PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KzlProj/Acd-Pres/CIRC/Dateoue.indd

 

 

 

A TALE OF TWO CITIES: DRINKING PRACTICES AND PROBLEMS IN
TWO METROPOLITAN CITIES IN CHINA, BEIJING AND SHANGHAI

Bv

Hui Cheng

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of requirements
for the degree of

DOCTOR OF PHYLOSOPHY
Epidemiology

2009

ABSTRACT

A TALE OF TWO CITIES: DRINKING PRACTICES AND PROBLEMS IN TWO
METROPOLITAN CITIES IN CHINA, BEIJING AND SHANGHAI

BY
Hui Cheng

Alcohol consumption has been common in China. Nonetheless,
drinking and drinking-related problems have not been sufficiently described.
In this dissertation project, data collected from two metropolitan cities in
China, Beijing and Shanghai, have been used to describe drinking behavior
and drinking-related problems in these two cities.

Previous studies have provided some basis for speculating that
childhood physical punishment (CPA) might be a causal influence on drinking
problems, but more research is needed before any causal inference is drawn.
In this dissertation project, this possible causal relationship is also inspected
in a Chinese context.

Multi-stage probability sampling was used to collect information from
5201 household-dwelling adults in Beijing (n=2633) and Shanghai (n=2568). A
version of the World Health Organization Composite International Diagnostic
Interview was used to assess drinking behavior and problems, childhood
experiences, as well as other relevant variables.

It was found that alcohol is highly accessible in these two metropolitan
cities; alcohol consumption is common; heavier drinking is not rare

(occurrence 2 7%); drinking problems (socially maladaptive drinking and

alcohol dependence) exist at a fairly low occurrence (occurrence <'Z%).
Males and younger people were more likely to be involved with drinking;
they are also more likely to have a history of drinking problems.

A positive association was found between CPA and drinking-related
problems after taking family history of drinking problems into account. The
assumption of exogeneity of covariates was tested using the recursive probit
regression method. Estimates from generalized linear models were corrected
when there was evidence of endogeneity. Good internal validity and model
stability was found by the bootstrap resampling approach. The strengths of
associations were stronger for “early onset drinking problem” variables as
compared to some other variables.

This dissertation project provided the first epidemiological description
of drinking behavior and drinking-related problems in the two biggest cities
in China, Beijing and Shanghai. Result implied that the priority of prevention
and intervention should be placed on males and young adults. A possible
causal relationship between CPA and riskier drinking and drinking problems
were found. Limitations of the study were discussed. Directions for future

research have been suggested.

ACKNOWLEDGEMENTS

I much appreciate the mentorship of my advisor, Dr. Jim Anthony, for
teaching me epidemiology from scratch, for nurturing independent and
critical thinking, for granting me the freedom and promoting creativity in
research activities, for pointing out directions in the labyrinth when I am lost.
He is such a role model for me to follow. His enthusiasm, his versatile way to
approach questions, and his vision of research integrity has taught me a lot. I

believe I will benefit from them for the rest of my research career.

I would also like to thank my committee members, Dr. Breslau, Dr.
Chung, and Dr. Kubiak for guiding me through this dissertation project and
providing their valuable insights. The dissertation will not see the light

without their help.

During the process of this dissertation, I have received generous help
from many other professors and colleagues. I deeply appreciate their help

and encouragement.

Thanks to Professor Yueqin Huang from the Institute of Mental Health,
Peking University for allowing me to use the data for this project, the World

Mental Health Survey initiative for making the data available.

Previous research has provided me with a shoulder to stand on.

iv

Words cannot express my gratitude to my parents. I can only say thank
you for their unconditional support to me throughout the years. This
dissertation is dedicated to them. My husband, Chris Green, has been an

extraordinary encouragement to me. I am grateful of all he did for me.

 

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... ix
LIST OF FIGURES ........................................................................ xii
CHAPTER 1 AIMS AND OBJECTIVES ................................................... 1
CHAPTER 2 BACKGROUND AND SIGNIFICANCE
2.1 Introduction ................................................................................ 4
2.2 Drinking-related problems - alcohol dependence and alcohol abuse or
harmful use: recent concepts ............................................................ 4
2.3 History of alcohol consumption, with a focus on China ...................... 10
2.4 The First Rubric of Epidemiology: Quantity .................................... 13
2.5 The Second Rubric of Epidemiology, Location
2.5.1 Stable characteristics .............................................................. 14
2.5.2 Cross-country or cross-region variation ..................................... 17
2.5.3 Time-varying characteristics .................................................... 23
2.6 The third rubric of epidemiology: cause
2.6.1 Causes in epidemiology .......................................................... 27
2.6.2 Macro—social influences ........................................................... 28
2.6.3 Meso—level influences .............................................................. 29
2.6.4 Micro-level influences ............................................................ 30
2.7 The fourth rubric: mechanism
2.7.1 Brief introduction of chemistry, pharmacology of ethanol ............. 33
2.7.2 Metabolism and biotransformation of ethanol .............................. 35
2.7.3 Reinforcing effect of alcohol .................................................... 36
2.7.4 Possible natural history of AUD ................................................. 39
2.7.5 Comorbid conditions .............................................................. 39
2.7.6 Alcohol-related disabilities and impairment, including secondary
social maladaptation and hazard-laden drinking ......................... 41
2.8 The fifth rubric: prevention and control ......................................... 43

2.9 Possible causal influence of childhood physical abuse (CPA) and
drinking-related problems

2.9.1 Strength of association and replication of findings ...................... 44
2.9.2 Consideration of alternate explanations .................................... 47
2.9.3 Temporal relationship ............................................................ 49
2.9.4 Dose-response relationship ..................................................... 52
2.9.5 Biological plausibility ............................................................. 52
2.9.6 Specificity of the association .................................................... 53
2.9.7 Possible mediating pathway .................................................... 54

2.10 Gaps in the epidemiological evidence

vi

2.10.1 Specific aim 1 ....................................................................... 56

2.10.2 Specific aim 2 ........................................................................ 57
2.10.3 Specific aim 3 ........................................................................ 58
CHAPTER 3 METHODS
3.1 Background ............................................................................... 63
3.2 Design ...................................................................................... 63
3.3 Sample selection ........................................................................ 63
3.4 Measures
3.4. 1 Assessments ......................................................................... 67
3.4. 2 Definition of drinking-related variables
3.4.2.1 Variables in the “drinking behavior” category ..................... 70
3.4.2.2 Variables in the “indicators of risky drinking” category ........ 77
3.4. 3 Covariates under study ........................................................... 89
3.5 Analysis Plan
Aim 1 ............................................................................................ 97
Aim 2 ............................................................................................ 99
Aim 3 ........................................................................................... 100
CHAPTER 4 RESULTS

4.1. Lifetime occurrence and 12 month prevalence of drinking-related
outcomes

4.1.1. Lifetime occurrence .............................................................. 108
4.1.2 lZ-month prevalence of drinking-related outcomes ..................... 121
4.2. Subgroup variations of lifetime occurrence and 12 month prevalence of
drinking-related outcomes
4.2.1. Male is associated with higher likelihood of drinking-related
outcomes ......................................................................................... 124
4.2.2. Subgroup variation in drinking outcomes with respect to age
group .............................................................................................. 127
4.2.3. The association between marital status and drinking-related
outcomes .......................................................................................... 129
4.2.4 The association between education attainment and drinking-related
outcomes .......................................................................................... 130
4.2.5. The association between personal income level and drinking
outcomes .......................................................................................... 132
4.2.6. The association between drinking outcomes and employment
status ............................................................................................... 134

4.3. The impact of childhood physical punishment on alcohol drinking

outcomes
4.3.1. The associations between childhood physical punishment (CPP) and

drinking outcomes
4.3.1.1 . Estimation using logistic regressions ................................... 136
4.3.1.2. Goodness-of-fit of logistic regressions and exploration of
endogeneity ................................................................................. 147

vii

4.3.1.3 Stability of estimate ......................................................... 153
4.3.2. Variations of associations between CPP and drinking across different

outcomes ........................................................................................ 157
4.3.3. The association of CPP and drinking outcomes with respect to the
earliest to the later stages of alcohol involvement .................................. 163

CHAPTER 5 DISCUSSION

5.1 The frequency of beverage alcohol involvement in two metropolitan
cities in China: Beijing and Shanghai

5.1.1 Summary of results ............................................................... 163
5.1.2 Strengths and limitations ....................................................... 164
5.1.3 Drinking practices in Beijing and Shanghai .............................. 167
5.1.4 Drinking problems in Beijing and Shanghai ............................. 168
5.2 Subgroup variation with respect to beverage alcohol involvement.
5.2.1 Summary of results .............................................................. 172
5.2.2 Strengths and limitations ...................................................... 173
5.2.3 Subgroup variations in drinking outcomes
5.2.3.1. Sex and age .................................................................. 174
5.2.3.2. Other variables ............................................................. 178
5.3. The association between childhood physical punishment and drinking
problems
5.3.1 Summary of results .............................................................. 179
5.3.2 Strengths and limitations ....................................................... 180
5.3.3 Possible causal inference ...................................................... 182
5.4. Future research ........................................................................... 188
APPENDIX
Questions about drinking ........................................................................ 192
Skip patterns ........................................................................................... 198
Questionnaire in Chinese ........................................................................ 201
Questions about childhood experiences ................................................. 212
Questions about conduct problems ......................................................... 222
Appendix tables ...................................................................................... 225
BIBLIOGRAPHY .............................................................................. 229

viii

LIST OF TABLES

Table 2.1. Nine guidelines and corresponding main questions ...................... 44

Table 3.1 Distribution of variables for drinking behavior. Data from the WMH-

mC, 2001-2002 ................................................................................................ 76
Table 3.2 Actual WMH-CIDI questions about socially maladaptive drinking
and clinical features of alcohol dependence .................................................. 83
Table 3.3 Distribution of variables for risky drinking. Data from the WMH-mC,
2001-2002 ....................................................................................................... 87
Table 3.4 Distribution of sociodemographic variables. Data from WMH-mC,
2001-2002 ....................................................................................................... 92
Table 3.5 Actual CIDI questions assessing conduct problems ........................ 95

Table 4.1.1. Lifetime cumulative occurrence of alcohol drinking related
variables. Data from the WMH-mC, 2001-2002 ............................................. 11 1

Table 4.1.2. Description of alcohol drinking related variables. Data from the
WMH-mC, 2001-2002 ................................................................................... 1 18

Table 4.1.3. Twelve-month prevalence of alcohol drinking related variables.
Data from WMH-mC, 2001-2002 ................................................................... 121

Table 4.2.1. The association between drinking-related outcomes and sex. Data
from WMH-mC, 2001-2002 ........................................................................... 124

Table 4.2.2. The association between drinking-related outcomes and age
categories. Data from WMH-mC, 2001-2002 ................................................. 126

Table 4.2.3. The association between drinking-related outcomes and marital
status. Data from WMH-mC, 2001-2002 ......................................................... 129

Table 4.2.4. The association between drinking-related outcomes and
education attainment. Data from WMH-mC, 2001-2002 ................................. 131

Table 4.2.5. The association between drinking-related outcomes and the
personal income level. Data from WMH-mC, 2001-2002 ............................... 133

ix

Table 4.2.7 . The association between drinking-related outcomes and current

employment status. Data from WMH-mC, 2001-2002 .................................... 135
Table 4.3.1.1. Associations between childhood physical punishment and
riskier drinking and problems. Data from WMH-mC, 2001-2002 .................. 138
Table 4.3.1.2. The association between childhood physical punishment and
alcohol drinking outcomes. Data from WMH-mC, 2001-2002 ........................ 140

Table 4.3.1.3. The association between childhood physical punishment and
alcohol drinking outcomes. Data from WMH—mC, 2001-2002 ........................ 141

Table 4.3.1.4.The association between childhood physical punishment and
alcohol drinking outcomes in males. Data from WMH-mC, 2001-2002 .......... 145

Table 4.3.1.5. The association between childhood physical punishment and
alcohol drinking outcomes in males. Data from WMH-mC, 2001-2002 .......... 146

Table 4.3.1.6. F-test results of the Goodness-of-fit. Data from WMH-mC, 2001-
2002 .............................................................................................................. 147

Table 4.3.1.7. p values from Wald tests for endogeneity. Data from WMH-mC,
2001-2002 ..................................................................................................... 149

Table 4.3.1.8 Estimates for CPP from recursive probit models after taking
endogeneity into account. Data from WMH-mC, 2001-2002 .......................... 151

Table 4.3.2.2. Variations of the association between childhood physical
punishment and drinking outcomes. Data from WMH-mC, 2001—2002 .......... 155

Table 4.3.2.3. Variations of the association between childhood physical
punishment and drinking outcomes. Data from WMH-mC, 2001-2002 .......... 159

Table 4.3.3.1. The association between childhood physical punishment and
drinking involvement with weight. Data from the WMH-mC, 2001-2002 ....... 162

Table A4.3.1.7. p values from Wald tests for endogeneity in males. Data from
WMH-mC, 2001-2002 .................................................................................... 225

Table A4.3.1.8 Estimates for CPP from recursive probit models after taking
endogeneity into account in males. Data from WMH-mC, 2001-2002 ............ 226

Table A4.3.2.3. Variations of the association between childhood physical
punishment and drinking outcomes among people who initiated drinking after
16 (with weight). Data from WMH-mC, 2001-2002 ........................................ 227

Table A4.3.3. l . The association between childhood physical punishment and
stages of drinking involvement without weight. Data from the WMH—mC, 2001-
2002 .............................................................................................................. 228

xi

LIST OF FIGURES

Figure 1.1 Burden of disease attributable to: ALCOHOL ................................. 14
Figure 2.1 A conceptual relationship between parental drinking problems,
CPA, and offspring drinking problems ........................................................... 49
Figure 2.2. Conceptual model of the relationship between childhood physical
abuse and drinking problems ........................................................................ 60
Figure 3.1 A map of China .............................................................................. 64
Figure 3.2 Sample geographic maps of the WMH-mC, Beijing (upper) and
Shanghai (lower) ............................................................................................ 65
Figure 3.3 Skip pattern of the WMH—CIDI alcohol assessment ........................ 69
Figure 3.4 Diagram of the MIMIC model ....................................................... 105

Figure 4.1.1 Number of drinks per day when drank the most. Data from WMH-
mC, 2001-2002 ............................................................................................. l 13

Figure 4.1.2 Kaplan-Meier failure function of the age of trying alcoholic
beverages .................................................................................................... 114

Figure 4.1.3 Kaplan-Meier failure function of the age of onset of MTM
drinking ........................................................................................................ l 15

Figure 4.1.4 Kaplan-Meier failure function of the first socially maladaptive
drinking ........................................................................................................ 116

Figure 4.1.5 Kaplan-Meier failure function of the first dependence
problems ...................................................................................................... 117

Figure 4.1.6 Lifetime occurrence of drinking-related problems. Data from
WMH-mC, 2001-2002 .................................................................................... 119

Figure 4.1.7 Number of drinks per day during the last year. Data from WMH-
mC, 2001-2002 .............................................................................................. 122

Figure 4.3.1.1 Lifetime occurrence of drinking outcomes stratified by
experience of childhood physical punishment in Beijing ............................. 137

xii

Figure 1.3.1.2 Lifetime occurrence of drinking outcomes stratified by
experience of childhood physical punishment in Shanghai .......................... 137

Figure 4.3.1.3 ORs of associations between childhood physical punishment
and drinking outcomes without covariate ..................................................... 142

Figure 4.3.1.4 ORs of associations between childhood physical punishment
and drinking outcomes holding sex, age, and parental alcohol/drug problems
constant ........................................................................................................ 143

Figure 4.3.1.5 Distribution of coefficients from bootstrap resampling
procedure .................................................................................................... 153

Figure 4.3.2.1 MIMIC model of riskier drinking. Data from WMH-mC, 2001-
2002 .............................................................................................................. 157

Figure 4.3.2.2 Lifetime occurrence of drinking problems. Data from WMH-mC,
2001-

2002 .............................................................................................................. 158

Figure 4.3.2.3 MIMIC model of alcohol dependence. Data from WMH-mC,

2001-2002 ..................................................................................................... 160
xiii

Chapter 1 Aims and objectives

This dissertation is focused on three aims.

Aim 1: To describe epidemiological facets of beverage alcohol involvement in

two metropolitan cities in China: Beijing and Shanghai

0 Estimation tasks:

1. For the community populations as a whole, to estimate the
cumulative occurrence of opportunity, the first chance to drink
alcohol, trying alcohol (ever), pre-teen trying alcohol, alcohol
drinking (ever), precocious onset of drinking (<20 years old),
heavier drinking, drinking-related social maladaption, clinical
features of alcohol dependence, and early onset of alcohol use
disorders (AUD, < 23 years old), in Beijing and Shanghai,
respectively.

2. For the community populations as a whole, to estimate the one-
year interval prevalence for recent alcohol drinking, heavy
drinking, drinking related social maladaption and alcohol
dependence, in Beijing and Shanghai, respectively;

3. For the community populations as a whole, to estimate time-to-
event parameters such as the mean and median age of first
drink, onset of drinking, and onset of drinking-related
problems, and to plot the comparative survival analysis

parameter estimates.

Aim 2: To estimate subgroup variation with respect to beverage alcohol

involvement.

0 Estimation tasks:

1. Estimation of subgroup-specific cumulative occurrence and
interval prevalence of drinking related outcomes with respect
to sex (male-female), age groups, and categories of marital
status, income level, education attainment, and employment
groups.

2. Estimation of association structure parameters that link
drinking and related problems back to their potential sources

of variation associated with membership in these subgroups.

Aim 3: To estimate a suspected causal association that links childhood
physical punishment to later drinking and drinking problems in order to shed
light on the suspected causes, as well as aspects of mechanisms that might

lead to alcohol drinking and associated problems.

0 Estimation tasks:
1. To estimate the association between childhood physical
punishment and drinking related outcomes, within the context of
a more comprehensive conceptual model, fit with the approach
of multiple logistic regression, and with a binvariate probit
model used to probe into assumptions of the logistic regression

model.

2. To estimate the variation in associations across different
drinking-related outcomes accounting for the correlation
between these outcomes.

3. To estimate the association between childhood physical
punishment and stages of alcohol involvement (with respect to
the earliest and later stages of alcohol drinking involvement),
also within the framework of a more comprehensive conceptual

model.

 

Chapter 2 Background and significance
2. 1 Introduction
In this dissertation project, the focus is on the epidemiology of drinking

and drinking-related problems in two metropolitan areas of China. In this
chapter, the aims are 1) to introduce the concept of alcohol drinking-related
problems; 2) to review selected aspects of the history of drinking-with a focus
on China; 3) to review each of the 5 rubrics of epidemiology (quantity,
location, cause, mechanism, and prevention and control) with respect to
drinking practices and related problems; 3) to identify gaps in current
knowledge about drinking practices and related problems; 4) to evaluate the
potential significance of the dissertation project.

2.2 Drinking-related problems - alcohol dependence and alcohol

abuse or harmful use: recent concepts
Alcohol drinking-related problems refer to negative consequences of
alcohol consumption, including social and interpersonal problems and a
dependence syndrome. In brief, the alcohol dependence syndrome involves
(a) neural adaptation to repeated drinking, (b) obsession-like disturbance of
the mental life (e.g. craving), and (c) compulsion-like disturbance of behavior
(e.g. inability to stop drinking).
Currently, there are two most commonly used mental disorder

classification systems, the fourth edition of the Diagnostic and Statistical

Manual of Mental Disorders (DSM-IV) developed by American Psychiatric

Association (APA, 1994) and the International Statistical Classification of

Diseases and Related Health Problems, the 10th edition (ICD-lO) developed

by the World Health Organization (WHO, 1992). Both clearly list alcohol as a
psychoactive drug that can cause a dependence syndrome. Two main
categories of alcohol use disorders (AUD) defined in the DSM-IV are alcohol
dependence and non-dependent alcohol abuse; 1CD experts avoided the
stigma-laden term ‘abuse’ and substituted ‘harmful use’ as a related
disturbance.

Alcohol dependence is defined in the ICD-l 0 glossary and in DSM-IV
shown in chart 2.1.
Chart 2.1
1CD-10 glossary
A cluster of behavioral, cognitive, and physiological phenomena that develop after repeated
substance use and that typically include
a strong desire to take the drug,
difficulties in controlling its use,
persisting in its use despite harmful consequences,
a higher priority given to drug use than to other activities and obligations,
increased tolerance,
and sometimes a physical withdrawal state.
The dependence syndrome may be present for a specific psychoactive substance (e.g. tobacco,

alcohol, or diazepam), for a class of substances (e.g. opioid drugs), or for a wider range of

phannacologically different subsances

DSM-I'V case definition

A maladaptive pattern of substance use, leading to clinically significant impairment or distress, as
manifested by three (or more) of the following, occurring at any time in the same 12-month period:
1. tolerance, as defined by either of the following:

a need for markedly increased amounts of the substance to achieve intoxication or desired effect
markedly diminished effect with continued use of the same amount of substance

2. withdrawal, as manifested by either of the following:

the characteristic withdrawal syndrome for the substance

the same (or a closely related) substance is taken to relieve or avoid withdrawal symptoms

3. the substance is often taken in larger amounts or over a longer period than was intended

4. there is a persistent desire or unsuccessful efforts to cut down or control substance use

5. a great deal of time is spent in activities to obtain the substance, use the substance, or recover
from its effects

6. important social, occupational or recreational activities are given up or reduced because of
substance use

7. the substance 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 the substance (e.g.,

continued drinking despite recognition that an ulcer was made worse by alcohol consumption)

In ICD-10, harmful use refers to “A pattern of psychoactive substance use
that is causing damage to health. The damage may be physical (as in cases of

hepatitis from the self-administration of injected psychoactive substances) or

mental (e. g. episodes of depressive disorder secondary to heavy consumption of
alcohol). ” Under DSM—IV, non-dependent abuse is largely characterized by
impairment in social functioning, such as socially maladaptive behaviors (e.g.,
drink-induced violence), or social role impairments (e.g., family or legal
troubles), as well as other hazard-laden alcohol-related behaviors such as
drunk driving (APA, 1994; WHO, 1992).

These concepts have been refined over the past 50 years. It may be useful
to provide some information on the history of the concept of alcohol
dependence that has prompted alcohol researchers to differentiate alcohol
dependence from harmful drinking or nondependent abuse. In the first and
second editions of the DSM, published in 1952 and 1968, respectively,
alcoholism (addiction), which was then the term for alcohol dependence, was
listed under ‘personality disorders and certain non-psychotic mental
disorders’ (APA, 1952, , 1968). Alcoholism also appeared as a category in the
eighth edition of ICD published in 1967 (WHO, 1967).

During the 19603, the WHO Expert Committee on Drug Addiction-
Producing Drugs decided that the term ‘addiction’ carried too much stigma-
laden connotation to be useful in scientific work. Therefore, they re-named
themselves the Expert Committee on Drug Dependence, and introduced the
definition of ‘dependence’ that was specific to each drug compound (e.g.
dependence of the amphetamine type, dependence of the alcohol/ethanol
type. This history is described by the report series of the WHO expert

committee on drug dependence, (WHO expert committee on addiction-

producing drugs, 1964)). In the 19703, diagnostic criteria for ‘alcoholism’ and
‘alcohol dependence’ were introduced by Feighner et a1. (Feighner et al.,
1972) and Edwards and Gross (Edwards 8: Gross, 1976), respectively, to
describe a number of physiological and psychological manifestations of
diminished control over alcohol use. In the Edwards-Gross conceptualization,
alcohol problems were dimensional; they did not formalize a concept of
‘alcohol abuse’. Ten years later, two discrete diagnostic categories specified
after the deliberation of a DSM-III committee of experts (Rounsaville, Spitzer,
8: Williams, 1986). These two conditions were: 1) alcohol dependence, and 2)
alcohol abuse, with dependence and abuse allowed to co-occur. In contrast,
the ICD-9, published in 1977, used a categorical approach when defining the
‘alcohol dependence syndrome’ and ‘non-dependent abuse’ to be mutually
exclusive (World Health Organization, 1977). The introduction of the ‘non—
dependence abuse’ category is in response to the clinical observation that
some non-alcohol-dependent clients sought out or were referred to medical
help for their non-dependent maladaptive behavior due to alcohol drinking.
This categorical approach was used in the current DSM-IV and ICD-10 as well
(American Psychiatric Association, 1994; WHO, 1992). It is noteworthy that in
ICD-10, the term ‘alcohol abuse’ was replaced by the term ‘harmful use’ in
order to better describe negative health consequences due to alcohol
consumption. Evidence has been mixed regarding the question of whether
alcohol abuse and alcohol dependence are best represented as distinct

dimensions or categories, whether alcohol abuse should be specified as an

intermediate stage along the progression to alcohol dependence, or whether
they are merely two labels for the same disorder. Results from a long term
follow-up study (60 years) showed there were people who stayed long-term
as cases of alcohol abuse without meeting any criterion of alcohol
dependence (V aillant, 2003). On the other hand, cross-sectional research
suggests that a noteworthy proportion of individuals with alcohol dependence
had never experienced clinical features of alcohol abuse (D. S. Hasin 8: Grant,
2004). Shorter term follow-up studies also provide evidence that the course of
alcohol abuse differs from the course of dependence (D. S. Hasin, Grant, 8:
Endicott, 1990; D. S. Hasin, Van Rossem, McCloud, 8: Endicott, 1997). Besides
evidence on differing natural history and clinical course of these two alcohol-
related disturbances, there have also been explorations on this question from
other angles. For example, a cross-sectional study in the US has suggested
that stronger association can be found between DSM-N alcohol dependence
and alcohol consumption, treatment seeking, alcohol intoxication, and
suicidal ideation, as well as family history of alcohol use problems, compared
to DSM-IV alcohol abuse (D. Hasin 8: Paykin, 1999).

