Iacono,
2011). Nonetheless, the study findings cannot be said to support these theoretical formulations of
the cannabis-depression associations. In addition, one might also expect the effect of EOCU to
be greater and more pronounced when cannabis smoking has started very very early in late
childhood or early adolescence, as compared to associations found when cannabis smoking has
started in later adolescence. However, in analyses that separated early-onset cannabis smoking
into year-by-year age categories, there was no strong support for any age-related gradient in the
strength of association after statistical adjustment. As such, early-onset cannabis smoking might
merely be a marker for more regular, chronic use in adulthood. For example, Georgiades and

135

Boyle (2007) found that cannabis use in adolescence only (but not in adulthood) had little
association with depression, but that occurrence of depression was associated with cannabis use
in adolescence and continued into adulthood, as well as with adult-onset cannabis smoking
without adolescent use of cannabis.
The effect of early-onset cannabis use on later depression spells, if there is any, might be
mediated through the influence of adolescent cannabis use on psychosocial factors associated
with depression. Most of the variables that were considered individually - educational
attainment, employment status, total family income, marital status, number of times married,
ever been arrested, population density - had little apparent effect on the early-onset cannabisassociated odds of depression spells. Only having ever used other illegal drugs besides cannabis
produced a noteworthy, but not full attenuation of the originally observed EOCU-depression
spell association. Statistical control of all considered covariates in a single model rendered the
EOCU-depression spell null, but only in the presence of illegal drug use, which might mediate
the association, and years of cannabis involvement, which showed to be highly correlated with
age of cannabis onset. Thus, the study findings are more consistent with a recent report that
concluded the increased risk of later depression for those with an adolescent-onset cannabis use
disorder was no more than partially mediated by adverse psychosocial consequences such as
educational failure, unemployment, and criminal behavior (Marmorstein & Iacono, 2011). This
recent work can be contrasted with earlier findings of others emphasizing the meditational role of
these factors (B. E. Green & Ritter, 2000; Kandel et al., 1986). Of course, it should be noted that
there is at least one limitation to our approach: since these variables were assessed as of the time
of the interview, it is possible that these variables do not fall within the causal chain – e.g., they
could be a result of depressed mood, rather than a cause.

136

Early-onset tobacco or alcohol use was also associated with an excess odds of a later
depression spell. The strength of the association was very modest for early tobacco cigarette
smoking, but about two-fold for early alcohol drinking. Adjustment for cannabis and alcohol use,
in theory, might attenuate estimates of the effect of early-onset tobacco cigarette use, which
typically has an earlier onset than other drugs (Degenhardt et al., 2008). A related rationale might
be used to explain the finding that adult-onset tobacco cigarette smoking was not associated with
depression spells. For alcohol drinking, both early- and adult-onset drinking were positively
associated with depression spells, but the odds ratio based on early-onset drinkers were slightly
greater and statistically different than the odds ratio estimates based on adult-onset drinkers.
While speculative, the pattern of findings might prompt a suggestion that it could be possible to
attempt prevention or delay of the onset of alcohol drinking, which in turn might have an impact
on occurrence of later mood disturbances. However, our aim was not to test this relationship, and
we should point out that our conceptual model did not account for lifetime cumulative exposure
of tobacco or alcohol, such as with elapsed time of cannabis involvement, which might explain
these findings. Recent reviews by Chaiton et al. (2009) and Boden and Fergusson (2011) draw
conclusions in favor of causal relationships that link tobacco smoking or alcohol drinking with
physical health, and also with depression.
This research should be viewed in light of several additional limitations, as noted below.
First, the NSDUH collects little information on the childhood experiences of adult respondents at
the time of interview. As a consequence, some potential confounders of the cannabis-depression
association, such as a history of childhood conduct problems, early family dysfunction, or other
mental disorders likely to have their onset in childhood/adolescence were not measured and
therefore are omitted variables (Degenhardt et al., 2003). As illustrated using the Greenland-

137

O’Rourke method, covariate adjustment most likely implies attenuated and perhaps null
cannabis-depression associations due to the omission of these potentially confounding variables
in the NSDUH measurement plan. Second, difficulties in recall or reporting of the age of onset of
cannabis use or depression spell sometimes can occur, although in this sample there often were
large differences in the age-of-onset of cannabis smoking in relation to age-of-onset of
depression spells, which implies that roughly accurate age-of-onset values should suffice in this
context. Third, the link from cannabis to depression might unfold within a matter of weeks or
months after adolescent onset of cannabis smoking. If so, a research design with week-to-week
or month-to-month time sequence granularity would be required to disclose any palpable
association under these circumstances, and the case-crossover design might be needed (e.g., see
O'Brien et al. 2005). This specific limitation does not undermine the value of the study estimates
just reported, which are based on a prevailing view that early-onset cannabis use might be a
signal of later incident depression spells occurring well after the interval of cannabis intoxication
(e.g., Chen et al. 2002).
Some strengths of this research include having large sample sizes that yield statistically
precise estimates even when modest-level associations are observed, plus the strengths of a
nationally representative sample that helps to promote the external validity of these results, as
well as standardized computer-assisted assessment methods to maximize validity and reliability
of the study measurements. That is, computer-assisted self-interviewing methods may have
helped ensure more honest, more complete, and perhaps more accurate responses to questions on
sensitive topics such as drug use and mood disturbances such as depression spells. These are
important strengths in a cross-sectional survey that is unburdened by the uncertainties that come
with sample attrition during longitudinal follow-up studies on relationships of this type, which

138

otherwise might be said to yield important evidence on a cannabis-depression association of
public health significance.
To conclude with an honest appraisal of the present findings, the accumulated evidence,
and their implications, mention must be made of a prominent assumption of all current
observational research on the suspected hazards associated with early-onset drug use, including
cannabis smoking. The rest of this concluding statement is offered with a spirit of mind
prompted by Professor C.F. Manski's description of a general problem in human behavioral and
social science research of an observational character. Namely, Manski has noted that in these
research areas we often fail "... to face up to the difficulty of [the] enterprise. Researchers
sometimes do not recognize that the interpretation of data requires assumptions. Researchers
sometimes understand the logic of scientific inference but ignore it when reporting their own
work. The scientific community rewards those who produce strong novel findings. The public,
impatient for solutions to its pressing concerns, rewards those who offer simple analyses leading
to unequivocal policy recommendations. These incentives make it tempting for researchers to
maintain assumptions far stronger than they can persuasively defend, in order to draw strong
conclusions" (Manski, 1999).
Accordingly, readers of this work deserve to appreciate a crucial assumption, if the goal
is to evaluate the current evidence and judge that the prevalence of adult-onset depression spells
might be reduced by preventing or delaying onset of cannabis smoking. Namely, as outlined
elsewhere (Anthony, 2013; O’Brien, Comment, Liang, & Anthony, 2012), the crucial
assumption is the 'no omitted variables' assumption in relation to this study's model
specifications. Here, there must be a strong assumption that there is no set of background
variables, lurking behind the scenes, either unknown or unmeasured or ineptly measured, that

139

might be functioning as a common vulnerability trait, accounting for both (a) an occasion of
adolescent-onset cannabis smoking before age 18 years, and (b) an adult-onset depression spell
after the 18th birthday. If there is an omitted variable, such as a gene that manifests its influence
in the form of a simple pleiotropism, as might be the case with a genetic mutation that gives rise
to an excess risk of both early-onset cannabis smoking and a later-onset depression spell, then it
might be controlled in a behavioral genetics research design (e.g., with monozygotic twins).
Nevertheless, if one suspects epigenesis or other forms of gene-environment interaction as a
common substrate, then a subject-as-own-control (SAOC) type design would be required, and
regrettably the primary SAOC design for large sample research is the epidemiologic casecrossover design -- which has little utility when there is a long induction interval from suspected
causal exposure to later excess risk (O’Brien et al., 2012).
Accordingly, in the next step toward confirmation of the suspected cannabis-depression
causal association, it might be best to turn attention toward randomized prevention trial evidence,
where an early drug use prevention program has prevented or delayed the adolescent-onset of
cannabis smoking. With a follow-up of adolescent prevention program participants into the adult
years, there might be evidence of a linked reduction in the form of later reduced risk of the
depression spells among those who responded well to the drug prevention program. A less
compelling alternative design might involve a prospective follow-up of adolescent-onset
cannabis smokers, stratified by whether initial cannabis smoking persisted and for how long it
persisted, with an expectation that there might be reduced risk of the depression spell among
adolescents who tried cannabis just once or twice and never again, or possibly a gradient of
increasing risk across strata defined by length of persistent cannabis smoking in adolescence.
Here again, the 'no omitted variables' assumption would surface, in that the earlier vulnerability

140

to reinforcing effects of cannabis smoking (with subsequent persistence of cannabis use) might
also be a cause or marker of later vulnerability for a depression spell. It is for this reason that the
follow-up of adolescent participants in randomized prevention experiments might be judged to
be a more compelling next step in this line of research.

5.2. Study 2: Cannabis Problems and Functional Impairment Attributed to Recently Active
Depression Spells
My aim for this study was to estimate the association between functional impairment
attributed to depression and the level of problematic cannabis use in a representative sample of
persons with depression in the past year. The guiding hypothesis was that persons experiencing
more cannabis problems would have a higher level of depression-related functional impairment –
a hypothesis that was motivated by the prior literature on the suspected causal relationship
between cannabis and depression. Findings from this study show a positive relationship between
cannabis problems and functional impairment attributed to depression. This association persisted
after accounting for the influence of sex, age, race/ethnicity, alcohol problems, and tobacco
dependence. Constraining the temporal sequencing between onset of cannabis use and onset of
first depression spell did not alter the results. This result mitigated some of the concern that these
findings could due to the ‘self-medication’ hypothesis, which has not been greatly supported by
the prior literature (Hofstra, van der Ende, & Verhulst, 2002; Kandel & Chen, 2000; Kandel &
Davies, 1986; Miller-Johnson, Lochman, Coie, Terry, & Hyman, 1998; Mueser et al., 1998;
Weissman et al., 1999).
Few studies illuminate the relationships linking cannabis smoking, depression, and
functional impairment, but results detailed in this research are consistent with what has been

141

published. For example, a recent study showed cannabis involvement was an independent
predictor of functional impairment among a clinical sample of youths seeking treatment for a
primary mood or anxiety disorder (Osuch et al., 2013). Another study following a sample of
college students found trajectories of increased, late onset, or chronic cannabis use during
college predicted higher depression scores and more days of impairment due to emotions
(Caldeira et al., 2012). Several studies are informative, but more equivocal. For instance, Roberts
and colleagues (2007) reported that in unadjusted models, cannabis dependence and abuse were
both associated with having a mood disorder, but only abuse was associated with more
impairment. Adjusted models did not have sufficient numbers to produce stable estimates.
Meanwhile, Flisher and others (2000) observed that cannabis smoking included in a scale of risk
behaviors (which also included cigarette smoking, alcohol use, sexual intercourse, violence, and
suicide) was independently associated with having a mood disorder and greater functional
impairment. However, over 20 risk factors were evaluated, and there was no indication of
correction for multiple comparisons, so there is the possibility their findings may have been due
to chance. One potentially contrasting finding was reported in a clinical sample of patients who
had recently experienced a manic episode (Goetz et al., 2007). The authors found that other drug
use, but not cannabis or alcohol, was associated with greater occupational impairment.
The current study in the dissertation research program assumed that the functional
impairment experience can be attributed to depression, and that the relationship with cannabis
problems is hypothesized to act through its influence on depressed mood. However, an
alternative explanation of these findings could be that the functional impairment experienced (or
part thereof) may be due to the experience of cannabis problems themselves, which are then
possibly misattributed to depressed mood. Like major depressive episode (MDE), clinically

142

significant impairment or distress is defined as part of the diagnosis of cannabis dependence in
the DSM-IV (American Psychiatric Association, 1994). Cannabis withdrawal is associated with
a clinically significant degree of distress and impairment, and predicts higher levels of cannabis
use after relapse (Allsop et al., 2012; Hasin et al., 2008). Thus, there is the potential for the
'chicken and egg problem' (which came first?), as well as the possibility of feedback mechanisms
that drive both levels of cannabis problems and functional impairment.
Another notable finding was the stronger relationship between the level of tobacco
dependence (TD) and functional impairment attributed to depression, as compared to level of
cannabis problems. Unlike the situation with cannabis, tobacco smoking itself seems unlikely to
directly produce much functional impairment, as many individuals are capable of being regular
smokers without much detrimental effect on occupational or interpersonal roles. Nonetheless,
this finding would be consistent with prior evidence that tobacco smokers are at an increased risk
for depression, and thus might experience more functional impairment attributed to depression
(Breslau, Peterson, Schultz, Chilcoat, & Andreski, 1998). Alternatively, tobacco smoking
inhibits monoamine oxidase A in pre-clinical and clinical samples similar to the action of many
antidepressants. It can be speculated that any observed higher level of tobacco consumption (and
dependence) might have post-dated the onset of depression, contrary to the hypothesized
direction in this investigation, and raising the chicken-egg problem (Fowler, Logan, Wang, &
Volkow, 2003).
To my knowledge, this is the first study to investigate the association between cannabis
problems and functional impairment attributed to depression, rather than diagnostic outcomes or
symptom severity, which are the constructs seen in most prior research on this relationship. In
addition, this study benefits from a number of other strengths. First, the samples used for this

143

research were drawn by sampling non-institutional dwelling units, and thus, the findings are
more likely to generalize to the source population and have increased external validity. Second,
the use of factor analysis and structural equation modeling in this research represent steps
forward beyond what typically has been used in prior studies on the topic of cannabis and
depression. These methods benefit from increased validity and reduced measurement error by
operationalizing constructs to be inferred from multiple observed variables; they allowed the
testing multiple relationships within a complex model (Anderson & Gerbing, 1988).
Despite these strengths, this study possesses a number of limitations. First, the use of
cross-sectional data creates a challenge in ensuring the proper temporal sequence between the
suspected exposure and outcomes, which is crucial if the goal is to infer causal relationships (i.e.,
the chicken-egg problem is unresolved). The time period for the experience of cannabis problems
and functional impairment attributed to depression in this study overlapped the same 12 month
period prior to interview. The study approach partially addressed this issue via exclusion of
exposed cases whose first depression spell preceded or was concurrent with the onset of cannabis
smoking. While no differences in the results were found, the possibility that depression-related
impairment influences problematic cannabis use, or the possibility of feedback mechanisms,
cannot be fully ruled out in these data. Second, this study used the brief measure of functional
impairment, the Sheehan Disability Scale (SDS), which contained only five indicators covering
four life domains. Alternative measures with a larger number of items and covering more life
domains might produce a better measure overall functional impairment, and potentially affect
results. A review by McKnight and Kashdan (2009) listed over 20 different measures of
impairment used in depression research, including at least three measures that have been used
more frequently than the SDS: the Short Form Health Survey (SF-36), the Dyadic Adjustment

144

Scale (DAS), and the Inventory of Interpersonal Problems (IIP) (Horowitz, Rosenberg, Baer,
Ureño, & Villaseñor, 1988; Spanier, 1976; Ware & Sherbourne, 1992). Nevertheless, the
psychometric properties of the SDS (i.e., internal consistency and reliability) compare favorably
to these other measures (McKnight & Kashdan, 2009). Third, the latent dimension of cannabis
problems was only measured in relation to past year experience. Therefore, those with a past
history of cannabis problems, but had quit more than a year prior to interview were treated
equally on the latent trait of cannabis problems as those who never used cannabis. This might
obscure important variation on the potential long-term effects of cannabis problems on later
impairment, and potentially underestimate the overall association. Furthermore, the data lacked
information on cannabis withdrawal, which has since been included under the DSM-V and was a
part of the ICD-10 diagnostic criteria for cannabis use disorders (American Psychiatric
Association, 2013; World Health Organization, 1993). Cannabis withdrawal has been associated
with both decreased mood and increased impairment (Allsop et al., 2012; Budney & Hughes,
2006; Hasin et al., 2008).

5.3. Study 3: Cannabis Problems and History of Blunt Smoking
The aim of this study was to estimate the degree to which a history of cannabis blunt
smoking might account for problematic cannabis use experienced in a large, representative
sample of cannabis users. Findings suggest that higher levels of cannabis problems are associated
with a more recent blunt smoking occurrence and with more frequent days of blunt smoking in
the past month when compared to their non-blunt cannabis-using peers. This result persisted in
statistical models accounting for differences in sex, age, race/ethnicity, population density,
alcohol problems, and tobacco dependence.