Recent advances in methods for latent variable analysis, including
computational software, have promoted application of Item Response Theory
(IRT), as well as Latent Class Analysis (LCA) models. This work has expanded
the scope of the probing into these latent structure questions. Since the early
1990s, a number of some researchers have explored the underlying latent

structure of AUD, using items from DSM and ICD constructs for alcohol

dependence, alcohol abuse, and harmful alcohol use, but these analyses have
not always yielded consistent evidence. For example, Nelson and colleagues
found that latent structures changed when analytic samples changed from the
entire sample (drinkers and non-drinkers) drinkers. In specific, a single
dimension sufficed as the best model fit to data from their entire sample,
whereas two dimensions were needed for the “drinkers only” sub-group.
Moreover, dimensions were not completely consistent with the alcohol
dependence and abuse conceptualizations (Nelson, Rehm, Ustun, Grant, 8:
Chatterji, 1999). Some investigations have extracted just one underlying
dimension (e.g. see (Saha, Chou, 8: Grant, 2006). Many studies have found
cross loadings of items from the dependence domain and the abuse domain
(Grant et al., 2007; Muthen, Hasin, 8: Wisnicki, 1993; Proudfoot, Baillie, 8:
Teesson, 2006). Research using LCA techniques found that items from both
DSM-IV alcohol abuse and alcohol dependence differentiate individuals into
class memberships (Bucholz et al., 1996; Smith 8: Shevlin, 2008). Summarized
from epidemiological studies, there is evidence of both similarities and
differences between DSM-IV alcohol abuse and dependence (D. S. Hasin et al.,
2003).
2.3 History of alcohol consumption, with a focus on China

Of course, the history of drinking and related problems did not begin in
the 20th century. Evidence of man-made alcohol has been found in pre-historic
time in Mesopotamia (e.g. current Iran and Iraq) and in other parts of the

world. In ancient Egypt, Greece, and Rome, alcohol had been part of people’s

10

daily diet, as well as served for medicinal and religious purposes (D. B. Heath,
1995). In ancient China, as early as the Shen Nong period of the New Stone
Age (approximately 7000 BC), traces of alcohol were found in a wine jar
discovered in Jiahu in Hunan province (McGovern 8: Patrick, 2003). Since then,
numerous fairy tales and legends have waltzed around alcohol throughout the
Chinese history. For example, there is a beautiful legend story about Du Kang,
a boy from a poor farmer family, which tells how alcohol was invented in
China. Today Du Kang is a brand of spirit and some Chinese still use the
character of his name to indicate alcohol. The invention of alcohol yeast
during the period of the Xia-Shang Age (1700 BC) and the reform of storage
techniques facilitated the production of alcohol beverages in Chinese history
(Cochrane, Chen, Conigrave, 8: Hao, 2003). Over the years, alcohol drinking
has been integrated deeply into the Chinese culture. From the ancient ritual
ceremonies to modern parties, alcohol has been regarded as a way to express
happiness and to exchange wishes of good luck. Besides daily activities,
alcohol has played an important role in Chinese medicine, and it has been
tightly associated with art and poetry as well (Hao, Chen, 8: Su, 2005). For
example, a famous Chinese poet Li Bai during the Tang dynasty (618-907) is
well-known for his inspiration after drinking alcohol.

Negative consequences of drinking also have been described since
ancient times. For example, the origins of present Dram Shop Laws can be
traced back to around 2000 BC from Hammurabi’s code in ancient Babylon (C.

B. Anthony, 1995; D. B. Heath, 1995). Problems induced by overdrinking have

11

been long recognized in China. For example, excessive drinking of the
emperor and his followers has been related to the fall of several Dynasties in
the Chinese history with the earliest being the Shang Dynasty in the 11th
century BC. An early epidemic of the use of a combination of alcohol and
hanshi has been documented in the Han Dynasty (second and third century) in
China. Users of this mixture described it as a cause of a mind-opening and
thought-clarifying effect from the psychoactive drug mixture, but the chemical
identity of hanshi remains unknown (Schutz, 1995).

Throughout the Chinese history, there have been multiple efforts at alcohol
control in response to drinking problems (Newman, 2002). For example,
overdrinking, together with overeating, gambling, and smoking, were
recognized as harmful and were listed as the “Four Vices” in China (Cochrane,
Chen, Conigrave, 8: Hao, 2003).

Against this background, it may be understood that social drinking often
has been highly accepted and even sometimes is encouraged in the Chinese
contexts. Drinking plays important roles in important events, such as Chinese
New Year Festival, wedding ceremonies, and birthday celebrations.
Ritualized drinking for special events still exists in some areas (Hao, Chen, 8:
Su, 2005).

Nonetheless, drinking behavior has changed markedly as the Chinese
market has opened to the outside world, with westernization. As a result, a
modified Chinese style of drinking has emerged from an intersection where

tradition and modernization meet up. For example, nowadays drinking is used

12

as a way to ease tension and to facilitate social exchange among Chinese
businessmen; it is also believed to help maintain good relationships between
employers and employees and among coworkers. On the other hand, the
Chinese culture tends to discourages solitary drinking (Williams, 1998) and
despite the custom of toasting as a common way to express friendliness,
Chinese drinkers tend to avoid overdrinking in social circumstances
(Cochrane, Chen, Conigrave, 8: Hao, 2003; Hao, Chen, 8: Su, 2005).

Of course, there is some evidence of beneficial health effects of drinking.
For example, light-to-moderate alcohol consumption has been associated with
lower occurrence of diabetes and cardiovascular diseases (Kloner 8: Rezkalla,
2007; Room, Babor, 8: Rehm, 2005; van de Wiel, 2004). Whereas these
beneficial effects should not be forgotten, it is the alcohol problems that
concern us in the present context. The extent of these problems is the topic of
the next section of this chapter.

2.4 The First Rubric of Epidemiology: Quantity

Alcohol drinking is common in many parts of the world, e.g. the
Americas, Europe, and Asia. There are approximately 2,000,000,000 alcohol
drinkers around the world (Anderson, 2006). Alcohol consumption has been
associated with substantial burden of disease, e.g. 1,800,000 deaths per year
according to the WHO estimate (WHO, 2004b). Following is a map from the
WHO website showing the estimated disease burden from alcohol

consumption in each WHO sub-region (WHO, 2009).

13

Figure 1.1 Burden of disease attributable to: ALCOHOL
(% DALYs in each subregion)

  
 

Proportion of DALYs

<0.5%
05-09%
1-1 9%
2-3.9%
4-7.9%
8-1 5.9%

0000.0

The disease burden comes from various negative health consequences,
such as unintentional injuries, AUD and other neuro-psychiatric conditions. It
was estimated that AUD accounts for approximately 40% of all disease burden
attributable to alcohol consumption (WHO, 2004b). According to the WHO,
approximately 125,000,000 people have AUD at any given time point
worldwide (WHO, 2004a).

2.5 The Second Rubric of Epidemiology, Location

2.5.1 Stable characteristics

There are stable and sometimes time-invariant characteristics that can be
used to map population subgroups variation in the occurrence of alcohol
dependence and related problems. In this section, the focus is upon three
stable characteristics: (a) sex, (b) year of birth, and (c) family—genetic

characteristics.

Consistently, evidence has shown that males are more likely to drink
(WHO, 2004b); males are more likely to experience AUD than females across
countries and cultures with no notable exceptions to date. However, the
degree of the male-associated excess risk can vary dramatically. For example,
studies from the USA and other European countries usually show that the
occurrence of AUD in males is an estimated 2 to 3 times higher than in females
(D. S. Hasin, Stinson, Ogburn, 8: Grant, 2007; Rehm, Taylor, 8: Patra, 2006;
Wilsnack et al., 2000). In eastern countries such as Korea Japan and China, the
male-female ratio is much larger (Hao et al., 2004; Higuchi, Matsushita,
Maesato, 8: Osaki, 2007; J. T. Park, Kim, 8: Jhun, 2008; Wei, Derson, Xiao, Li, 8:
Zhang, 1999). Accordingly, the odds ratio and other statistical indices of the
strength of association between sex and AUD is larger in these eastern
countries as compared to values observed in European countries and North
America (Keyes, Grant, 8: Hasin, 2007; Rehm, Taylor, 8: Patra, 2006). One
argument in the literature is that the male-female difference in AUD is largely
due to the smaller amounts of alcohol consumed in women, and that at the
same levels of consumption, women drinkers might experience as many or
more problems than men drinkers (Ely, Hardy, Longford, 8: Wadsworth, 1999;
Fillmore et al., 1995; Miller, Plant, 8: Plant, 2005). Contrary to this argument is
an observation that many females metabolize ethanol less efficiently than
males, which means the dose-response curve is left-shifted in females
compared to males. Therefore, adverse effects may result from smaller

amounts of alcohol in females as in males. The male excess of AUD is seen in

15

China, which is the location of this dissertation research. In China, the male
excess is especially pronounced (Hao et al., 2004; Wei, Derson, Xiao, Li, 8:
Zhang, 1999).

The association between year of birth (as expressed in age strata) and
drinking and AUD is not consistent in the literature. For example, some
studies have found that people in younger age strata are more likely to have a
history of drinking, binge drinking, and AUD (Degenhardt, Chiu, Sampson,
Kessler, 8: Anthony, 2007; Higuchi, Parrish, Dufour, Towle, 8: Harford, 1994;
Kessler et al., 1994; Kim et al., 2008; Naimi et al., 2003; Serdula, Brewer,
Gillespie, Denny, 8: Mokdad, 2004); while some other studies find higher
likelihood of drinking and AUD in middle age groups (Rehm, Room, van den
Brink, 8:]acobi, 2005; Wilsnack et al., 2000; X. Zhou et al., 2006). Explanations
for these age-related variations include: chronicle age of the person, period
effects, and cohort effects. For example, AUD usually starts to emerge during
adolescent to early adulthood. Surveys among pre-teen population may find
very low occurrence of AUD because they have not started to drink alcohol
yet. During some periods in the history, policies and regulations may
influence the availability of alcohol, e.g. the US National prohibition of alcohol
(1920-33). The occurrence of AUD may be different for these periods
compared to others. People in some cohorts may be more or less likely to be
abstainers, e.g. “baby boomers”. However, these three factors are highly
intertwined with each other. It is especially difficult to tease them out from

cross-sectional studies when survival of drinkers may also play a role in

16

estimates. Nevertheless, up-to-date data on AUD in different age strata
provide information about the distribution of disease burden of AUD in
population subgroups. In China, the patterns of drinking also are seen to vary
across age strata (Hao et al., 2004): current drinking increases with age
peaking in middle-age group (36-50), and then declined in older age groups.
For centuries, it has been observed that the AUD (including alcoholism)
tend to show familial aggregation (Merikangas, 1990; Radouco-Thomas et al.,
1979; Schuckit et al., 2001). Recently, studies equipped with advanced
techniques have found that AUD cases differ from controls in selected
genotypes. Some studies have been able to pinpoint mechanisms of the family
influence down to the level of Single nucleotide polymorphism (SNP)
(Edenberg et al., 2004; Schuckit, Smith, 8: Kalmijn, 2004; Zlotnick et al., 2006).
These AUD-associated genotypes and SNPs are involved in various biological
and pharmacological functions including neuro transmitters, ethanol
metabolism, cell adhesion, etc (Schuckit, Smith, 8: Kalmijn, 2004). As the
technology and knowledge about molecular genetics of AUD continues to
evolve, our understanding of these observed locational differences will be
clarified to the point that we will regard some genotypes as causal influences
on AUD. However, at the present time, these observations remain
associational in nature, and the evidence of causal influence is not yet fully
developed. Notwithstanding the molecular genetic pathways of family-genetic
influence, there also are other mechanisms of note. For example, social

learning mechanisms can foster drinking behavior (e.g. offspring of

17

abstainers are more likely to be abstainers as well (Harburg, DiFranceisco,
Webster, Gleiberman, 8: Schork, 1990)). In addition, studies have provided
evidence that some cultures tolerate drinking-related misbehavior more than
others (Donovan 8: Molina, 2008; Room, 2006).

2.5.2. Cross-country or cross-region variation

Theoretically, country or region is time-variant. However, compared to
the rate of occurrence of AUD, country and region are relatively stable
characteristics. Thus, we treat these characteristics as time-invariant in this
section of the background.

Summarized from 55 studies from around the world before 2000, the
point prevalence estimate of AUD varied widely from country to country and
region to region. For example, estimates of point prevalence in countries in
the America (North and South) are typically higher than those in Islamic
countries. In some of African countries (e.g. Nigeria and Ethiopia), the
estimate is close to zero while in other African countries (e.g. South Africa,
Zambia, and Zinbabwe). The same pattern can be seen in incidence as well (C.
Mathers 8: Ayuso-Mateos, 2000). There is an obvious unbalance in literatures
written in English regarding the occurrence of AUD from different regions or
countries in the world. For example, the US population is much more
frequently studied than populations in some other regions and countries,
especially lower-income and non-English speaking countries (e.g. countries
in Africa, the middle-east, and Asia). For this reason, we summarize studies in

the US first and then expand to evidence from other countries.

18

In the US, the tradition of community surveys of drinking dates back to
the 19503 and 19608 (e.g. see W. B. Clark 8: Hilton, 1991). Using these surveys
and national sales and tax records about alcohol, in some countries the
alcohol use disorders (AUD) have become quite common. For example, the
2001-2002 US National Epidemiologic Survey on Alcohol and Related
Conditions (NESARC) is a cross-sectional study assessing the occurrence of
alcohol drinking and related problems based upon a nationally
representative sample of household-dwelling individuals. Estimates suggest
that an estimated 30% of the US population age 18 and older had a lifetime
history of AUD, defined to include both DSM-IV alcohol dependence and
DSM-IV alcohol abuse (D. S. Hasin, Stinson, Ogburn, 8: Grant, 2007).

As noted by Anthony and colleagues (].C. Anthony 8: Van Etten, 1998),
epidemiological measures based upon the lifetime history include lifetime
cumulative incidence proportion (of those who survived to be assessed). This
same proportion is sometimes referred to as “lifetime prevalence”, but, it
deviates from epidemiological definition of a prevalence measure as one that
varies with both the incidence of the condition and the duration of the
condition (Gordis, 2004). Hereinafter, this measurement will be called by the
term “cumulative occurrence”, which avoids the conundrum where the
“lifetime prevalence” term is used, and is in keeping with a contemporary

proposal to send the concept of “lifetime prevalence” into retirement (LC.

Anthony 8: Van Etten, 1998).

19

Whereas Hasin and colleagues found that the lifetime occurrence of AUD
was about 30% overall, the estimated 12 month prevalence of AUD was 8.5%
(D. S. Hasin, Stinson, Ogburn, 8: Grant, 2007). Two prior large-scale
epidemiological studies in the US, the Epidemiologic Catchment Area (ECA)
program and the National Comorbidity Study (NCS), yielded AUD 12 month
prevalence estimates of about six percent and seven percent, respectively
(Kessler et al., 1994; Robins, Locke, 8: Regier, 1991), not appreciably distant
from the corresponding NESARC estimate.

Besides the occurrence of AUD in the entire population, another
pertinent set of estimates involves the occurrence of AUD in the sub-
population of drinkers. For many reasons, (e.g. religion, personal choice),
many people in the population are lifetime abstainers, who thereby remain
not at risk of developing AUD. Estimated from the 1994 NCS, about 15% of
those who ever consumed alcohol had developed alcohol dependence (J. C.
Anthony, Warner, 8: Kessler, 1994). Studying current drinkers who
participated in the 2001 National Household Survey on Drug Abuse in the US,
age 12 years old and above, Harford and colleagues found that among young
adult users (18-23 years old), one in three male current drinkers and one in
five female current drinkers had experienced at least one of the clinical
features of alcohol dependence during the 12 months prior to the assessment.
“Tolerance” and “a great deal of time spent on alcohol” were the most
common clinical features among current drinkers. With respect to the DSM-IV

alcohol abuse, one in four male and one in eight female current drinkers, age

20

18—23, had experienced at least one of the clinical features. Hazard-laden use
was the most common clinical feature in this group (Harford, Grant, Yi, 8:
Chen, 2005).

As for other countries of the world, a number of research groups have
contributed to an evidence base showing that alcohol problems are not rare
(Borges et al., 2005; Demyttenaere et al., 2004; McGovern et al., 2004; Rehm,
Room, van den Brink, 8: Jacobi, 2005; Rehm, Taylor, 8: Patra, 2006; Soueif,
Yunis, 8: Taha, 1986). In some Latin American countries (e.g. Brazil, Chile, and
Mexico), the estimated cumulative occurrence of AUD ranged from 5% to 20%,
12-month prevalence from 4% to 10% (Andrade, Walters, Gentil, 8: Laurenti,
2002; Barros, Botega, Dalgalarrondo, Marin-Leon, 8: de Oliveira, 2007;
Medina-Mora, Borges, Benjet, Lara, 8: Berglund, 2007; Vicente etal., 2006).
Rehm and colleagues (Rehm, Room, van den Brink, 8: Jacobi, 2005) attempted
to summarize all pertinent epidemiological studies from European countries,
as published since 1990. The cumulative occurrence of DSM-IV alcohol
dependence ranged from 2% to 14% in males, with substantially smaller
estimates for females. As for DSM-IV alcohol abuse, the corresponding ranges
were 1% to 14% (males), and 1% to 3% (females). Even in countries of Africa,
such as Nigeria, where there are many of the Islamic faith who abstain from
alcohol, Gureje and colleagues used a similar methodology as the NCS
approach. In their Nigerian Mental Health Survey (NMHS), they found a
lifetime cumulative occurrence estimate of 2.8% in a general population

sample (Gureje, Lasebikan, Kola, 8: Makanjuola, 2006). In Australia, estimates

21

measured for the occurrence of alcohol dependence have been similar to the
US estimates (McBride et al., 2008). In Japan and China, the values have
tended to be lower than those in Europe, the US, and Australia (Demyttenaere
et al., 2004; Kawakami, Shimizu, Haratani, Iwata, 8: Kitamura, 2004). It has been
reported that in eastern countries, there is a specific alcohol-metabolizing
pathway, involving ADH and ALDH, rests upon a genetic scaffold that
arguably affects drinking behavior (Couzigou, Coutelle, Fleury, 8: Iron, 1994;
Peng, Chen, Tsao, Wang, 8: Yin, 2007). More details of the pathway can be
found in section 2.6.2. In brief, It has been found that polymorphisms in genes
encoding enzymes to metabolize ethanol causes a larger proportion of Asians
to have flushing effects at a lower amount of alcohol intake (Schuckit, 2009a; Y.
C. Shen et al., 1997). In Japanese and Chinese, it has been argued to be a
protective factor against excessive drinking. In Korean, however, the drinking
culture is believed to be “drinking through flushing” (R. C. Johnson et al.,
1984;]. Y. Park et al., 1984). The occurrence of AUD is almost equivalent to
Europe, the US, and Australia (J. T. Park, Kim, 8: Jhun, 2008).

Since the population under this study belongs to the Chinese population,
this paragraph provides a summary of previous studies of alcohol drinking
and related problems in China. Hao et a1. summarized nine epidemiological
studies on alcohol-related disorders in China from 1984 to 1994. The overall
occurrence of alcohol dependence varied substantially from virtually zero
percent up to the 5—6% level (Hao, Chen, 8: Su, 2005). Although it is difficult to

conduct meta-analysis of estimates from these studies due to the differences

22

in sampling methodology and diagnostic criteria, the occurrence of AUD has
tended to be greater in studies conducted more recently compared to earlier
studies. One recent WHO sponsored survey on alcohol use conducted in five
areas during 2001, (n=24992), found that DSM-III-R defined alcohol used
disorders (AUD) may be more common than other studies in China. The
prevalence estimate for the AUD was 9% among males (7% alcohol
dependence, 2% alcohol abuse) but was indistinguishable from zero in
females (Huang, Zhang, Momartin, Cao, 8: Zhao, 2006). A study conducted in a
metropolitan city in Hebei province yielded 22% overall point prevalence of
alcohol abuse with 30% and 5% for males and females, respectively (Jiafang,
Jiachun, Yunxia, Xiaoxia, 8: Ya, 2004). However, results from these studies are
not directly comparable because of differences in assessments. For example,
in two studies conducted by Hao et al., DSM-III criteria were used to assess
cases of AUD. The Hebei study by Jiafang et al. used a screening test that
included a combination of quantity of alcohol consumption and some alcohol
induced problems: a score of 8 and above qualified the drinker as a case of
alcohol abuse. These differences in case definition preclude direct
comparison of results between these studies and other studies.

With respect to secular trends in China, summarizing the three studies
conducted during 1993 to 2001 by Hao et al., thre appears to be a quite stable
trend of AUD (Hao et al., 2004; Wei, Derson, Xiao, Li, 8: Zhang, 1999; Wei et al.,
1995). In summary, for China, the occurrence of AUD is around 8% in males

and is quite rare among females.

23

2.5.3 Time-varying characteristics

There are time-varying characteristics that have been found to be
associated with drinking-related problems as well, including marital status,
occupation, income level, educational attainment, and religion. The
association between AUD and these characteristics has not been entirely
consistent (Andrews, Henderson, 8: Hall, 2001; J. C. Anthony, Warner, 8:
Kessler, 1994; Crum, Chan, Chen, Storr, 8: Anthony, 2005; Crum, Helzer, 8:
Anthony, 1993; Crum, Storr, 8: Anthony, 2005; Gureje, Lasebikan, Kola, 8:
Makanjuola, 2006; D. S. Hasin, Stinson, Ogburn, 8: Grant, 2007; Huang, Zhang,
Momartin, Cao, 8: Zhao, 2006). In addition, the real meaning of these
characteristics can vary under different contexts. As such, special caution is
required when comparing these characteristics from one society to another.
Islamic religion has been associated with abstinence from alcohol in both
individual level and etiological level studies (Michalak, Trocki, 8: Bond, 2007;
WHO, 2004b).

Despite inconsistencies, some general patterns emerge in specific
contexts. For example, with respect to marital status, AUD seem to occur more
frequently among the separated-divorced, and among the never married;
there is some evidence of late-life incidence of AUD, perhaps in connection
with the experience of becoming a widower (Power, Rodgers, 8: Hope, 1999;
Prescott 8: Kendler, 2001). With respect to educational attainment, the work of
Professor Rosa Crum indicates that dropout in high school is associated with

higher incidence of AUD (Crum, Chan, Chen, Storr, 8: Anthony, 2005; Crum,

24

Helzer, 8: Anthony, 1993; Crum et al., 2006). With respect to occupation,
Mandell et al, in the US, found several occupations with especially high
prevalence of AUD (e.g. construction and transportation), and some with
especially low prevalence (e.g. white-collar occupations). Reed et a1. linked
drug problems to psychosocial dimensions of the work environment (Reed,
Anthony, 8: Breslau, 2007). With respect to income, in the US, there is an
inverse association between income level and alcohol dependence (D. S.
Hasin, Stinson, Ogburn, 8: Grant, 2007; Keyes 8: Hasin, 2008). In China, there is
little evidence on these associations, and the present investigation will be one
of the first to provide empirical estimates on these topics. However, Zhao et al.
found that being married and divorced (compared to never married), workers
and government officials (compared to students) are more likely to be a
current drinker (X. Zhou et al., 2006).

Numerous cross-section studies have found that earlier age of the first
drink is associated with higher occurrence of later negative drinking-related
consequences, e.g. heavy drinking, socially maladaptive drinking, and
alcohol dependence (Chou 8: Pickering, 1992; O'Grady, Arria, Fitzelle, 8: Wish,
2008; Rothman, DeJong, Palfai, 8: Saitz, 2008). Follow-up studies also found that
earlier age of drinking is associated with higher AUD incidence (Dawson,
Goldstein, Patricia Chou, June Ruan, 8: Grant, 2008; Kendler, Prescott, Neale,
8: Pedersen, 1997). One limitation of these studies is from the fact that
individuals with earlier initiation of drinking are exposed to alcohol effects for

a longer time period compared to later-onset individuals, and therefore have

25

 

had more time to develop AUD. Time-to-event analytical tools serve better in
this context, and have been used in several studies to confirm that earlier
onset of drinking is associated with more rapid development of alcohol
dependence (DeWit, Adlaf, Offord, 8: Ogborne, 2000; Hingson, Heeren, 8:
Winter, 2006). DeWit and colleagues found a graded inverse association
between age of first drink and the occurrence of alcohol dependence. In a 10
years span after the first drink, the estimated incidence of alcohol
dependence was one percent in those who had their first drink after 19,
compared to 16% in those who had the first drink when they were 11 or 12
years old (DeWit, Adlaf, Offord, 8: Ogborne, 2000). It can be argued that the
observed higher occurrence of AUD in earlier onset drinkers is due to some
background or predisposition not well-controlled in these studies, e.g., family
history, childhood adversities, or early mental disturbances. Mixed evidence
has resulted in studies that controlled for these possible confounding such as
family history (FH). As might be expected, studies with larger sample sizes
find a statistically robust age of onset associations with FH controlled (e.g.
Dawson, Goldstein, Patricia Chou, June Ruan, 8: Grant, 2008), but studies with
smaller samples have not (e.g. King 8: Chassin, 2007; Warner 8: White, 2003).
Using latent variable analysis techniques, Kuo and colleagues found in twin-
pair data that earlier onset of drinking was associated with greater alcohol
problems, and found it predicted younger ages of onset of regular drinking,
as well as the first alcohol dependence clinical feature (Kuo, Aggen, Prescott,

Kendler, 8: Neale, 2008).

26

Z'r‘

 

In some research, AUD and other drinking-related outcomes have been
found to occur more often in individuals with adverse experiences during
their childhood years, as compared to individuals whose childhood did not
include these adversities. Because childhood adversity is one of the main
topics of this dissertation, a more detailed literature review on this topic
appears in a later section of this dissertation. To the best of the author’s
knowledge, there have been no China studies on the topic of early age of
drinking or childhood adversity and occurrence of AUD.

2.6 The third rubric of epidemiology: cause

2.6.1 Causes in epidemiology

There are four main types of causal inference approaches that are
commonly utilized in epidemiological studies, namely graphical models,
counterfactual models, sufficient-component cause models, and structural-
equations models (Greenland 8: Brumback, 2002). These four methods are not
mutually exclusive. In many circumstances, they are transferable (Flanders,
2006). Each method has its own advantages and disadvantages compared to
the others (Parascandola 8: Weed, 2001). The Scottish philosopher David
Hume was the first to explicitly define a cause from the counterfactual
perspective as “if the first object had not been, the second never had existed” (Hume,
1977). Although this original statement is deterministic, the counterfactual
definition of cause can be modified to be probabilistic: “if the first object had
not been, the probability of the second would have changed” (Parascandola 8:

Weed, 2001). Over the past century, the counterfactual approach of causal

27

inference has been serving as the foundation of many quantitative methods
used in epidemiological research (Greenland 8: Brumback, 2002). Guided by
this operational definition, this section reviews previous studies on causes of
drinking-related problems.

As with many prevalent human diseases that surface to prominence, the
causes of drinking-related problems include many genetic and environmental
factors, and their interplay. Among various possible causes, alcohol is a
necessary cause, but is not sufficient. In fact, none of possible causes will be
sufficient by itself, as exposure to alcohol will be a necessity, either as a self-
administered exposure or a passive exposure (e.g. drinking by one’s mother
with a subsequent fetal alcohol spectrum disorder).

2.6.2 Macro-social influences

There is widespread agreement that socially shared macro-social
influences can contribute to the individual-level risk and the population-level
occurrence of drinking-related problems. For example, Professor Harold
Holder maintains that alcohol problems can be controlled via careful
manipulation of community-level variables such as alcohol price (or taxation),
policies (Andreasson, Holder, Norstrom, Osterberg, 8: Rossow, 2006; Holder,
2007). These are socially shared “policy instruments” that would have
functional significance as “causes of incidence” (Rose, 2001) in that they can
account for population-level variation in the occurrence of drinking-related
problems, even if they cannot be studied at the level of individuals. Some

other examples of macro-social “causes of incidence” variables are divorce

28

proportion (Fillmore, Golding, Leino, Ager, 8: Ferrer, 1994), unemployment
proportion, crime in the society (Ager et al., 1996), and strictness of law
enforcement (Sloan, Reilly, 8: Schenzler, 1994). A woman’s position in the
social structure of society has been found to be associated with proportion of
drinking problems in women and macro-level forces that influence equality of
access and opportunity in the workplace may qualify as causal influences at
this level (Rahav, Wilsnack, Bloomfield, Gmel, 8: Kuntsche, 2006). Recently
Room, Schmidt, Rehm, and Makela have argued that increasing affluence at
the national level (e.g. due to globalization) will cause increased alcohol
consumption and increased incidence of alcohol related hazards (Room,
Schmidt, Rehm, 8: Makela, 2008).
2.6.3 Meso-level influences
Besides macro-level variables, some mesa-level variables, which lie
between social structural and individual level, have also been suggested as
potential influences on drinking behavior and drinking-related problems.
Derived from social learning theory (Petraitis, Flay, 8: Miller, 1995), peer
influence has been one of the most commonly studied meso-level variables in
adolescent populations. Numerous studies have found that peer alcohol
drinking, peer encouragement of drinking, and peer deviance all have
possibly causal influences on adolescent drinking behaviors, early-onset
drinking, and associated problems (Ary, Tildesley, Hops, 8: Andrews, 1993;

Blackson 8: Tarter, 1994; Coombs, Paulson, 8: Richardson, 1991; S. C. Duncan,

29

Duncan, 8: Strycker, 2006; Hawkins, Catalano, 8: Miller, 1992; Quine 8:
Stephenson, 1990; Wu, Lu, Sterling, 8: Weisner, 2004).

Apparently, this peer influence on drinking behavior is not completely
attributable to levels of parental drinking problems, or individual-level
predispositions to experience of drinking problems (Bahr, Marcos, 8:
Maughan, 1995; Barnow, Schuckit, Lucht, John, 8: Freyberger, 2002; Wood,
Read, Mitchell, 8: Brand, 2004). One especially compelling study of peer
influence on heavy drinking involved a randomized experiment for which
incoming university freshmen were assigned at random to the dormitory
roommate. The randomization created pairs of roommates, two freshmen,
sometimes both, sometimes one, and sometimes neither of whom entered
university with a prior history of heavy drinking. They found that male
students with a history of heavy drinking had higher levels of alcohol
consumption when pairing with a roommate with a history of heavy drinking
as well, as compared to pairing with one without such a history (G. J. Duncan,
Boisjoly, Kremer, Levy, 8: Eccles, 2005). Although there have been
suggestions for a more appropriate analytic strategy and more fine-grained
methods to better understand the mechanism, this study provided empirical
evidence of peer influence on drinking. In addition, several intervention
studies have found that peer-led intervention can be effective in reducing
adolescent drinking. Inclusion of peers in intervention sessions optimized
estimated effects of alcohol reduction programs (Perry et al., 1989; Rowe et al.,

2007; Tevyaw, Borsari, Colby, 8: Monti, 2007).

30

Other suspected meso-level variables include sibling influence (Trim,
Leuthe, 8: Chassin, 2006), school policies (Desousa, Murphy, Roberts, 8:
Anderson, 2008), and living in dormitories while in residence at college
(Barnes, Welte, 8: Dintcheff, 1992). In the absence of experimental evidence, it
is possible that these just-listed researches should be reviewed under the
second rubric of epidemiology (location)-that is, until more compelling
evidence for causal influence has been gathered. The evidence on these
researches is not as compelling as the evidence on peer influence.

2.6.4 Micro-level influences

As reviewed under the heading of the second rubric, it is widely agreed
that there is family-genetic predisposition for AUD. For instance, twin studies
have found that the concordance of alcohol dependence is greater in
monozygotic twins, who share approximately the same individual-level
genome, than that in same-sex dizygotic twins, who on average share half of
that genome (Kessler, Davis, 8: Kendler, 1997). In general, twin studies and
adoption studies have yielded heritability in the range of 50% to 60% for
alcohol dependence (Dick 8: Bierut, 2006; Merikangas, 1990; Prescott et al.,
2005; Schuckit, 2009b). These studies estimate the proportion of AUD variance
that can be attributed to genetic factors. However, they do not pinpoint
specific loci or region in the human genome that might account for the
occurrence of AUD, and that might become future targets for intervention.

Recent advances in genetic engineering have made it possible to

manipulate gene polymorphisms in mice. For example, to study functions of

31

specific genes, the knock-out technique can turn off a specific gene locus to
study functional changes (Hooper, Hardy, Handyside, Hunter, 8: Monk, 1987).
In contrast, the knock-in technique can insert a specific gene (Kuehn, Bradley,
Robertson, 8: Evans, 1987). Using various gene targeting techniques, animal
studies have found that some specific polymorphisms in selected genes
increase or decrease alcohol intake in mice. The most extensively studied
genes are those encoding neurotransmitters, such as GABA, dopamine, and
serotonin; cell adhesion genes; and protein kinase genes (Crabbe, Phillips,
Harris, Arends, 8: Koob, 2006; Hishimoto et al., 2007; Newton 8: Messing, 2006;
Racz et al., 2003; Werner et al., 2006). It is believed that some of these genes
account for loci identified by GWA studies, and there is a convergence in
evidence between animal studies and human GWA studies (Uhl et al., 2008).
However GWA studies have revealed alcohol dependence associated loci on
17 out of the 23 human chromosomes (Ehlers et al., 2004; Zlotnick et al., 2006).
Furthermore, recent discoveries in epigenetics and gene expression have
made the understanding of causes of alcohol dependence even more
complicated. Therefore, multi—disciplinary effort is needed to draw the
complete picture of alcohol dependence, and to trace specific genetic
meditational pathways that account for intergenerational “transmission” of
susceptibilities for alcohol dependence (J. Liu et al., 2006; Uhl et al., 2008).
Besides genetic factors, environmental factors play important roles in
AUD as well. Estimated from twin studies and adoption studies, environmental

factors account for more than 40% of the variance (Agrawal 8: Lynskey, 2008;

32

Kendler, Myers, 8: Prescott, 2007; J. Liu et al., 2006; Prescott et al., 2005).
Various environmental factors have been investigated. However, due to
underlying heterogeneities and unobserved confounding variables, it is not
easy to infer definite cause-effect relationships from observational studies.
Many (but not all) studies have found evidence of interactions between
genetic and environmental factors, e.g. the 5-H'I'I' genotype and stressful life
events (Caspi 8: Moffitt, 2006; Dai, Thavundayil, Santella, 8: Gianoulakis, 2007;
Dick, Rose, Viken, Kaprio, 8: Koskenvuo, 2001; Schuckit 8: Smith, 2006;
Schuckit et al., 2005). Among many possible environmental causes, child
abuse and neglect has been investigated as a suspected “cause of cases” and
determinant of the individual-level risk of becoming a case of AUD by some
research teams. The dissertation returns to this topic in section 2.7, which
covers the possible role of child abuse/neglect as an early life condition that
might account for variation in individual-level risk of AUD and related
problems.
2.7 The fourth rubric: mechanism

This section is organized in relation to the following topics: 1) chemistry
and pharmacology of ethanol, 2) metabolism and biotransformation of ethanol,
3) reinforcing effects of alcohol, 4) natural history of alcohol problems, 5)
comorbid condition, and 6) alcohol-related disabilities and impairment,
including secondary social maladaptation and hazard-laden drinking.

2.7.1 Brief introduction of chemistry, pharmacology of ethanol

33

In chemistry, alcohol can be said to be any organic compound where “a
hydroxyl group (-OH) is bound to a carbon atom of an alkyl or substituted
alkyl group” (Kopnisky 8: Hyman, 2002). In simple English, an alcoholic
beverage refers to a drink containing the chemical drug compound known as
ethanol. Ethanol is the principal, active ingredient in alcoholic beverages,
traditionally produced by fermentation (fermentation involving the f
metabolism of carbohydrates by certain species of yeast under anaerobic
conditions). The chemical composition of ethanol is written as

H H

l |
H-C-C-H

I |

H OH

 

Ethanol generally is colorless, volatile, and water soluble with a mild odor.
Ethanol in humans often shows biphasic responsesin CNS (central
nervous system)-mediated behavioral functions, with a more prominent
depressant effect to the CNS at higher doses, and sometimes with
disinhibition of behavior that can simulates ‘stimulant’ outcomes at lower
doses. These effects seem to be mediated via certain subtypes of the gamma-
aminobutyric acid A (GABA-A) receptors and inhibition of NMDA (N-methyl-
D—aspartic acid) glutamate receptors. These neurotransmitter mechanisms are
intermediaries for relaxation, release from anxiety, sedation, and lowering of
inhibitions, with apparent ‘stimulation’ of behavior-hence, the biphasic
response. Nevertheless, higher doses, some other targets are also involved,

such as sodium channels, serotonin, as well as dopamine receptors and other

34

psychopathological pathways, and the CNS depressant response become
more salient. Whether there are two separate mediating mechanisms is not
yet clear (Boehm et al., 2004; Bowirrat 8: Oscar-Berman, 2005; Gorwood et al.,
2000; Kopnisky 8: Hyman, 2002; Mihic et al., 1997). At the higher doses,
ethanol impairs sensory and motor functions and can slow cognition. An
extremely high dose of ethanol can cause unconsciousness and possible
death. Meta analysis suggests that the LD50 (half lethal dose) of ethanol is
10,300 mg/kg for rats, 6800mg/kg for mouse and the LD50 in humans is
estimated to be 330g (276-455) for a 70kg healthy adult, which translates into
approximately 16 bottles of half liter beer with 5% alcohol by volume (Gable,
2004). Thus, ethanol can be lethal; the lethal risk is usually higher for people
with a smaller volume of body water, such as children, women, and low body
mass individuals. However, clinical significant CNS depressant effects take
place at sub-lethal levels; these effects happen much more commonly than
lethal overdose, and are more relevant to health conditions than is the lethal
threat directly from ethanol. With respect to dos-response relationships,
Goldberg evaluated the behavioral and physiological changes in 160 healthy
volunteers. Evidence from regression models supports the idea that
intoxication can appeare as low as 22mg/DL of blood alcohol concentration in
humans (GOLDBERG, 1966). The legal threshold for blood alcohol
concentration typically ranges from 50 to 80 mg/DL for most Western

countries.

2.7.2 Metabolism and biotransformation of ethanol

35

Through a series of reactions, complete metabolism of ethanol produces
water and carbon dioxide. The major reactions can be depicted as,

CZHGO(Ethanol)->CZH4O(Acetaldehyde) -+CZH4OZ(Acetic Acid) —>Acetyl-
CoA—>H20+COz (Kopnisky 8: Hyman, 2002)

The enzyme alcohol dehydrogenase (ADH) oxidizes ethanol into
acetaldehyde, which is then converted into the relatively harmless acetic acid
(vinegar) by acetaldehyde dehydrogenase (ALDH) (Kopnisky 8: Hyman, 2002).
Acetaldehyde, an intermediate product of ethanol metabolism, also is toxic,
with negative health effects. Acetaldehyde is listed as a probable human
carcinogen by the US Environmental Protection Agency (U.S.EPA, 1994).
Ingestion of too much acetaldehyde can cause a cluster of unpleasant effects
characterized by facial flushing, dehydration, headache, palpitations, nausea,
and vomiting, often characterized to as a “hangover” (I-larada, Agarwal,
Goedde, Tagaki, 8: lshikawa, 1982). As described above, the concentration of
acetaldehyde depends on the ethanol intake, the amount of ADH (synthesizing
acetaldehyde from ethanol), and the amount of ALDH (degrading
acetaldehyde into acetic acid). There is wide variation in the level of ADH and
ALDH across individuals, which results in different rates of ethanol
metabolism. Low levels of ALDH or high levels of ADH cause acetaldehyde to
accumulate through increased synthesis and/ or decreased metabolism.

Genetic research has disclosed that the functional variants in genes
encoding ADH and ALDH can account for variations in levels of ADH and/or

ALDH (Edenberg et al., 2006; Kuo, Aggen, Prescott, Kendler, 8: Neale, 2008). It

36

has been widely reported that in a larger proportion of Asian populations (e.g.
Chinese, Japanese) these allele variants can cause more rapid and longer-
lasting accumulations of acetaldehyde, which in turn can discourage
additional alcohol intake due to the soon-appeared unpleasant effects (C. C.
Chen et al., 1999; Y. C. Shen et al., 1997; Thomasson et al., 1991). On the other
hand, regardless of individual variations in enzymes to digest ethanol,
excessive intake of ethanol can cause accumulation of acetaldehyde in the
human body.
2.7.3 Reinforcing effect of alcohol
Alcohol is a psychoactive drug that serves reinforcing functions. Via the
neurotransmitter mechanisms already discussed, alcohol may disturb the
reward circuit by interfering with neurotransmitters and their receptors, such
as the D2 dopamine receptors (DRDZ), glutamate, serotonin, and the GABA-A
receptors (Kopnisky 8: Hyman, 2002; Lewis, 1996). Whereas the exact
mechanisms underlying the reward system are not yet fully understood, it is
known that these neurotransmitter systems can work interactively to yield
reinforcement of the drinking behavior (Lewis, 1996). This reinforcement
mig ht be distinct from so called “natural rewards”- those sought be avoid
de‘étth from starvation, e.g. associated with hunger. Whereas some scholars
have described this reward effect as an acquired “pleasure” from alcohol
intalte, from the behaviorism perspective, ethanol can function as a positive

reiI‘tforcer for sustained drinking behavior, operationally defined as an event

37

that increases the probability of a subsequent event, with no appeal to the lay
concept of “pleasure”.

Besides positive reinforcement, alcohol also can serve a negative
reinforcing function. This function was described by Solomon and Corbit in
their ‘opponent process theory,’ which hypothesizes that the initial drug
intake induces a ‘hedonic’ state. In response, the CNS automatically seeks
homeostasis, and with a counteraction to reduce the intensity of the ‘hedonic’
effect. After the drug wears off, this CNS-mediated counteraction persists,
inducing a negative emotional state (Koob, 2006; Solomon 8: Corbit, 1974).
Subsequently, individuals may increase consumption or drug-seeking
behavior in order to relieve this acquired negative emotional state. According
to the theory, the positive reinforcement that can cause decreased reward

thresholds, and with the negative reinforcement, there is an increased self-

administration of alcohol. Thereafter, positive and negative reinforcement
work together to preoccupy the individual with drug-related activities.
Guided by this theory, animal studies find consistent evidence (Koob, 2006).

Alcohol’s reinforcing function maybe stronger for some individuals than

the drive state required by hunger for food. Similar behavioral mechanisms
may influence dependence syndromes that involve other drugs (e.g., cocaine)
and behavior functions of the non-drug behavioral repertoire, such as sports
activities, or musical performance and practice, and gambling (C. Y. Chen et
31, a 2004; Wightman 8: Robinson, 2002). As the exogenous alcohol reinforcers

become established, a drinker may develop alcohol dependence.

38

The “internal” functions and biochemical changes, which coincide with
alcohol’s reinforcing functions, cannot be observed by unaided naked eyes.
They can be studied with brain imaging techniques and can be reflected in
behavioral manifestations, such as compulsion-like drinking behavior,
tolerance, withdrawal, and the other facets of the alcohol dependence
syndrome. The positive reinforcement is believed to be a crucial mechanism
behind alcohol tolerance, with negative reinforcement in a similar position
with respect to alcohol withdrawal. Compulsive drinking and lost of control
over alcohol is a manifestation of both positive and negative reinforcement
(Koob, 2006). In addition to neuroadaptational changes that coincide with
pharmacological tolerance and withdrawal, there also may be alcohol-related
social maladaptation and interpersonal and social problems as well (e.g.,

drink-induced violence, family or legal troubles, and drunk driving).
2.7.4 Possible natural history of AUD

Our understanding of the natural history of alcohol use and related
PrOblems has been advanced with long term longitudinal studies in which
attrition has been limited. Due to logistical difficulties, these studies are rare.

Nonetheless, these studies have found that the remission of alcohol use in

ADDS occurs frequently (V aillant, 1996; Vaillant 8: Milofsky, 1982), with little

e"'iC‘Ience of male-female differences in the course of alcohol dependence

(Schuckit, Daeppen, Tipp, Hesselbrock, 8: Bucholz, 1998). Well designed
c370 ss-sectional studies can provide valuable insight into the natural history

albeit there can be limitations such as recall bias and incomplete reporting of

39

past events. For example, findings from the cross-sectional component of the
ECA study estimated a median age of onset of AUD to be 21 years and 90% of
all AUD cases had experienced their first clinical feature of AUD before the
age of 38 years (Helzer JE, 1991). The estimated prevalence of recent AUD
decreased across age strata, from about 4% in 18-24 year olds to about 1%
among those age 65 years old and above (Regier et al., 1993). A similar trend
was found for cumulative occurrence in Caucasians, but not for African
Americans. In African Americans, peak prevalence of AUD presented in the
middle age stratum, which was 45 to 64 years old (Regier et al., 1993).
2.7.5 Comorbid conditions
Various comorbid conditions have been observed to co-occur with
alcohol dependence, including depression, anxiety disorder, and tobacco
dependence (Regier et al., 1990; Schuckit, 1985). Many studies have found
that childhood conduct disorder, cognitive problems, and attention problems
forecast later onset of fully expressed alcohol dependence (Elkins, McGue, 8:
Iacono, 2007; Giancola 8: Moss, 1998; Gorenstein, 1987; Looby, 2008; Molina,
Pelharn, Gnagy, Thompson, & Marshal, 2007; Moss 8: Kirisci, 1995; Myers,
Br OWn, 8: Mott, 1995). It is also widely documented that antisocial personality
is associated with alcohol dependence (Harford 8: Parker, 1994; Stabenau,
1 984). Estimated from clinical patients, onset of antisocial personality occurs
about four years earlier than the onset of alcohol dependence (Bahlmann,
Ifireuss, 8: Soyka, 2002; Stabenau, 1984). There has been evidence that genetic

fautors play a role in the observed association between these pre—existing

4o

conditions and later alcohol dependence, and might function as confounders
(Kendler et al., 2006; Stallings et al., 1997).

Epidemiological studies have also found there is elevated occurrence of
mood disorders (e.g. depression and anxiety disorders) in people with
alcohol and other drug dependence compared to expectations based upon
the general population (J .C. Anthony 8: Petronis, 1989; Grant 8: Harford, 1995;
Kessler et al., 1994; Regier et al., 1990). The self-medication theory
hypothesizes that people drink alcohol to cope with emotional stress or to
release their unhappiness (Quitkin, Riflcin, Kaplan, 8: Klein, 1972). There has
been some evidence supporting the self-medication theory (Bolton, Robinson,
8: Sareen, 2008; Carrigan 8: Randall, 2003; Robinson, Sareen, Cox, 8: Bolton,
2009). There are also scholars who argue for common genetic vulnerability

underlying these comorbid conditions (Merikangas, Leckman, Prusoff, Pauls,
8: Weissman, 1985; Merikangas, Risch, 8: Weissman, 1994; Prescott, Aggen, 8:
Kendler, 2000). In summary, these conditions serve as pre-conditions of
alcohol dependence in some groups of people.
2. 7.6 Alcohol-related disabilities and impairment, including
secondary social maladaptation and hazard-laden drinking
Alcohol dependence accounts for substantial disease burden via
chhol-related mortality and various health consequences including physical,
ernotional, and social consequences. The WHO global burden of disease
project (GBD) found that AUD is one of the leading causes of disease burden

111 the more established market economies of the world. According to the GBD

41

 

data, in 2002, AUD claimed almost 1,800,000 lives worldwide. The peak
alcohol-related mortality occurs in the age group 15-44, and the substantial
AUD disease burden is due mainly to these deaths (WHO, 2004b).
Nevertheless, mortality does not give the complete picture of disease burden
because diseases such as AUD also cause disability and dysfunction. The
Disability Adjusted Life Years index (DALYs) measures disease burden so as
to reflect both premature death and disability (C.D. Mathers et al., 2003). Each
AUD DALY represents one lost year of healthy life, either to AUD-caused
premature death or to an “AUD-attributable” disability. Despite some
limitations, the DALY index has been widely used and has been one of the
main measurements of disease burden nowadays (Bastian, 2000; Reidpath,
Allotey, Kouame, 8: Cummins, 2003). According to the WHO, in 2002, AUD was
responsible for 58,300,000 DALYs (C.D. Mathers et al., 2003; WHO, 2004b).
It must be mentioned that various other physical and mental health
conditions caused by ethanol exposure also contribute to the alcohol-
attributable disease burden. Studies have found that alcohol consumption is a
Possible cause for various physical and mental conditions including cirrhosis
0f the liver, motor vehicle accidents, drowning, falls, poisonings, self-inflicted
inj uries and homicide, low birth weight, some cancers, depression, epilepsy,
hyipertensive disorders (Bazzano et al., 2007; Gu et al., 2007; Huang et al.,
2008; Lin et al., 2005; Rehm, Taylor, 8: Patra, 2006; Ruixing et al., 2006; H. Zhou
et al., 2003). As such, alcohol consumption is one of the top determining

11"-fl-uences on the burden of disease globally each year. The WHO GBD

42

project estimated that in 2002, 1,800,000 deaths are attributable to alcohol
consumption (WHO, 2004b). One third of these deaths were due to un-
intentional injury, e.g. drunk driving, drowning, etc. Some of these deaths are
attributable to alcohol dependence, as when persistence of drinking is
explained by the presence of alcohol dependence. Moreover, alcohol
consumption accounted for 58,300,000 (3.7% of total) DALYs in the same year
(Lopez, Mathers, Ezzati, Jamison, 8: Murray, 2006; WHO, 2004b). In European
countries, where alcohol consumption is higher than the global average, in
2002 alcohol drinking appears to be responsible for 10-1 1% of total DALYs
(Rehm, Taylor, 8: Patra, 2006). Unless something occurs to change the current
trend, AUD alone will climb up to become the fourth most burdensome
disorder within the high-income countries accounting for 4.7% of the total
DALYs in 2030 for those countries (C. D. Mathers 8: Loncar, 2006). As for China,
the main focus of this dissertation, alcohol consumption poses large burden of
disease as well. According to the WHO estimates, as measured in relation to
determinants of DALYs, alcohol consumption ranks high, accounting for 4-8%
of DALYs in China (Grimm, 2003).
2 .8 The fifth rubric: prevention and control
Some community trials found that alcohol consumption, incidence of
drunk driving and assault decreased after the implementations of more
restrict local regulations of alcohol, such as encouraging responsible
lbQVerage service; limiting access to alcohol, especially to adolescents; and

1h-Q:reasing local enforcement of drinking and driving laws (Holder et al., 2000;

43

Stafstrom, Ostergren, Larsson, Lindgren, 8: Lundborg, 2006; Treno,
Gruenewald, Lee, 8: Remer, 2007). Other research teams employed
intervention strategies aiming at improving social skills, e.g. the good
behavior game (Barrish, Saunders, 8: Wolf, 1969), in school kids. Multiple
studies have found that these strategies not only enhance social skills, but also
delay the onset of alcohol drinking and reduce drinking-related problems (E.
' C. Brown, Catalano, Fleming, Haggerty, 8: Abbott, 2005; Kellam et al., 2008;
Poduska et al., 2008; van Lier, Huizink, 8: Crijnen, 2008). There are reports of
other prevention and intervention strategies that was suggested to be
effective in reducing drinking and drinking-related problems. These
strategies include parent-targeted education (Koutakis, Stattin, 8: Kerr, 2008),
peer-led motivational intervention (Fromme 8: Corbin, 2004; Tevyaw, Borsari,
Colby, 8: Monti, 2007), social norm education (Turner, Perkins, 8: Bauerle,
2008), incentive reward (Glindemann, Ehrhart, Drake, 8: Geller, 2007), etc.
Some researchers also found incorporating these strategies with computer-
based survey and feedback to be effective (Bewick, Trusler, Mulhern,
Barkham, a Hill, 2008; Schinke, Schwinn, Di Noia, 8. Cole, 2004).
2.9 Possible causal influence of childhood physical abuse (CPA) and
drinking-related problems
This section of the dissertation is focused upon a sub-topic of the research,
I"*al't‘lely, the possibility that childhood physical abuse (CPA) might influence

drinking-related problems. Current evidence is reviewed under the nine

44

guidelines that are used when evaluating the potential causal significance of

exposure-disease associations in epidemiology (Gordis, 2004; United States

Department of Health, 1964). Table 2.1 lists the nine guidelines and the main

issues or questions under each guideline.