145

Findings such as these are concordant with that of Timberlake (2009), who observed a
two-fold excess odds of experiencing a cannabis use disorder for those who primarily smoked
blunts. Another study found frequency of blunt smoking to be uniquely associated with more
cannabis problems, craving, and negative affect (Ream et al., 2008). These results would appear
to contrast qualitative observations of blunt smokers, their rituals, and group consumption norms,
which indicated blunt smoking promoted moderation, rather than excessive consumption
(Dunlap et al., 2006). Further, an experimental study did report that smoking joints produced
higher levels of plasma THC and ratings of intoxication than smoking blunts (Cooper & Haney,
2009). However, the authors disclosed a possible methodological reason for the differences:
participants (who were blindfolded) smoked blunts through a cigarette holder, which may have
made inhaling smoke through cigar paper more difficult compared to the cigarette paper used for
joints.
This study’s finding may also be relevant to the prior literature on the co-use of tobacco
and cannabis. Much of this work has been done on the lifetime co-use or concurrent use (i.e.,
both drug compounds used during the same overlapping time periods). These studies have
generally found tobacco and cannabis co-use to be associated an increased risk for cannabis use
disorders, although there have been a few null findings (Agrawal & Lynskey, 2009; Agrawal et
al., 2009; Coffey, Carlin, Lynskey, Li, & Patton, 2003; Degenhardt & Hall, 2001; Patton, Coffey,
Carlin, Sawyer, & Wakefield, 2006; Swift, Coffey, Carlin, Degenhardt, & Patton, 2008). Studies
that have looked at simultaneous use (i.e., used together or on the same occasion) include one by
Ream and colleagues (2008), which found that those who often ‘chased’ cannabis with tobacco
(i.e., smoked tobacco soon after cannabis) were 25%-50% more likely to experience cannabis
problems. Work by Agrawal and others (2009) also found women who simultaneously used

146

tobacco and cannabis were more likely to be a daily cannabis user, to have used cannabis more
than 40 times, and to have been diagnosed with DSM-IV cannabis abuse, but not with
dependence. However, the degree to which blunt smoking can be generalized in this manner
remains unclear. There has been no systematic study of the nicotine content of blunts after much
of the tobacco has been removed from the cigar. Therefore, the amount of exposure may be low.
Before a more detailed discussion of the study’s findings, it is worth noting some of the
more central limitations. Of central concern, with these cross-sectional data, issues of temporal
sequencing between blunt smoking and cannabis problems cannot be determined with certainty
(i.e., the previously mentioned chicken-egg problem of which came first). Thus, it is possible
that problematic cannabis use could have contributed to the incidence or frequency of blunt use.
Another potential limitation is the degree to which these findings can be generalized to other
populations outside the United States. Blunt smoking appears to be a uniquely American
phenomenon, against a background of USA cannabis smoking in which the typical practice is to
not mix cannabis and tobacco. This USA situation is in contrast with how people in most other
countries consume cannabis (i.e., together with tobacco products), and one might expect these
differences would translate into greater risk of cannabis use disorders in these countries.
However, this appears to not be the case; greater risk is not always seen elsewhere outside the
USA (Perkonigg et al., 1999).
These limitations should be viewed in light of several strengths. First, this study is
perhaps the largest investigation of blunt smoking practices among cannabis users, and as such
the estimates produced benefited from greater precision and external validity to the U.S.
population. Second, blunt smoking appears to be gaining popularity over time as evidenced by
this study and previously by Timberlake (2013). Thus, there is an advantage to using data

147

collected very recently to describe current relationships. Further, given the annual replication of
the NSDUH, it is possible to not only quickly and easily replicate these results, but also to
characterize population changes over time. Third, the coordinated sampling designs and
standardized, computer-assisted survey methods aided in reducing between survey variability
and promoting honest, accurate reporting on drug use behaviors and other sensitive topics.
Finally, it should be mentioned that this research applied more advanced statistical analysis
methods to the measurement and study of cannabis problems (methods not often seen in prior
research on cannabis problems). Dimensional approaches to the study of cannabis problems
provide a complementary approach to the typical diagnostic or categorical approach.

5.4. Study 4: Time to Cannabis Onset and Cannabis Problems
In this cross-national investigation of cannabis-only users, the aim was to study the
degree to which the risk of becoming cannabis dependent could be predicted by how quickly (or
slowly) individuals transitioned from having a chance to try drugs to eventual cannabis onset.
These findings appear to support the hypothesis that those who delay cannabis onset are less
likely to be dependent later on and experience less cannabis problems. My results show that, on
average, each year increase in delaying cannabis onset after the first opportunity might reduce
the risk of cannabis dependence by 20%. An inverse relationship was similarly observed for
cannabis problems measured as a continuous latent trait, suggesting a possibly protective effect
of TCO across the range of problematic cannabis use experience. These results could not be
accounted for by differences in background characteristics (i.e., sex, age, education, income,
marital status, and country), age of first internationally regulated drug (IRD) opportunity, current

148

cigarette smoking status, level of alcohol problems, parent’s education, and lifetime history of
major depressive episode (MDE).
These findings extend the prior work on identifying who are the most vulnerable to
progressing sooner from drug opportunity to use, or from first use to dependence (Chilcoat &
Anthony, 1996; Wagner & Anthony, 2002a, 2002b, 2007). Initiating drug or alcohol use later in
life has been consistently shown to be associated with a lower risk of experiencing disordered
use (Anthony & Petronis, 1995; Behrendt, Wittchen, Höfler, Lieb, & Beesdo, 2009; Breslau,
Kilbey, & Andreski, 1993; C. Y. Chen, Storr, & Anthony, 2009; Grant & Dawson, 1997;
Lynskey et al., 2003). However, this may be the first study to show that delaying cannabis onset
in general, subsequent to the timing of their first chance to use the drug, signals reduced risk of
later cannabis problems. One prior study did report an 8% decrease in the odds of drug abuse for
each year increase in the time from age of first IRD opportunity to use, controlling for age of
first IRD use, type of drug used, and other background factors (Swendsen et al., 2008). A general
adverse effect of rapid drug stage transition was observed in a study of cigarette smokers, where
quickly progressing from less than weekly to weekly smoking signaled an risk of becoming a
daily smoker, and quickly progressing from weekly to daily smoking predicted nicotine
dependence (Dierker et al., 2008).
Several mechanisms, individually or in conjunction, might help explain these findings.
First, time to cannabis onset could serve as a proxy measure for impulsiveness or risktaking/sensation-seeking personality traits, which have been shown to be associated with drug
use and disorders (Clark, Roiser, Robbins, & Sahakian, 2009; Hoyle, Stephenson, Palmgreen,
Lorch, & Donohew, 2002; M. W. Johnson et al., 2010; C. A. Martin et al., 2002; Simons &
Carey, 2002; Whitlow et al., 2004). This subgroup may be more likely to use drugs given the

149

chance, and/or seek out an opportunity to do so. Drug use in turn might affect impulsivity,
initiating possible feedback loops and increasing the risk for disordered drug use (Jentsch &
Taylor, 1999; McDonald et al., 2003). Second, time to cannabis onset might be a function of the
number of cannabis opportunities, which could be affected by environmental and social factors.
For example, living in disadvantaged neighborhoods is associated with higher rates of drug
opportunity and use conditional on opportunity (Crum et al., 1996; Storr, Arria, Workman, &
Anthony, 2004; Storr, Chen, et al., 2004). Parental and peer drug use also appears to have effects
on drug use, although debate on this topic continues (J. S. Brook et al., 1998; Buckner, Crosby,
Silgado, Wonderlich, & Schmidt, 2012; Chabrol, Mabila, Chauchard, Mantoulan, & Rousseau,
2008; Fergusson & Horwood, 1997; Kandel, 1996; Kuntsche & Jordan, 2006).
The observations from this study should be viewed within the light of several important
limitations. When cross-sectional data is used, there is always the concern with selection biases,
in which the most seriously involved cannabis users are left out of the sampling frame. In this
regard, most WMH surveys did not target youths under age 18 years, population members who
lacked a permanent household (e.g., homeless shelters and institutionalized group quarters),
those experiencing premature mortality due to cannabis smoking, and rural populations in some
surveys (i.e., Colombia, Mexico, Japan, Brazil, and China sampled major metropolitan areas).
This research also relied on retrospective self-report data, which is subject to recall bias. While
lifetime indicators of drugs use have good reliability, the degree to which respondents could
accurately recall the age of their first IRD opportunity and/or cannabis use may be of some
concern.

150

With respect to the size of the sample, many WMH countries lacked sufficient numbers
of cannabis-only users and/or cases of cannabis dependence to produce country-specific
estimates.
With respect to the measurement of time to cannabis onset, these estimates would have
been more informative from a specific measure of the age of first cannabis opportunity, rather
than of IRDs in general. However, cannabis tends to be used earlier than other IRDs, and one
might reasonably conclude that it is likely the first IRD persons are given a chance to try.
Further, the time to cannabis onset variable was measured relatively coarse-grained, and thus it
was not possible to probe into potential variation in the risk of later cannabis problems for those
who might have quickly transitioned from first opportunity to cannabis use in a matter of days,
weeks, or months, which might be more informative.

151

CHAPTER 6. CONCLUSIONS
6.1. Cannabis Problems and Functional Impairment Attributed to Depression
The results of this study may have important implications as we seek to account for the
role cannabis smoking might play in exogenous modulation of the natural history of depression.
This research highlights the need and opportunities for investigating alternative clinical
endpoints for depression that have been largely overlooked with respect to cannabis smoking.
Studies that focus exclusively on identifying risk factors for incidence of depression tend to
neglect important issues of severity, duration, recurrence, and impairment that contribute to the
overall burden of depression. Given that the totality of evidence suggests cannabis smoking to be
at best a very modest correlate or predictor of incident depression, now may be time for
researchers focus on alternative explanations to account for co-morbidity findings. Should these
findings be replicated, it may signal that individuals with co-occurring depression and cannabis
use are particularly vulnerable to more debilitating decrements in the home, work, and
interpersonal relationship domains that in turn might exacerbate their depression, and thus be in
more need of effective treatment. Likewise, clinicians might be concerned that depressed patients
experiencing significant impairment, and who might be also using cannabis, may additionally be
in need of treatment for their cannabis smoking. This is illustrated by a study by Green and
colleagues (2011), who found that parents were more likely to be aware of their child’s drug or
alcohol problem if the youth exhibited signs of a high level of functional impairment. General
measures of functional impairment (without asking the person to attribute their impairment to
any specific cause) might serve as a sensitive indicator for an underlying psychiatric problem the
patient may be hesitant to disclose directly, especially if the problem involves illegal drug use.

152

In summary, individuals with depression and co-occurring cannabis problems experience
more functional impairment attributed to their depressed mood. This finding could be the result
of cannabis smoking contributing to more severe depression. Well-designed prospective studies
would be better equipped to demonstrate the causal sequence and development of these two
experiences over time, although there is reason to suspect cannabis smoking having an adverse
influence on depression, rather than the reverse. Future studies will need to determine whether
cannabis smoking worsens impairment in depression, or if impairment related to cannabis
problems is additive, and perhaps misattributed to depression. Perhaps the best evidence might
come from randomized experimental treatment trials, where treatment for either depression or
cannabis problems might disclose whether there is a corresponding improvement in the other.

6.2. Cannabis Problems and History of Blunt Smoking
This research highlights the need for researchers, clinicians, and public health officials
interested in the prevention and treatment of cannabis problems to pay particular attention to the
phenomena of blunt smoking. This study provides novel evidence that blunt smoking may be a
more potent indicator of problematic cannabis smoking as compared to alternative predictors or
covariates such as being male or younger. It may be time to reject notions that all forms of
cannabis use have similar etiology, use patterns, and consequences, and begin to focus on
cannabis subgroups with important distinctions. Qualitative research on blunt smokers has
detailed distinctions in the rituals, normative behavior, motivations for use, social identity,
symbols, language, and group settings that set them apart from previous generations of cannabis
users (Dunlap et al., 2006, 2005; B. D. Johnson et al., 2006; Kelly, 2005; Schensul et al., 2000;
Sifaneck et al., 2005; Soller & Lee, 2010). These factors may have important implications for

153

who among this newer generation start to use cannabis, persist in use, and are at risk for harmful
or problematic use.
Blunt smoking may be indicative of cannabis users who are especially in need of
treatment services. If so, there might be implications for how prevention/treatment messages are
targeted to minority groups. Screening and measurement of individuals in need of treatment
could be problematic if blunt smokers experience a different profile of features compared to
other cannabis users, as in results shown here. Blunts may be perceived to not be addictive, and
if they are perceived as such, the addictive qualities of the preparation might be attributed to the
tobacco content, rather than to the cannabis (Dunlap et al., 2006). Some even report the ritual
surrounding constructing a blunt to be habit-forming (Dunlap et al., 2006). This may lead some
users to disregard the need for treatment after ceasing blunt use, but to continue their cannabis
use, possibly concurrently with cigarette smoking. Even within treatment settings, blunt users
may have more difficulty quitting given evidence that current tobacco use is associated with
poorer treatment outcomes for cannabis use disorders (de Dios, Vaughan, Stanton, & Niaura,
2009; Gray et al., 2011; B. A. Moore & Budney, 2002). Co-treatment for both nicotine and
cannabis dependence may be needed. Additionally, interventions could be better targeted to blunt
smokers by highlighting the potential risks of both cannabis and tobacco, and focusing messages
that appeal to the language, symbols, and culture of blunt smokers.
Public health officials need to be more aware of the role cigars and cigar manufactures
play in the marketing, sale, and distribution of products that promote cannabis consumption,
which particularly appeal to youths, women, and minorities. These groups are more likely to
smoke menthol-flavored cigarettes, which has been shown to affect risk of developing nicotine
dependence and abstinence (Dauphinee, Doxey, Schleicher, Fortmann, & Henriksen, 2013;

154

Reitzel et al., 2013; Rosenbloom, Rees, Reid, Wong, & Kinnunen, 2012; Villanti et al., 2012).
Similarly, flavored cigars and ‘blunt wraps’ may help initiate these groups into cannabis
smoking. However, there have been no quantitative studies on the degree to which blunts are the
first form of cannabis used. Nevertheless, tighter regulation of the sale of cigars used for blunts
may indirectly affect cannabis consumption and problematic use.
In conclusion, cannabis users involved with blunt smoking experience more cannabis
problems than their peers who abstain from blunts. Prospective, longitudinal data would be best
suited to clarify the degree to which prior blunt smoking might be influencing the development
of later cannabis problems. However, the author is unaware of any ongoing prospective studies
of cannabis use that ask separately about blunt smoking. New longitudinal studies may be
required. In order to separate out the role of blunt smoking from non-blunt cannabis use, and it
will also be necessary for questionnaires to be more specific in asking about consumption that is
strictly cannabis-only versus blunts. For instance, the NSDUH asked about cannabis use during
the past month, and then separately about blunt use, but it could not be determined from these
responses whether participants considered days of blunt smoking to also be days of cannabis use.
This point may be important to identify groups who use blunts as their primary or sole mode of
cannabis consumption versus those whose blunt use may be intermittent within their total
cannabis use. Finally, the role blunt smoking plays in the promotion and maintenance of
cigarettes, cigars, and nicotine dependence (and possible feedback mechanisms) are of interest
and require further investigation.

155

6.3. Time to Cannabis Onset and Cannabis Problems
The results of this study may have important implications as we seek to understand the
progression from opportunity, to use, to problematic use of cannabis, and how earlier stages of
the process might predict transitions to later stages. Elaboration of the mechanisms behind why
some drug users progress very quickly from opportunity to use may develop important new
insights into reducing the incidence of cannabis smoking. Conversely, investigations into factors
that delay, but not necessarily prevent cannabis onset, may reap new knowledge about the
development of the cannabis use disorder process.
These findings may also have implications for the design and evaluation of prevention
programs. Drug prevention strategies or programs currently designed for a ‘one-sized fits all’
approach may benefit from targeting individuals at different stages from pre-opportunity to postopportunity. For example, resources might also be better allocated to those who have only
relatively recently been exposed to a chance to try drugs, rather than those who have already
shown a resilience. Further, strategies designed to build cognitive, self-esteem, coping, or peersocial skills targeted to the prevention of drug use (e.g., “Just Say No” campaign or D.A.R.E.),
might additionally benefit from novel strategies that seek to push potential drug experimentation
into later young adult years, where assuming more adult roles and responsibilities could help
prevent drug use escalation. Such strategies could be conceived similar to messages that seek to
delay adolescent sexual intercourse and/or reinforce safe-sex practices.
Finally, this research program concludes that cannabis users who have longer lag times
from the first chance to try drugs to onset of cannabis smoking may be protected from
experiencing later cannabis problems. Future work that might build off these findings include the
following: 1) Studies that measure time from first cannabis opportunity to use in more fine-

156

grained intervals of days, weeks, or months, rather than years; 2) Studies could extend these
findings to other drugs, including alcohol, cocaine, heroin, extra-medical prescription drug use,
ecstasy, etc.; 3) Studies that incorporate the potential influence of prior opportunities and
intervals to use of tobacco and alcohol on time to cannabis onset and later problematic outcomes;
4) Studies that search for subgroup variation by sex or nationality; 5) Studies that investigate
whether time to cannabis onset also affects time from cannabis use to onset of the first cannabis
problem or disordered use syndrome.

157

APPENDIX

158

Table A.1. Spearman rank order correlation matrix between model covariates.