 

Table 2.1. Ningguidelines and corresponding main questions

 

 

 

_guidelines main question
Temporal relationship Does the exposure occur before the disease/condition?
Strength of association How strochis the association?

 

Dose-response relationship
fiplication of the findims
Biologic plausibility

Does the risk of disease/condition increase when the dose of
exposure increases?

Do different studies yield the same results?

Is it coherent with biologic knowledge?

 

 

Consideration of alternate

explanations Could the observed association been explained by confounders?
Could the observed association been explained by model
misspecification?
Does the risk of disease/condition decline when exposure is

Cessation of exposure reduced or eliminated?

 

Consistency with other
knowledge
Specificityof the association

Is the finding consistent with findings from other data?
Is the association specific to the disease?

2.9. 1 Strength of association and replication of findings

Retrospective case-control studies with clinical samples have found that

Odds of AUD are elevated among patients with history of CPA as compared to

controls, and that CPA-associated cases had experienced more drinking-

related problems (G. R. Brown 8: Anderson, 1991; Downs, Capshew, 8: Rindels,

3004; Kunitz, Levy, McCloskey, 8: Gabriel, 1998; Swett, Cohen, Surrey,

C<>lrnpaine, 8: Chavez, 1991). This association also was found in comparisons of

AUD adolescents and community controls (D. B. Clark, Lesnick, 8: Hegedus,

1 997). These studies provide initial evidence of the association between CPA

45

and AUD. Regrettably, these studies may have suffered from a major limitation,
usually referred as “Berkson’s bias,” as can happen when cases are recruited
from clinical settings or from intervention programs. The clinical population,
consisting of treatment seeking individuals, may differ from the general
population in many different ways including demographic characteristics,
such as sex, age, and ethnicity, as well as aspects of personal history, such as
comorbid illnesses or CPA histories. Unless this “transition bias” or “selection
bias” can be taken into account, these differences may lead to biased
estimates of the CPA-AUD association. Additionally, individuals with both
childhood abuse history and alcohol problems may be more likely to seek
treatment than those with only one or none, in a realization of potential
Berksonian bias. As such, there is need for more general population-based
research on this issue.

In one of the earliest population based studies on the CPA-AUD
associations, there was some evidence of a tangible association between
earlier CPA and later AUD (Holmes 8: Robins, 1987, 1988). Later on, estimates
from the NCS, with its nationally representative adult sample of US household
residents, indicated a weak but statistically robust CPA-AUD association
(OR=1.3; 95% CI, 1.1, 1.6; Afifi, Brownridge, Cox, 8: Sareen, 2006). Based on
similar survey methodology, the Ontario Health Survey also found that people

Vvith a history of CPA had an elevated odds of AUD (OR=1.8, 95% CI: 1 .4, 2.3;
MacMillan et al., 2001). Besides CPA, a history of slapping and spanking, and

milder forms of physical punishment, were also associated with excess odds

46

of AUD (MacMillan et al., 1999). A smaller CPA-AUD association also has been
observed in more restricted non-clinical populations, such as college
freshmen, prisoners, lesbians, and Marine recruits (Carrigan 8: Randall, 2003;
Sher, Gershuny, Peterson, 8: Raskin, 1997; Trent, Stander, Thomson, 8: Merrill,
2007).

As now can be summarized from case-control and cross-sectional
research with clinical and non-clinical population samples, the strength of the
association can be characterized as weak to moderate, with ORs generally in a
range from 1.2 to 2.5 (few study provided estimates for Relative Risk and its
standard error). Based upon evidence of this type, Simpson 8: Miller have
already concluded that there is evidence of a possible causal relationship
between CPA and AUD in females. In males, findings were inconsistent
(Simpson 8: Miller, 2002).

2.9.2 Consideration of alternate explanations

The consideration of alternative explanations guideline is attached to
many names. In epidemiology, these “alternative explanations” often are
groped under the heading of “confounding” variable. In econometrics, it
might be said that alternative explanations are sources of unspecified
heterogeneity in the outcome. One of the main plausible confounding
variables or source of heterogeneity on the outcome is parental drinking,
which might account for the CPA as well as AUD susceptibility. Children from
alcoholism-affected families are more likely to be victims of childhood abuse

(DiLalla 8: Gottesman, 1991; Dube, Anda, Felitti, Croft et al., 2001; Widom 8:

47

"u
i" '
”0'

.tv

ss.‘

‘IJ

Hiller-Sturmhofel, 2001); AUD is a condition known to aggregate within
families and to show heritability (Dick 8: Bierut, 2006).

In a few of the just cited studies conducted in general population samples,
there has been a null association between CPA and risky drinking or AUD,
once parental drinking problems are taken into account. The population
subgroups studied in this research has included Marine recruits, lesbians,
prisoners, and US Indian tribes (Hughes, Johnson, Wilsnack, 8: Szalacha, 2007 ;
Koss et al., 2003; Libby et al., 2004; Mullings, Hartley, 8: Marquart, 2004;
Young, Hansen, Gibson, 8: Ryan, 2006). Only one study found with a sample
recruited from a primary care setting, found a possibly non-null association
between CPA and self-defined alcoholism. In this study, parental drinking
problems were collected from offspring (Carrigan 8: Randall, 2003).

In these studies which took parental drinking problems into account, the
generalized linear model (GLM) has been used to estimate a regression
coefficient linking CPA to parental drinking problems. One possible problem
in this approach comes from an assumption of GLM that covariates are
independent. In other words, there might be a violation of an “exogeneity”
assumption: error terms are supposed to be independent for each covariate.
One plausible relationship between parental drinking problems, CPA, and
drinking problems in the offspring is depicted in figure 2.1, such that the

exogeneity assumption might be violated (Engle, Hendry, 8: Richard, 1983).

48

Figure 2.1 A conceptual relationship between parental drinking
problems, CPA, and offspring drinking problems

Parental
drinking

pr0b1ems \

Child _ offs ring
abuse ' drin ing
problems

 

Therefore, CPA is reasonably an endogenous variable with respect to
parental drinking problems, and a simultaneously regression of offspring
drinking problem on both parental drinking problems and CPA might have
violated the assumption of independence. Violation of this assumption can
cause bias as well as inconsistency in estimation (Briscoe, Akin, 8: Guilkey,
1990; Felitti et al., 1998). Resolutions of this endogeneity problem may include
use of instrumental variable methods, which originated and are commonly
used in econometrics, as well as structural equation modeling (SEM)
(Cameron 8: Trivedi, 2009; Greenland 8: Brumback, 2002).

2.9.3 Temporal relationship

In much of the prior research, the retrospective design made causal
inference difficult due to uncertainty about of temporal sequencing. This
uncertainty can be traced to use of cross-sectional study designs and
retrospective methods subject to differential recall biases and differential
survivorship (left-censoring). In the few available prospective studies

investigating the association between CPA and AUD, the evidence is mixed.

49

For example, Jasinski et al. showed that CPA predicted heavy drinking in 113
African American childhood victims (OR=8.7, 95% CI, 1.9, 40.0) holding
parental relationships, parental drinking, and sexual abuse constant (Jasinski,
Williams, 8: Siegel, 2000). In Horwitz et a1. (2001), 908 abused children and
667 non-abused children were identified from court records. The “abused”
exposure group and a court-referred control group were matched on sex, age,
race/ ethnicity, and socio—economic status (SES). Among the 61% followed up
for approximately 20 years, the abused children had increased risk of AUD
development, statistically robust for females only. In deed, with stressful
lifetime events taken into account via regression models, the obserd male
childhood experience a lower risk of AUD (Horwitz, Widom, McLaughlin, 8:
White, 2001; Widom, White, Czaja, 8: Marmorstein, 2007). It is noteworthy that
in this study, although children in the control group were referred to the court
for reasons other than childhood abuse or neglect, it is possible that some of
them also had suffered from childhood abuse or neglect, which might have
biased estimates toward the null for both males and females. Jackson and Sher
also completed longitudinal research, based on a sample of 489 incoming
college freshmen followed for 11 years. They found that the association
between childhood stressors and adulthood AUD was attenuated at p>0.10
when family history of drinking problems was included in their SEM: the
estimate of an effect for childhood stressor on AUD diminished considerably
and was not statistically robust after the family history of alcoholism was taken

into account. Although the attrition level (22%) was fairly low in this study,

50

AUD predicted attrition (Jackson 8: Sher, 2003). In these two studies,
childhood abuse (sexual or physical) and neglect were combined into one
variable. Thus, the effect of each specific type of experience is not known.

Another longitudinal study with a baseline sample of school-recruited
students (n= 1634), but substantial attrition (>60%), also suggested a null
association between childhood physical or emotional abuse and AUD after
taking other childhood adversities, such as parent divorce, family support,
and childhood sexual abuse, into consideration (Galaif, Stein, Newcomb, 8:
Bernstein, 2001). However, besides the limitation of high level of attrition, the
history of childhood maltreatment in this research was based on retrospective
recall at the time of the follow-up assessment, which compromises validity of
this study.

The study by Galaif and colleagues raises issues of note. Since the
assessment of parenting is concurrent with the assessment of AUD, even if the
study design is longitudinal or prospective, this study might be envisioned as
a case-control design or cross-sectional design regarding the CPA-AUD
association. Nonetheless, an AUD generally develops over a long time, such
that follow-up over long spans of time is a requirement if the CPA assessment
is to precede the AUD assessment, and this makes the study vulnerable to
attrition, with differential attrition as a potentially severe complication.

For this reason, epidemiologists typically will conduct a series of case-
control studies, with incidence cases matched to non-cases who passed

through the same interval of risk without developing AUD. Then, in accord

51

with the case-control design, there is a look back (among both cases and
controls) to whether an early-life exposure might be observed more
frequently among the cases as compared to controls. This design is especially
powerful when the early-life experience has a discrete quality and can be
placed in time early in life, well before development of the outcome. For
example, it is possible to ask cases and controls about their childhood
experiences with interview methods that make the study subjects unaware
that these experiences will be studied in relation to a specific outcome such as
AUD. Then, the occurrence of the outcome can be evaluated for post-
childhood years, with knowledge that AUD rarely start during the childhood
years. Accordingly, it may be best to postpone longitudinal and perspective
research on the CPA and AUD association until after case-control research has
been completed to gauge the size of the association, which must be estimated
with some fidelity.

Because there are no prospective or longitudinal studies of CPA and AUD
in China, and because the logistical problems of epidemiological research on
CPA-AUD associations have not been studies, this dissertation involves use of
one of the case-control design protocols with an attempt to sort out the
temporal sequencing issue by employing reference to time frames in
assessments of CPA and AUD: CPA experience “when the respondent was
growing up” and the age at onset of AUD and related drinking behaviors.
More details are provided in the methods section.

2.9.4 Dose-response relationship

52

The author is not aware of any study showing the dose-response
relationship specifically for CPA to AUD. However, there is evidence that as
the number of childhood adversities increases, the risk of alcoholism
increases (Felitti et al., 1998).

2.9.5 Biological plausibility

Neurobiological studies have shown the biological plausibility of the
long-term effect of childhood adversities. For example, animal studies in
monkeys showed that being raised in isolation induces abnormal activities in
their hippocampus (R. G. Heath, 1972) and reduces corpus callosum volume
(Holder et al., 2000). Studies in rats showed that rats subjected to low levels of
maternal care showed alterations in the structure and function of GABA-A
receptors (Caldji, Diorio, 8: Meaney, 2003), and suppression of neurogenesis
(Teicher, Tomoda, 8: Andersen, 2006). People with a history of childhood
abuse showed similar changes, such as abnormal electroencephalogram
(EEG), smaller volume of hippocampus and prefrontal cortex, altered cortical
symmetry in frontal lobes, reduced neuronal density in the anterior cingulate,
etc. (Bremner et al., 1997; Teicher, Tomoda, 8: Andersen, 2006). These
changes have also been shown in people with drinking problems, especially
the frontal lobes, the lirnbic system (including hippocampus), and the
cerebellum (Oscar-Berman 8: Marinkovic, 2007). Although the causal
relationship cannot be established from these observations in AUD patients,
they provided evidence for the biological plausibility of the association

between childhood stressors and AUD.

53

2.9.6 Specificity of the association

With respect to the specificity, Brown 8: Anderson showed that in clinical
patients, AUD is more common in CPA victims compared to sexual abuse
victims, while no such difference was found in Axis II disorders and suicidality
(G. R. Brown 8: Anderson, 1991). Green proposed that, compared with victims
of sexual abuse, victims of physical abuse had more problems in aggression
modulation (Green, 1988). However, population-based studies have shown
that CPA is associated with a widerange of mental conditions, including mood
disorder, anxiety disorders, suicide ideation, antisocial behaviors, and
personality disorders (Dube, Anda, Felitti, Chapman et al., 2001; Kessler,
Davis, 8: Kendler, 1997; MacMillan et al., 1999; MacMillan et al., 2001; Pollock
et al., 1990; Windle, Windle, Scheidt, 8: Miller, 1995). Due to the high
comorbidity and the overlap in etiology of mental disorders, it is difficult to
infer the specificity of the effect of CPA. And animal studies suggested that
childhood stressors cause changes in multiple brain regions and
neurotransmitters. To our knowledge, there has been no study showing a
specific association between CPA and a single trait or biomarker.

2.9.7 Possible mediating pathway

Several studies have examined the mediating pathway from CPA to
drinking and AUD. Based on data from the follow-up study of court record
recruits as mentioned above, Schuck 8: Widom explored the mediating
pathway using SEM from CPA to AUD. Results suggested mediation through

depression and using alcohol/ drug to cope with difficulties. No such

54

mediation was found through worthless, isolation/loneliness, and low self-
esteem (Schuckit et al., 2001). Analysis based upon the NOS showed that the
association between CPA and AUD is completely explained by childhood
conduct disorder (Kessler, Davis, 8: Kendler, 1997). Zlotnick et al.’s study in
clinical patients suggested that the effect was potentially mediated by
posttraumatic stress disorder (PTSD) (Zlotnick et al., 2006). Tarter et al.
proposed that early adverse environment causes neurobiological
deregulations in children and causes alcohol and drug problems through
series manifestations during childhood and adolescent, such as
emotional/behavioral deregulations, externalizing difficulties, and antisocial
personality (Blackson 8: Tarter, 1994). Furthermore, studies in adolescents
and young adults showed CPA was associated with earlier onset of drinking
and heavy drinking (Bensley, Spieker, Van Eenwyk, 8: Schoder, 1999; Brems,
Johnson, Neal, 8: Freeman, 2004; Riggs, Alario, 8: McHorney, 1990; Rothman,
DeJong, Palfai, 8: Saitz, 2008).

In summary, with respect to the inference of a causal association links
CPA to AUD, it is plausible to investigate CPA as a possible cause of AUD,
within the context of a conceptual model in which there is an attempt to
specify temporal sequencing with CPA occurring before the onset of the AUD
and with attention to parental drinking and other covariates that might
structure a biased CPA-AUD association. This work builds from prior
evidence of a modest to moderate strength of the CPA-AUD association,

mixed evidence on the temporal sequencing, support for biological

55

plausibility and consistency with other knowledge, replicability, and a
possible dose-response or gradient relationship. There is no evidence of
specificity, such that CPA causes AUD and only AUD, but this guideline may
not be applicable in the context of psychosocial research.

In China, few studies have examined the association between
childhood adversities and drinking and related problems. Fairly low
occurrence of CPA was shown in one study of factory workers in Shanghai,
while high CPA occurrence was shown in another study of high school
students in Henan province. A recent publication showed an occurrence of
4.2% of childhood sexual abuse from a representative urban Chinese sample
(3.3% in women, 5.1% in men), which is considerably lower than that from US
samples (e.g. 32.3% in women, 14.2% in men; Briere 8: Elliott, 2003; J. Chen,
Dunne, 8: Han, 2006; Luo, Parish, 8: Laumann, 2008; Ross et al., 2005). In the
same research in Chinese high school studnets, childhood sexual abuse was
associated with higher likelihood of recent drinking (OR=2.7, 95% CI, 1.5,
5.1), history of being drunk (OR=3.6, 95% CI, 1.9, 6.8), and history of being
accidentally injured while drunk (OR=5.36, 95% CI, 2.1, 13.7), but the CPA-
AUD association was not investigated.

2. 10 Gaps in the epidemiological evidence
In this section, the dissertation returns to the three specific aims under
investigation, and places them in a larger context so that the potential

significance of the dissertation research may be appreciated. Each aim

56

addresses a gap in the epidemiological evidence on alcohol dependence and
related problems.

2.10.1 Specific aim 1

China is currently the most populated country in the world, hosting one-
fifth of the people on Earth. In China, drinking is a common behavior in social
contexts and it has been argued that drinking imposes substantial burden of
disease (Grimm, 2008). There have been concerns that with the increasing
contact with the western drinking culture and the increasing number of
automobiles, alcohol related problems are likely to increase (Newman, 2002).
In some research, the estimates have suggested that alcohol problems might
be on the rise, especially in the urban parts of China in association with
China’s increasing prosperity (Zhou, et a1. 2006; Hao, et a1. 1995; Zhang, et a1.
1999; Yang, et a1. 1999; Cal et al., 1998). Although some studies have studied
drinking practices and problems within China, these data may be outdated,
and often have not been based upon DSM-IV or ICD-10 criteria to assess AUD
(Hao et al., 2004; Jiafang, Jiachun, Yunxia, Xiaoxia, 8: Ya, 2004; Y. C. Shen et al.,
2006; Wei, Derson, Xiao, Li, 8: Zhang, 1999). Furthermore, in past surveys
involving multi-stage probability sampling, with individuals nested within
sampled households, households nested within sampled cities or villages, the
data have been analyzed data as they were collected with a simple random
sampling plan. By treating clustered data as simple randomly sampled data,
the estimation of variances, standard errors, as well as confidence intervals,

can be erroneous (often smaller than they should be), which can disrupt

57

statistical inference. This problem is especially pertinent in studies on alcohol
and drug use since these behaviors are found with significant geographical
and local area clustering (Bobashev 8: Anthony, 2000).

Moreover, in the published literature, the author found no
epidemiological description of drinking patterns and behaviors or problems
with representative sample from two of the biggest cities in China, Beijing and
Shanghai. Nevertheless, it is important to have population-based estimates of
drinking practices and problems because the drinking patterns in Beijing and
Shanghai today might be the ones that are followed in other cities of China
during later years. Of more public health importance is the description of
riskier drinking behavior, such as heavier drinking, early onset of alcohol
involvement, and socially maladaptive drinking and manifestations of alcohol
dependence, topics rarely studied in past research on drinking in China.

2.10.2 Specific aim 2

There is one prior study in China on the prevalence of drinking in
relation to population subgroups, such as males vs. females, but there is no
prior research in China on the prevalence or occurrence of alcohol problems
in these subgroups. This dissertation, for the first time in China, will present
epidemiological estimates for these public health problems. In addition to the
estimation of the sub-group specific parameters, the dissertation involves
estimation of the strength of association that links membership in the
population subgroups to the occurrence of alcohol problems, first based upon

a bivariate analysis and then based upon the multiple logistic regression

58

model, which is used to evaluate which of the observed associations might be
statistically independent of the others.

2.10.3 Specific aim 3

In this section, the dissertation returns to the topic of childhood physical
abuse (CPA) as a possible causal influence on the risk of developing drinking-
related problems. In many studies, there is evidence that childhood
adversities are associated with higher occurrence of adverse drinking
outcomes (Galaif, Stein, Newcomb, 8: Bernstein, 2001; Horwitz, Widom,
McLaughlin, 8: White, 2001; Widom, White, Czaja, 8: Marmorstein, 2007), but
these studies did not focus upon the CPA-AUD association specifically. Here,
we argue though that it is of importance to estimate the specific CPA-problem
drinking relationship because, as a potentially modifiable characterization,
CPA might in theory be manipulated or prevented to relieve some of the
disease burden of drinking and related problems. Additionally, as in a “cycle
of violence”, AUD runs within families as well. Childhood physical abuse may
be an important mediating factor in the transmission of AUD from parents to
the offspring; it may also be a possible cause of AUD independent of family
history of AUD. Figure 2.2 presents a heuristic and conceptual model (figure
2.2) in which CPA is specified to play a role as a cause in a larger
multivariable system. This model is presented not as a specification for an
econometric or structural equations model, although it conveys how multiple

variables might possibly work together to cause alcohol dependence.

59

Figure 2.2 Conceptual model of the relationship between childhood

physical abuse and drinking problems

 

 

 

 

 

 

      

Alcohol price,
law enforcement,
regulation...
genefic
vulnerability
e.g. family history,
specific genotype
~.\_ mediating factors: ' . .
other e.g.conduct/ drlngtlrng
childhood —’ 3mm" pro ems
adversitie problems, peer
influence...
\a
CPA

 

 

 

As depicted in figure 2.2, CPA may be positioned as a possible cause of
drinking problems, either independent of family history, or mediating the
pathway from family history to offspring drinking problems, or both. Many
studies showed a null association between CPA and AUD after introducing
parental drinking problems as an independent variable into the model
(Hughes, Johnson, Wilsnack, 8: Szalacha, 2007; Jackson 8: Sher, 2003; Koss et
al., 2003; Libby et al., 2004; Mullings, Hartley, 8: Marquart, 2004; Young,
Hansen, Gibson, 8: Ryan, 2006), while some others showed a positive
association (Carrigan 8: Randall, 2003; Jasinski, Williams, 8: Siegel, 2000).
However, these studies suffered from some major methodological limitations.

First, they were based on samples from special populations, such as prisoners,

6O

US Indian tribes, and marine recruits. The results cannot be generalized to the
non-institutional population, where the majority of the AUD cases come from
and where drinking behaviors and characteristics of AUD might differ. Second,
most studies used the logistic regression to estimate a slope coefficient
linking CPA to drinking problems. As stated in section 2.7, many of these
studies did not test the independence assumption of logistic regression, and
did not take the possible endogeneity problems into account. Moreover,
some studies included potential mediators (e.g. lifetime diagnosis of PTSD,
depression, individual’s education attainment, being a victim of abuse during
adulthood) in the multivariable model. The inclusion of these endogenous
variables might have artificially biased estimates of CPA toward the null
(Libby et al., 2004; Young, Hansen, Gibson, 8: Ryan, 2006). In summary, there
is some basis for speculating that CPA might be a causal influence on drinking
problems, but there is reason to complete more research on this topic before
any causal inference is drawn.

This dissertation contributes new evidence on the possibility that CPA
might merit interpretation as a causal influence on alcohol problems. The
dissertation cannot produce definitive evidence on this topic, or settle the
question. Nonetheless, its results will help guide future research of a more
definitive character and will provide study estimates needed to plan this more
definitive research. In addition to estimating the size of the suspected causal

association linking CPA with AUD, this dissertation research will help to

61

clarify whether the association is independent of other associations under
study.

Previously published studies are based on samples drawn from Western
countries. However, exploring the association between CPA and AUD in non-
Western countries is of both theoretical and public health pertinence.

In summary, under these specific aims, the main contribution of this
dissertation research will be to add new epidemiological evidence that is
pertinent to these guidelines for causal influence about observed associations:
strength of the association, consideration of alternate explanations
(endogeneity), replication of findings (in the Chinese context), and
consistency with other knowledge.

If successful, this dissertation will have scientific and public health
significance to the extent that it fills the gaps in evidence outlined in sections
2.10.1 to 2.10.3 of this chapter. In particular, the research will add new
estimates on drinking problems in two cities of China. It will contribute new
estimates of the size of association that link suspected background
characteristics with occurrence and prevalence of drinking problems. Finally,
it also will probe into the suspected causal association linking CPA with

drinking-related problems.