Model Covariate
1 Cannabis Onset
2 Years of Cannabis Involvement
3 Cigarette Onset
4 Alcohol Onset
5 Sex
6 Age
7 Race/Ethnicity
8 Survey Year
9 Education
10 Employment
11 Marital Status
12 # of Times Married
13 Family Income
14 Arrested (Lifetime)
15 Other Illegal Drug Use (Lifetime)
16 Population Density
17 Anxiety Disorder (Lifetime)

1
1.0
0.9
0.3
0.0
-0.1
0.0
-0.1
0.0
0.1
-0.1
0.1
0.1
0.0
0.2
0.4
0.0
-0.1

2

3

4

5

6

7

8

9

10

11

12

1.0
0.3
-0.1
-0.1
-0.1
-0.1
0.0
0.0
-0.1
0.1
0.1
0.0
0.3
0.6
0.0
-0.1

1.0
0.2
-0.1
0.1
-0.1
0.0
0.0
-0.1
0.0
0.0
0.0
0.1
0.2
0.0
-0.1

1.0
0.0
0.1
0.0
0.0
0.1
-0.1
-0.1
-0.1
0.0
0.0
-0.1
0.0
0.0

1.0
0.0
0.0
0.0
0.1
0.2
-0.1
-0.1
-0.1
-0.2
-0.1
0.0
-0.1

1.0
-0.1
0.0
0.1
0.0
-0.6
-0.6
0.2
0.0
-0.1
0.0
-0.2

1.0
0.0
-0.2
0.0
0.1
0.1
-0.2
0.0
-0.1
-0.2
0.1

1.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

1.0
-0.2
-0.1
-0.1
0.3
-0.1
0.0
-0.1
0.0

1.0
0.1
0.1
-0.3
0.0
-0.1
0.0
-0.1

1.0
0.9
-0.3
0.1
0.1
-0.1
0.1

1.0
-0.2
0.1
0.1
-0.1
0.1

159

13 14 15 16 17

1.0
-0.1
0.0
-0.1
0.0

1.0
0.3 1.0
0.0 0.0 1.0
0.0 -0.1 0.0 1.0

REFERENCES

160

REFERENCES

Abel, E. L. (1980). Marihuana: The First Twelve Thousand Years. New York: Plenum Press.
Agrawal, A., & Lynskey, M. T. (2009). Tobacco and cannabis co-occurrence: Does route of
administration matter? Drug and Alcohol Dependence, 99(1-3), 240–247.
doi:10.1016/j.drugalcdep.2008.08.007
Agrawal, A., Lynskey, M. T., Hinrichs, A., Grucza, R., Saccone, S. F., Krueger, R., … Bierut, L.
J. (2011). A genome-wide association study of DSM-IV cannabis dependence. Addiction
Biology, 16(3), 514–518. doi:10.1111/j.1369-1600.2010.00255.x
Agrawal, A., Lynskey, M. T., Madden, P. A. F., Pergadia, M. L., Bucholz, K. K., & Heath, A. C.
(2009). Simultaneous cannabis and tobacco use and cannabis-related outcomes in young
women. Drug and Alcohol Dependence, 101(1-2), 8–12.
doi:10.1016/j.drugalcdep.2008.10.019
Agrawal, A., Madden, P. A. F., Bucholz, K. K., Heath, A. C., & Lynskey, M. T. (2014). Initial
reactions to tobacco and cannabis smoking: a twin study. Addiction (Abingdon, England),
109(4), 663–671. doi:10.1111/add.12449
Agrawal, A., Madden, P. A. F., Martin, N. G., & Lynskey, M. T. (2013). Do early experiences
with cannabis vary in cigarette smokers? Drug and Alcohol Dependence, 128(3), 255–
259. doi:10.1016/j.drugalcdep.2012.09.002
Agrawal, A., Pergadia, M. L., Saccone, S. F., Lynskey, M. T., Wang, J. C., Martin, N. G., …
Madden, P. A. F. (2008a). An autosomal linkage scan for cannabis use disorders in the
nicotine addiction genetics project. Archives of General Psychiatry, 65(6), 713–722.
doi:10.1001/archpsyc.65.6.713
Agrawal, A., Pergadia, M. L., Saccone, S. F., Lynskey, M. T., Wang, J. C., Martin, N. G., …
Madden, P. A. F. (2008b). An autosomal linkage scan for cannabis use disorders in the
nicotine addiction genetics project. Archives of General Psychiatry, 65(6), 713–722.
doi:10.1001/archpsyc.65.6.713
Aldington, S., Harwood, M., Cox, B., Weatherall, M., Beckert, L., Hansell, A., … Beasley, R.
(2008). Cannabis use and risk of lung cancer: a case-control study. European Respiratory
Journal, 31(2), 280–286. doi:10.1183/09031936.00065707
Allsop, D. J., Copeland, J., Norberg, M. M., Fu, S., Molnar, A., Lewis, J., & Budney, A. J.
(2012). Quantifying the Clinical Significance of Cannabis Withdrawal. PLoS ONE, 7(9),
e44864. doi:10.1371/journal.pone.0044864
Alpert, J. E., Maddocks, A., Rosenbaum, J. F., & Fava, M. (1994). Childhood psychopathology
retrospectively assessed among adults with early onset major depression. Journal of
Affective Disorders, 31(3), 165–171.
161

American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental
Disorders (4th Ed.). Washington, DC.
American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental
Disorders (5th Ed.). Washington, DC.
Anderson, J. C., & Gerbing, D. W. (1988). Structural Equation Modeling in Practice - A Review
and Recommended 2-Step Approach. Psychological Bulletin, 103(3), 411–423.
doi:10.1037/0033-2909.103.3.411
Anthony, J. C. (2006). The Epidemiology of Cannabis Dependence. In R. A. Roffman & R. S.
Stephens (Eds.), Cannabis Dependence: Its Nature, Consequences, and Treatment (pp.
58–105). New York: Cambridge University Press.
Anthony, J. C. (2010). Novel phenotype issues raised in cross-national epidemiological research
on drug dependence. Annals of the New York Academy of Sciences, 1187(1), 353–369.
Anthony, J. C. (2013). Understanding Psychiatric Comorbidities and Addictions. In J. MacKillop
& H. de Wit (Eds.), The Wiley-Blackwell Handbook of Addiction Psychopharmacology
(pp. 83–109). Wiley-Blackwell. Retrieved from
http://onlinelibrary.wiley.com/doi/10.1002/9781118384404.ch4/summary
Anthony, J. C., & Petronis, K. R. (1991). Suspected risk factors for depression among adults 1844 years old. Epidemiology, 2(2), 123–132.
Anthony, J. C., & Petronis, K. R. (1995). Early-Onset Drug-Use and Risk of Later Drug
Problems. Drug and Alcohol Dependence, 40(1), 9–15.
Anthony, J. C., & Van Etten, M. L. (1998). Epidemiology and Its Rubrics. In Comprehensive
Clinical Psychology (pp. 355–390). Oxford, U.K.: Elzevier Science Publication.
Arbuckle, R., Frye, M. A., Brecher, M., Paulsson, B., Rajagopalan, K., Palmer, S., & Degl’
Innocenti, A. (2009). The psychometric validation of the Sheehan Disability Scale (SDS)
in patients with bipolar disorder. Psychiatry Research, 165(1–2), 163–174.
doi:10.1016/j.psychres.2007.11.018
Arendt, M., Munk-Jørgensen, P., Sher, L., & Jensen, S. O. W. (2011). Mortality among
individuals with cannabis, cocaine, amphetamine, MDMA, and opioid use disorders: a
nationwide follow-up study of Danish substance users in treatment. Drug and Alcohol
Dependence, 114(2-3), 134–139. doi:10.1016/j.drugalcdep.2010.09.013
Arendt, M., Rosenberg, R., Fjordback, L., Brandholdt, J., Foldager, L., Sher, L., & MunkJorgensen, P. (2007). Testing the self-medication hypothesis of depression and
aggression in cannabis-dependent subjects. Psychological Medicine, 37(7), 935–945.
doi:10.1017/S0033291706009688

162

Arseneault, L., Cannon, M., Witton, J., & Murray, R. M. (2004). Causal association between
cannabis and psychosis: examination of the evidence. British Journal of Psychiatry,
184(2), 110.
Ashton, C. (2001). Pharmacology and effects of cannabis: a brief review. British Journal of
Psychiatry, 178, 101–106.
Baba, T., Ganai, A., Qadri, S., Margoob, M., Iqbal, Q., & Khan, Z. (2013). An Epidemiological
study on Substance Abuse among college students of north India Kashmir valley.
International Journal of Medical Science and Public Health, 2(3), 540.
doi:10.5455/ijmsph.2013.080420131
Babor, T. F. (2006). The Diagnosis of Cannabis Dependence. In R. A. Roffman & R. S. Stephens
(Eds.), Cannabis Dependence: Its Nature, Consequences, and Treatment (pp. 21–36).
New York: Cambridge University Press.
Bambico, F. R., Duranti, A., Tontini, A., Tarzia, G., & Gobbi, G. (2009). Endocannabinoids in
the Treatment of Mood Disorders: Evidence from Animal Models. Current
Pharmaceutical Design, 15(14), 1623–1646.
Bambico, F. R., Nguyen, N. T., Katz, N., & Gobbi, G. (2010). Chronic exposure to cannabinoids
during adolescence but not during adulthood impairs emotional behaviour and
monoaminergic neurotransmission. Neurobiology of Disease, 37(3), 641–655.
Bardone, A. M., Moffitt, T. E., Caspi, A., Dickson, N., Stanton, W. R., & Silva, P. A. (1998).
Adult physical health outcomes of adolescent girls with conduct disorder, depression, and
anxiety. Journal of the American Academy of Child & Adolescent Psychiatry, 37(6), 594–
601.
Beautrais, A. L., Joyce, P. R., & Mulder, R. T. (1999). Cannabis abuse and serious suicide
attempts. Addiction, 94(8), 1155–1164.
Bedard, M., Dubois, S., & Weaver, B. (2007). The impact of cannabis on driving. Canadian
Journal of Public Health-Revue Canadienne De Sante Publique, 98(1), 6–11.
Behrendt, S., Wittchen, H.-U., Höfler, M., Lieb, R., & Beesdo, K. (2009). Transitions from first
substance use to substance use disorders in adolescence: Is early onset associated with a
rapid escalation? Drug and Alcohol Dependence, 99(1-3), 68–78.
doi:10.1016/j.drugalcdep.2008.06.014
Benjet, C., Borges, G., Medina-Mora, M. E., Blanco, J., Zambrano, J., Orozco, R., … Rojas, E.
(2007). Drug use opportunities and the transition to drug use among adolescents from the
Mexico City metropolitan area. Drug and Alcohol Dependence, 90(2-3), 128–134.
doi:10.1016/j.drugalcdep.2007.02.018
Berthiller, J., Straif, K., Boniol, M., Voirin, N., Benhaim-Luzon, V., Ben Ayoub, W., … Sasco,
A. (2008). Cannabis Smoking and Risk of Lung Cancer in Men A Pooled Analysis of
Three Studies in Maghreb. Journal of Thoracic Oncology, 3(12), 1398–1403.
163

Bey, H., & Zug (Eds.). (2004). Orgies of the Hemp Eaters: Cuisine, Slang, Literature, & Ritual
of Cannabis Culture. Brooklyn, NY: Autonomedia.
Black, B. (2012, October 2). To Dab Or Not To Dab? High Times. Retrieved April 28, 2014,
from http://www.hightimes.com/read/dab-or-not-dab
Blazer, D. G., Kessler, R. C., McGonagle, K. A., & Swartz, M. S. (1994). The Prevalence and
Distribution of Major Depression in a National Community Sample - The National
Comorbidity Survey. American Journal of Psychiatry, 151(7), 979–986.
Blows, S., Ivers, R. Q., Connor, J., Ameratunga, S., Woodward, M., & Norton, R. (2005).
Marijuana use and car crash injury. Addiction, 100(5), 605–611. doi:10.1111/j.13600443.2005.01100.x
Bobashev, G. V., & Anthony, J. C. (1998). Clusters of marijuana use in the United States.
American Journal of Epidemiology, 148(12), 1168–1174.
Bobashev, G. V., & Anthony, J. C. (2000). Use of alternating logistic regression in studies of
drug-use clustering. Substance Use & Misuse, 35(6-8), 1051–1073.
doi:10.3109/10826080009148432
Boden, J. M., & Fergusson, D. M. (2011). Alcohol and depression. Addiction, 106(5), 906–914.
Bohnert, K. M., Anthony, J. C., & Breslau, N. (2012). Parental Monitoring at Age 11 and
Subsequent Onset of Cannabis Use Up to Age 17: Results From a Prospective Study.
Journal of Studies on Alcohol and Drugs, 73(2), 173–177.
Bovasso, G. B. (2001). Cannabis abuse as a risk factor for depressive symptoms. American
Journal of Psychiatry, 158(12), 2033–2037.
Breslau, N., Kilbey, M. M., & Andreski, P. (1993). Nicotine dependence and major depression:
new evidence from a prospective investigation. Archives of General Psychiatry, 50(1),
31.
Breslau, N., Peterson, E. L., Schultz, L. R., Chilcoat, H. D., & Andreski, P. (1998). Major
depression and stages of smoking - A longitudinal investigation. Archives of General
Psychiatry, 55(2), 161–166. doi:10.1001/archpsyc.55.2.161
Briscoe, C. W., & Smith, J. B. (1973). Depression and Marital Turmoil. Archives of General
Psychiatry, 29(6), 811–817.
Brook, D. W., Brook, J. S., Zhang, C., Cohen, P., & Whiteman, M. (2002). Drug use and the risk
of major depressive disorder, alcohol dependence, and substance use disorders. Archives
of General Psychiatry, 59(11), 1039.
Brook, J. S., Adams, R. E., Balka, E. B., & Johnson, E. (2002). Early adolescent marijuana use:
risks for the transition to young adulthood. Psychological Medicine, 32(1), 79–91.

164

Brook, J. S., Brook, D. W., De la Rosa, M., Duque, L. F., Rodriguez, E., Montoya, I. D., &
Whiteman, M. (1998). Pathways to marijuana use among adolescents:
Cultural/ecological, family, peer, and personality influences. Journal of the American
Academy of Child and Adolescent Psychiatry, 37(7), 759–766. doi:10.1097/00004583199807000-00016
Brook, J. S., Lee, J. Y., Brown, E. N., Finch, S. J., & Brook, D. W. (2011). Developmental
trajectories of marijuana use from adolescence to adulthood: personality and social role
outcomes. Psychological Reports, 108(2), 339–357.
Browne, M. W., Cudeck, R., Bollen, K. A., & Long, J. S. (1993). Alternative ways of assessing
model fit. Sage Focus Editions, 154, 136–136.
Brylcreem. (2014, April 26). In Wikipedia, the free encyclopedia. Retrieved from
http://en.wikipedia.org/w/index.php?title=Brylcreem&oldid=590493567
Buckner, J. D., Crosby, R. D., Silgado, J., Wonderlich, S. A., & Schmidt, N. B. (2012).
Immediate antecedents of marijuana use: An analysis from ecological momentary
assessment. Journal of Behavior Therapy and Experimental Psychiatry, 43(1), 647–655.
doi:10.1016/j.jbtep.2011.09.010
Budney, A. J., Higgins, S. T., Radonovich, K. J., & Novy, P. L. (2000). Adding voucher-based
incentives to coping skills and motivational enhancement improves outcomes during
treatment for marijuana dependence. Journal of Consulting and Clinical Psychology,
68(6), 1051–1061. doi:10.1037/0022-006X.68.6.1051
Budney, A. J., & Hughes, J. R. (2006). The cannabis withdrawal syndrome. Current Opinion in
Psychiatry, 19(3), 233–238. doi:10.1097/01.yco.0000218592.00689.e5
Budney, A. J., Hughes, J. R., Moore, B. A., & Vandrey, R. (2004). Review of the validity and
significance of cannabis withdrawal syndrome. American Journal of Psychiatry, 161(11),
1967–1977. doi:10.1176/appi.ajp.161.11.1967
Budney, A. J., Moore, B. A., Rocha, H. L., & Higgins, S. T. (2006). Clinical trial of abstinencebased vouchers and cognitive-behavioral therapy for cannabis dependence. Journal of
Consulting and Clinical Psychology, 74(2), 307–316. doi:10.1037/0022-006X.74.2.307
Burns, C. B., Ivers, R. G., Lindorff, K. J., & Clough, A. R. (2000). Cannabis: a Trojan horse for
nicotine? Australian and New Zealand Journal of Public Health, 24(6), 637–637.
doi:10.1111/j.1467-842X.2000.tb00533.x
Caldeira, K. M., O’Grady, K. E., Vincent, K. B., & Arria, A. M. (2012). Marijuana use
trajectories during the post-college transition: Health outcomes in young adulthood. Drug
and Alcohol Dependence, 125(3), 267–275. doi:10.1016/j.drugalcdep.2012.02.022
Callaghan, R. C., Allebeck, P., & Sidorchuk, A. (2013). Marijuana use and risk of lung cancer: a
40-year cohort study. Cancer Causes & Control, 24(10), 1811–1820.
doi:10.1007/s10552-013-0259-0
165

Callaghan, R. C., Gatley, J. M., Veldhuizen, S., Lev-Ran, S., Mann, R., & Asbridge, M. (2013).
Alcohol- or drug-use disorders and motor vehicle accident mortality: A retrospective
cohort study. Accident Analysis and Prevention, 53, 149–155.
doi:10.1016/j.aap.2013.01.008
Caplan, A. L. (2006). Ethical issues surrounding forced, mandated, or coerced treatment. Journal
of Substance Abuse Treatment, 31(2), 117–120. doi:10.1016/j.jsat.2006.06.009
Carroll, K. M., Easton, C. J., Nich, C., Hunkele, K. A., Neavins, T. M., Sinha, R., …
Rounsaville, B. J. (2006). The use of contingency management and motivational/skillsbuilding therapy to treat young adults with marijuana dependence. Journal of Consulting
and Clinical Psychology, 74(5), 955–966. doi:10.1037/0022-006X.74.5.955
Castañé, A., Valjent, E., Ledent, C., Parmentier, M., Maldonado, R., & Valverde, O. (2002).
Lack of CB1 cannabinoid receptors modifies nicotine behavioural responses, but not
nicotine abstinence. Neuropharmacology, 43(5), 857–867.
Center for Behavioral Health and Statistics and Quality. (2012). Comparison of NSDUH Mental
Health Data and Methods with Other Data Sources. Substance Abuse and Mental Health
Services Administration.
Centers for Disease Control and Prevention. (2012). Youth Risk Behavior Surveillance - United
States, 2011 (No. 61, No. SS-4). Atlanta, GA.
Chabrol, H., Mabila, J. D., Chauchard, E., Mantoulan, R., & Rousseau, A. (2008). Contributions
of parental and social influences to cannabis use in a non-clinical sample of adolescents.
Encephale-Revue De Psychiatrie Clinique Biologique Et Therapeutique, 34(1), 8–16.
doi:10.1016/j.encep.2007.01.002
Chait, L. D., & Zacny, J. P. (1992). Reinforcing and subjective effects of oral Δ9-THC and
smoked marijuana in humans. Psychopharmacology, 107(2-3), 255–262.
Chaiton, M. O., Cohen, J. E., O’Loughlin, J., & Rehm, J. (2009). A systematic review of
longitudinal studies on the association between depression and smoking in adolescents.
BMC Public Health, 9, 356. doi:10.1186/1471-2458-9-356
Chen, A., Chen, T., Braverman, E., Acuri, V., Kerner, M., Varshavskiy, M., … Blum, K. (2008).
Hypothesizing that Marijuana Smokers are at a Significantly Lower Risk of
Carcinogenicity Relative to Tobacco-Non-Marijuana Smokers: Evidenced Based on
Statistical Reevaluation of Current Literature. Journal of Psychoactive Drugs, 40(3),
263–272.
Chen, C. Y., Dormitzer, C. M., Bejarano, J., & Anthony, J. C. (2004). Religiosity and the earliest
stages of adolescent drug involvement in seven countries of Latin America. American
Journal of Epidemiology, 159(12), 1180–1188.

166

Chen, C. Y., Dormitzer, C. M., Gutierrez, U., Vittetoe, K., Gonzalez, G. B., & Anthony, J. C.
(2004). The adolescent behavioral repertoire as a context for drug exposure: behavioral
autarcesis at play. Addiction, 99(7), 897–906. doi:10.1111/j.1360-0443.2004.00774.x
Chen, C. Y., O’Brien, M. S., & Anthony, J. C. (2005). Who becomes cannabis dependent soon
after onset of use? Epidemiological evidence from the United States: 2000-2001. Drug
and Alcohol Dependence, 79(1), 11.
Chen, C. Y., Storr, C. L., & Anthony, J. C. (2005). Influences of parenting practices on the risk
of having a chance to try cannabis. Pediatrics, 115(6), 1631–1639.
Chen, C. Y., Storr, C. L., & Anthony, J. C. (2009). Early-onset drug use and risk for drug
dependence problems. Addictive Behaviors, 34(3), 319–322.
doi:10.1016/j.addbeh.2008.10.021
Chen, C. Y., Wagner, F. A., & Anthony, J. C. (2002). Marijuana use and the risk of Major
Depressive Episode Epidemiological evidence from the United States National
Comorbidity Survey. Social Psychiatry and Psychiatric Epidemiology, 37(5), 199–206.
Chilcoat, H. D., & Anthony, J. C. (1996). Impact of parent monitoring on initiation of drug use
through late childhood. Journal of the American Academy of Child and Adolescent
Psychiatry, 35(1), 91–100. doi:10.1097/00004583-199601000-00017
Clark, L., Roiser, J. P., Robbins, T. W., & Sahakian, B. J. (2009). Disrupted “reflection”
impulsivity in cannabis users but not current or former ecstasy users. Journal of
Psychopharmacology, 23(1), 14–22.
Cochran, J. K., Wood, P. B., & Arneklev, B. J. (1994). Is the Religiosity-Delinquency
Relationship Spurious? A Test of Arousal and Social Control Theories. Journal of
Research in Crime and Delinquency, 31(1), 92–123. doi:10.1177/0022427894031001004
Coffey, C., Carlin, J. B., Lynskey, M., Li, N., & Patton, G. C. (2003). Adolescent precursors of
cannabis dependence: findings from the Victorian Adolescent Health Cohort Study. The
British Journal of Psychiatry, 182(4), 330–336.
Colman, I., & Ataullahjan, A. (2010). Life Course Perspectives on the Epidemiology of
Depression. Canadian Journal of Psychiatry-Revue Canadienne De Psychiatrie, 55(10),
622–632.
Compton, W. M., Grant, B. F., Colliver, J. D., Glantz, M. D., & Stinson, F. S. (2004). Prevalence
of marijuana use disorders in the United States - 1991-1992 and 2001-2002. JamaJournal of the American Medical Association, 291(17), 2114–2121.
doi:10.1001/jama.291.17.2114
Cooper, Z., & Haney, M. (2009). Comparison of subjective, pharmacokinetic, and physiological
effects of marijuana smoked as joints and blunts. Drug and Alcohol Dependence, 103(3),
107–113. doi:10.1016/j.drugalcdep.2009.01.023

167

Copeland, J., & Maxwell, J. C. (2007). Cannabis treatment outcomes among legally coerced and
non-coerced adults. BMC Public Health, 7(1), 111. doi:10.1186/1471-2458-7-111
Copeland, J., Swift, W., Roffman, R., & Stephens, R. (2001). A randomized controlled trial of
brief cognitive-behavioral interventions for cannabis use disorder. Journal of Substance
Abuse Treatment, 21(2), 55–64. doi:10.1016/S0740-5472(01)00179-9
Crowley, T. J., Macdonald, M. J., Whitmore, E. A., & Mikulich, S. K. (1998). Cannabis
dependence, withdrawal, and reinforcing effects among adolescents with conduct
symptoms and substance use disorders. Drug and Alcohol Dependence, 50(1), 27–37.
doi:10.1016/S0376-8716(98)00003-9
Crum, R. M., Lillie-Blanton, M., & Anthony, J. C. (1996). Neighborhood environment and
opportunity to use cocaine and other drugs in late childhood and early adolescence. Drug
and Alcohol Dependence, 43(3), 155–161. doi:10.1016/S0376-8716(96)01298-7
Dauphinee, A. L., Doxey, J. R., Schleicher, N. C., Fortmann, S. P., & Henriksen, L. (2013).
Racial differences in cigarette brand recognition and impact on youth smoking. BMC
Public Health, 13, 170. doi:10.1186/1471-2458-13-170
Davstad, I., Allebeck, P., Leifman, A., Stenbacka, M., & Romelsjö, A. (2011). Self-reported drug
use and mortality among a nationwide sample of Swedish conscripts - a 35-year followup. Drug and Alcohol Dependence, 118(2-3), 383–390.
doi:10.1016/j.drugalcdep.2011.04.025
Day, A. M., Metrik, J., Spillane, N. S., & Kahler, C. W. (2013). Working memory and
impulsivity predict marijuana-related problems among frequent users. Drug and Alcohol
Dependence, 131(1-2), 171–174. doi:10.1016/j.drugalcdep.2012.12.016
De Dios, M. A., Vaughan, E. L., Stanton, C. A., & Niaura, R. (2009). Adolescent tobacco use
and substance abuse treatment outcomes. Journal of Substance Abuse Treatment, 37(1),
17–24.
De Graaf, R., Radovanovic, M., van Laar, M., Fairman, B., Degenhardt, L., Aguilar-Gaxiola, S.,
… Anthony, J. C. (2010). Early cannabis use and estimated risk of later onset of
depression spells. Epidemiological evidence from the population-based WHO World
Mental Health Survey Initiative. American Journal of Epidemiology, 172(2), 149–159.
Degenhardt, L., Chiu, W. T., Sampson, N., Kessler, R. C., Anthony, J. C., Angermeyer, M., …
Huang, Y. (2008). Toward a global view of alcohol, tobacco, cannabis, and cocaine use:
Findings from the WHO World Mental Health Surveys. Public Library of Science
Medicine, 5(7), 1053–1067.
Degenhardt, L., Coffey, C., Carlin, J. B., Swift, W., Moore, E., & Patton, G. C. (2010).
Outcomes of occasional cannabis use in adolescence: 10-year follow-up study in
Victoria, Australia. The British Journal of Psychiatry: The Journal of Mental Science,
196(4), 290–295. doi:10.1192/bjp.bp.108.056952
168

Degenhardt, L., Coffey, C., Romaniuk, H., Swift, W., Carlin, J. B., Hall, W. D., & Patton, G. C.
(2013). The persistence of the association between adolescent cannabis use and common
mental disorders into young adulthood. Addiction, 108(1), 124–133. doi:10.1111/j.13600443.2012.04015.x
Degenhardt, L., & Hall, W. D. (2001). The relationship between tobacco use, substance-use
disorders and mental health: results from the National Survey of Mental Health and Wellbeing. Nicotine & Tobacco Research, 3(3), 225–234.
Degenhardt, L., Hall, W., & Lynskey, M. (2001). The relationship between cannabis use,
depression and anxiety among Australian adults: findings from the National Survey of
Mental Health and Well-Being. Social Psychiatry and Psychiatric Epidemiology, 36(5),
219–227.
Degenhardt, L., Hall, W., & Lynskey, M. (2003). Exploring the association between cannabis
use and depression. Addiction, 98(11), 1493–1504.
Delva, J., Van Etten, M. L., Gonzalez, G. B., Cedeno, M. A., Penna, M., Caris, L. H., &
Anthony, J. C. (1999). First opportunities to try drugs and the transition to first drug use:
Evidence from a national school survey in Panama. Substance Use & Misuse, 34(10),
1451–1467. doi:10.3109/10826089909029392
Dennis, M., Godley, S. H., Diamond, G., Tims, F. M., Babor, T., Donaldson, J., … Funk, R.
(2004). The Cannabis Youth Treatment (CYT) Study: Main findings from two
randomized trials. Journal of Substance Abuse Treatment, 27(3), 197–213.
doi:10.1016/j.jsat.2003.09.005
Desbois, A. C., & Cacoub, P. (2013). Cannabis-Associated Arterial Disease. Annals of Vascular
Surgery, 27(7), 996–1005. doi:10.1016/j.avsg.2013.01.002
Devane, W. A., Dysarz, F. A., & others. (1988). Determination and characterization of a
cannabinoid receptor in rat brain. Molecular Pharmacology, 34(5), 605.
Devane, W. A., Hanus, L., Breuer, A., Pertwee, R. G., Stevenson, L. A., Griffin, G., …
Mechoulam, R. (1992). Isolation and structure of a brain constituent that binds to the
cannabinoid receptor. Science (New York, NY), 258(5090), 1946.
Dierker, L., He, J., Kalaydjian, A., Swendsen, J., Degenhardt, L., Glantz, M., … Merikangas, K.
(2008). The Importance of Timing of Transitions for Risk of Regular Smoking and
Nicotine Dependence. Annals of Behavioral Medicine : A Publication of the Society of
Behavioral Medicine, 36(1), 87–92. doi:10.1007/s12160-008-9051-x
Dishion, T., & Loeber, R. (1985). Adolescent Marijuana and Alcohol-Use - the Role of Parents
and Peers Revisited. American Journal of Drug and Alcohol Abuse, 11(1-2), 11–25.
doi:10.3109/00952998509016846

169

Dornbusch, S. M., Erickson, K. G., Laird, J., & Wong, C. A. (2001). The relation of family and
school attachment to adolescent deviance in diverse groups and communities. Journal of
Adolescent Research, 16(4), 396–422. doi:10.1177/0743558401164006
Drummer, O. H., Gerostamoulos, J., Batziris, H., Chu, M., Caplehorn, J. R. M., Robertson, M.
D., & Swann, P. (2003). The incidence of drugs in drivers killed in Australian road traffic
crashes. Forensic Science International, 134(2-3), 154–162. doi:10.1016/S03790738(03)00134-8
Dunlap, E., Benoit, E., Sifaneck, S., & Johnson, B. (2006). Social constructions of dependency
by blunts smokers: Qualitative reports. INTERNATIONAL JOURNAL OF DRUG
POLICY, 17(3), 171–182. doi:10.1016/j.drugpo.2006.01.004
Dunlap, E., Johnson, B. D., Benoit, E., & Sifaneck, S. (2005). Sessions, cyphers, and parties:
settings for informal social controls of blunt smoking. Journal of Ethnicity in Substance
Abuse, 4(3-4).
Edwards, G., Arif, A., & Hodgson, R. (1981). Nomenclature and Classification of Drug-Related
and Alcohol-Related Problems - a Who Memorandum. Bulletin of the World Health
Organization, 59(2), 225–242.
Ehlers, C. L., Gizer, I. R., Vieten, C., & Wilhelmsen, K. C. (2010). Linkage Analyses of
Cannabis Dependence, Craving, and Withdrawal in the San Francisco Family Study.
American Journal of Medical Genetics Part B-Neuropsychiatric Genetics, 153B(3), 802–
811. doi:10.1002/ajmg.b.31050
Elliott, D. S., Huizinga, D., & Ageton, S. S. (1985). Explaining delinquency and drug use.
Beverly Hills, CA: Sage. Retrieved from http://www.getcited.org/pub/102380429
Elliott, D. S., Huizinga, D., & Menard, S. (1989). Multiple problem youth: Delinquency,
substance use, and mental health problems. In This book is an extended version of a
paper presented at a research conference held in conjunction with the Alcohol, Drug
Abuse, Mental Health Administration and the Office of Juvenile Justice and Delinquency
Prevention in April 1984. Retrieved from http://doi.apa.org/psycinfo/1989-98128-000
Fairman, B. J., & Anthony, J. C. (2012). Are early-onset cannabis smokers at an increased risk of
depression spells? Journal of Affective Disorders, 138(1-2), 54–62.
doi:10.1016/j.jad.2011.12.031
Fergusson, D. M., Boden, J., & Horwood, L. (2008). The developmental antecedents of illicit
drug use: Evidence from a 25-year longitudinal study. DRUG AND ALCOHOL
DEPENDENCE, 96(1-2), 165–177. doi:10.1016/j.drugalcdep.2008.03.003
Fergusson, D. M., & Boden, J. M. (2008). Cannabis use and later life outcomes. Addiction,
103(6), 969–976. doi:10.1111/j.1360-0443.2008.02221.x
Fergusson, D. M., & Horwood, L. J. (1997). Early onset cannabis use and psychosocial
adjustment in young adults. Addiction, 92(3), 279–296.
170

Fergusson, D. M., Horwood, L. J., & Boden, J. M. (2008). Is driving under the influence of
cannabis becoming a greater risk to driver safety than drink driving? Findings from a
longitudinal study. Accident; Analysis and Prevention, 40(4), 1345–1350.
doi:10.1016/j.aap.2008.02.005
Fergusson, D. M., Horwood, L. J., & Swain-Campbell, N. (2002). Cannabis use and
psychosocial adjustment in adolescence and young adulthood. Addiction, 97(9), 1123–
1135.
Fergusson, D. M., Lynskey, M. T., & Horwood, L. J. (1996). The short-term consequences of
early onset cannabis use. Journal of Abnormal Child Psychology, 24(4), 499–512.
Ferrari, A. J., Charlson, F. J., Norman, R. E., Patten, S. B., Freedman, G., Murray, C. J. L., …
Whiteford, H. A. (2013). Burden of Depressive Disorders by Country, Sex, Age, and
Year: Findings from the Global Burden of Disease Study 2010. PLoS Med, 10(11),
e1001547. doi:10.1371/journal.pmed.1001547
Fincham, F. D., Beach, S. R., Harold, G. T., & Osborne, L. N. (1997). Marital satisfaction and
depression: Different causal relationships for men and women? Psychological Science,
8(5), 351–356.
Fleming, C. B., Mason, W. A., Mazza, J. J., Abbott, R. D., & Catalano, R. F. (2008). Latent
growth modeling of the relationship between depressive symptoms and substance use
during adolescence. Psychology of Addictive Behaviors: Journal of the Society of
Psychologists in Addictive Behaviors, 22(2), 186–197. doi:10.1037/0893-164X.22.2.186
Fletcher, J. M. (2008). Adolescent depression: diagnosis, treatment, and educational attainment.
Health Economics, 17(11), 1215–1235. doi:10.1002/hec.1319
Flisher, A. J., Kramer, R. A., Hoven, C. W., King, R. A., Bird, H. R., Davies, M., … Shaffer, D.
(2000). Risk behavior in a community sample of children and adolescents. Journal of the
American Academy of Child and Adolescent Psychiatry, 39(7), 881–887.
doi:10.1097/00004583-200007000-00017
Fombonne, E. (1994). Increased Rates of Depression - Update of Epidemiologic Findings and
Analytical Problems. Acta Psychiatrica Scandinavica, 90(3), 145–156.
doi:10.1111/j.1600-0447.1994.tb01571.x
Fournier, J. C., DeRubeis, R. J., Hollon, S. D., Dimidjian, S., Amsterdam, J. D., Shelton, R. C.,
& Fawcett, J. (2010). Antidepressant Drug Effects and Depression Severity A PatientLevel Meta-analysis. Jama-Journal of the American Medical Association, 303(1), 47–53.
Fowler, J. S., Logan, J., Wang, G.-J., & Volkow, N. D. (2003). Monoamine Oxidase and
Cigarette Smoking. NeuroToxicology, 24(1), 75–82. doi:10.1016/S0161-813X(02)001092
Gaoni, Y., & Mechoulam, R. (1964). Isolation, structure, and partial synthesis of an active
constituent of hashish. Journal of the American Chemical Society, 86(8), 1646–1647.
171