62

Chapter 3 Methods

3.1 Background: The author of this dissertation research was working in
Professor Yueqin Huang’s mental health and psychiatry research unit during
the period of conceptualization, planning, and field work preparations for the
Beijing and Shanghai field surveys of the WMHS initiative. In that context, the
author became familiar with the broad outlines of the research and was able
to gain the permission of Professor Huang and Professor Shen to make use of
the Chinese WMH data for this dissertation research. In addition, the author
has developed a deep familiarity with the WMHS research design and
analysis issues (e.g. problems of survey design effects) by working between
2005 and 2009 as a data analyst in Professor Jim Anthony’s WMHS research
group at Michigan State University. As such, this dissertation research project
builds from the primary fieldwork experiences that the author gained while
completing a master’s degree in Peking University, as well as additional data
gathering experiences between 2005 and 2009.
3.2 Design: This dissertation involves an analysis of data collected from the
World Mental Health Survey-metropolitan China initiative (WMHS-mC), a
cross-sectional survey in household-dwelling adults in Beijing and Shanghai,
China.
3.3 Sample selection

The WMHS-mC used a stratified multi-stage probability sampling

method to select household-dwelling non-institutionalized adults between 18

63

to 70 years old. In the first stage, the primary sampling unit (PSU) was
neighborhood (jumin weiyuanhui) within each metropolitan area. Figure 3.1 is
a map of China, showing the location of Beijing and Shanghai. Figure 3.2
shows geographic locations of sampling unit in Beijing and Shanghai.

Figure 3.1 A map of China.

 

Beijing

 

 

 

Shanghai

CHINA 1‘

 

 

64

Figure 3.2 Sample geographic maps of the WMH-mC, Beijing (upper) and

Shanghai (lower)

 

65

There were 47 PSUs in Beijing and 44 PSUs in Shanghai identified
through the Demographic Data for Neighborhoods, 1999, published by the
Statistics Bureau of Beijing and Shanghai, respectively. PSUs were selected
using the probability proportional to size sampling method. For the second
stage, lists of households within each neighborhood were obtained from
neighborhood committees. Then, households within each neighborhood were
randomly selected. In the final stage, one adult from each identified
household was randomly selected to be the respondent. The Research Center
for Contemporary China (RCCC) at Peking University directed the data
collection in both cities through two designated field managers, one for each
city. The field manager of each site organized a team to implement the field
work following the same survey protocol. Before data collection commenced,
two training sessions were carried out by each of the field managers under
the supervision of RCCC. Data were collected through face-to-face interviews,
which were conducted by trained lay interviewers between November 2001
and February 2002. All respondents were informed about the study and
provided written informed consent prior to the interview, using a study
protocol approved by the designated Institutional Review Board. Response
levels were 74.8% and 74.6% in Beijing and Shanghai, respectively. The final
sample used for analysis consists of 2633 participants from Beijing and 2568
participants from Shanghai. Complete details about the sampling and field
procedures can be found in the field survey final report (M. Shen, Chai, Yang,

Huang, 8: Yan, 2003).

66

3.4 Measures
3.4.1 Assessments

The assessment instrument used in WMH-mC is the World Mental
Health Initiative version of the World Health Organization Composite
International Diagnostic Interview (WMH-CIDI; Kessler 8: Ustun, 2004). The
WMH-le is a comprehensive, fully structured diagnostic interview designed
to be administered by trained lay interviewers to assess clinical features of

psychiatric illnesses, and symptoms of mental disorders according to criteria

in both the International Classification of Disease, the 10th edition (ICD-10)

and the Diagnostic and Statistical Manual of Mental Disorders, the fourth
edition (DSM-IV). The Chinese version of the WMH-le used in WMH-mC
was derived using standard protocols of iterative translation, back translation,
and harmonization conducted by panels of bilingual experts.

The interview was administered in two parts. Part I included the core
diagnostic assessment. Part 11 included questions about suspected correlates
or determinants as well as additional topics including tobacco use and extra-
medical psychoactive drug use. Part II was administered to all respondents
who were suspected to have a history of past or recent core mental disorders,
assessed in Part 1, plus a 25% random sample drawn from the rest of
respondents. A total of 5201 participants completed Part I; 1628 completed
Part II.

The WMH-le consists of modules on various topics. Each module

contains standardized questions relevant to the specific topic. Questions about

67

lifetime history of alcohol drinking were located at the beginning of the
“substance” module and were administered to all respondents. For drinkers
ever in lifetime, follow-up questions were asked about recent drinking
behavior (in the prior 12 months), the occurrence of socially maladaptive
drinking and other clinical features associated with alcohol dependence, as
well as the recency of these problems. Figure 3.3 provides a brief description
of the logical skip pattern employed in the CIDI. Actual questions in the
original English version of the WMH-CIDI, as well as the final Chinese
versions of these English language items, and a more detailed diagram of the
skip pattern can be found in appendix materials. Two major assumptions are
made in the CIDI assessment of alcohol problems. The first is that if an 3
individual has never drunk more than monthly (once a month), then he/she
would never qualify as a case of DSM-IV or ICD-10 defined alcohol use
disorders (AUD). The second is that if an individual never had experienced
socially maladaptive or hazard-laden drinking, he/ she never would meet
DSM-IV diagnostic criteria for “clinically significant alcohol dependence”,
with the “clinical significance” criteria as presented in Degenhardt et a1.
(2007) in their discussion of the DSM-IV concept (Degenhardt, Bohnert, 8:
Anthony, 2007). More detailed discussion about these assumptions can be

found in chapter 5 of this dissertation.

68

 

 

mam"... dense...
3.328% _o:oo_a Sega 2.2.2.5 .motom 23m

 

 

mm>

 

 

 

 

.amcaue.
22.33 x5...
2 32.3.2.3

cacao>m

 

 

 

 

 

2.60:. a
.28... 2.33m do.

 

 

.iz/

$.95»
.. es. 2 9.23

 

 

 

28."... oz

«8.3» «S... 2.. 3 Sign and? >5.

 

 

 

 

 

$3": $556 53.6.32
.8 5.3322. .38» 33.. «5:830 .828 «Sm

 

 

mm;

 

 

as... o c. 9.56
8...: .82 5 x55 55 $583.33 3.5

 

 

Leo—oceans.
2.. moon

 

 

 

among. nm+xo+oz 2 2.0 m. >833..— msxctu 8......on

 

: «255:. so... 29.. seat .o>m

 

 

dm+xo+mm>

 

 

 

 

.mmEuE .962

Sew—n5 oz

5mm": too» near. a... e. 9.5:. 2. sea... .o>w

 

 

 

mm+xo+mo>

 

 

 

 

ram": $0500? xcgb hm>m

 

 

.EoEnnommc _2_oo_a 5.0.12.5 9: .0 E33 aim Qm 2.6.“.

69

s2

1!)

There was a WMH-CIDI question about the first opportunity to drink
alcohol in Part II of the assessment, so its responses were available for the Part
11 sample of 1628 respondents. Information about childhood experience was
located in the “Childhood” module, which was administered to the Part 11
sample of 1628 respondents. Information about demographic characteristics
was obtained through two separate modules. The “core” demographic
module was administered to all participants in Part I (n=5201), while the other
more detailed demographic module was administered to the Part 11 sample
only. As a result, data on sex, age, marital status, personal income, and
employment status were available for all participants, while data on education
attainment were available for the sub-sample only.

3.4.2 Definition of drinking-related variables

In this study, the self-report drinking related variables fall into two
main categories, the “drinking behavior” and “indicators of risky drinking”.

3.4.2.1 Variables in the “drinking behavior” category

1) Ever had opportunity to drink alcohol;

2) Ever tried alcohol (even a sip);

3) More than minimum (MTM) drinking (>= 12 drinks in a given year);

4) Being a MTM drinker during the year prior to the assessment;

5) Frequency of drinking when drank the most;

6) Frequency of drinking during the year prior to the assessment;

7) Age of first try;

8) Age of onset of drinking;

7O

9) Number of drinks during a typical drinking day when drank the most;

10) Number of drinks during a typical drinking day during the year

prior to the assessment.

Opportunity to drink alcohol was assessed by means of the following
WMH-le question, which was asked in the first series of Part 11 questions
about drinking experiences of the 1628 Part 11 respondents: “The next
questions are about the first time you had an opportunity to drink alcohol or to
use drugs, whether or not you used them. By “an opportunity to use ”I mean
someone either offered you alcohol or 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 responses to the item were
coded as follows:

the actual number <100 =_years old (n=1167)

997 =never (n=366)

998=don’t know (n=92)

999=refused (n=3)

For the present study, respondents who gave an age value in response
to the question were coded as “yes”. Respondents who answered “Never”
were firstly coded as “no”; then, if they indicated that they had tried alcohol
(from another question, SUI , in the “substance” module), they were recoded

as “yes” (n=32). Among 95 respondents who answered “don’t know” or

71

“refused” to the original question, 59 indicated they had tried alcohol. These
59 were then coded as “yes”. The remaining 36 were coded as missing. A
description of missing values is given in Table 3.1.

The variable “ever tried alcohol” was assessed by means of the
following WMH-le question: “The next questions are about your use of
alcoholic beverages, including beer, wine, wine coolers, and hard liquor like
vodka, gin or whiskey. How old were you the very first time you ever drank an

alcoholic beverage ?” The responses to the question were coded as follows:

the actual number < 100 =____ YEARS OLD (n=3 193)
997=NEVER (n=1660)
998=DON’T KNOW (n=348)
999=REFUSED (n=0)

Responses who gave an age value in response to the question were
coded as “yes”; respondents who answered “Never” were coded as “no”.
Among 348 respondents who gave “don’t know” or “refused” responses to
the original question, 155 of them indicated they have had drunk >=12 drinks
in a year. These 155 were therefore recoded as “yes”. The remaining 201
respondents were coded as missing.

The variable “ever being an MTM drinker” was assessed by means of
the following WMH-le question: “When I use the word "drink" in the next
questions, I mean either a glass of wine, a can or bottle of beer, or a shot or
jigger of liquor either alone or in a mixed drink. How old were you when you

first started drinla‘ng at least 12 drinks in a year?

72

IF “ALL MY LIFE” OR “AS LONG AS I CAN RENEMBER, ” PROBE: Was it before
your teens?
H-‘NO/DK, PROBE: Was it before your twenties?”

The responses to the question were coded as follows:

 

the actual number <100= YEARS OLD \
12=BEFORE TEENS

> (n=2049)
19=BEFORE 205
20=NOT BEFORE 205 J
997=NEVER (n= 1433)
998=DON’T KNOW (n=55)
999=REFUSED (n=2)

Respondents who gave an age value in response to the question were
coded as “MTM drinkers”; respondents who answered “Never” were coded
as “non-MTM drinkers”. Among 57 respondents who gave “don’t know” or
“refused” to the original question, 28 indicated that they had drank at least
once a month during a year in their lifetime from later questions; these 28
were recoded as “yes”. The remaining 29 were coded as missing.

MTM drinking during the year prior to the assessment was assessed by
means of the following WMH-le question: “Think about the past 12 months.
In the past 12 months, how often did you usually have at least one drink- nearly
every day, three to four days a week, one to two days a week, one to three days a
month, or less than once a month? ”

The responses to the question were coded as follows:

73

1=NEARLY EVERY DAY (n=459)

2:3 - 4 DAYS PER WEEK (n=200)
3=l - 2 DAYS PER WEEK (n=371)
4:1 - 3 DAYS PER MONTH (n=389)

5=LESS THAN ONCE A MONTH (INCLUDING NEVER DRINK) (n=651)

8=DON’T KNOW (n=34)

9=REFUSED (n=2)

Respondents who gave a value of one through four in response to the
question were coded as “yes”; responses of option five were coded as “no”.
“Don’t know” and “refused” were coded as missing values.

Drinking frequency during the year prior to the assessment was
assessed by the same question. Options one through five were maintained as
they were. “Don’t know” and “refused” were coded as missing values.

If the respondent did not have a period of time when they drank more
than they did during the 12 months prior to the assessment, their drinking
frequency when drank the most was the same as the drinking frequency
during the last 12 months. If they indicated that there was a period of time
when they drank more than they did during the last 12 months, their drinking
frequency when drank the most was assessed by means of the following
WMH-le question “Think about the years in your life when you drank most.
During those years, how often did you usually have at least one drink— nearly
every day, three to four days a week, one to two days a week, one to three days a

month, or less than once a month?”

74

The responses to the question were coded as follows:

1=NEARLY EVERY DAY (n=333)
2:3 - 4 DAYS PER WEEK (n=165)
3=1 - 2 DAYS PER WEEK (n=189)
4:1 -3DAYS PERMON'I‘H (n=161)

5=LESS THAN ONCE A MONTH (n=369)

8=DON’T KNOW (n=39)

9=REFUSED (n=2)

For this variable, the same logic was used as the past year drinking
frequency variables.

A value for the number of drinks per day has been obtained from this
question: “On the days you drank, about how many drinks you usually had per
day?” Separate questions were asked regarding the year prior to the
assessment and the period of time when they drank the most. Numbers of
drinks were maintained. “Don’t know” and “refused” were coded as missing
values. For number of drinks per day during the year prior to the assessment,
there were 105 missing values, accounting for 9.0%of all answers. For number
of drinks per day when drank the most, there were 116 missing values,

accounting for 7.6% of all answers.

75

 

Table 3.1 Distribution of variables for drinking behavior. Data from the WMH-mC, 2001-2002.

 

 

 

Entire Sample Beijing Shanghai
n % w1°/02 n % W’t%2 n % Wl°/02
Opportunity’ Yes 1258 77.3 77.9 722 79.0 80.4 536 75.1 74.7
No 334 20.5 19.8 180 19.7 18.6 154 21.6 21.3
Missing 36 2.2 2.7 12 1.3 0.9 24 3.4 4.0
Evertrying
alcohol Yes 3340 64.2 67.2 1756 66.6 70.1 1582 61.6 54.2
No 1660 31.9 29.1 832 31.6 28.1 828 32.2 30.2
Missing 201 3.9 3.7 43 1.6 1.8 158 6.2 5.7
MTM drinking Yes 2077 39.9 41.7 1146 43.5 45.0 931 36.3 38.4
No 3095 59.5 57.7 1480 56.2 54.7 1515 65.9 60.8
Missing 29 0.6 0.6 7 0.3 0.3 22 0.9 0.8
Past year MTM Yes 1419 27.3 28.0 787 29.9 30.1 632 24.6 25.8
drinking No 3744 72.0 71.3 1831 69.5 69.4 1913 74.6 73.4
Missing 36 0.7 0.6 15 0.6 0.5 21 0.8 0.8
Frequency of drinking
Nearlysveryday 610 11.7 10.3 347 13.2 10.9 263 10.2 9.6
“days/week 277 5.3 5.8 146 5.6 6.0 131 5.1 5.5
1-2 days/week 417 8.0 9.0 253 9.6 11.1 164 6.4 6.9
1-3 days/month 395 7.6 8.6 182 6.9 7.7 213 8.3 9.5
<month|y 1817 34.9 36.8 861 32.7 35.9 956 37.2 37.9
No drink 1660 31.9 29.1 832 31.6 28.1 828 32.2 30.2
Missing 25 0.5 0.5 12 0.5 0.4 13 0.5 0.6
Past year drinking
frequency
Nearly everyday 459 8.8 7.3 259 9.8 7.6 200 7.8 7.0
34 days/week 200 3.9 4.0 115 4.4 4.4 85 3.3 3.6
1-2dayslweek 371 7.1 7.9 218 8.3 9.2 153 6.0 6.7
1-3 days/month 389 7.5 8.8 195 7.4 9.1 194 7.6 8.6
Less than once per month 2084 40.1 42.2 999 37.9 41.3 1085 42.3 43.1
Neverdrank 1660 31.9 29.1 832 31.6 28.1 828 32.2 30.2
Missing 38 0.7 0.7 15 0.6 0.5 23 0.9 0.9

 

1- Variable available in Part II only

 

2' Due to rounding, some values do not sum to exactly 100%

 

76

3.4.2.2 Variables in the “indicators of risky drinking” category
Indicators of “risky drinking” in this study were:
1) Early trying alcohol (before teens (<13 years old));
2) Early onset of MTM drinking (before 20 years old);
3) Ever heavier drinking;
4) Heavier drinking during the year prior to the assessment;
5) Past year socially maladaptive or hazard-laden drinking problems;
6) Ever had any alcohol dependence clinical feature;
7) Past year alcohol dependence clinical feature;
8) Early onset of socially maladaptive or hazard-laden drinking (before 23
years old);
9) Early onset of alcohol dependence clinical feature (before 23 years old);
10) The occurrence of five socially maladaptive or hazard-laden drinking
problems: responsibility interference, social-interpersonal problems,
drinking despite social problems, hazardous use, legal problems;
11) The occurrence of eight clinical features of alcohol dependence: tolerance,
withdrawal, difficulty cutting down, giving up activities because of drinking, a
great deal of time spent on drinking-related activities, drinking despite
physical/ emotional problems, and irresistible desire.

Respondents who gave an age value that was less than 13 years, when
asked about first trying alcoholic beverages were assigned a “yes” for “early
trying alcohol”. Respondents, who answered “don’t know” or “refused” were

deemed as missing values; all others were coded as “no”. (The actual

77

question is shown in section 2.1.) A description of missing values is given in
Table 3.3.

A similar procedure was used to recode the “early onset of MTM
drinking” values, for which the actual question is shown in section 2.1. The
cutoff age of 13 for the first consumption of alcohol is based on previous
literature about preteen initiation of drinking (Dube et al., 2006; Hamburger,
Leeb, 8: Swahn, 2008). A cutoff age of 20 was used to designate early onset of
MTM drinking. Since there has been no legal age for drinking in China and
many children sip alcoholic beverages in celebrations, the age of the first sip
does not necessarily mean higher risk of negative consequences in Chinese
cultures. Due to this concern, the author created another marker for earlier
onset of drinking, “early onset of MTM drinking”. No reference age cutoff for
MTM drinking was found in the literature. The cutoff point of 20 years old for
“MTM drinking” is mainly based on the way of how data were collected. As
described above, if the respondent could not recall the exact age to answer
the WMH-le question “How old were you when you first started drinla'ng at
least 12 drinlc in a year?”, a follow-up question “Was it before your twenties?”
was asked. If the response is “yes”, the age in the dataset is marked as 19.
From the de-identified data, there is no way to separate the 19s meaning
“before twenties” and the 195 meaning the exact recalled age of 19.
Nonetheless, there is evidence that earlier onset of drinking is associated with

more negative drinking consequences. This variable serves as a marker for

78

earlier onset of MTM drinking. Description of these variables can be found in
Table 3.3.

This study’s “heavier drinking” assessment was derived from questions
about number of drinks consumed in a day, for which the actual questions can
be found in section 2.1. Lifetime history of heavier drinking and past year
heavier drinking were assessed through two separate questions regarding
the period when the respondent drank the most and 12 months prior to the
assessment, respectively. Heavier drinking was defined as at least five drinks
in a typical drinking day for males and at least four drinks for females. This
cut-off threshold of five and four is in accordance with the definition of “binge
drinking” used by the US National Institute of Alcohol Abuse and Alcoholism
(NIAAA) and with the most commonly used definitions for “binge drinking”
and “heavy episodic drinking” in previous studies (Jasinski, Williams, 8:
Siegel, 2000; NIAAA, 2004; Trent, Stander, Thomsen, 8: Merrill, 2007). “Don’t
know” and “Refused” were coded as missing values. The author is aware of an
agreement among alcohol research journal editors to avoid the NIAAA
concept of “binge drinking” so that the traditional concept of a “binge” can
be retained (e.g. a ‘binge’ or ‘bender’ involving multiple days of continuous
intoxication). For this reason, the term “heavier drinking” is used in reports
on this dissertation research.

Socially maladaptive or hazard-laden drinking was assessed via five
consecutive questions. These five questions are based upon criteria of

“alcohol abuse” as outlined in DSM-IV. Actual questions appear in table 3.2. In

79

order to ease the discussion, we labeled them as “interfere with
responsibility”, “social problems”, “drink despite social problems”,
“hazard-laden drinking”, and “legal problems”. For all the five questions,
respondents who answered “yes” to the original question were coded as
“yes”; respondents who answered “no” were coded as “no”; respondents
who answered “don’t know” or “refused” were coded as missing.

For this dissertation research endorsement of any of the five questions
qualified the respondent as a case of “socially maladaptive drinking”; “no” to
all five questions qualified the respondent as a non-case of “socially
maladaptive drinking”. If the respondents answered a combination of “no”
and missing (“don’t know” or “refused”), he/She was assigned a missing
value for “socially maladaptive drinking”. Standard algorithms for the WMH-
CIDI were used to label the diagnosis of DSM-IV alcohol abuse. Essentially,
endorsement of any of the above clinical features, except for “social
problems”, qualified the respondent as a case of “lifetime DSM-IV alcohol
abuse”. If any of the clinical features occurred during the 12 months prior to
the assessment, the respondent was qualified for a case of “past year DSM-IV
alcohol abuse.”

Eight clinical features of alcohol dependence, as outlined by DSM-IV
and ICD-10, were assessed via 11 consecutive questions. Actual questions
appear in table 3.5. Although multiple questions were asked for the clinical
features of “withdrawal” and “drink more than intended”, the “once endorsed,

skip the rest” logic embedded in the assessment testlet precluded us from

80

studying items within one criterion individually. Therefore, we can only
provide estimates for each diagnostic criterion and not for these individual
items. In order to ease discussion, we labeled these eight clinical features as
“strong desire”, “tolerance”, “withdrawal”, “more than intended”, “cut
down”, “great deal of time”, “give up activities”, and “drink despite
physical/mental problems”. For each of the eight criteria, respondents who
answered “yes” to the original question were coded as “yes”; respondents
who answered “no” were coded as “no”; respondents who answered “don’t
know” or “refused” were coded as missing. Two questions were asked
(SU19b and SUl9c) to assess “withdrawal” criterion, a “yes” answer to either
question qualified the respondent as a case of “withdrawal”; the same logic
was used to recode “more than intended” criterion, where three questions
were asked (SU19d to SU19f). One more question (SU32) besides SU19g were
used to recode the “cut down” criterion. The question is: “Starting from the
time you first began having any of these problems, how many different times did
you ever make a serious attempt to quit drinking?” If the respondents gave a
value of greater than two, they were recoded as “yes” to the “cut down”
criterion as well.

Endorsement of any of the eight criteria qualified the respondent as a
case of “any clinical feature of dependence”; “no” to all eight criteria
qualified the respondent as a non-case. If the respondent answered a
combination of “no” and missing (“don’t know” or “refused”), he/ she was

assigned a missing value.

81

A standardized computerized algorithm for the WMH-CIDI was used to
sort drinkers in relation to the diagnosis of DSM-IV alcohol dependence. If the
respondents had met at least three of the above criteria, except for “strong
desire”, and indicated that they had at least three problems in a same year,
they qualified as cases of “DSM-IV alcohol dependence”; if the respondents
met none of the seven criteria, they qualified as non-cases; if the respondents
endorsed a combination of “no” and missing (“don’t know” or “refused”),
they were assigned a missing value for “DSM-IV alcohol dependence”.

Age of the first occurrence of socially maladaptive drinking was
obtained from two CIDI questions. The first question is: “You just reported that
your (1ng (IGY PHRASE FOR “YES” RESPONSE 1N SU12 SERES). Can you
remember your exact age the very first time you had this problem? If so, how old
were you ?” Respondent who gave an age value less than 23 to this question
was identified as a case of “early onset of socially maladaptive drinking
problems” (n=52). The second question is for when the respondent cannot
recall the exact age. The question is “About how old were you (the first time
you had [this problem/ (either/ any) of these problems] because of drinking)?

If “All My Life ” Or “As LongAs I Can Remember, ” Probe: Was It Before
Your Teens? H’NO/DK, PROBE: ' Was It Before Your Twenties?”

Responses to this question are:

82

 

00000.

0. 00

:02. .00. .00 .mméxn. Ev. 0050.090 00000.00 .0 000.000 05 00000.0 8000000.. 00000.
00... 00.0.0.0..0 0000000000 005 000 000.000 505.; .00.... .0 .0300 .00 00000.0 00.. 000.... 00.0..
0.00 .0>0 00.. 0.0. 00.60000 .0000 0... 005.0000. .00. 000 000 0000.0. 0.00. 0.0 005.00. 0000 0

0.00. .00. .00 >05 00... 00.00 .0. 0.08.0 000.000 .000 000.000 00.0 .0 02.00 .00 00.... 0.0000 .0050.

 

.0000.0 .0 000.00 W00. 0 0.0.00 .mméxn. >00.
00.0.00 0. 0000 00.. .00050 05 00 00.0 0 .0 .N00. 0 .00 .0000. 00 0.08 00..
.05 0.5 .0>0 00.. 0.0 .0 00000 00 .00 0. 008.0 .0 5000.0 .090. 0 0.000 0. 0000 .0>0 00. 0.0 .0050.

 

0.0..0 0. 0..80 00000 0 .mmérn. Ev.

000.0 00.5.00 .0 0.0.0. 0. 0005.0 .. 0000. .0 0.000 0 00.0.0. .08. ..00.000

00.. .00. 0.000 0. 0.80 009.0 0 0000 000 0000 00.. 00...... 0... .00.. 0. 00... 0 .0>0 0.00. 00.5
0000.000 .0 0000000 000 0>00 >00. 00.. 0.00.090 .050 00.00 .0000 0.0 00000000 0.00 00 b 0.3m.

 

A8000 00. 0...... 0.00.090 0. 000000. ”mm/0.110. >50

0.00.0000
0.00.0 .0 00.>.00 .000 .0 0000000 00:00 00. >0 00000.0 .0 00.800 0000 000. 0.0... 00.. 0.0.5 .0059.

 

.00.. .00 0.08
00.. 0.00.... 00000000 0. 000.0 8050.000 00.. 0000000 0.0.00 .00.. 08.0.0000. 505...... >000
000.0 00.00.00
.0 00.000... 0 0000.000 00.0.0 0.0.6.0 0 00.0.0 000.... 0.0.0000 .0. .000 .00 0.000 00.. 0.00....
00000000 0. 6000.0 .0 0000000. 05 .0000 0000 0.0; 00.. 000.... 0... .00.. 0. 000... 0.05 0.03 .0050.