Gardner, E. L., & Vorel, S. R. (1998). Cannabinoid transmission and reward-related events.
Neurobiology of Disease, 5(6), 502–533. doi:10.1006/nbdi.1998.0219
Georgiades, K., & Boyle, M. H. (2007). Adolescent tobacco and cannabis use: young adult
outcomes from the Ontario Child Health Study. Journal of Child Psychology and
Psychiatry, 48(7), 724–731.
Gfroerer, J. C., Wu, L.-T., & Penne, M. A. (2002). Initiation of Marijuana Use: Trends,
Patterns, and Implications. Rockville, MD: Substance Abuse and Mental Health Services
Administration, Office of Applied Studies.
Gobbi, G., Bambico, F. R., Mangieri, R., Bortolato, M., Campolongo, P., Solinas, M., …
Duranti, A. (2005). Antidepressant-like activity and modulation of brain monoaminergic
transmission by blockade of anandamide hydrolysis. Proceedings of the National
Academy of Sciences of the United States of America, 102(51), 18620.
Goel, A., & Chakrabarti, A. (2010). Prevalence and socio-demographic correlates of substance
use in a rural community in Sikkim, North East India: Results from a pilot population
survey. Journal of Substance Use, 15(1), 13–23. doi:10.3109/14659890902964005
Goetz, I., Tohen, M., Reed, C., Lorenzo, M., Vieta, E., & Board, the E. A. (2007). Functional
impairment in patients with mania: baseline results of the EMBLEM study. Bipolar
Disorders, 9(1-2), 45–52. doi:10.1111/j.1399-5618.2007.00325.x
Gold, M. S. (1991). Marijuana. In N. S. Miller (Ed.), Comprehensive Handbook of Drug and
Alcohol Addiction (pp. 353–382). New York: Marcel Dekker.
Golub, A., & Johnson, B. D. (1999). Cohort changes in illegal drug use among arrestees in
Manhattan: From the heroin injection generation to the blunts generation. Substance Use
& Misuse, 34(13), 1733–1763. doi:10.3109/10826089909039425
Golub, A., Johnson, B. D., & Dunlap, E. (2005). The growth in marijuana use among American
youths during the 1990s and the extent of blunt smoking. Journal of Ethnicity in
Substance Abuse, 4(3-4), 1–21.
Golub, A., Johnson, B. D., Dunlap, E., & Sifaneck, S. (2004). Projecting and monitoring the life
course of the marijuana/blunts generation. Journal of Drug Issues, 34(2), 361–388.
Goode, E. (1989). Drugs in American society. Knopf New York. Retrieved from
https://www.ncjrs.gov/App/Publications/abstract.aspx?ID=257689
Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford University Press.
Retrieved from
http://books.google.com/books?hl=en&lr=&id=loNPs7n94p0C&oi=fnd&pg=PR13&dq=
Gottfredson+Hirschi+General+Theory+of+Crime&ots=H4rx5Jn_3&sig=nLzQZ5UcSpswH8MKzU475RNQ1m0

172

Grant, B. F. (1995). Comorbidity between DSM-IV drug use disorders and major depression:
results of a national survey of adults. Journal of Substance Abuse, 7(4), 481–497.
Grant, B. F., & Dawson, D. A. (1997). Age at onset of alcohol use and its association with DSMIV alcohol abuse and dependence: Results from the National Longitudinal Alcohol
Epidemiologic Survey. Journal of Substance Abuse, 9, 103–110. doi:10.1016/S08993289(97)90009-2
Grant, B. F., & Pickering, R. (1998). The relationship between cannabis use and DSM-IV
cannabis abuse and dependence: Results from the national longitudinal alcohol
epidemiologic survey. Journal of Substance Abuse, 10(3), 255–264.
Gray, K. M., Riggs, P. D., Min, S.-J., Mikulich-Gilbertson, S. K., Bandyopadhyay, D., &
Winhusen, T. (2011). Cigarette and cannabis use trajectories among adolescents in
treatment for attention-deficit/hyperactivity disorder and substance use disorders. Drug
and Alcohol Dependence, 117(2), 242–247.
Green, A. E., Bekman, N. M., Miller, E. A., Perrott, J. A., Brown, S. A., & Aarons, G. A. (2011).
Parental Awareness of Substance Use Among Youths in Public Service Sectors. Journal
of Studies on Alcohol and Drugs, 72(1), 44–52.
Green, B. E., & Ritter, C. (2000). Marijuana use and depression. Journal of Health and Social
Behavior, 41(1), 40–49.
Green, B., Kavanagh, D., & Young, R. (2003). Being stoned: a review of self-reported cannabis
effects. Drug and Alcohol Review, 22(4), 453–460. doi:10.1080/09595230310001613976
Greenberg, P. E., Kessler, R. C., Birnbaum, H. G., Leong, S. A., Lowe, S. W., Berglund, P. A.,
& Corey-Lisle, P. K. (2003). The economic burden of depression in the United States:
How did it change between 1990 and 2000?. Journal of Clinical Psychiatry, 64(12),
1465–1475.
Greenland, S., & O’Rourke, K. (2008). Meta-Analysis. In Modern Epidemiology. Lippincott
Williams & Wilkins.
Griffith-Lendering, M. F. H., Huijbregts, S. C. J., Mooijaart, A., Vollebergh, W. A. M., &
Swaab, H. (2011). Cannabis use and development of externalizing and internalizing
behaviour problems in early adolescence: A TRAILS study. Drug and Alcohol
Dependence, 116(1-3), 11–17. doi:10.1016/j.drugalcdep.2010.11.024
Groth, S. W., & Morrison-Beedy, D. (2010). Smoking, Substance Use, and Mental Health
Correlates in Urban Adolescent Girls. Journal of Community Health, 36(4), 552–558.
doi:10.1007/s10900-010-9340-8
Gupta, S., Sarpal, S. S., Kumar, D., Kaur, T., & Arora, S. (2013). Prevalence, Pattern and
Familial Effects of Substance Use Among the Male College Students -A North Indian
Study. Journal of Clinical and Diagnostic Research : JCDR, 7(8), 1632–1636.
doi:10.7860/JCDR/2013/6441.3215
173

Haertzen, C., Kocher, T., & Miyasato, K. (1983). Reinforcements from the 1st Drug Experience
Can Predict Later Drug Habits and or Addiction - Results with Coffee, Cigarettes,
Alcohol, Barbiturates, Minor and Major Tranquilizers, Stimulants, Marijuana,
Hallucinogens, Heroin, Opiates and Cocaine. Drug and Alcohol Dependence, 11(2), 147–
165. doi:10.1016/0376-8716(83)90076-5
Hall, W. D., & Degenhardt, L. (2009). Adverse health effects of non-medical cannabis use.
Lancet, 374(9698), 1383–1391.
Hall, W. D., & Swift, W. (2007). The THC content of cannabis in Australia: evidence and
implications. Australian and New Zealand Journal of Public Health, 24(5), 503–508.
Hall, W. D., Teesson, M., Lynskey, M., & Degenhardt, L. (1999). The 12-month prevalence of
substance use and ICD-10 substance use disorders in Australian adults: findings from the
National Survey of Mental Health and Well-Being. Addiction, 94(10), 1541–1550.
Hankin, B. L., Abramson, L. Y., Moffitt, T. E., Silva, P. A., McGee, R., & Angell, K. E. (1998).
Development of depression from preadolescence to young adulthood: Emerging gender
differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107(1),
128–140. doi:10.1037//0021-843X.107.1.128
Hansen, W. b., Graham, J. W., Sobel, J. L., Shelton, D. R., Flay, B. R., & Johnson, C. A. (1987).
The Consistency of Peer and Parent Influences on Tobacco, Alcohol, and Marijuana Use
Among Young Adolescents. Journal of Behavioral Medicine, 10(6), 559–579.
doi:10.1007/BF00846655
Harder, V. S., Morral, A. R., & Arkes, J. (2006). Marijuana use and depression among adults:
testing for causal associations. Addiction, 101(10), 1463–1472.
Harrison, L. D., Martin, S. S., Enev, T., & Harrington, D. (2007). Comparing drug testing and
self-report of drug use among youths and young adults in the general population (DHHS
Publication No. SMA 07-4249, Methodology Series M-7). Rockvill: Substance Abuse
and Mental Health Services Administration, Office of Applied Studies.
Hart, C. L., Ksir, C., & Ray, O. (2009). Drugs, Society, and Human Behavior (13th ed.). New
York: McGraw-Hill.
Hashibe, M., Straif, K., Tashkin, D., Morgenstern, H., Greenland, S., & Zhang, Z. (2005).
Epidemiologic review of marijuana use and cancer risk. ALCOHOL, 35(3), 265–275.
doi:10.1016/j.alcohol.2005.04.008
Hasin, D. S., Goodwin, R. D., Stinson, F. S., & Grant, B. F. (2005). Epidemiology of major
depressive disorder: results from the National Epidemiologic Survey on Alcoholism and
Related Conditions. Archives of General Psychiatry, 62(10), 1097.
Hasin, D. S., Keyes, K. M., Alderson, D., Wang, S., Aharonovich, E., & Grant, B. E. (2008).
Cannabis withdrawal in the United States: Results from NESARC. Journal of Clinical
Psychiatry, 69(9), 1354–1363.
174

Hayatbakhsh, M. R., Najman, J. M., Jamrozik, K., Mamun, A. A., Alati, R., & Bor, W. (2007).
Cannabis and anxiety and depression in young adults: a large prospective study. Journal
of Amer Academy of Child & Adolescent Psychiatry, 46(3), 408.
Heatherton, T. F., Kozlowski, L. T., Frecker, R. C., & Fagerstrom, K.-O. (1991). The Fagerström
test for nicotine dependence: a revision of the Fagerstrom Tolerance Questionnaire.
British Journal of Addiction, 86(9), 1119–1127.
Helzer, J. E., Robins, L. N., & Davis, D. H. (1976). Antecedents of narcotic use and addiction. A
study of 898 Vietnam veterans. Drug and Alcohol Dependence, 1(3), 183–190.
doi:10.1016/0376-8716(76)90028-4
Herkenham, M., Lynn, A. B., Johnson, M. R., Melvin, L. S., de Costa, B. R., & Rice, K. C.
(1991). Characterization and localization of cannabinoid receptors in rat brain: a
quantitative in vitro autoradiographic study. Journal of Neuroscience, 11(2), 563.
Hill, M. N., & Gorzalka, B. B. (2005). Is there a role for the endocannabinoid system in the
etiology and treatment of melancholic depression? Behavioural Pharmacology, 16(5-6),
333–352. doi:10.1097/00008877-200509000-00006
Hill, M. N., & Gorzalka, B. B. (2005). Pharmacological enhancement of cannabinoid CB1
receptor activity elicits an antidepressant-like response in the rat forced swim test.
European Neuropsychopharmacology: The Journal of the European College of
Neuropsychopharmacology, 15(6), 593–599. doi:10.1016/j.euroneuro.2005.03.003
Hill, S. A. . (1965). The Environment and Disease: Association or Causation? Proceedings of the
Royal Society of Medicine, 58, 295–300.
Hirschi, T. (1969). Causes of delinquency. Berkeley, CA: University of California Press.
Retrieved from
http://books.google.com/books?hl=en&lr=&id=i13b00vhluoC&oi=fnd&pg=PR10&dq=H
irschi+Causes+of+delinquency&ots=as2QBxOZTf&sig=JFTveTra3bKJt0w6frMPDnp4
D1Y
Hirvonen, J., Goodwin, R. S., Li, C.-T., Terry, G. E., Zoghbi, S. S., Morse, C., … Innis, R. B.
(2012). Reversible and regionally selective downregulation of brain cannabinoid CB1
receptors in chronic daily cannabis smokers. Molecular Psychiatry, 17(6), 642–649.
doi:10.1038/mp.2011.82
Hofler, M., Lieb, R., Perkonigg, A., Schuster, P., Sonntag, H., & Wittchen, H. U. (1999).
Covariates of cannabis use progression in a representative population sample of
adolescents: a prospective examination of vulnerabilitly and risk factors. Addiction,
94(11), 1679–1694. doi:10.1046/j.1360-0443.1999.941116796.x
Hofstra, M. B., van der Ende, J., & Verhulst, F. C. (2002). Child and adolescent problems predict
DSM-IV disorders in adulthood: A 14-year follow-up of a Dutch epidemiological sample.
Journal of the American Academy of Child and Adolescent Psychiatry, 41(2), 182–189.
doi:10.1097/00004583-200202000-00012
175

Holsboer, F. (2000). The corticosteroid receptor hypothesis of depression.
Neuropsychopharmacology, 23(5), 477–501. doi:10.1016/S0893-133X(00)00159-7
Hopfer, C. J., Lessem, J. M., Hartman, C. A., Stallings, M. C., Cherny, S. S., Corley, R. P., …
Crowley, T. J. (2007). A genome-wide scan for loci influencing adolescent cannabis
dependence symptoms: Evidence for linkage on chromosomes 3 and 9. Drug and Alcohol
Dependence, 89(1), 34–41. doi:10.1016/j.drugalcdep.2006.11.015
Horowitz, L. M., Rosenberg, S. E., Baer, B. A., Ureño, G., & Villaseñor, V. S. (1988). Inventory
of interpersonal problems: psychometric properties and clinical applications. Journal of
Consulting and Clinical Psychology, 56(6), 885.
Horwood, L. J., Fergusson, D. M., Coffey, C., Patton, G. C., Tait, R., Smart, D., … Hutchinson,
D. M. (2012). Cannabis and depression: an integrative data analysis of four Australasian
cohorts. Drug and Alcohol Dependence, 126(3), 369–378.
doi:10.1016/j.drugalcdep.2012.06.002
Hosmer, D. W., & Lemeshow, S. (2000). Model-Building Strategies and Methods for Logistic
Regression. Applied Logistic Regression, Second Edition, 91–142.
Hoyle, R. H., Stephenson, M. T., Palmgreen, P., Lorch, E. P., & Donohew, R. L. (2002).
Reliability and validity of a brief measure of sensation seeking. Personality and
Individual Differences, 32(3), 401–414. doi:10.1016/S0191-8869(01)00032-0
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling: A
Multidisciplinary Journal, 6(1), 1–55.
Jager, G., Block, R. I., Luijten, M., & Ramsey, N. F. (2010). Cannabis Use and Memory Brain
Function in Adolescent Boys: A Cross-Sectional Multicenter Functional Magnetic
Resonance Imaging Study. Journal of the American Academy of Child & Adolescent
Psychiatry, 49(6), 561–572.e3. doi:10.1016/j.jaac.2010.02.001
Jentsch, J. D., & Taylor, J. R. (1999). Impulsivity resulting from frontostriatal dysfunction in
drug abuse: implications for the control of behavior by reward-related stimuli.
Psychopharmacology, 146(4), 373–390. doi:10.1007/PL00005483
Jessor, R., Chase, J. A., & Donovan, J. E. (1980). Psychosocial correlates of marijuana use and
problem drinking in a national sample of adolescents. American Journal of Public
Health, 70(6), 604–613.
Jessor, R., & Jessor, S. (1980). A Social-Psychological Framework for Studying Drug Use. In D.
J. Lettieri, M. Sayers, & H. W. Pearson (Eds.), Theories on Drug Abuse (pp. 102–109).
Rockville, MD: Department of Health and Human Services.
Jessor, R., Jessor, S. L., & Finney, J. (1973). Social Psychology of Marihuana Use - Longitudinal
Studies of High-School and College Youth. Journal of Personality and Social
Psychology, 26(1), 1–15. doi:10.1037/h0034214
176

Johnson, B. D. (1980). Toward a Theory of Drug Subcultures. In D. J. Lettieri, M. Sayers, & H.
W. Pearson (Eds.), Theories on Drug Abuse (pp. 110–119). Rockville, MD: Department
of Health and Human Services.
Johnson, B. D., Bardhi, F., Sifaneck, S. J., & Dunlap, E. (2006). Marijuana argot as subculture
threads - Social constructions by users in New York City. British Journal of Criminology,
46(1), 46–77. doi:10.1093/bjc/azi053
Johnson, M. W., Bickel, W. K., Baker, F., Moore, B. A., Badger, G. J., & Budney, A. J. (2010).
Delay discounting in current and former marijuana-dependent individuals. Experimental
and Clinical Psychopharmacology, 18(1), 99.
Johnston, L. D., O’Malley, P. M., Bachman, J. G., & Schulenberg, J. E. (2012). Monitoring the
Future National Results on Adolescent Drug Use: Overview of Key Findings, 2011. Ann
Arbor, MI: Institute for Social Research, The University of Michigan.
Judd, L. L., Paulus, M. J., Schettler, P. J., Akiskal, H. S., Endicott, J., Leon, A. C., … Keller, M.
B. (2000). Does incomplete recovery from first lifetime major depressive episode herald
a chronic course of illness? American Journal of Psychiatry, 157(9), 1501–1504.
doi:10.1176/appi.ajp.157.9.1501
Kadden, R. M., Litt, M. D., Kabela-Cormier, E., & Petry, N. M. (2007). Abstinence rates
following behavioral treatments for marijuana dependence. Addictive Behaviors, 32(6),
1220–1236. doi:10.1016/j.addbeh.2006.08.009
Kandel, D. B. (1980). Developmental Stages in Adolescent Drug Involvement. In D. J. Lettieri,
M. Sayers, & H. W. Pearson (Eds.), Theories on Drug Abuse (pp. 110–119). Rockville,
MD: Department of Health and Human Services.
Kandel, D. B. (1996). The parental and peer contexts of adolescent deviance: An algebra of
interpersonal influences. Journal of Drug Issues, 26(2), 289–315.
Kandel, D. B., & Chen, K. (2000). Types of marijuana users by longitudinal course. Journal of
Studies on Alcohol, 61(3), 367–378.
Kandel, D. B., & Davies, M. (1986). Adult sequelae of adolescent depressive symptoms.
Archives of General Psychiatry, 43(3), 255.
Kandel, D. B., Davies, M., Karus, D., & Yamaguchi, K. (1986). The Consequences in Young
Adulthood of Adolescent Drug Involvement - an Overview. Archives of General
Psychiatry, 43(8), 746–754.
Kathuria, S., Gaetani, S., Fegley, D., Valino, F., Duranti, A., Tontini, A., … Piomelli, D. (2003).
Modulation of anxiety through blockade of anandamide hydrolysis. Nature Medicine,
9(1), 76–81. doi:10.1038/nm803
Keller, M., Lavori, P., Mueller, T., Endicott, J., Coryell, W., Hirschfeld, R., & Shea, T. (1992).
Time to Recovery, Chronicity, and Levels of Psychopathology in Major Depression - a 5177

Year Prospective Follow-up of 431 Subjects. Archives of General Psychiatry, 49(10),
809–816.
Kelly, B. C. (2005). Bongs and blunts: notes from a suburban marijuana subculture. Journal of
Ethnicity in Substance Abuse, 4(3-4).
Kendler, K. S., Jacobson, K. C., Prescott, C. A., & Neale, M. C. (2003). Specificity of genetic
and environmental risk factors for use and abuse/dependence of cannabis, cocaine,
hallucinogens, sedatives, stimulants, and opiates in male twins. American Journal of
Psychiatry, 160(4), 687–695. doi:10.1176/appi.ajp.160.4.687
Kendler, K. S., Karkowski, L. M., Neale, M. C., & Prescott, C. A. (2000). Illicit psychoactive
substance use, heavy use, abuse, and dependence in a US population-based sample of
male twins. Archives of General Psychiatry, 57(3), 261.
Kendler, K. S., & Prescott, C. A. (1998). Cannabis use, abuse, and dependence in a populationbased sample of female twins. American Journal of Psychiatry, 155(8), 1016–1022.
Kessler, R. C., Angermeyer, M., Anthony, J. C., DE GRAAF, R., Demyttenaere, K., Gasquet, I.,
… others. (2007). Lifetime prevalence and age-of-onset distributions of mental disorders
in the World Health Organization’s World Mental Health Survey Initiative. World
Psychiatry, 6(3), 168.
Kessler, R. C., Avenevoli, S., Costello, J., Green, J. G., Gruber, M. J., McLaughlin, K. A., …
Merikangas, K. R. (2012). Severity of 12-Month DSM-IV Disorders in the National
Comorbidity Survey Replication Adolescent Supplement. Archives of General
Psychiatry, 69(4), 381–389.
Kessler, R. C., Berglund, P., Demler, O., Jin, R., Koretz, D., Merikangas, K. R., … Wang, P. S.
(2003). The epidemiology of major depressive disorder - Results from the National
Comorbidity Survey Replication (NCS-R). Jama-Journal of the American Medical
Association, 289(23), 3095–3105. doi:10.1001/jama.289.23.3095
Kessler, R. C., McGonagle, K. A., Swartz, M., Blazer, D. G., & Nelson, C. B. (1993). Sex and
depression in the National Comorbidity Survey I: Lifetime prevalence, chronicity and
recurrence* 1. Journal of Affective Disorders, 29(2-3), 85–96.
Kessler, R. C., Mcgonagle, K. A., Zhao, S. Y., Nelson, C. B., Hughes, M., Eshleman, S., …
Kendler, K. S. (1994). Lifetime and 12-Month Prevalence of Dsm-III-R PsychiatricDisorders in the United-States - Results from the National-Comorbidity-Survey. Archives
of General Psychiatry, 51(1), 8–19.
Kessler, R. C., Ormel, J., Petukhova, M., McLaughlin, K. A., Green, J. G., Russo, L. J., …
Ustün, T. B. (2011). Development of lifetime comorbidity in the World Health
Organization world mental health surveys. Archives of General Psychiatry, 68(1), 90–
100. doi:10.1001/archgenpsychiatry.2010.180

178

Kessler, R. C., & Ustun, T. B. (2008). The WHO World Mental Health Surveys: Global
Perspectives on the Epidemiology of Mental Disorders. New York: Cambridge
University Press.
Khantzian, E. J. (1997). The self-medication hypothesis of substance use disorders: a
reconsideration and recent applications. Harvard Review of Psychiatry, 4(5), 231–244.
Kirsch, I., Deacon, B. J., Huedo-Medina, T. B., Scoboria, A., Moore, T. J., & Johnson, B. T.
(2008). Initial severity and antidepressant benefits: A meta-analysis of data submitted to
the food and drug administration. Plos Medicine, 5(2), 260–268.
doi:10.1371/journal.pmed.0050045
Klerman, G. L., & Weissman, M. M. (1989). Increasing Rates of Depression. JAMA: The
Journal of the American Medical Association, 261(15), 2229 –2235.
doi:10.1001/jama.1989.03420150079041
Kong, G., Smith, A. E., McMahon, T. J., Cavallo, D. A., Schepis, T. S., Desai, R. A., …
Krishnan-Sarin, S. (2013). Pubertal Status, Sensation-Seeking, Impulsivity, and
Substance Use in High School-Aged Boys and Girls. Journal of Addiction Medicine,
7(2), 116–121. doi:10.1097/ADM.0b013e31828230ca
Krueger, R. F., Hicks, B. M., Patrick, C. J., Carlson, S. R., Iacono, W. G., & McGue, M. (2002).
Etiologic connections among substance dependence, antisocial behavior and personality:
Modeling the externalizing spectrum. Journal of Abnormal Psychology, 111(3), 411.
Kuntsche, E., & Jordan, M. D. (2006). Adolescent alcohol and cannabis use in relation to peer
and school factors - Results of multilevel analyses. Drug and Alcohol Dependence, 84(2),
167–174. doi:10.1016/j.drugalcdep.2006.01.014
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical
data. Biometrics, 159–174.
Le Merrer, J., Befort, K., Gardon, O., Filliol, D., Darcq, E., Dembele, D., … Kieffer, B. L.
(2012). Protracted abstinence from distinct drugs of abuse shows regulation of a common
gene network. Addiction Biology, 17(1), 1–12. doi:10.1111/j.1369-1600.2011.00365.x
Lemstra, M., Bennett, N., Nannapaneni, U., Neudorf, C., Warren, L., Kershaw, T., & Scott, C.
(2010). A systematic review of school-based marijuana and alcohol prevention programs
targeting adolescents aged 10-15. Addiction Research & Theory, 18(1), 84–96.
doi:10.3109/16066350802673224
Leon, A. C., Olfson, M., Portera, L., Farber, L., & Sheehan, D. V. (1997). Assessing psychiatric
impairment in primary care with the Sheehan disability scale. International Journal of
Psychiatry in Medicine, 27(2), 93–105.
Leon, A. C., Shear, M., Portera, L., & Klerman, G. (1992). Assessing Impairment in Patients
with Panic Disorder - the Sheehan Disability Scale. Social Psychiatry and Psychiatric
Epidemiology, 27(2), 78–82. doi:10.1007/BF00788510
179

Lettieri, D. J., Sayers, M., & Pearson, H. W. (Eds.). (1980). Theories on Drug Abuse: Selected
Contemporary Perspectives. Research Monograph 30. Rockville, MD: Department of
Health and Human Services. Retrieved from
http://www.eric.ed.gov/ERICWebPortal/recordDetail?accno=ED203259
Lev-Ran, S., Roerecke, M., Le Foll, B., George, T. P., McKenzie, K., & Rehm, J. (2013). The
association between cannabis use and depression: a systematic review and meta-analysis
of longitudinal studies. Psychological Medicine, 1–14. doi:10.1017/S0033291713001438
Lewinsohn, P., Hops, H., Roberts, R., Seeley, J., & Andrews, J. (1993). Adolescent
Psychopathology .1. Prevalence and Incidence of Depression and Other Dsm-Iii-R
Disorders in High-School-Students. Journal of Abnormal Psychology, 102(1), 133–144.
doi:10.1037/0021-843X.102.1.133
Lindesmith, A. R. (1968). Addiction and opiates. Transaction Publishers. Retrieved from
http://books.google.com/books?hl=en&lr=&id=4kN_Vza2gGQC&oi=fnd&pg=PR5&dq=
Lindesmith+1968&ots=GydA7pB_4s&sig=JyHYeZTnEG3T9z4UDbTpQN8aqy4
Lopez, A. D., Mathers, C. D., Ezzati, M., Jamison, D. T., & Murray, C. J. L. (2006). Global and
regional burden of disease and risk factors, 2001: systematic analysis of population health
data. Lancet, 367(9524), 1747–1757. doi:10.1016/S0140-6736(06)68770-9
Luo, Z., Cowell, A. J., Musuda, Y. J., Novak, S. P., & Johnson, E. O. (2010). Course of Major
Depressive Disorder and Labor Market Outcome Disruption. The Journal of Mental
Health Policy and Economics, 13(3), 135–149.
Lux, V., Aggen, S. H., & Kendler, K. S. (2010). The DSM-IV definition of severity of major
depression: inter-relationship and validity. Psychological Medicine, 40(10), 1691–1701.
doi:10.1017/S0033291709992066
Lynskey, M. T., Fergusson, D. M., & Horwood, L. J. (1998). The origins of the correlations
between tobacco, alcohol, and cannabis use during adolescence. Journal of Child
Psychology and Psychiatry, 39(7), 995–1005.
Lynskey, M. T., Glowinski, A. L., Todorov, A. A., Bucholz, K. K., Madden, P. A. F., Nelson, E.
C., … Heath, A. C. (2004). Major depressive disorder, suicidal ideation, and suicide
attempt in twins discordant for cannabis dependence and early-onset cannabis use.
Archives of General Psychiatry, 61(10), 1026.
Lynskey, M. T., Heath, A. C., Bucholz, K. K., Slutske, W. S., Madden, P. A. F., Nelson, E. C.,
… Martin, N. G. (2003). Escalation of Drug Use in Early-Onset Cannabis Users vs Cotwin Controls. JAMA: The Journal of the American Medical Association, 289(4), 427 –
433. doi:10.1001/jama.289.4.427
Manrique-Garcia, E., Zammit, S., Dalman, C., Hemmingsson, T., & Allebeck, P. (2012).
Cannabis use and depression: a longitudinal study of a national cohort of Swedish
conscripts. Bmc Psychiatry, 12. doi:10.1186/1471-244X-12-112
180

Manski, C. F. (1999). Identification problems in the social sciences. Harvard Univ Pr.
Mariani, J., Brooks, D., Haney, M., & Levin, F. (2011). Quantification and comparison of
marijuana smoking practices: Blunts, joints, and pipes. DRUG AND ALCOHOL
DEPENDENCE, 113(2-3), 249–251. doi:10.1016/j.drugalcdep.2010.08.008
Marmorstein, N. R., & Iacono, W. G. (2011). Explaining associations between cannabis use
disorders in adolescence and later major depression: A test of the psychosocial failure
model. Addictive Behaviors, 36(7), 773–776. doi:10.1016/j.addbeh.2011.02.006
Martin, B. R., Sim-Selley, L. J., & Selley, D. E. (2004). Signaling pathways involved in the
development of cannabinoid tolerance. Trends in Pharmacological Sciences, 25(6), 325–
330. doi:10.1016/j.tips.2004.04.005
Martin, C. A., Kelly, T. H., Rayens, M. K., Brogli, B. R., Brenzel, A., Smith, W. J., & Omar, H.
A. (2002). Sensation seeking, puberty, and nicotine, alcohol, and marijuana use in
adolescence. Journal of the American Academy of Child and Adolescent Psychiatry,
41(12), 1495–1502. doi:10.1097/01.CHI.0000024864.60748.9D
Martin, M., Ledent, C., Parmentier, M., Maldonado, R., & Valverde, O. (2002). Involvement of
CB1 cannabinoid receptors in emotional behaviour. Psychopharmacology, 159(4), 379–
387.
Martin, W. (1980). Emerging Concepts Concerning Drug Abuse. In D. J. Lettieri, M. Sayers, &
H. W. Pearson (Eds.), Theories on Drug Abuse (pp. 278–285). Rockville, MD:
Department of Health and Human Services.
McAuliffe, W. E., & Gordon, R. A. (1980). Reinforcement and the Combination of Effects:
Summary of a Theory of Opiate Addiction. In D. J. Lettieri, M. Sayers, & H. W. Pearson
(Eds.), Theories on Drug Abuse (pp. 110–119). Rockville, MD: Department of Health
and Human Services.
McDonald, J., Schleifer, L., Richards, J. B., & de Wit, H. (2003). Effects of THC on behavioral
measures of impulsivity in humans. Neuropsychopharmacology, 28(7), 1356–1365.
doi:10.1038/sj.npp.1300176
McKnight, P. E., & Kashdan, T. B. (2009). The importance of functional impairment to mental
health outcomes: a case for reassessing our goals in depression treatment research.
Clinical Psychology Review, 29(3), 243–259.
Mehmedic, Z., Chandra, S., Slade, D., Denham, H., Foster, S., Patel, A. S., … ElSohly, M. A.
(2010). Potency Trends of Δ9-THC and Other Cannabinoids in Confiscated Cannabis
Preparations from 1993 to 2008*. Journal of Forensic Sciences, 55(5), 1209–1217.
Mehra, R., Moore, B., Crothers, K., Tetrault, J., & Fiellin, D. (2006). The association between
marijuana smoking and lung cancer - A systematic review. ARCHIVES OF INTERNAL
MEDICINE, 166(13), 1359–1367.

181

Mendelson, J. H., & Mello, N. K. (1984). Reinforcing properties of oral Δ 9tetrahydrocannabinol, smoked marijuana, and nabilone: influence of previous marijuana
use. Psychopharmacology, 83(4), 351–356.
Miles, D. R., van den Bree, M. B. ., Gupman, A. E., Newlin, D. B., Glantz, M. D., & Pickens, R.
W. (2001). A twin study on sensation seeking, risk taking behavior and marijuana use.
Drug and Alcohol Dependence, 62(1), 57–68. doi:10.1016/S0376-8716(00)00165-4
Miles, D. R., van den Bree, M. B. M., & Pickens, R. W. (2002). Sex differences in shared
genetic and environmental influences between conduct disorder symptoms and marijuana
use in adolescents. American Journal of Medical Genetics, 114(2), 159–168.
doi:10.1002/ajmg.10178
Miller, N. S., & Flaherty, J. A. (2000). Effectiveness of coerced addiction treatment (alternative
consequences) - A review of the clinical research. Journal of Substance Abuse Treatment,
18(1), 9–16. doi:10.1016/S0740-5472(99)00073-2
Miller-Johnson, S., Lochman, J. E., Coie, J. D., Terry, R., & Hyman, C. (1998). Comorbidity of
conduct and depressive problems at sixth grade: Substance use outcomes across
adolescence. Journal of Abnormal Child Psychology, 26(3), 221–232.
doi:10.1023/A:1022676302865
Mittleman, M. A., Lewis, R. A., Maclure, M., Sherwood, J. B., & Muller, J. E. (2001).
Triggering myocardial infarction by marijuana. Circulation, 103(23), 2805.
Moore, B. A., & Budney, A. J. (2002). Abstinence at intake for marijuana dependence treatment
predicts response. Drug and Alcohol Dependence, 67(3), 249–257.
Moore, T. H. M., Zammit, S., Lingford-Hughes, A., Barnes, T. R. E., Jones, P. B., Burke, M., &
Lewis, G. (2007). Cannabis use and risk of psychotic or affective mental health
outcomes: a systematic review. Lancet, 370(9584), 319–328.
Moreau, J. J. (1845). Hashish and Mental Illness.
Mueser, K. T., Drake, R. E., & Wallach, M. A. (1998). Dual diagnosis: A review of etiological
theories. Addictive Behaviors, 23(6), 717–734.
Munro, S., Thomas, K. L., & Abu-Shaar, M. (1993). Molecular characterization of a peripheral
receptor for cannabinoids. Retrieved from
http://www.nature.com/nature/journal/v365/n6441/abs/365061a0.html
Murphy, K. R., Barkley, R. A., & Bush, T. (2002). Young adults with attention deficit
hyperactivity disorder: Subtype differences in comorbidity, educational, and clinical
history. Journal of Nervous and Mental Disease, 190(3), 147–157.
doi:10.1097/00005053-200203000-00003
Murray, C. J. L., & Lopez, A. D. (1997). Alternative projections of mortality and disability by
cause 1990-2020: Global Burden of Disease Study. The Lancet, 349(9064), 1498–1504.
182