 

m

0

00.0000 0000. 0...... 0.00.090 000000.. 00000. 00>0 0.000 0. 000008 00.. 0... .0059.

 

0. 00

00:0.

0. 00
0

ON Sm...

00.00
0. 00
0

0.050 .0 0000... 00.0.0. 0...... 0.00.020 000000 MEET... Ex.
00.00.0200 .0 00000.00 .0000... 00.0.0. .00.. 5.0. 8.0.00.0 00.0000.
.0 000000 .050 .0 0.00.0090 000000 000.000 .00» 000.... 0.... .00. 0. 0.0.. 0 .0>0 0.00. 00.... 0.0.5.0...

 

0.0; so. 5; 09000.... 00.0.00 .00..
000.00 .0 .0 .00. 0 00 000000 .0 8.....0_00008. .0
000; .00.. 0...... 8.0.00.0. 500000.. .0>0 0000 00.00 .0 000.000 .00.. 000.... 0... .00.. 0. 0.0.. 0 .0>0
0.00. 00; ..0..n. .00....000 .0 0000000 000 0000 .0... 00.. 0.00.090 .0000 0.0 00000000 0.00 00 0 .050.

 

.0.
n...

I

.0.

 

v.0

.0.

 

oz

 

E
00..

 

.059. 0.0 00 oz< .0. wzo.._.mm50 mmew ~59 0mmm>>mz<23
.....< moon. szmeSC O>>... mmwznmm mo whwmhomm m 0.. 20.003252. mm>>mSmm._.z.

 

 

0000000000 .280 .0 8.0.00. 00.5.0 000 00.0.00 050000.00. 50.000 .0000 00000000 5.0.1.2.; .020... N. m 0.00 ..

 

 

83

 

.000...

.0 00.0... 000.0 .0 .0 000000 E0_00.0 .000..0.00 .0 .00.0..00 0.0.... .000 0. 0000000 .mmézn. >mv.

0000.000 .0 00.0... 000... .0 .0 000000 0000 0>00 .00.... .00.
0.0.00.0 .000..00.0 .0 00.0.00 000000 0 000 00.. 0.00.. 00.. 000.... .000 0. 000008 .0>0 00.. 0.0 .050.

 

000.000 .0050 0000000 0000.000 .000000.. 000000. 0.00.0 .0 00 9.00 MEET... .m... .
0.0.0.0. 000 0000... 00.000 .0 .00... .00000 0.... I 000.000 .00.. .0 0000000 80.0.00 00000....

000000. 0.00.0 .0 00 0>00 00.. 000... .0000. .0 0.00... 0 .0 00000 0 0>00 .0>0 00.. 0.0. .050.

 

00.0 00.00.00 .0.. 0.0.. 0.0.. ”09.0.... 50.
000.0 00.00.00 .0. 0.0.. 0.0.. 000 00.. .00. .0000.0 .0 0.0000 00. .00.. 0000.68.
.0 000.000 0.0.. 000.0 00 .0000 00.. 000.... 0.0.0 .0 0.00 .0.0>00 .0 000000 0>00 .0>0 00.. 0.0. .0050.

 

0.00 .00 .00 000.000 .00.. 00 00.00 .00 .0 00.0 0. 000. .mmézn. F5.
000 00 0. 0.00.00
0.0... 00.. .00. 0000. 000 000.000 .00.. 00 0....00 .00 .0 00.0 0. 000. 00.. 000.... 00.0.. 0.00. 0.0.... .0050.

 

0. .00.... ..00.0 00.. 000.... ..00.0 U00.0.0.0... .6..
00. .003 ._00.0 00.. 000.... ..00.0 0.00000 000 000.0..0 0000.0 00.. 000.... 00.0.. 0>00 00.. 0.0. .050.

 

 

00.00.. 00000.0. 00. 000. 0.0. 0 0. 0.00 0.0.0 “$5.10 5...
0. 00 000000.0.
. 00.. 00... .00. 0 0. 0.00 0.0.0 .0. .0 0.000000 0.0.0 .000 00.. 000.... 00.0.. .0>0 0.00. 0.0.5 .0050.
00.00.. 00000.0. 00.. 000. 0.0.0 .0. 0 000.0 .mmézn. Ev.
0. 00 00008.0. 00.. 000. 0.0.0.0. 0 000.0 00.. 000... .0

..00.00.... 00.. 000.00.. 08.0.0.0 00.. 00000. 00>0 000.000 0000.0 00.. 000.... 8.0.. 0>00 00.. 0.0 .0050.

 

 

 

 

 

000... 0... 00.00.... 00.0.. 0.2. 08.. 00%.... .00.
00000. 0.... 0.00.090 000.00 .00.. 000.. 0. .000 0 ..00. 00.. 000.... 00.0.. 0>00 .0>0 00.. 0.0 .0050.

 

 

. .008. 0.0 0.00.

 

 

 

84

a number <100= YEARS OLD

12= BEFORE TEENS

 

19: BEFORE 20s >(n=28)
20: NOT BEFORE 20s J

998: DON’T KNOW (n=8)
999: REFUSED (n= 1)

Respondents who gave an age value less than 23, but not 20, were
coded as cases of “early onset of social maladaptive problems” (n=6).
Respondents who had a value of 20 were coded as non-cases because their
age of onset was likely to be after 23 (not before the 203). Age of first
occurrence of clinical features of alcohol dependence was recoded in the
same manner.

The cutoff age of 23 is based upon previous literature. Previous studies
often defined early onset of alcohol problems as before 20 or 25 years old (B.
A. Johnson, Cloninger, Roache, Bordnick, 8c Ruiz, 2000; Watson et al., 1997).
Brown and colleagues found that 22.5 is the mean onset age of problem
drinking in the more severe group, and 24.3 in the less severe group (I.
Brown, Babor, Litt, 8: Kranzler, 1994). Moss and colleagues’ post-hoc analysis
with latent class modeling found that the mean age of onset of alcohol
dependence clinical features is approximately 23 in the most severe classes,
and older in other classes (Moss, Chen, 8: Yi, 2008). Thus, for this dissertation
research, an age of 23 was specified to be cutoff point for early onset of

alcohol problems in this study.

85

 

 

8 o No to N 3 3 N ER:
«.8 «.8 3mm «.8 3 $8 Q8 35 R5 oz 2832
ed md om 3 08 we o4 we mm mm; .308 33% 3.55
no to 2 no no a no 2 2 822
m8 is 8mm 5m «.3 88 Em 2m 88 oz
2 3 mm on 2 8 3 3 3 8> 2.285 E08
3 3 a no 2 o. to to 2 922
m8 «.8 ER :8 3m 3mm «.3 m5 2% oz a___n_w§me
2 3 mm 3 3 E 2 E E R; 5; $55
no E 2 S 3 2 mo 3 mm 32:
3m as $3. 35 0.3 at.“ 3m 8m $9, oz
Qm Qm on no m6 m3 5. me FNN mm> 83m .286 2.28
R 3 on em 3 a, 3 3 2: ER:
3m :3 «SN 03 N8 33 20 mg was 02 3er
3 R E 3 3 t 3 3 E R; 52.33%;
mm mm a no 2 m E 3 8 322
SN 3N § 3: we 2: me SN «8 02
m2 QB mmm mom v.2 owe o3. NR momF mm; 9.256 >28...
2 3 8 co mo 5 3 E B 322
as Sm 8% m2 is 8mm «.8 m8 8:, oz 92%
we 0.: 8m 2: n: «mm as :2 5o R; Ezasacoemw
2: we «8 2 2 8 we 3 as” 822
wk 2: 88 :8 Ba 88 0.; 3e 82 oz .283
3 v: 8m 2: 3 EN 0.: E E. R; 95:23:25
Foss as c as: as c as; as c

5.955 3.3% magma 2:5

 

.mooméoom .OE-I_>_>> 9.: E0: 98 .mcictu Em: .0. 339.9 do :o_SnEm_o 0m 3an

 

86

 

fiwm

o.o
Nmm

:m

gm

mo
Nmm

vwmm
3

Emu
om

Bow
5

5mm
mm

mwmm
N9

88
mw

mmmm
mm
F F
m — mm

o v
nmmw
mm

mm 5
mm

mm 5
mm

mm B
om

mmom
co F

95m
om F

min
mv

min
om
E
SE
OF
om
omom
2;

oz

mm;
9.622
02

mm;
9:85
02

mm;
mags.
02

mm;
33:2
02

mm;
9.622
02

mm;
385
oz

mm;
9:32
02

mm;
mags.
02

mm;

5285 Esme B
Rumba £38 £5

mmaéom a: go

we; do 6% 690

:38 So

8225 :9: 222

.m3m_u£_>>

8:928

3285 3mg

9.256 584-2,“..me

 

628v on 038

 

87

 

£09 2888 2 E3 woo oo moE:_o> 0E8 .325 2 one .omscoeoq 8292s .F

 

F.o 0.0 F F.o No v F.o 3 m mc_wm__z

Rmm mmm 3mm 5mm m.mm momm o.mm me mm Fm oz 8coocoaoo _o;8_m
m.o mm 3 NF m.o om m.o mm ow mo> 2-2mm 59» “mod
mm «.0 S No md 2 od v.0 mm 535

m.mm omm vam mNm Qmm 8mm F.mm mewm FNFm oz 83m _o;oo_m
g No NF 0N m. F ow m. F F. F mm mm; >_-_2mo so» “mom
F.o om F F.o F.o m 3 F.o F. 9622

0.8 mmm mem me mmm Sow vmm v.mm 2.5 oz 882.8%
mm to 2 m.o mm oF om no em mo> _o;oo_o do “8:0 25m
No mm vF mm md vF md no mm 9.622

F.m® 0mm mNmN Vow mNm wmmm NNm aka mmom 02 9.25.6 gang—NE
F. F F. F mm mm m. F Fm cm 2 cm mm; >__m_oom Fo flowco From
No No o 3 F o N 3 E N 3me

mmm mmm 3mm Ymm N mm mmmm m.mm me mmFm oz 852.8%
vd v.0 m QF N F mm 3 mm Fv mo> _oF_8_o 2.28
0.0 0.0 o 0.0 F o N 0.0 od N 9:83

Qmm m.mm 0va Nmm m mm mmmm 3 m.mm mmFm oz

o.F F.F mm m F 3 mo 9mm 3 K we, 25% 955
o.o F.o N F o No m F.o F.o F 9:35

 

face no 2an

88

3.4.3. Covariates under study

With respect to the aim 1, there is no covariate of interest.

As for aim 2, subgroup variations in drinking behavior and outcomes,
covariates of interest are sex, age, marital status, education attainment,
income level, and employment status.

Sex was dichotomized into male and female based upon the
observation of the interviewer. Age was based upon self-report by
participants. Participants were further divided into four age groups: “born
before 1949”, “1949-1965”, “1966-1977”, and “born after 1977”. The division
is based on significant events in contemporary Chinese history. The year 1949
is the year of the establishment of the government of the People’s Republic of
China; the “great leap and culture revolution” started in 1966; the open policy
and the only child policy were implemented in 1978. Table 3.4 provides a
description of these variables.

Current marital status was categorized as married or cohabitating, no
longer married (separated/divorced/widowed), and never married, based on
the self-report of participants. Original questions in WMH-le arez” Are you
currently married?”, “Are you currently living with someone in a marriage-like
relationship?”, and “Have you ever been married?”

Responses to these items were coded as following:

I: currently married or cohabiting (n=4035)
2: have been married but currently not (n=318)
3: never married (n=848)

89

Current education attainment was grouped into four categories based
on self-report of years of education received by the time of assessment. The
original WMH-CIDI question is:” How many years of school have you completed?
(IF'NEC: Please include any years of higher education. )”

Responses to this item were coded as following:

1: “0—6 years” (n=l68)
2: “7-9 years” (n=425)
3= “IO-12 years” (n=580)
4: “> 12 years” (n=455)

These four categories correspond to “less than or finished elementary
school”, “some or finished middle school”, “some or finished high school”,
and “some college and above”, respectively.

Information about income was obtained from respondent self-report
about personal income in the last month. The original question is: “How much
Might your own personal earnings income be in last month? including wage,
P1129, extra income from another job, present from relatives or friends, yield
from investment such as stocks etc and all other sources of income) ”

Responses to this item were coded as following:

 

a number <99990= Yuan (¥) (n=5101)
99998= DON’T KNOW (n=14)
99999: REFUSED (n=77)

Personal income was obtained for 5101 out of 5201 respondents (2605

out of 2633 in Beijing and 2496 out of 2568 in Shanghai). Respondents who

90

answered “don’t know” or “refused” were coded as missing values. Personal
income was further categorized into high (> 1500 Yuan), high average (1001-
1500 Yuan), low average (601-1000 Yuan), and low income (<=600 Yuan)
based on the percentile distribution of the personal income variable.

The employment status variable has two categories, “working” and
“not working”, based upon the respondents’ self-report of current
employment status at the time of assessment. The original WMH-le question
is:” What about your current employment situation -- are you worla'ng now for

pay, self-employed, loola'ng for work, disabled, temporarily laid off, retired, a
homemaker, a full-time or part-time student, or something else?”

Responses to this item were coded as following:

1: working now or self-employed (n=293 l)

2: unemployed (looking for work), temporarily laid off, retired,
homemaker, student, maternity leave, illness/sick leave, disabled (n=2266)

98: don’t know (n=0)

99= refused (n=4)

Table 3.4 provides a description of these variables.

With respect to aim three, the main suspected causal variable is
“childhood physical punishment” (CPP). It is assessed by a question in the
“childhood” module and was assessed for all Part 11 respondents (n=1628).
The question is:“ When you were growing up, how often did someone in your
household do any of the things (on the list on page 38 in your booklet) to you -

often, sometimes, rarely, or never?”

91

 

Table 3.4 Distribution of sociodemographic variables. Data from WMH-mC, 2001-2002.

 

Entire Sample Beijing Shanghai
' h ' hted
wig ted weighted wearg
n % % n % % n % %

 

Sex

Female 2668 54.3 47.4 1366 51.9 47.2 1302 50.7 47.6
Male 2533 48.7 52.6 1267 48.1 52.8 1266 49.3 52.4

Age groups
Born before 1949 1385 26.6 18.7 767 29.1 17.8 618 24.1 19.6
1949-1965 2493 47.9 40.1 1264 48.0 36.4 1229 47.9 43.8
1966-1977 811 15.6 25.2 400 15.2 27.3 411 16.0 23.0
Born after 1977 512 9.8 16.05 202 7.7 18.5 310 12.1 13.6

Education attainment 1
<= 6 years 168 10.3 6.7 105 11.5 6.4 63 8.8 7.1
7-12 years 425 26.1 22.7 237 25.9 21.9 188 26.3 23.9
13-15 years 580 35.6 38.1 302 33.0 34.9 278 38.9 42.2
>15 years 455 28.0 32.5 270 29.6 36.9 185 25.9 26.8

Personal income level
Low 1396 27.4 29.0 672 25.8 29.3 724 29.0 28.7
Low-average 1581 31.0 27.0 815 31.3 25.6 766 30.7 28.5
High-average 871 17.1 17.0 478 18.4 17.7 396 15.9 16.3
High 1250 24.5 27.0 640 24.6 27.5 610 24.4 26.5

Marital Status

Married/cohabiting 4035 77.6 69.1 2144 81 .4 68.2 1891 73.6 70.0

No longer married 318 6.1 4.2 164 6.2 3.8 154 6.0 4.5
Never married 848 16.3 26.8 325 12.3 28.1 523 20.4 25.5
Employment

Working 2931 56.4 61.4 1489 56.6 63.0 1442 56.2 59.8
Notworking 2266 43.6 38.6 1141 43.4 37.1 1125 43.8 40.3

 

1' Variable available in part 2 only

 

2' Due to rounding, percentages may not add up to 100%.

 

Responses to the question are:

1: often (n=50)
2= sometimes (n= 185)
3= rarely (n=330)
4: never (n= 1046)
8: don’t know (n= 12)

92

9= refused (11:5)
Deeds of physical punishment included “pushed, grabbed or shoved”,

“threw something”, or “slapped, hit, or punched.” More severe forms of
physical punishment, such as burning and hanging, were not assessed due to
a concern about upsetting the respondent. Thus, this variable is slightly
different from a variable that appears in some of the previous literature on
physical abuse in childhood. For this reason, we labeled this variable
“childhood physical punishment”. In order to gain statistical efficiency, we
created a new binary variable for CPP. Respondents who answered “often”
and “sometimes” were coded as “yes”; “rarely” and “never” were coded as
“no”; “don’t know” and “refused” were coded as missing.

Other covariates as possible confounders include “parental
alcohol/drug problem”, “parental mental problems” and “childhood conduct
problems” as well as sex and age of the respondent. “Parental alcohol/drug
problem” was derived from two separate questions about alcohol/drug
problems regarding mother and father, respectively. Original questions are:
“Did (MAN WHO RAISED R) ever have a problem with alcohol or drugs?” and
“Did ('W OMAN WHO RAISED R) ever have a problem with alcohol or drugs?”
Due to the low frequency of mother alcohol/ drug problems, we combined
maternal problems and paternal problems into one variable to represent
alcohol/drug problems of both parents. “Don’t know” and “refused” were
coded as missing values. Responses were coded as the following

1: either father or mother or both had alcohol or drug problems (n= 100)

93

0= neither father or mother had alcohol or drug problems (n= 1494)
missing values (n=34)

“Parental mental problems” included mother/father depressive mood,
anxiety, and suicide attempt. Original questions for the woman who raised the
respondent are: “During the years you were growing up, did (WOMAN WHO
RAISED R) ever have periods lasting 2 weeks or more where she was sad or
depressed most of the time 2’”, “during the time you were growing up, did
(WOMAN WHO RAISED R) ever have periods of a month or more when she was
constantly nervous, edgy, or anxious?”, and “Did (WOMAN WHO RAISED R)
ever attempt to commit suicide?” Three such questions were asked regarding
the man who raised the respondent as well. Due to the low occurrence of the
individual condition, we coded the “parental mental problem” variable as
“yes” when the respondent reported any of the three conditions for either
mother or father. “Don’t know” and “refused” were coded as missing values.
Responses were coded as the following:

1: either mother or father had above conditions (n=303)

0= neither mother or father had above conditions (n=1135)

missing (n= 190)

“Childhood conduct problems” were obtained from two sets of
consecutive questions about conduct problems. These questions about
conduct problems were asked to respondents younger than 40 years among
the sub-sample of 1628 respondents (n=570). Actual questions are listed in the

following table.

94

 

 

Table 3.5 Actual CIDI questions assessing conductproblems

 

*CD1a. As a child or teenager, did you often tell lies to trick people into giving you things or doing
what you wanted them to do?

 

*CD1 b. As a child or teenager, did you often get out of doing things you were supposed to do by
fooling people or lying to them?

 

*CD1c. As a child or teenager, did you often stay out much later at night than your parents
wanted?

 

*CDld. As a child or teenager, did you often skip school without permission?

 

*CDle. As a child or teenager, did you ever shoplift or steal something worth at least $10?

 

*Cle. As a child or teenager, did you ever steal money or other things from your parents or the
other people you lived with?

 

*CD1g. As a child or teenager, did you ever break into someone’s locked car, or a locked home or
building?

 

*CD1 h. As a child or teenager, did you ever set a fire to try to cause serious damage?

 

*CD1 i. (Other than by setting fires,) As a child or teenager, did you ever deliberately damage
someone's property by doing something like breaking windows, slashing tires, vandalizing,
or writing graffiti on buildings?

 

*CDl j. As a child or teenager, did you ever run away from home and stay away for at least four
days?

 

*CD1k. As a child or teenager, did you run away from home overnight more than once?

 

*C016b. As a child or teenager, did you often get involved in physical fights?

 

*CD16c. As a child or teenager, did you ever use a weapon on another person, like a baseball
bat, glass bottle, knife, gun, or brick?

 

*CD16d. As a child or teenager, were you ever physically cruel to an animal and hurt it on
purpose? (IF NEC: This does not include hunting or getting rid of pests like rodents or
insects.)

 

*CDlBe. As a child or teenager, were you ever physically cruel to a person and hurt them on
purpose?

 

*CDl6f. As a child or teenager, did you ever force someone to give you something like money,
jewelry, or clothing by threatening them or causing them injury?

 

*CD169. As a child or teenager, did you ever steal someone’s purse, wallet, luggage, package or
bag by grabbing it from them? (IF NEC: This does not include stealing from someone
who wasn’t aware of the theft, such as stealing a piece of luggage when the owner
wasn’t watching.)

 

*C016h. As a child or teenager, did you ever make anyone do something sexual by either
forcing, intimidating, or threatening them?

 

 

 

Respondents who answered “yes” to any of the above question were
coded as “yes”. Respondents who answered “no” to all of above questions
were coded as “no”. “Don’t know” and “refused” answers were coded as

missing values. Responses were coded as the following:

95

1= ever had conduct problem(s) (n=88)

0= never had above conditions (n=458)

missing (n=24)
3.5 Analysis Plan

In order to address the study aims and fulfill corresponding estimation
tasks, several statistical techniques have been used for this project. A
standard “explore, analyze/ estimate, explore” three step cycle was used for

analysis. Described below are the statistical methods for each study aim.

Aim 1: To describe beverage alcohol involvement in two metropolitan cities
in China: Beijing and Shanghai

1. Frequency tables and estimated proportions were used to study the
lifetime occurrence of drinking-related variables including opportunity to use
alcohol, trying alcohol, alcohol drinking (>= 12 drinks/year), frequency of
drinking when drank the most, heavier drinking, as well as diagnosis of
alcohol abuse/harmful use and alcohol dependence as outlined in DSM-N and
ICD-10. Additionally, the estimated occurrence of clinical features of socially
maladaptive drinking and alcohol dependence as outlined in DSM-N and
ICD-10 were plotted. Further, lifetime occurrence estimates wsere derived for
early onset of trying alcohol (< 13 years), early onset of MTM drinking (< 20
years), early onset of socially maladaptive drinking (< 23 years), and early

onset of the first clinical feature of alcohol dependence (< 23 years).

96

2. In the next, interval prevalence was estimated for the period of 12
months prior to the assessment with respect to alcohol drinking, heavier
drinking, frequency of drinking, as well as diagnosis of alcohol abuse or
harmful use and alcohol dependence as outlined in DSM-1V and ICD-10.
Frequencies and estimated proportion have been used to describe the
recency of the clinical features of alcohol abuse or harmful use and alcohol
dependence.

3. Finally, Kaplan-Meier methods have been used to describe the age of
first trying alcoholic beverages, the age of onset of MTM drinking, the age of
first socio—maladaptive drinking problem, and the age of the first clinical
feature of alcohol dependence. Range, means, and medians have been
presented for these variables as well as number of drinks during a typical
drinking day when the respondent drank the most and during the 12 months
prior to the assessment, respectively.

These estimates have been presented for both specific sites and for the
entire sample.
Aim 2: Estimation of subgroup variations with respect to alcohol drinking and
related problems

Subgroup specific estimation of lifetime occurrence and 12-month
prevalence of drinking-related variables have been presented with respect to
sex, age groups, marital status, education attainment, income level, and
employment status. Subgroup variations in drinking behaviors and drinking-

associated problems of these variables were addressed via generalized linear

97

models with a logit link (logistic regression). The strength of association is
estimated in the form of odds ratios (OR). The statistical robustness is
evaluated by the p value with a p<0.05 considered to be statistically

significant by conventional standards. The following equation depicts logistic

 

)= flo + Zflixi

regression:

log( p
1-p

p is the probability of the occurrence of the outcome;

30 is the log odds of the occurrence of the outcome when all covariates, X, are
zero;

13,- is the log odds of the occurrence of the outcome for each unit increase in x,
relative to the log odds of the outcome when xi=0.

Multiple covariates can be entered into logistic regression in order to
hold others constant. In this study, both unadjusted OR (uOR) and adjusted OR
(aOR) are presented. To estimate the aOR for sex, age was held constant; to
estimate the aOR for age groups, sex was held constant; and to estimate the
aOR for all other variables, sex and age were held constant.

Aim 3: To estimate the association between childhood physical punishment
(CPP) in riskier drinking and associated problems in order to shed light on
the suspected causes of drinking problems.

1. First, estimates wre produced for a line chart description of the
occurrence of drinking associated outcomes for individuals with CPP

experience and without CPP experience, respectively.

98

Next, logistic regressions were used to estimate the association
between CPP and drinking-related outcomes with a focus on indicators of
riskier drinking behavior. These indicators included lifetime occurrence of
“heavier drinking”, “socially maladaptive drinking”, “DSM-IV alcohol
dependence”, “any alcohol dependence clinical feature”, “early try (<13)”,
“early MTM drinking (<20)”, “early socio-maladaptive problems (<23)”,
“early dependence problems (<23)”, and “early any problem (<23)”. A
series of logistic regression models was fit to assess the strength of
association between CPP and drinking-related outcomes as well as the
statistical independency of the association from other covariates: the
participant’s sex, age, parental alcohol/drug problems, parental mental
problems, and history of conduct problems.

Due to the cross-sectional nature of the study design, it is difficult to pin
down temporal sequencing. For example, parent may punish children
physically when children start to drink too early or had problems with
drinking. Therefore, an attempt was made in order to probe into this issue.
First, the variable “onset of problems after 16” was created, where individuals
with an onset of problems before 16 years old were excluded. Second, in one
of the models for each outcome, we restricted the analytical sample to
individuals who started MTM drinking after 16 years old.