Muthén, L. K., & Muthén, B. O. (2012). Mplus (Version 7). Los Angles, CA: Muthén and
Muthén.
Navarro, M., Hernández, E., Mu\ noz, R. M., del Arco, I., Villanúa, M. A., Carrera, M. R. ., &
Rodríguez de Fonseca, F. (1997). Acute administration of the CB1 cannabinoid receptor
antagonist SR 141716A induces anxiety-like responses in the rat. Neuroreport, 8(2), 491.
Nemeroff, C. B. (2007). The burden of severe depression: A review of diagnostic challenges and
treatment alternatives. Journal of Psychiatric Research, 41(3-4), 189–206.
doi:10.1016/j.jpsychires.2006.05.008
Neumark, Y. D., Lopez-Quintero, C., & Bobashev, G. (2012). Drug use opportunities as
opportunities for drug use prevention: Bogota, Colombia a case in point. Drug and
Alcohol Dependence, 122(1-2), 127–134. doi:10.1016/j.drugalcdep.2011.09.022
Neumark, Y. D., Van Etten, M. L., & Anthony, J. C. (2000). “Drug dependence” and death:
survival analysis of the Baltimore ECA sample from 1981 to 1995. Substance Use &
Misuse, 35(3), 313–327.
Ningombam, S., Hutin, Y., & Murhekar, M. V. (2011). Prevalence and pattern of substance use
among the higher secondary school students of Imphal, Manipur, India. Retrieved from
http://imsear.hellis.org/handle/123456789/139166
Norberg, M. M., Kezelman, S., & Lim-Howe, N. (2013). Primary Prevention of Cannabis Use: A
Systematic Review of Randomized Controlled Trials. Plos One, 8(1).
doi:10.1371/journal.pone.0053187
Nyhlén, A., Fridell, M., Bäckström, M., Hesse, M., & Krantz, P. (2011). Substance abuse and
psychiatric co-morbidity as predictors of premature mortality in Swedish drug abusers: a
prospective longitudinal study 1970-2006. BMC Psychiatry, 11, 122. doi:10.1186/1471244X-11-122
O’Brien, M. S., Comment, L. A., Liang, K. Y., & Anthony, J. C. (2012). Does cannabis onset
trigger cocaine onset? A case-crossover approach. International Journal of Methods in
Psychiatric Research. doi:10.1002/mpr.359
O’Brien, M. S., Wu, L. T., & Anthony, J. C. (2005). Cocaine use and the occurrence of panic
attacks in the community: a case-crossover approach. Substance Use & Misuse, 40(3),
285–297.
Oetting, E. R., & Beauvais, F. (1987). Peer cluster theory, socialization characteristics, and
adolescent drug use: A path analysis. Journal of Counseling Psychology, 34(2), 205.
Oetting, E. R., & Donnermeyer, J. F. (1998). Primary socialization theory: The etiology of drug
use and deviance. I. Substance Use & Misuse, 33(4), 995–1026.
doi:10.3109/10826089809056252

183

Osuch, E., Vingilis, E., Ross, E., Forster, C., & Summerhurst, C. (2013). Cannabis use, addiction
risk and functional impairment in youth seeking treatment for primary mood or anxiety
concerns. International Journal of Adolescent Medicine and Health, 1–6.
doi:10.1515/ijamh-2013-0067
Otten, R., Barker, E. D., Maughan, B., Arseneault, L., & Engels, R. C. M. E. (2010). Self-control
and its relation to joint developmental trajectories of cannabis use and depressive mood
symptoms. Drug and Alcohol Dependence, 112(3), 201–208.
doi:10.1016/j.drugalcdep.2010.06.007
Pahl, K., Brook, J. S., & Koppel, J. (2010). Trajectories of Marijuana Use and Psychological
Adjustment Among Urban African American and Puerto Rican Women. Psychological
Medicine, 41(8), 1775–1783. doi:10.1017/S0033291710002345
Patel, S., & Hillard, C. J. (2006). Pharmacological evaluation of cannabinoid receptor ligands in
a mouse model of anxiety: Further evidence for an anxiolytic role for endogenous
cannabinoid signaling. Journal of Pharmacology and Experimental Therapeutics, 318(1),
304–311. doi:10.1124/jpet.106.101287
Paton, S., Kessler, R., & Kandel, D. (1977). Depressive Mood and Adolescent Illicit Drug-Use Longitudinal Analysis. Journal of Genetic Psychology, 131(2), 267–289.
Patton, G. C., Coffey, C., Carlin, J. B., Degenhardt, L., Lynskey, M., & Hall, W. (2002).
Cannabis use and mental health in young people: cohort study. British Medical Journal,
325(7374), 1195.
Patton, G. C., Coffey, C., Carlin, J. B., Sawyer, S. M., & Wakefield, M. (2006). Teen smokers
reach their mid twenties. Journal of Adolescent Health, 39(2), 214–220.
Pedersen, W. (2008). Does cannabis use lead to depression and suicidal behaviours? A
population-based longitudinal study. Acta Psychiatrica Scandinavica, 118(5), 395–403.
Perkonigg, A., Lieb, R., Hofler, M., Schuster, P., Sonntag, H., & Wittchen, H. U. (1999).
Patterns of cannabis use, abuse and dependence over time: incidence, progression and
stability in a sample of 1228 adolescents. Addiction, 94(11), 1663–1678.
doi:10.1046/j.1360-0443.1999.941116635.x
Petraitis, J., Flay, B., & Miller, T. (1995). Reviewing Theories of Adolescent Substance Use Organizing Pieces in the Puzzle. Psychological Bulletin, 117(1), 67–86.
doi:10.1037//0033-2909.117.1.67
Piccinelli, M., & Wilkinson, G. (2000). Gender differences in depression - Critical review.
British Journal of Psychiatry, 177, 486–492. doi:10.1192/bjp.177.6.486
Pinchevsky, G. M., Arria, A. M., Caldeira, K. M., Garnier-Dykstra, L. M., Vincent, K. B., &
O’Grady, K. E. (2012). Marijuana Exposure Opportunity and Initiation during College:
Parent and Peer Influences. Prevention Science, 13(1), 43–54. doi:10.1007/s11121-0110243-4
184

Pope, H. G., Gruber, A. J., Hudson, J. I., Huestis, M. A., & Yurgelun-Todd, D. (2001).
Neuropsychological performance in long-term cannabis users. Archives of General
Psychiatry, 58(10), 909–915. doi:10.1001/archpsyc.58.10.909
Pope, H. G., & YurgelunTodd, D. (1996). The residual cognitive effects of heavy marijuana use
in college students. Jama-Journal of the American Medical Association, 275(7), 521–
527. doi:10.1001/jama.275.7.521
Porath-Waller, A. J., Beasley, E., & Beirness, D. J. (2010). A Meta-Analytic Review of SchoolBased Prevention for Cannabis Use. Health Education & Behavior, 37(5), 709–723.
doi:10.1177/1090198110361315
Rasic, D., Weerasinghe, S., Asbridge, M., & Langille, D. B. (2012). Longitudinal associations of
cannabis and illicit drug use with depression, suicidal ideation and suicidal attempts
among Nova Scotia high school students. Drug and Alcohol Dependence.
doi:10.1016/j.drugalcdep.2012.09.009
Raykov, T., & Marcoulides, G. A. (2008). An Introduction to Applied Multivariate Analysis.
New York: Routledge.
Ream, G., Benoit, E., Johnson, B., & Dunlap, E. (2008). Smoking tobacco along with marijuana
increases symptoms of cannabis dependence. DRUG AND ALCOHOL DEPENDENCE,
95(3), 199–208. doi:10.1016/j.drugalcdep.2008.01.011
Ream, G., Johnson, B., Sifaneck, S., & Dunlap, E. (2006). Distinguishing blunt users from joint
users: A comparison of marijuana subcultures. In S. Cole (Ed.), New Research on Street
Drugs (pp. 245–273). Hauppauge, NY: Nova Science Publishers.
Reitzel, L. R., Li, Y., Stewart, D. W., Cao, Y., Wetter, D. W., Waters, A. J., & Vidrine, J. I.
(2013). Race moderates the effect of menthol cigarette use on short-term smoking
abstinence. Nicotine & Tobacco Research: Official Journal of the Society for Research
on Nicotine and Tobacco, 15(5), 883–889. doi:10.1093/ntr/nts335
Repetto, P. B., Zimmerman, M. A., & Caldwell, C. H. (2008). A longitudinal study of depressive
symptoms and marijuana use in a sample of inner-city African Americans. Journal of
Research on Adolescence, 18(3), 421–447.
Rey, J. M., Sawyer, M. G., Raphael, B., Patton, G. C., & Lynskey, M. (2002). Mental health of
teenagers who use cannabis - Results of an Australian survey. British Journal of
Psychiatry, 180, 216–221.
Rhee, S. H., Hewitt, J. K., Young, S. E., Corley, R. P., Crowley, T. J., & Stallings, M. C. (2003).
Genetic and environmental influences on substance initiation, use, and problem use in
adolescents. Archives of General Psychiatry, 60(12), 1256–1264.
doi:10.1001/archpsyc.60.12.1256

185

Rice, D., & Barone, S. (2000). Critical periods of vulnerability for the developing nervous
system: Evidence from humans and animal models. Environmental Health Perspectives,
108, 511–533.
Ridenour, T. A., Lanza, S. T., Donny, E. C., & Clark, D. B. (2006). Different lengths of times for
progressions in adolescent substance involvement. Addictive Behaviors, 31(6), 962–983.
doi:10.1016/j.addbeh.2006.03.015
Rios-Bedoya, C. F., Wilcox, H. C., Piazza, M., & Anthony, J. C. (2008). Children taking risks:
The association with cocaine and other drug use by young adulthood. Addictive
Behaviors, 33(9), 1154–1161. doi:10.1016/j.addbeh.2008.04.016
Roberts, C. (2013, March 13). Thanks to “Dabbing,” It Is Possible to Overdose on Marijuana.
The Snitch. Retrieved April 28, 2014, from
http://blogs.sfweekly.com/thesnitch/2013/03/medical_marijuana_overdose_dabbing.php
Roberts, R. E., Roberts, C. R., & Xing, Y. (2007). Comorbidity of substance use disorders and
other psychiatric disorders among adolescents: Evidence from an epidemiologic survey.
Drug and Alcohol Dependence, 88, S4–S13. doi:10.1016/j.drugalcdep.2006.12.010
Robins, L. N. (1980). The natural history of drug abuse. Acta Psychiatrica Scandinavica,
62(s284), 7–20.
Robins, L. N., & Regier, D. A. (1991). Psychiatric disorders in America: the epidemiologic
catchment area study. Free Press.
Roffman, R. A., Schwartz, S., & Stephens, R. S. (2006). Themes in the History of Cannabis
Dependence. In R. A. Roffman & R. S. Stephens (Eds.), Cannabis Dependence: Its
Nature, Consequences, and Treatment (pp. 3–20). New York: Cambridge University
Press.
Rosenberg, M. F., & Anthony, J. C. (2001a). Aggressive behavior and opportunities to purchase
drugs. Drug and Alcohol Dependence, 63(3), 245–252. doi:10.1016/S03768716(00)00213-1
Rosenberg, M. F., & Anthony, J. C. (2001b). Early clinical manifestations of cannabis
dependence in a community sample. Drug and Alcohol Dependence, 64(2), 123–131.
doi:10.1016/S0376-8716(00)00229-5
Rosenbloom, J., Rees, V. W., Reid, K., Wong, J., & Kinnunen, T. (2012). A cross-sectional
study on tobacco use and dependence among women: Does menthol matter? Tobacco
Induced Diseases, 10(1), 19. doi:10.1186/1617-9625-10-19
Rowe, M. G., Fleming, M. F., Barry, K. L., Manwell, L. B., & Kropp, S. (1995). Correlates of
depression in primary care. The Journal of Family Practice, 41(6), 551.
Royston, P., & Sauerbrei, W. (2008). Multivariable model-building: a pragmatic approach to
regression anaylsis based on fractional polynomials for modelling continuous variables
186

(Vol. 777). West Sussex, UK: Wiley. Retrieved from
http://books.google.com/books?hl=en&lr=&id=mRLaob58098C&oi=fnd&pg=PR5&dq=
Royston+model+building&ots=3RRO0oGojS&sig=_Ajwm3woggBBwnt7-r5jEgDaQQE
Rubino, T., Daniela Vigano, N. R., Guidali, C., Braida, D., Capurro, V., Castiglioni, C., … Sala,
M. (2008). Chronic 9-tetrahydrocannabinol during adolescence provokes sex-dependent
changes in the emotional profile in adult rats: behavioral and biochemical correlates.
Neuropsychopharmacology, 33(11), 2760–2771.
Rubino, T., Realini, N., Castiglioni, C., Guidali, C., Vigano, D., Marras, E., … Parolaro, D.
(2008). Role in anxiety behavior of the endocannabinoid system in the prefrontal cortex.
Cerebral Cortex, 18(6), 1292–1301. doi:10.1093/cercor/bhm161
Scheen, A. J., Finer, N., Hollander, P., Jensen, M. D., & Van Goal, L. F. (2006). Efficacy and
tolerability of rimonabant in overweight or obese patients with type 2 diabetes: a
randomised controlled study. Lancet, 368(9548), 1660–1672. doi:10.1016/S01406736(06)69571-8
Schensul, J. J., Huebner, C., Singer, M., Snow, M., Feliciano, P., & Broomhall, L. (2000). The
high, the money, and the fame: The emergent social context of “new marijuana” use
among urban youth. Medical Anthropology, 18(4), 389–414.
doi:10.1080/01459740.2000.9966164
Schildkraut, J. (1965). The Catecholamine Hypothesis of Affective-Disorders - a Review of
Supporting Evidence. American Journal of Psychiatry, 122(5), 509–522.
Schneider, M. (2008). Puberty as a highly vulnerable developmental period for the consequences
of cannabis exposure. Addiction Biology, 13(2), 253–263.
Schneider, M., & Koch, M. (2003). Chronic pubertal, but not adult chronic cannabinoid
treatment impairs sensorimotor gating, recognition memory, and the performance in a
progressive ratio task in adult rats. Neuropsychopharmacology, 28(10), 1760.
Schubiner, H., Tzelepis, A., Milberger, S., Lockhart, N., Kruger, M., Kelley, B., & Schoener, E.
(2000). Prevalence of attention-deficit/hyperactivity disorder and conduct disorder among
substance abusers. JOURNAL OF CLINICAL PSYCHIATRY, 61(4), 244–251.
Schuckit, M. A. (1980). A Theory of Alcohol and Drug Abuse: A Genetic Approach. In D. J.
Lettieri, M. Sayers, & H. W. Pearson (Eds.), Theories on Drug Abuse (pp. 297–302).
Rockville, MD: Department of Health and Human Services.
Schulenberg, J. E., Merline, A. C., Johnston, L. D., O’Malley, P. M., Bachman, J. G., & Laetz,
V. B. (2005). Trajectories of marijuana use during the transition to adulthood: The big
picture based on national panel data. Journal of Drug Issues, 35(2), 255–279.
Shapiro, D., Barkham, M., Rees, A., Hardy, G., Reynolds, S., & Startup, M. (1994). Effects of
Treatment Duration and Severity of Depression on the Effectiveness of Cognitive-

187

Behavioral and Psychodynamic Interpersonal Psychotherapy. Journal of Consulting and
Clinical Psychology, 62(3), 522–534. doi:10.1037/0022-006X.62.3.522
Shearman, L. P., Rosko, K. M., Fleischer, R., Wang, J., Xu, S., Tong, X. S., & Rocha, B. A.
(2003). Antidepressant-like and anorectic effects of the cannabinoid CB1 receptor inverse
agonist AM251 in mice. Behavioural Pharmacology, 14(8), 573–582.
doi:10.1097/01.fbp.0000104880.69384.38
Sheehan, D. V., Harnett-Sheehan, K., & Raj, B. A. (1996). The measurement of disability.
International Clinical Psychopharmacology, 11, 89–95. doi:10.1097/00004850199606003-00015
Shiffman, S., Waters, A. J., & Hickcox, M. (2004). The Nicotine Dependence Syndrome Scale:
A multidimensional measure of nicotine dependence. Nicotine & Tobacco Research,
6(2), 327–348. doi:10.1080/1462220042000202481
Sidney, S., Beck, J. E., Tekawa, I. S., Quesenberry, C. P., & Friedman, G. D. (1997). Marijuana
use and mortality. American Journal of Public Health, 87(4), 585–590.
Sifaneck, S. J., Johnson, B. D., & Dunlap, E. (2005). Cigars-for-blunts: choice of tobacco
products by blunt smokers. Journal of Ethnicity in Substance Abuse, 4(3-4), 23–42.
Simons, J. S., & Carey, K. B. (2002). Risk and vulnerability for marijuana use problems: The
role of affect dysregulation. Psychology of Addictive Behaviors, 16(1), 72–75.
doi:10.1037//0893-164X.16.1.72
Soldz, S., Huyser, D., & Dorsey, E. (2003). The cigar as a drug delivery device: youth use of
blunts. ADDICTION, 98(10), 1379–1386.
Soller, B., & Lee, J. P. (2010). Drug-Intake Methods and Social Identity: The Use of Marijuana
in Blunts Among Southeast Asian Adolescents and Emerging Adults. Journal of
Adolescent Research, 25(6), 783–806. doi:10.1177/0743558410376828
Solowij, N., Stephens, R. S., Roffman, R. A., Babor, T., Kadden, R., Miller, M., … Vendetti, J.
(2002). Cognitive functioning of long-term heavy cannabis users seeking treatment.
Jama-Journal of the American Medical Association, 287(9), 1123–1131.
doi:10.1001/jama.287.9.1123
Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for assessing the quality of
marriage and similar dyads. Journal of Marriage and the Family, 15–28.
StataCorp. (2011). Stata Statistical Software: Release 12. College Station, TX: StataCorp, LP.
Stenbacka, M., Allebeck, P., & Romelsjö, A. (1993). Initiation into drug abuse The pathway
from being offered drugs to trying cannabis and progression to intravenous drug abuse.
Scandinavian Journal of Public Health, 21(1), 31–39. doi:10.1177/140349489302100106

188

Stephens, R. S., Roffman, R. A., Copeland, J., & Swift, W. (2006). Cognitive-Behavioral and
Motivational Enhancement Treatments for Cannabis Dependence. In R. A. Roffman & R.
S. Stephens (Eds.), Cannabis Dependence: Its Nature, Consequences, and Treatment (pp.
133–150). New York: Cambridge University Press.
Stephens, R. S., Roffman, R. A., & Curtin, L. (2000). Comparison of extended versus brief
treatments for marijuana use. Journal of Consulting and Clinical Psychology, 68(5), 898–
908. doi:10.1037//0022-006X.68.5.898
Stephens, R. S., Roffman, R. A., & Simpson, E. E. (1994). Treating Adult Marijuana
Dependence - a Test of the Relapse Prevention Model. Journal of Consulting and
Clinical Psychology, 62(1), 92–99. doi:10.1037/0022-006X.62.1.92
Stinson, F. S., Ruan, W. J., Pickering, R., & Grant, B. F. (2006). Cannabis use disorders in the
USA: prevalence, correlates and co-morbidity. Psychological Medicine, 36(10), 1447–
1460.
Storr, C. L., Arria, A. M., Workman, R. L., & Anthony, J. C. (2004). Neighborhood environment
and opportunity to try methamphetamine (“ice”) and marijuana: Evidence from Guam in
the western Pacific region of Micronesia. Substance Use & Misuse, 39(2), 253–276.
doi:10.1081/JA-120028490
Storr, C. L., Chen, C. Y., & Anthony, J. C. (2004). “Unequal opportunity”: neighbourhood
disadvantage and the chance to buy illegal drugs. Journal of Epidemiology and
Community Health, 58(3), 231–237.
Storr, C. L., Wagner, F. A., Chen, C.-Y., & Anthony, J. C. (2011). Childhood predictors of first
chance to use and use of cannabis by young adulthood. Drug and Alcohol Dependence,
117(1), 7–15. doi:10.1016/j.drugalcdep.2010.12.023
Substance Abuse and Mental Health Services Administration. (2007). Use of Marijuana and
Blunts among Adolescents: 2005. Office of Applied Studies.
Substance Abuse and Mental Health Services Administration. (2010). Reliability of Key
Measures in the National Survey on Drug Use and Health (Office of Applied Studies,
Methodology Series M-8, HHS Publication No. SMA 09-4425). Rockville, MD.
Sutherland, E. H. (1939). Principles of Criminology (3rd ed.). Chicago: Philadelphia, Lippincott.
Swendsen, J., Anthony, J. C., Conway, K. P., Degenhardt, L., Dierker, L., Glantz, M., …
Merikangas, K. R. (2008). Improving targets for the prevention of drug use disorders:
Sociodemographic predictors of transitions across drug use stages in the national
comorbidity survey replication. Preventive Medicine, 47(6), 629–634.
doi:10.1016/j.ypmed.2008.09.009
Swift, W., Coffey, C., Carlin, J. B., Degenhardt, L., & Patton, G. C. (2008). Adolescent cannabis
users at 24 years: trajectories to regular weekly use and dependence in young adulthood.
Addiction, 103(8), 1361–1370. doi:10.1111/j.1360-0443.2008.02246.x
189

Tanda, G., & Goldberg, S. R. (2003). Cannabinoids: reward, dependence, and underlying
neurochemical mechanisms—a review of recent preclinical data. Psychopharmacology,
169(2), 115–134.
Terry-McElrath, Y. M., O’Malley, P. M., & Johnston, L. D. (2008). Saying No to Marijuana:
Why American Youth Report Quitting or Abstaining. Journal of Studies on Alcohol and
Drugs, 69(6), 796–805.
Timberlake, D. S. (2009). A Comparison of Drug use and Dependence Between Blunt Smokers
and Other Cannabis Users. SUBSTANCE USE & MISUSE, 44(3), 401–415.
doi:10.1080/10826080802347651
Timberlake, D. S. (2013). The changing demographic of blunt smokers across birth cohorts.
Drug and Alcohol Dependence, 130(1-3), 129–134.
doi:10.1016/j.drugalcdep.2012.10.022
Tobler, N. S., Lessard, T., Marshall, D., Ochshorn, P., & Roona, M. (1999). Effectiveness of
school-based drug prevention programs for marijuana use. School Psychology
International, 20(1), 105–137. doi:10.1177/0143034399201008
Tsuang, M. T., Lyons, M. J., Meyer, J. M., Doyle, T., Eisen, S. A., Goldberg, J., … Eaves, L.
(1998). Co-occurrence of abuse of different drugs in men - The role of drug-specific and
shared vulnerabilities. Archives of General Psychiatry, 55(11), 967–972.
doi:10.1001/archpsyc.55.11.967
Tucker, J. S., Ellickson, P. L., Orlando, M., Martino, S. C., & Klein, D. J. (2005). Substance use
trajectories from early adolescence to emerging adulthood: A comparison of smoking,
binge drinking, and marijuana use. Journal of Drug Issues, 35(2), 307–331.
Tzavara, E. T., Davis, R. J., Perry, K. W., Li, X., Salhoff, C., Bymaster, F. P., … Nomikos, G. G.
(2003). The CB1 receptor antagonist SR141716A selectively increases monoaminergic
neurotransmission in the medial prefrontal cortex: implications for therapeutic actions.
British Journal of Pharmacology, 138(4), 544–553.
U.S. Department of Health and Humans Services. (2013). Results from the 2012 National Survey
on Drug Use and Health: Summary of National Findings (NSDUH Series H-46, HHS
Publication No. (SMA) 13-4795). Rockville, MD: Substance Abuse and Mental Health
Services Administration.
United Nations Office on Drugs. (2006). World Drug Report 2006, Volume 2: Statistics (No.
E.06.XI.10). Vienna: United Nations.
United Nations Office on Drugs and Crime. (2011). World Drug Report 2011 (No. E.11.XI.10).
Vienna: United Nations.
United Nations Office on Drugs and Crime. (2012). World Drug Report 2012 (No. E.12.XI.1).
Vienna: United Nations.

190

United States Department of Health and Humans Services. (2012). 2011 National Survey on
Drug Use and Health: Public Use File Codebook (No. ICPSR). Ann Arbor, MI: Interuniversity Consortium for Political and Social Research [distributor].
United States of America, Substance Abuse and Mental Health Services Administration, &
Office of Applied Studies. (2002a). The NHSDA Report: Marijuana Use Among Youths.
Rockville, MD: Substance Abuse and Mental Health Services Administration.
United States of America, Substance Abuse and Mental Health Services Administration, &
Office of Applied Studies. (2002b). The NHSDA Report: Neighborhood Characteristics
and Youth Marijuana Use. Rockville, MD: Substance Abuse and Mental Health Services
Administration.
United States of America, Substance Abuse and Mental Health Services Administration, &
Office of Applied Studies. (2004). The NSDUH Report: Daily Marijuana Users.
Rockville, MD: Substance Abuse and Mental Health Services Administration.
US Department of Health and Human Services. (2004). The health consequences of smoking: a
report of the surgeon general. US Department of Health and Human Services, Centers
for Disease Control and Prevention, National Center for Chronic Disease Prevention and
Health Promotion, Office on Smoking and Health Atlanta^ eGeorgia Georgia.
Ustun, B., & Kennedy, C. (2009). What is “functional impairment”? Disentangling disability
from clinical significance. World Psychiatry, 8(2), 82–85.
Valjent, E., Mitchell, J. M., Besson, M. J., Caboche, J., & Maldonado, R. (2002). Behavioural
and biochemical evidence for interactions between Delta-9-tetrahydrocannabinol and
nicotine. British Journal of Pharmacology, 135(2), 564–578.
Van Beurden, E. K., Zask, A., Passey, M., & Kia, A. M. (2008). The mull hypothesis: is
cannabis use contributing to high tobacco use prevalence among young North Coast
males? NSW Public Health Bulletin, 19(3/4), 72–74. doi:10.1071/NB07052
Van Etten, M. L., & Anthony, J. C. (1999). Comparative epidemiology of initial drug
opportunities and transitions to first use: marijuana, cocaine, hallucinogens and heroin.
Drug and Alcohol Dependence, 54(2), 117–125. doi:10.1016/S0376-8716(98)00151-3
Van Etten, M. L., Neumark, Y. D., & Anthony, J. C. (1997). Initial opportunity to use marijuana
and the transition to first use: United States, 1979-1994. Drug and Alcohol Dependence,
49(1), 1–7. doi:10.1016/S0376-8716(97)00127-0
Van Etten, M. L., Neumark, Y. D., & Anthony, J. C. (1999). Male-female differences in the
earliest stages of drug involvement. Addiction, 94(9), 1413–1419.
Van Gaal, L. F., Rissanen, A. M., Scheen, A. J., Ziegler, O., & Rossner, S. (2005). Effects of the
cannabinoid-1 receptor blocker rimonabant on weight reduction and cardiovascular risk
factors in overweight patients: 1-year experience from the RIO-Europe study. Lancet,
365(9468), 1389–1397. doi:10.1016/S0140-6736(05)66374-X
191

Van Laar, M., van Dorsselaer, S., Monshouwer, K., & de Graaf, R. (2007). Does cannabis use
predict the first incidence of mood and anxiety disorders in the adult population?
Addiction, 102(8), 1251–1260.
Vandrey, R. G., Budney, A. J., Hughes, J. R., & Liguori, A. (2008). A within-subject comparison
of withdrawal symptoms during abstinence from cannabis, tobacco, and both substances.
Drug and Alcohol Dependence, 92(1-3), 48–54. doi:10.1016/j.drugalcdep.2007.06.010
Verweij, K. J. H., Vinkhuyzen, A. A. E., Benyamin, B., Lynskey, M. T., Quaye, L., Agrawal, A.,
… Medland, S. E. (2013). The genetic aetiology of cannabis use initiation: a metaanalysis of genome-wide association studies and a SNP-based heritability estimation.
Addiction Biology, 18(5), 846–850. doi:10.1111/j.1369-1600.2012.00478.x
Verweij, K. J. H., Zietsch, B. P., Lynskey, M. T., Medland, S. E., Neale, M. C., Martin, N. G., …
Vink, J. M. (2010). Genetic and environmental influences on cannabis use initiation and
problematic use: a meta-analysis of twin studies. Addiction, 105(3), 417–430.
doi:10.1111/j.1360-0443.2009.02831.x
Villanti, A. C., Giovino, G. A., Barker, D. C., Mowery, P. D., Sevilimedu, V., & Abrams, D. B.
(2012). Menthol brand switching among adolescents and young adults in the National
Youth Smoking Cessation Survey. American Journal of Public Health, 102(7), 1310–
1312. doi:10.2105/AJPH.2011.300632
Viveros, M.-P., Marco, E. M., & File, S. E. (2006). Nicotine and cannabinoids: parallels,
contrasts and interactions. Neuroscience and Biobehavioral Reviews, 30(8), 1161–1181.
doi:10.1016/j.neubiorev.2006.08.002
Von Sydow, K., Lieb, R., Pfister, H., Hofler, M., & Wittchen, H. U. (2002). What predicts
incident use of cannabis and progression to abuse and dependence? A 4-year prospective
examination of risk factors in a community sample of adolescents and young adults.
Drug and Alcohol Dependence, 68(1), 49–64. doi:10.1016/S0376-8716(02)00102-3
Wagner, F. A., & Anthony, J. C. (2002a). From first drug use to drug dependence:
Developmental periods of risk for dependence upon marijuana, cocaine, and alcohol.
NEUROPSYCHOPHARMACOLOGY, 26(4), 479–488.
Wagner, F. A., & Anthony, J. C. (2002b). Into the world of illegal drug use: exposure
opportunity and other mechanisms linking the use of alcohol, tobacco, marijuana, and
cocaine. American Journal of Epidemiology, 155(10), 918.
Wagner, F. A., & Anthony, J. C. (2007). Male-female differences in the risk of progression from
first use to dependence upon cannabis, cocaine, and alcohol. Drug and Alcohol
Dependence, 86(2-3), 191–198.
Walton, R. P. (1938). Marihuana: America’s New Drug Problem. Philadelphia: J.B. Lippincott.

192

Ware, J., & Sherbourne, C. (1992). The Mos 36-Item Short-Form Health Survey (sf-36) .1.
Conceptual-Framework and Item Selection. Medical Care, 30(6), 473–483.
doi:10.1097/00005650-199206000-00002
Weich, S., & Lewis, G. (1998). Poverty, unemployment, and common mental disorders:
population based cohort study. British Medical Journal, 317(7151), 115–119.
Weidenfeld, J., Feldman, S., & Mechoulam, R. (1994). Effect of the Brain Constituent
Anandamide, a Cannabinoid Receptor Agonist, on the Hypothalamo-Pitutary-Adrenal
Axis in the Rat. Neuroendocrinology, 59(2), 110–112. doi:10.1159/000126646
Weissman, M. M., Bland, R. C., Canino, G. J., Faravelli, C., Greenwald, S., Hwu, H. G., … Yeh,
E. K. (1996). Cross-national epidemiology of major depression and bipolar disorder.
JAMA, 276(4), 293–299.
Weissman, M. M., Wolk, S., Wickramaratne, P., Goldstein, R. B., Adams, P., Greenwald, S., …
Steinberg, D. (1999). Children with prepubertal-onset major depressive disorder and
anxiety grown up. Archives of General Psychiatry, 56(9), 794–801.
doi:10.1001/archpsyc.56.9.794
Wells, J. E., Degenhardt, L., Bohnert, K. M., Anthony, J. C., & Scott, K. A. (2009).
Geographical clustering of cannabis use: Results from the New Zealand Mental Health
Survey 2003-2004. Drug and Alcohol Dependence, 99(1-3), 309–316.
doi:10.1016/j.drugalcdep.2008.09.002
Wells, J. E., Maria Haro, J., Karam, E., Lee, S., Lepine, J.-P., Elena Medina-Mora, M., …
Gureje, O. (2011). Cross-National Comparisons of Sex Differences in Opportunities to
Use Alcohol or Drugs, and the Transitions to Use. Substance Use & Misuse, 46(9),
1169–1178. doi:10.3109/10826084.2011.553659
West, R. (2001). Theories of addiction. Addiction, 96(1), 3–13. doi:10.1046/j.13600443.2001.96131.x
Whitlow, C. T., Liguori, A., Brooke Livengood, L., Hart, S. L., Mussat-Whitlow, B. J.,
Lamborn, C. M., … Porrino, L. J. (2004). Long-term heavy marijuana users make costly
decisions on a gambling task. Drug and Alcohol Dependence, 76(1), 107–111.
Wilcox, H. C., & Anthony, J. C. (2004). The development of suicide ideation and attempts: an
epidemiologic study of first graders followed into young adulthood. Drug and Alcohol
Dependence, 76 Suppl, S53–67. doi:10.1016/j.drugalcdep.2004.08.007
Wilson, N., Syme, S. L., Boyce, W. T., Battistich, V. A., & Selvin, S. (2005). Adolescent
alcohol, tobacco, and marijuana use: The influence of neighborhood disorder and hope.
American Journal of Health Promotion, 20(1), 11–19.
Windle, M., & Wiesner, M. (2004). Trajectories of marijuana use from adolescence to young
adulthood: predictors and outcomes. Development and Psychopathology, 16(4), 1007–
1027.
193

Wittchen, H.-U., Behrendt, S., Hoefler, M., Perkonigg, A., Lieb, R., Buehringer, G., & Beesdo,
K. (2008). What are the high risk periods for incident substance use and transitions to
abuse and dependence? Implications for early intervention and prevention. International
Journal of Methods in Psychiatric Research, 17, S16–S29. doi:10.1002/mpr.254
World Health Organization. (1993). The ICD-10 classification of mental and behavioural
disorders: diagnostic criteria for research. World Health Organization. Retrieved from
http://books.google.com/books?hl=en&lr=&id=QiWPj_23ei4C&oi=fnd&pg=PR5&dq=I
CD-10&ots=Enr87Zw1TF&sig=5fdTOCLEFG-l8OL4ELJ8e1dFXgg
Xian, H., Scherrer, J. F., Grant, J. D., Eisen, S. A., True, W. R., Jacob, T., & Bucholz, K. K.
(2008). Genetic and environmental contributions to nicotine, alcohol and cannabis
dependence in male twins. Addiction, 103(8), 1391–1398.
Zanettini, C., Panlilio, L. V., Alicki, M., Goldberg, S. R., Haller, J., & Yasar, S. (2011). Effects
of endocannabinoid system modulation on cognitive and emotional behavior. Frontiers in
Behavioral Neuroscience, 5, 57. doi:10.3389/fnbeh.2011.00057

194