Subsequently, possible misspecification of logistic regression models
was explored via an adapted version of the Lemeshow-Hosmer test for

goodness-of-fit for complex survey data (KJ. Archer 8: Lemeshow, 2006). The

99

Lemeshow-Hosmer test evaluates the departure of predicted probability from
observed probability. The first step of Lemeshow-Hosmer test consists of
sorting observations into g (e.g. g=10) groups according to their predicted
probability. The observed number of cases in the in the dth decile is given by
0 1d; the observed number of non-cases in the dth decile is given by 00d;
similarly, expected numbers of cases and non-cases are given by Em and EM,

respectively. The Lemeshow-Hosmer test is calculated by

1

_ 2
6g zzzg:{0thE/1d) }

h—O d_1 hd . The test statistic follows a chi—

 

square distribution with g-2 degree of freedom (DF; (Hosmer & Lemeshow,
1980). However, this test is only suitable for random samples. It is not suitable
for data collected from multi-stage sampling, where within cluster variations
are usually smaller than between cluster variations and a probability weight is
involved. To overcome this, Archer and Lemeshow proposed a F-corrected
Wald test to test goodness-of-fit for multi-stage survey data (KJ. Archer &
Lemeshow, 2006; K. ]. Archer, Lemeshow, 8: Hosmer, 2007). The Wald test

statistic for g groups is given by

QM = (f_g+2)MtI}(M)—1M

(fg ) . where f =(number of PSU-

 

number of strata), g =number of groups, it is the sum of mean residuals by

deciles, l? is estimated variance-covariance matrix, the estimation of which

100

involves a Taylor series linearization. The statistic follows an F distribution
with g-l numerator DF and f-g+2 denominator DF (KJ. Archer 8: Lemeshow,
2006).

As described in the introduction, CPP might be influenced by parental
drinking/ drug problems and parental mental disturbances. Thus, estimates
from logistic regression might have been biased or inconsistent due to a
possible endogeneity problem. In other words, the assumption of
independent error terms might have been violated. In this context, recursive
probit regression and instrumental variable methods were used to explore
the possible endogeneity of CPP (Avery, 2005). The bivariate probit with
dummy endogenous variable, or recursive probit model, was first introduced
by Heckman (Heckman, 1978). In this model, two probit equations are fit

simultaneously to the data:

4: r
yz =x1fl1+51 , (4.1)

:r: r
yr : 352,32 +72J’2 +52 ,(4.2)

C0V[£1,£2|x1,x2]=p.

Thus, the parameter rho represents the correlation between error
terms of the two equations. Non-zero rho indicates the existence of correlation.
Using this approach, possible endogeneity of CPP from parental variables as

well as “conduct problem” have been explored. The log-likelihood function is

101

LnL = 2w,- 1n<1>2<q15f425;,p;)

 

given by j— 1 (StataCorp LP.,
1 1+ p
2005). Instead of directly estimating p, a tanh )0 = 21n(1— )0) is estimated

from this maximum likelihood estimation. Although the likelihood function
treats the two equations as if they are exogenous, it can be shown that when

y2 is observed as binary, the likelihood function is the same (Wooldridge,

2002). Thus, the Va is a legitimate estimate of the CPP-drinking outcome

association taking the endogeneity into account (if there is any).

Therefore, a bootstrap re-sampling approach with replacement was
used to investigate the internal validity and stability of the model-based
estimates. This approach was encouraged by Royston and Sauerbrei (2008)
in their text on multivariable model building and evaluation. As part of the
third step in the analysis approach for this dissertation research, the bootstrap
was used to draw 600 bootstrap samples of size 1628 from the WMH-mC
dataset. Then, the final model was fit to these 600 bootstrap samples in order
to produce a distribution of estimates of the CPP-alcohol association. The
actual empirical distribution of estimates was used to generate the 95%
confidence interval. Two models were evaluated for internal validity and
stability: estimates for the occurrence of socially maladaptive drinking when

adjusting for sex, age, parental alcohol/drug problems; the occurrence of

102

early onset of socially maladaptive drinking when adjusting for sex, age,
parental alcohol/drug problems.

Next, to investigate this possible causal relationship between CPP and
drinking problems, there was probing into questions of whether there might
be variations in the associations across different outcomes. Comparing the
confidence intervals from logistic regression leads to loss of power because it
ignores the correlation between outcomes. For example, a person with history
of heavier drinking has a higher probability to have socially maladaptive
problems than a person with no history of heavier drinking. Hence,
comparing confidence intervals is too conservative and may lead to
erroneous inference when correlated outcomes are present. The Generalized
Estimating Equations (GEE) proposed by Liang and Zeger (Zhao et al., 2004)
takes both within-cluster and between cluster information into account in
estimating coefficients. GEE also provides direct statistical comparison in
order to answer the question proposed above. Thus, in the next steps of
analysis, the GEE was used to examine variations across different drinking
outcomes in associations between CPP and drinking outcomes, taking the
inter-correlation of these outcomes into account. The GEE model is a marginal
model and does not require specification of a specific joint distribution. When
the study interest is the population mean, GEE provides an efficient and valid
tool for this estimation problem.

The GEE model requires independency between clusters (individuals

in this case), which might not be met due to the complex study design of this

103

study. Thus, a variant of GEE, the Alternating Logistic Regression (ALR) is
used. ALR is capable of modeling multiple levels of clustering by alternating
between the marginal logistic regression and the log odds ratio within
clusters (Carey, Zeger, 8: Diggle, 1993; Katz, Carey, Zeger, 8: Sommer, 1993).
Here we treated neighborhood as the level of geographical clustering
because it is the final unit of sampling of the study. The ALR approach can be
time-consuming and sometimes has difficulty in convergence. The GEE has
the advantage of computational efficiency over the ALR.

Lastly, when robust variations in these associations were found, the
Multiple Indicator Multiple Cause (MIMIC) model has been used to probe
into whether the observed variation might be due to differential item—level
responses, holding the level of problematic drinking constant. The MIMIC
model is within the framework of latent variable analysis (Muthen 8: Muthen,
2007). The diagram of the model is depicted in Figure 3.4. In the MIMIC
model, information was gathered from multiple clinical features of alcohol
dependence to create a dimensional latent variable. The existence of CPP-
associated differential item functioning can be tested by simultaneously
regressing the latent alcohol dependence variable and specific clinical
features on CPP. If the estimated path from CPP to an item response is non-
null, it indicates reporting biases. The technique has been previously used in
the study of psychiatric conditions including alcohol and drug dependence (C.
Y. Chen 8: Anthony, 2003; Gallo, Anthony, 8: Muthen, 1994; Harford 8: Muthen,

2001).

104

oEoBoa _oEoE\_oo_oEo
£38 9256 2 85:60

Iv _o;oo_m co zoom 9:: do fioo :85 <
Iv 9255 do 823 E 8228 a: 2,5
I :28 =5 3.855

I 955%

Iv
Iv 8:922.
Iv 26% 22662.. /
Iv mEoBoa .mmob /

Iv 9.25.6 8888:

Iv msozoa _o_ooo 28% 9.255

Iv mEozoa _m_oom

Iv ooceotog b___o_ocoomom

mEozoa

Eoefica .8528 ”Fx

 

/
9.255 \1

\/I\\

 

 

@o

\\

 

 

.\\

88. 0.2.2 2. .6 5385 3. 2:3

105

One goal has been to find out whether CPP affects every stage of
alcohol involvement, or only a subset of stages. The statistical tool is logistic
regression. Stages of alcohol involvement include

1) the opportunity to drink alcohol,

2) ever tried alcohol given the opportunity

3) ever drinking at least 12 drinks per year given tried

4) ever heavier drinking given drinking

5) ever socio-maladaptive problems given drinking

6) ever clinical features of alcohol dependence given drinking

Analyses consist of a series of logistic regressions to show the crude OR
and adjusted ORs after holding potential confounders constant. Since the
occurrence of the first stages of alcohol involvement is quite high (>50%), the

OR does not simulate relative risk (RR) well. To ease interpretation, we used

OR
the formula RR = (1_ P0 ) + (P0 X OR) to convert OR to RR (Zhang 8: Yu, 1998).

 

Special issues regarding variance estimate

As a consequence of the multi-stage sampling procedure, participants
are not completely independent. That is, an individual is more likely to be
similar to an individual from the same neighborhood than one from another
neighborhood. In addition, in this study, participants had different
probabilities to enter part II of assessment depending on their response to
part I. To address these issues, sample weights were generated to account for

differential participation and part I/part II selection. A Taylor series

106

linearization method was used for variance estimation, with due attention to
variation in sample selection probabilities, post-stratification adjustment
factors, and the nested strata created as part of the multi—stage probability
sampling procedures. The purpose of doing this is to minimize possible
estimation bias and make results more generalizable to the source population,
non-institutional household dwelling adults living in Beijing and Shanghai,
China.

Analyses have been implemented using statistical software Stata
(version 9.2, StataCorp LP), Mplus (version 5.1, Muthén and Muthén), and SAS

(version 9.1, SAS Institute Inc).

107

Chapter 4 Results

This chapter presents the results of analyses corresponding to each of
the study aims.
4. 1 . Lifetime occurrence and 12 month prevalence of drinking-related
outcomes
4.1.1. Lifetime occurrence

Lifetime occurrence of drinking behavior and drinking-related
outcomes are provided in table 4.1.1. Results show that, for the population
under study as a whole, alcohol is very accessible. Only a small fraction of
people had never had a chance to drink alcohol and the majority of people
have tried alcohol at least once (71.9% in Beijing and 69.9% in Shanghai). And
less than half of people had ever consumed 12 drinks in a given year during
their lifetime up until the assessment (45.3% in Beijing and 39.2% in Shanghai).
For the third of people who drank, the frequency of drinking is quite evenly
distributed across four categories of frequencies. For both sites, about 10% of
people drank nearly everyday when drank the most. The median number of
drinks consumed is two for the periods when the respondent drank the most.
Histogram of number of drinks in a day is presented in figure 4.1.1 for the
period when the respondent drank the most.

More serious drinking, by contrast, is relatively rare. Less than 10% of
people had ever drunk heavily (8.0% in Beijing and 6.9% in Shanghai). Less

than 7% of people ever had socially maladaptive problems because of

108

drinking and only 2% of people met criteria of alcohol dependence posed by
DSM—IV or ICD-10. One thing merits mention is that the occurrence of
drinking-induced problems is consistently higher in Beijing than in Shanghai.
The occurrence of drinking-related problems in people living in Beijing is
approximately twice of that in those living in Shanghai. The difference is
statistically significant at a 0.05 level. This difference remains robust even
after holding drinking frequencies, history of heavier drinking, as well as
smoking status, constant (OR=0.5, 95%, 0.3, 0.8).

With respect to the age of onset of alcohol involvement, the median age
of the first sip of alcohol is 18 years old; the median age is 20 for the onset of
MTM drinking (>=12 drinks in a year) among MTM drinkers. The median age
of onset of drinking-related problems is in the early to mid 203. Details about
the distribution of these variables are presented in table 4.1.2 and figure 4.1.3
to figure 4.1.6. For the population as a whole, slightly more than 10% of
people tried alcohol before teens (11.3% in Beijing and 13.7% in Shanghai);
and 19.5% of people started drinking before 20 years old in Beijing, 12.3% in
Shanghai. In Beijing, 2% of people initiated socially maladaptive problems
before 23 years old, 1.1% in Shanghai; and 0.9% had the first dependence
clinical feature before 23 in Beijing, 0.3% in Shanghai.

We also provided line chart for a site-specific criterion-wise
description of the lifetime occurrence for each socially maladaptive problem
and clinical features of dependence (figure 4.1.6). The lifetime occurrence is

consistently lower in people living in Shanghai than that in people living in

109

Beijing. “Hazard-laden drinking” and “drink more than intended” are the
most common clinical features for both sites. “Interfere with responsibility" is
also one of the most common clinical features in Beijing. The least occurred

clinical feature is “legal problems”.

110

 

 

 

mom gm EN F.mm o.vm Bow 3m Em Nvmv oz 83%
FodvQ no Wm Nm 5 No mm om BF «.0 Fm. 3. mmm oo> _o;8_o 2-00.
osm Sm «mom 5mm 93 Bow 0mm 3m Bmv oz 83o
Fodvo to Qm o.m mu No 3 m6. Q: to 3 3. me 8; _ocoo_o>_-_2mo
mF mdm tum mum 3 gm NE «8 NF 0mm Fan 82 xcco o:
QF me Em omm «F gm mNm F8 3 Sm .F.mm th >_£coEv
3 mm 3 Eu No S m.m «2 9o 3 m: mmm 58.5.18 3
No me F3 3: m.o F.FF mm mmm md Fm Fm to 5.833.533
ed 3 F.m E No Fm Qm m: md Qm mm RN 5oozm>ooxm
modvq mm mm 2: mmm o.F m.oF NmF 3m mm 02 Q: 05 830558:
9.555 6 583mm
F.mm 5mm Emu on mNm Fmvm gm Nmm 32: oz
md mm no 2: ad om us 2: mm 3 3 m3 mo> 85:558.;
Em «.8 mmmm new v.3 ommm 35 Q8 moss oz 9.556
5.0VQ mm QNF mFF mmm 2 02 3: «mm m.o mg m? F8 mm; 25528853
05 two mFoF m3 v.8 0va 9mm mmm mmom oz
5.0 m: m.wm Sm me NF F93 mm... 82 FF 9% Nov Rom 8> 9556 .25
m8 Sm omom Nam FNm comm Sm odm moms oz
0F 52 «NF mmm FF 0: mg Fom No QNF 0.9 So oo> _o;oo_o 852:8
can v.3 mum 9mm FNm «8 Now Nmm oooF oz
9.0 mF o8 Qmo $9 3 v: m8 3: NF mmo m3 comm mo> _o;oo_o BEBE
«Nu 2N EF 3F Now 8F now 0.8 Em oz
mud ma QR NR 08 QF «.5 odw «mu 3 Nmn omN mmmF 8; 5.22.88
N o .o.m Fees, 5. : .o.m Foss, 5. c do F055, 5. :
cm? $5, 55.;
5355 m5__om 5958 52m

 

.NoowFoom 68-123 9: £0: £8 803%? 866.. 9555.6 _ocoo_m Fo 8:388 m>=m5E=o meson—3 .F. FF. 2an

 

111

 

82.88 36 C858 B was, a N

{code 2 a: 8m 6: >8. 88.528 8.2.8. o. 80 88.8.8 839%., F

 

o8 m8 mmmN v.3 Em momN N8 08 8% oz 8.38.

.8...__o 8c%=8%

No No 2 we 5 3 N. F N. _‘ 8 NNA _o..8_m

5.08 to md to or md ad ad 2 No ed no 0N NNuv 8338.5
Em Sm EVN #8 N8 NoVN N8 98 8% oz 8.38.5256

: 3 2. mm mm 5— NN mN m3 NNA 83m 68%

Edva no I E mN md ozN o.N Fm md o.N m: cm NNuv 2338.5
38 m8 mmmN 2m Sm momN 38 28 8% oz 238.

_8_c__o 859.8%

65%

dea N... Z : mN md NN VN No 2. we 3 S 8> .8 85.26
0.8 Nam EmN 38 m8 mamN 0.8 N8 85 oz 8..%:8%

dea to to ed a to 3 Ne Nm N.o o4 ed 3 8> _o;8_m 2.28
0.8 mdm mmmw Nam Qmm 30m me mdm mm 5 oz sewncmamv

Edva to we to 9 Go 3 o; N No 2. No B 8> _o;8_m 3-09

 

328. P. E. 83

112

 

 

 

 

 

 

 

 

 

 

 

 

Bvamwmwvm—VNVOOOGOOOONNVNFNQQNwOOmo
..1.1wr. . . . F _ .I. . 51.81%... _. . .0
LEE .
-m
19d
-99 w
1 m w
L TQNnuu 5
1..- UI
.mm m.
:8
-mm
1. [ll 0
SFE .5
IQ.
,9d 8
1 a e
I -on m
e U
-mww. Ru
1,8
[mm

 

 

 

NOON-FOON .05.:23 EC; Guan— JmOE or: scant €2.25 >NU nan 9.5.6 *0 swag—:2 WV? 239“—

113

 

.2?me M 3:38 IIIII 9.5%. n .358

 

28> E mum
on om ov ow o

_ _ _ _ _

1
00'0

SZ'O

I
09'0

 

 

l
00'l

88.33 0:28? 9...... No man 2: no cases: 22.3 .o_os_.=a_aav_ N._..v 959".

114

 

ficmcmcm u .558 IIIII mar—Em u .558

 

28> 5 0mm
ow om ov 0N

_ _ _ —

 

9.25% 2:... No 828 ho was 2: .o 5.8:... 23:3 .m_os_.§_%x .45. 2:2".

 

SL'O 09'0 SZ'O OO'O

001

115

ow

 

.8925 n >.E:oo IIIII 9.2mm. u >..c:oo

 

28> :. mom
om ow ON 0

1
00'0

 

 

l
09'0 SZ'O

SL'O

Ir
0
l 0

 

95.5... 8.8828. >=~.oom .2.. m... .o 8.85.. 22.2 3.0.2.533. v.5. 2.5.“.

116

 

.mcmcmcw u >258 lllll 922mm. u >258

 

28> c. wow
00 CV ON

 

 

 

r

 

2.3303 882.8% .2.. 2.. .o 5.3:... 2:5. 332.533. 3... 2:9...

SL'O 09'0 SZ'O 00'0

OO'L

117

 

Table 4.1.2. Description of alcohol drinking related variables. Data from the WMH-mC, 2001-2002

 

Entire Sample

 

weighted weighted

 

 

 

age n range mean median mean (s.d.) median
first try 3193 0, 66 17.7 18 16.8 (7.8) 18
onset of drinking 2049 3, 67 22.9 20 21.8 (7.3) 20
1st social maladaptive problem 200 10, 52 27.5 26 25.6 (7.0) 23
1st dependence problem 73 15, 48 27.8 28 26.6 (4.3) 24
# of drinks/day when most 1593 0, 60 3.3 3 3.5 (4.5) 2
# of drinks/day during last year 1314 0, 60 2.4 2 2.5 (2.7) 2
Beijing
first try 1715 0. 66 18.2 18 16.8 (8.6) 18
onset of drinking 1132 3, 65 23.3 20 21.7 (7.3) 20
1st social maladaptive problem 134 10, 52 27.3 26.5 25.0 (7.2) 23
1st dependence problem 49 15, 40 28.3 30 26.5 (3.4) 24
# of drinks/day when most 886 0, 60 3.2 2 3.4 (4.1) 2
# of drinks/day during last year 738 0, 60 2.4 2 2.5 (2.8) 2
Shanghai
first try 1478 1, 54 16.9 18 16.8 (6.5) 18
onset of drinking 917 3, 67 22.3 20 22.0 (7.2) 20
1st social maladaptive problem 66 10, 48 27.7 25 27.1 (6.2) 23
1st dependence problem 24 17, 48 26.8 24.5 26.8 (6.1) 23
# of drinks/day when most 707 0, 54 3.4 3 3.6 (4.9) 2
# of drinks/day during last year 576 0, 54 2.4 2 2.5 (2.5) 2

 

118

 

 

 

 

 

~I-j- Shanghai:

 

 

+ Beijing

1
1

Figure 4.1.6. Lifetime occurrence of drinking-related problmes.
Data from WMH-m0, 2001-2002

 

«o
.9%
cc
.8.
o . 3
”we 8%
3%.er 8%
80 90.,
./ . on, as a
2 sum. 0%
G» e
a 2%
\ - 01w: %

/

/ \\
/\
a2:

 

 

 

 

1:.

119

 

3“

f3.-

4.1.2 lZ-month prevalence of drinking-related outcomes

Drinking in general is not an overly common phenomenon during the
12 months prior to the assessment. Approximately one third of the population
did not drink alcohol during the past 12 months (28.2% in Beijing and 30.4%
in Shanghai), and another 40% drank less than monthly (41.5% in Beijing and
43.5% in Shanghai). An estimated one in thirteen people in Beijing and one in
fourteen people in Shanghai drank everyday during the 12 months prior to the
assessment. Only 3% of people drank heavily (> =5 drinks in a day for males
and >=4 drinks in a day for females) during the past 12 months. The
prevalence proportions of drinking-related problems are also low. The 12-
month prevalence of any socially maladaptive problem is 2.1% in Beijing and
0.7% in Shanghai; the prevalence of alcohol dependence is less than 1%
according to either DSM-IV criteria or ICD-10 criteria. Details can be found in
table 4.1.3. The histogram of the number of drinks in a typical drinking day is

provided in figure 4.1.7.

120

 

 

 

Naa Naa aaam Naa a.aa tam a.aa Naa ata oz 8cmc=8md
addvQ Pd ad ad a ad ad ad 3 rd ad ad am 99> 658.3700.
Naa Naa aaam —.aa vaa aram vda a.aa Ea oz .852me
addva Pd ad ad a ad ad ad a? ad ad vd am 8> _o;8_m>_-s_ao
a.aa Naa daam asa vda raga ada dda aEa oz
Edvd Pd ad ad am ad E 3 av ad 3 3 aa 8> 3325.23-09
vda a.aa 3am dda ada ddam nda ada RE 02
deQ d ad ad 2 ad ad a; av ad ad 2 Na am; 33223,.“ 2.28
dd a.a~ tan aFaP we Naa o2 ad? ad a: ad“ Ska ascoEv
ad ad as «2 ad a as aw ad ad ad ada 5:25:33
ad Na dd aaF ad Na ad dFN ad ad ad Ka xmméfimcm;
vd Na ad ad ad 3. 3. a: ad 3. ad dam xmméfimc Ya
addAQ ad a.“ as com ad ad ad aam ed ad ad aav >855 2.8:
accent do 58:95
ada 2a Neva ada d.\.a Sam Nda :a avav oz
308 ed ta ad on ad ad ad R ad ad ad E mm; asxédbaoz
d3 Nam aFaF Naa add 52 a: adn «Ea 02
ded d._ tam dew «ad ad ada da an ad Nda OR at; 9.; 9.256 E2
.3 has, 0a c .3 «OE ea c .3 $3, 0a :
o\o§> oats oats
codmcg 5:955 9.58 238 9:5
27.. 5228
do mam:

 

.Noofioom dead; sea 98 85%? 852 9.255 28.“ do 8:2?2385205 .a. E. 2de

 

121

8vamrm®vmvmvoo©mooomnmvmwwwamaww o o o

k — — _ _ — p b _ h ir — P _ p — _

iriirF

 

[___-O

 

 

r

assasséébossssaawébo
iueoied

I

leufiueus

 

 

T

 

 

 

 

iuemed
Bunileg

 

 

I

I

I

 

 

 

Nooméoow .0E-I_2>> Eo: 3mm. .59» ammo. 9.: 95% >3 .8 9.56 do .8832 n. F .v 05am.

122

4.2. Subgroup variations of lifetime occurrence and 12 month prevalence
of drinking-related outcomes

4.2.1 . Male is associated with higher likelihood of drinking-related
outcomes.

Being male is associated with higher likelihood of having experienced
all drinking-related outcomes studied (table 4.2.1). The strongest associations
were found in more serious and problematic drinking, such as heavier
drinking, socially maladaptive drinking, and alcohol dependence clinical
features. For example, compared with about 1% females who ever drank
heavily (>= 4 drinks in a day), approximately 13% males ever drank heavily
(>=5 drinks per day); less than 10% females drank at least monthly, while
more than 40% males drank at least monthly during the 12 months prior to the
assessment.

Interestingly, males are more likely to ever have opportunity to drink
alcohol (OR=2.5, 95% C.I. 1.8, 3.6). Further analysis showed there are age-
related subgroup variations in the male-female difference in opportunity to
drink alcohol. The sex-specific proportions of opportunity to drink alcohol in
each age group (from the oldest to the youngest) are 85.7% in males vs.
64.9% in females (aOR=3.2 after holding age constant, p<0.01), 92.6% in
males vs. 70.8% in females (aOR=5.3, p<0.01), 85.8% in males vs. 85.8% in
females (aOR=l.0, p>0.05), and 82.9% vs. 84.8% in females (aOR=0.85,

p>0.05).

123

There is no between site variation in associations between sex and
these drinking outcomes except for ever tried alcohol. The association is

slightly stronger in Shanghai compared with that in Beijing.

 

Table 4.2.1. The association between drinking-related outcomes and sex. Data from WMH-mC,
2001-2002

 

 

nof
cases wt%2 OR 95%CI 80R3 95% Cl

opportunity to drink alcohol female 549 71.9
1 male 709 86.8 2.6 1.8 3.6 2.5 1.8 3.6

ever tried Beijing female 733 61.4
male 1025 80.1 2.5 1.8 3.5 2.5 1.8 3.5

. female 612 52.8
Shangha' male 970 81.5 3.9 3.3 4.6 4.0 3.3 4.7

MTM drinking female 491 19.6
male 1586 62.0 6.7 5.6 8.1 6.9 5.7 8.3

ever heavy drinking female 29 1.2
male 320 13.1 13.2 9.0 19.5 11.9 7.6 18.4

ever socially maladaptive female 15 0-7
problem male 223 9.1 17.1 10.1 29.0 15.0 7.0 32.3

ever dependence clinical female 5 0-2
feature male 85 3.4 15.4 6.7 35.3 14.7 5.5 39.4

past year MTM drinking female 228 8.7
male 1191 45.7 9.5 8.2 11.2 9.3 7.6 11.2

past year heavy drinking female 11 0.4
male 136 6.0 14.0 7.6 26.0 14.1 7.3 27.5

past year socially female 5 0-2
maladaptive problem male 58 2.5 12.5 5.0 31.3 10.9 3.9 30.7

early try female 172 9.3
male 315 15.2 2.1 1.7 2.5 1.7 1.4 2.2

early MTM drinking female 94 4.8
male 587 26.0 8.3 6.6 10.4 7.3 5.8 9.2

early alcohol abuse or female 5 0.3
dependence male 79 3.8 17.3 7.0 42.7 11.3 4.1 30.6

 

1' Variable available in part 2 only

2 Proportion of cases of all female/male participants.

3

aOR: OR after holding age constant

 

124

4.2.2. Subgroup variation in drinking outcomes with respect to age
groups

Due to small numbers of cases in sub-group in drinking outcomes
“past-year socially maladaptive problems” and “early onset of drinking
problems”, we combined the two older groups and the two younger groups.
Also, analyses were restricted in the subsample of males for outcomes “ever

socially maladaptive problems”, “ever dependence clinical feature , past-
year socially maladaptive problems”, and “early onset of drinking problems”
due to too few female cases in each sub-group. Compared with the oldest
group (born before 1949), later groups presented higher occurrence of most
of the drinking outcomes studied here (table 4.2.2). More individuals in the
youngest group tried alcohol before teens and initiated MTM drinking before
20, as compared to the two oldest groups (p<0.01). Twenty percent of
individuals in the youngest group tried alcohol before teens, and 28.3%
started to drink at least 12 drinks in a year before age 20. After collapsing two
younger groups and two older groups, the younger group had higher
occurrence of “early onset of alcohol problems”: 5.7% in the younger group

vs. 2.4% in the older group. No cross-sites variations were found in any

outcome studies here.

125

 

 

ad ad dd ad 2 Ratmdm EB

dd ad ad ad da Radaaad

ad ad ad ad an aaadavad
dd. dd avad 988 m 9.256 >52 .8» 7.8

ad ad ad ad ad dd dd dd Radars E8

dd ad ad dd ad ad d.a da Radaaad

dd ad ad dd ad ad ad add aaadavad
Va da add 223 95.56 >52 55

ad ad ad dd ad ad add aa Ratmcm E8

dd ad ad dd ad ad dda dad Radaaad

ad ad ad dd ad ad a.aa 3d aaadavad
d. d dud avad 988 9.256 5E2 .8.» “mad

ad ad dd ad ad dd a.aa at. Ratmcm E8

ad ad dd ad ad ad adv daa Radaaad

ad ad ad dd ad ad dad. aa: aaa—-avad
a. da adv avad 928 acicca 2:2

ad dd ad ad dd dd ada 3a Radium E8

ad ad ad vd ad ad a.aa Ea Radaaad

ad dd dd ad ad ad add dd: aaadavad
a.aa won avad 983 8E 55

ad ad ad Na ad ad add aad Ratmfi Eon

d.a ed ad ad 3 ad a.aa dad Radaaad

ad d... ad dd _.d ad add aaa aaadavad
add dad avad 288 _ _o;8_a 8.56 2 Ecdtoddo

_o oaaa «mom .0 0\..aa 10 gas, 88
do:

 

88-38 .OE-I_2>> E9: Ema

.mmaoaofio mam Em mmE8So 322-92%“. c853 c2388.... 2:. ddd was.

 

126

 

.mm_mE do maEmanza 05 :0 Emma 05 agar—Emu .a

E228 x8 asgon 5am mo ”mod .N
28 d :8 E o_na__a>m o_nmca> ..

 

add
ad
ad
fia
Na
d. _.

ad
ad
ad
ad
ad
ad

ad
ad
ad
ad
ad
a. d

56

N. 2.
ad
ad
ad
a.a
a. _.

n. _.
Na
ad
ad

ad
ad
ad
ad
ad
ad
ad
ad
dd
ad
dd
ad

ad
ed
_..a
ad
ad
ad
a. d
ad
d. d
dd
vi
ad

Na
vd
add
a. dd
add
d.a
ddd
a.ad
ad
a.a
Na
ad
ad
ad
ad
add
ddd
dd

aa

a3
ad
add
aa

aa

dad
Rd

aaad Eta Eom
aaad onoaon
Rad norm Eon
Radaaad
aaadavad
avad anodon
Rad norm Eon
Radaaad
aaadavad
avad mnoaon
aaad 5cm Eon
aaad onoaon
aaad norm Eon
aaad mnoaon
Rad norm Eon
Radaaad
aaadavad
avad mnoaon

m 85238 no manna _onoo_m 2.8

952% SE). a8

E Ema
n 5.85
o>_amoa_mE £28m do» damn

m 238.. _mo_:__o 85288 .96

m Eo_noa ozamagae >__a_ooa .96

 

3.53 d2. 833

 

127

4.2.

Ollll

We

We

4.2

011'

lit

4.2.3. The association between marital status and drinking-related
outcomes

Some subgroup variations in drinking-related outcomes are found
across different marital status at the time of assessment (table 4.2.3).
Compared with people who were married, those who were no longer married
(divorced/widowed/separated) are more likely to have a history of heavier
drinking. Between-site variations were found for the variable “ever tried
alcohol”. The difference lies in that “no longer married” people in Beijing
were more likely to have opportunity and have tried alcohol than people who
were married (table 4.2.3).
4.2.4 The association between education attainment and drinking-related
outcomes

Table 4.2.4 presents the association between education attainment by
the day of assessment and drinking outcomes. After holding sex and age
constant, people with higher education levels are more likely to have tried
alcohol and have a history of MTM drinking. In contrast, for riskier drinking
behavior and drinking problems, no subgroup variations were found across
education levels. No between-site variation was found in any of drinking-

associated outcomes.

128

 

 

 

dd dd ad ad ad ad ad aa 855.96: ”5285
ad dad nmn8\85m5 98835
.96 .o 255.8 >__m_ooa .o>o
dd ad dd ad ad dd aa dd 853.98
a... 8d @835... n 9.25..
.o>o 5 252.3 >>mon .mod 58
ad ad ad ad ad ad ad aa 8.53.96:
ad. dd ad ad ad ad add ad aoEm5.oano_on 9.255
dd Ed nanooduoEmE .98.. .26
dd ad ad ad ad ad 3d add $53.95:
ad ad ad dd ad ad add ad 85559.29. 9.255
dad a: nmnSBoEmE 2:2 5o.» 68
ad ad ad dd ad ad ddv daa 855.96:
ad ad dd ad ad ad a.aa dad ao5m5.oano_on
odd. daad nanoodaoEmE 9.255 2.52
dd ad ad ad dd ad add ad 855.22
ad ad ad dd ad ad ada an do5m5.oano_o_. finanmna
a.aa a3 d 8.62855
dd ad ad ad ad ad add add $55.96..
dd ad ad ad ad ad a.aa td ao5m5.oano_on an_.__om
a. E a S; nanooBoEmE 85 .26
dd ad ad ad ad ad a.aa add 355.99.
Na 2. ad ad a... 2 as 02 85.2.8522. :2835...
at «ca nanoodaoEmE 2 352.88
5 oaaa “mod _0 naaa mo cad; mowmodon
dddddddd

65-12; 5o: a.aa .mamda _£_.m5 and $58.8 8.99-9256 nooéon 8.568% on» ddd. o_nm._.

 

129

 

.865 do 635688 65 no woman 6.6 865.8 .a

589.8 885 6am and .oaa .xoa anion 6% mo ”mom .

d

zoo d :8 5 6336.6 6_nm_.a> d

 

ad

ad

dd
ad

ad

vd

dd

ad

a.a
ad

_..a

ad

a.a

ad

ad
ad

a.a

vd

a.a

dd

vd

Nd
a. d

ad

ad

ad

ad

ad

dd
ad

ad

ad

ad

ad

ad

ad
ad

dd

ad

a.a

Nam
a. dd
add
a. dd
a.a
a.a
ad
ad

ad
a.a

34v
aad
a d
aaa
a _.
av

ad
ca

.6565 6.6:
nanoodaoEmE
.66 6 25656
8565 6.6..

8:65 692 on
nanoodnoEmE
85.5 6.6:
“.6565 69.2 on
nmn8.8565
8565 6.6:
nanoodaoEmE
6.6 .o 25658
3565 .66..

nmn8\BEmE
6.6 6 25658

M 856288
6 manna
_onoo_m 2.66

5.5....
Es. a.a...

an 2.66

n 5285
6253365
a__m_ooa .moa and

a 6.363 6256
856288 6.6

5.83 add. 638

 

130

 

Table 4.2.4. The association between drinking-related outcomes and education attainment.
Data from WMH-m0, 2001-2002

 

#cases wt% OR 95% CI 30R2 95% CI
opportunity to drink 16 yrs 107 64.4
alcohol‘ 78 yrs 335 81.5 2.4 1.4 4.2 1.6 0.9 2.9
10-12yrs 455 82.3 2.6 1.6 4.1 1.6 0.8 3.0
>12yrs 361 81.7 2.5 1.4 4.3 1.6 0.9 2.9
ever tried 1-6 yrs 81 47.1
7-9 yrs 278 69.0 2.5 1.6 4.0 1.6 1.0 2.5
10-12yrs 386 70.7 2.7 1.7 4.2 1.6 0.9 2.9
>12yrs 314 71.6 2.8 1.7 4.6 1.8 1.1 3.1
MTM drinking 1-5 yrs 47 24,9
7-9yrs 181 40.0 2.0 1.1 3.5 1.1 0.6 2.0
10-12 yrs 253 44.4 2.4 1.5 3.9 1.4 0.7 2.6
>12yrs 197 41.1 2.1 1.3 3.4 1.4 0.7 2.5
past year MTM drinking 1-6 yrs 35 13,5
7-9yrs 133 27.4 1.6 0.9 3.0 0.8 0.4 1.7
10-12yrs 188 31.9 2.0 1.2 3.4 1.1 0.6 2.1
>12yrs 132 27.9 1.7 1.0 2.9 1.0 0.5 2.1
ever heavy drinking 1-6 yrs 10 4.3
7-9 yrs 42 7.6 1.6 0.5 5.0 0.7 0.2 2.7
10-12yrs 66 12.3 2.8 1.0 8.0 1.1 0.3 3.5
>12 yrs 35 7.7 1.7 0.6 4.7 0.7 0.2 2.4
past year heavy drinking3 0-9 yrs 13 33
>9yrs 41 8.4 2.7 1.1 6.8 2.3 0.9 5.6
ever socially maladaptive 1-9 yrs 39 92
Problem3 >9yrs 75 9.0 1.0 0.5 1.9 0.7 0.4 1.5
ever dependence clinical 0-9 yrs 25 3,9 .
feature" >9yrs 38 3.2 0.8 0.4 1.5 0.6 0.3 1.2
past year socially 0-9 yrs 15 2,0
maladaptive Pmb'em3 >9yrs 23 3.1 1.6 0.6 3.9 1.1 0.4 3.0
early try 08 yrs 36 6.1
>9yrs 126 13.5 2.4 1.1 5.0 1.5 0.7 3.3
early MTM drinking 1-5 yrs 9 4,5
7-9yrs 62 13.9 3.4 1.1 9.9 1.0 0.3 3.1
10-12yrs 103 19.5 5.1 1.7 15.0 1.2 0.4 3.7
>12yrs 86 17.5 4.4 1.5 13.2 1.3 0.4 3.9
early alcohol abuse or 0-9 yrs 11 2,5
dependence3 >9 yrs 33 4.2 1.7 0.6 5.2 1.2 0.5 2.5

 

1' Variable available in part 2 only; 2' aOR: OR after holding sex and age constant; 3- Estimates are based on

the subsample of males.

131

4.2.5. The association between personal income level and drinking
outcomes

Table 4.2.5 presents associations between drinking outcomes and
categorical personal income during the month prior to the assessment. After
holding sex and age constant, individuals in the higher personal income
categories were more likely to be involved in drinking, such as opportunity to
drink alcohol, ever tried alcohol, and ever MTM drinking, compared with
people in the low personal income category (less than 600 Yuan). They are
also more likely to drink during the past year; and people in the highest
personal income category are more likely to drink heavily during the past
year (OR=2.0, 95% CI, 1.3, 3.3). In contrast, income level does not appear to

be associated with riskier drinking and drinking problems.

4.2.6 . The association between drinking outcomes and employment
status

Individuals who had a job at the time of the assessment were associated
With higher likelihood of drinking involvement, including experiences during
lifetime and past year. For riskier drinking and drinking problems, no

employment-status-associated subgroup variations were found (table 4.2.7).

132

 

Table 4.2.5. The association between drinking-related outcomes and the personal income level.
# cases wt% OR 95% Cl aOR 2 95% Cl

 

opportunity to Low 333 74.4
drinkalcohO' Low-average 355 80.3 1.4 0.8 2.4 1.8 1.1 3.0
1 High-average 221 77.8 1.2 0.7 2.2 1.2 0.7 2.1
High 324 86.4 2.2 1.3 3.7 2.0 1.2 3.4

ever tried Low 789 61.6
Low-average 945 66.6 1.2 1.0 1.5 1.5 1.3 1.8
High-average 603 73.8 1.8 1.4 2.2 1.8 1.4 2.3
High 938 79.0 2.4 1.8 3.1 2.0 1.5 2.7

MTM drinking Low 462 34.3
Low-average 576 37.8 1.1 0.9 1.4 1.3 1.1 1.6
High-average 382 46.8 1.7 1.3 2.1 1.6 1.2 2.1
High 616 50.4 1.9 1.5 2.4 1.5 1.2 1.9

past year Low 168 23.8
MTM drinking Beijing Low-average 230 27.2 1.2 0.8 1.7 1.3 0.9 1.9
High-average 142 30.9 1.4 0.9 2.2 1.5 0.9 2.5
High 240 39.4 2.1 1.5 3.0 1.7 1.2 2.4

Low 119 16.6

Shan hai Low-average 173 23.5 1.5 1.2 1.9 1.7 1.3 2.2
9 High-average 126 35.4 2.7 2.1 3.6 2.4 1.3 3.1

High 196 33.3 2.5 2.0 3.2 1.3 1.4 2.4

ever heavy Low 52 8.8
drinking Beijing Low-average 48 6.0 0.7 0.4 1.1 0.7 0.4 1.2
High-average 31 8.5 1.0 0.5 1.7 0.9 0.5 1.6
High 54 8.2 0.9 0.5 1 0.7 0.4 1.2

Low 32 4.9
. Low-average 39 5.7 1.2 0.7 2.1 1.3 0.7 2.3
Shangha' High-average 28 7.7 1.6 0.8 3.0 1.2 0.7 2.4
High 59 10.2 2.2 1.4 3.4 1.5 0.9 2.5

past year Low 31 4.6
heavy Low-average 26 4.1 0.9 0.5 1.6 0.9 0.4 1.9
drinki093 High-average 25 6.0 1.3 0.7 2.5 1.3 0.6 2.7
High 51 8.6 1.9 1.3 3.0 1.7 1.0 3.1

ever socially Low 53 10,1
maladaplfve Low-average 58 8.0 0.8 0.5 1.3 0.9 0.5 1.5
problem3 High-average 30 8.5 0.8 0.5 1.4 0.8 0.5 1.4
High 70 9.8 1.0 0.6 1.5 0.8 0.5 1.4

ever Low 31 5.2
dependence Low-average 18 2.4 0.5 0.2 1.0 0.5 0.2 1.0
3253;, High-average 10 2.1 0.4 0.1 1.1 0.3 0.1 1.0
_ High 25 3.4 0.6 0.3 1.4 0.5 0.2 1.2

 

133

 

Table 4.2.5. (cont'd)

 

past year Low

socially , Low-average

maladaptive High-average

problem 3 High

early try Low
Low-average
High-average

High

early MTM Low

drinking Low-average
High-average
High

early alcohol Low

abuse 0’ Low-average

gependence High-average

High
1. Variable available in part 2 only;

21
14

18
123
108

79
161
185
158
114
214

28

10

12

29

2. aOR: OR after holding sex and age constant;
3. Estimates are based on the subsample of males.

134

3.8
1.7
1.0
2.8
12.5
9.1
10.6
16.1
16.8
11.0
16.6
19.8
5.1
1.7
4.7
4.0

0.4
0.3
0.7

0.7
0.8
1.3

0.6
1.0
1.2

0.3
0.9
0.8

0.2
0.1
0.3

0.5
0.6
1.0

0.5
0.7
0.9

0.1
0.4
0.4

0.9
0.8
1.8

1.0
1.2
1.9

0.8
1.4
1.7

0.8
2.0
1.4

0.7
0.3
0.7

1.0
1.0
1.4

1.0
1.3
1.3

0.5
1.4
1.1

0.3
0.1
0.2

0.7
0.7
1.0

0.7
0.9
0.9

0.2
0.6
0.5

1.4
1.0
2.1

1.5
1.4
1.9

1.4
1.8
1.7

1.3
3.4
2.2

 

Table 4.2.7. The association between drinking-related outcomes and current employment
status. Data from WMH-m0, 2001-2002

 

 

nof

cases wt% OR 95% Cl aOR2 95% Cl

opportunity to drink not working 517 732
eleehe" working 740 85.1 2.1 1.4 3.1 1.6 1.0 2.4

ever tried not working 1239 60.3
working 2099 75.7 2.0 1.8 2.4 1.4 1.2 1.7

MTM drinking not working 697 32.4
working 1379 48.0 1.9 1.7 2.2 1.4 1.1 1.7

past year MTM drinking not working 436 20.0
working 983 33.7 2.1 1.3 2.4 1.5 1.2 1.9

ever heavy drinking not working 105 5.6
working 244 8.6 1.6 1.2 2.2 1.1 0.7 1.7

past year heavy drinking not working 39 4,4
3 working 97 6.7 1.6 1.0 2.4 1.3 0.7 2.3

ever socially not working 67 8.1
maladaptive emblem" working 156 9.6 1.2 0.8 1.8 0.7 0.5 1.2

ever dependence not working 29 3,8
Cliniee'feer'e3 working 56 3.1 0.8 0.5 1.5 0.4 0.2 0.8

past year socially not working 20 3.0
meledepfivepreb'em3 working 38 2.2 0.7 0.4 1.5 0.5 0.2 1.1

early try not working 183 11.3
working 304 13.2 1.2 0.9 1.5 0.9 0.7 1.1

early MTM drinking not working 207 12.3
working 474 18.2 1.6 1.3 2.0 1.1 0.9 1.5

early alcohol abuse or not working 27 4.0
dependence" working 52 3.7 0.9 0.5 1.6 0.8 0.4 1.6

 

1' Variable available in part 2 only

2' aOR: OR after holding sex and age constant

3' Estimates are based on the subsample of males.

135

4.3. The impact of childhood physical punishment on alcohol drinking
outcomes
4.3.1 . The associations between childhood physical punishment (CPP)
and drinking outcomes.
4.3. 1 . l . Estimation using logistic regressions

As shown in the figure 4.3.1.1, figure 4.3.1.2, and table 4.3.1.1,
individuals, who suffered childhood physical punishment (CPP), are more
likely to have experienced undesirable drinking outcomes in both cities. ORs
were mostly moderate. No between site variation in the association between
CPP and drinking outcomes were found. After adjusting for sex and age,
estimates are not appreciably different; all estimates remained statistically

robust (table 4.3.1.1).

136

Figure 4.3.1.1. Lifetime occurrence of drinking outcomes stratified by
experience of childhood physical punishment in Beijing.

40.0 1

35.04

30.0 ; +non-CPP +CPP
25.0 7
20.0 i
15.0 4
10.0
5.0
0.0 d -

 

9:190}, I7 elf/e

0:7(79/2910 . 0)
flag?) "945/'0 02/“? 3/09 ’l’ee 07% 8’?
6b/ 0’0 all“

[globe ea; “:3“ 06/9/72

.9

Ta

4 ‘33}?-
aye/9

Mei/1,641 6‘77
fly/(é C%:6;9SZI:/:yoé0

019/
379% a

#6),qu 0'

Figure 4.3.1.2. Lifetime occurrence of drinking outcomes stratified by
experience of childhood physical punishment in Shanghai.

30.0

25.0 +non-CPP +CPP
20.0 «
15.0 l
10.0
5.0
0.0 i , -

 

137

 

.8538 mom can .mmm .xmm .8 8632 .N

8:95, gm $228 so 02? a .F

 

«No
05o
Ed
vmd
$6
26
2.0
mmd
mmd
Rd

Fas

m6 s...
ed v...
a.a o...
ed 54
Ev 0.?
Q2 _..~
5mm m.~
fit. NN
ad «é
m.” E.
._.0 $8

a.a
a.a

.3

Eu
:6

5K

m.~
ad

NmOm

 

who o.» 3 a.a 5.9 mm 8»

Wm E o: 2 5% 5285 be $80
E q: 3 N... 3 a m;

me 3. 0: 23mm. _S_:__o 88288 Em
NS «5 3 3. mm 3 ma

No R c: 88288 65% >228 .05
8.0 S 3 2 3 mm mm;

o6 mm o: 9556 023893 2.98m 55
m; 5. 3 3 v.2 9. 8»

E NF F o: 9.255 53mm; 55
v2 ”.8 3 3 S 2 mm;

3 mm o: a; 8qu 5285 .Em 2L8
3o Q: 2 N.» E m m;

md NF 0: Am; mmquEman 88288 2.8
So I: 3 E E 2 ma Eémuvv

N— mm o: 3:56 9:33ng 2.308 2.8
8.0 a.a 3 tm mam 8 mm;

mi. 2: 0: Am; Envy 9:55 5:2 2.8
8.0 3 2 E 2: on 9.:

a.a RF 0: Am; 9qu 31:3
2 ._.0 $8 mo: o\° mmwmo EU

82%; so:

 

88-58 dais; so: 28 €sz 25 3x55 5%: Em assign .813 82220 $sz 2258.02 .F. :2. was

 

138

0R5 remained robust after holding sets of potential confounders
(parental alcohol or drug problem, parental mental disturbances, and conduct
problems of the respondent) constant (table 4.3.1.2 and 4.3.1.3), although
points estimates decreased. In the subpopulation of individuals, who initiated
drinking (at least 12 drinks in a year) after 16, estimates were not as precise
and a few associations were not statistically robust at 0.05 level (possibly due
to a lack of power); nonetheless many associations remain robust and point
estimates are not distant from estimates based on the entire part II sample.

No between-site variations in these associations were found. Since no
evidence of site variations was found from above analysis, samples from the

two sites were combined for later analysis in order to gain statistical efficiency.

139

 

was 925.2 .2828 3285 .235 3:98 22 2.2868 m .262 €328 3285 9:26:86 5:28 new .motodmfio mom .mdm .xom 22 :08: .

cedar? 25 comics do 8.? a .

N

.N Enos 5 8 05:98 do Em 08mm o5 £3 3 5% 9.256 BEE on; 22365 do 833933 05 c_ 8858.

—

 

 

 

dd 8 o: or
and E v... «d NNN mm 8» 3d ed 3 m.~ ed to o; tn 5% 528.3238
dd mm o: 05de
mod md dd Nd dd 3 8» med 3 3 ed Ed dd 3. md _mo_:__o 8coccmamc Em
: mm 8
Rd ed 5. ed ad 3 8.» mad 3. 3 3 Ed o.» 3 ed 8528855
rd E o: 3285 $3823
Rd «.5 3 «6 «Nu mm mm» id Nm 2. ed 8d ed 3 _..m 256855
ddm mm o:
mmd 5m dd 5 add an me» mad 3 I. a.a Ed 3 E a.a dExEEmSmmfgm
m._. E o: a;
mod Ev m.~ «.2 NE 9 m2 mmd 3: m... S. mmd mi 5 ed mmuvdsoaoa Em 2.8
dd 5 0: Am; mwnvv 5283
8.0 gm N... ad Pd m 9% end «.2 ad ad 2d 93. .3 3 882.8% 25m
F. _. NP 0: N Am; mmuvv Eozoa
vmd adv a.a «.2. mm? 2 mm; Nvd 92. ~._. 3. Ed 92 5 _..m m>_am8_m§__m_8i=8
mom 5 8 a; 2i
Ed ed dd 5 mdv t. 8» dvd 3. 3 ed mmd Nm 3 2 9.256 2.22 2.8
I; F. 0:
add 3 ad md 59 2 8» dad 3. ed 3 Ed Nm 3 d4 Edamruvvblcmm
«a ._.0 $8 mom $9; 8mg ado N ._.0 $8 mom an ._.0 $8 men
do:
ammuci .082 .N 382 2 .822

 

88.58 55-12; :5: 98 $2850 9.255 .283 2m $2523 323 Boise :853 seasons 2: .N. :3. was

140

 

.808. 2 a: can do: >9: 83:8:8 96:29 9 one 63:8:8 8.363 d

:ode; 0% :852 .6 029 a .m

dv :9: Edge» 9w; 9.3 manusis .o 3328 389: :8 29:8 .832: 9:. 2228 8:28: _8_:__o .882: 58:8 2:: 26828: $82 .—

 

 

3 mm 0:

odd 3 Rd 3 3; om 8» 9 sum 52:9: do #8::
3 mm 0:

Ed 2 3 d.~ dd 6.: 8» 23:2 $253 85288 Em
dd 2 o:

«Pd ed dd Nu dd d 8.» 8:85am: :26
.3 mm 0:

mod 3 Rd 3 rd— mm mm; 32:2: 2883? >__m_oom :96
I: d: 0:

mdd Nd de mm m. 5 mm 8.» 92:2: 53$: 55
3 t 0:

dvd 2 md 3 ad 9 mg Am; mmuvv Emsoa >:m 3:8
vd d 0:

5d mde v... I d: d m2 Am; mmuvv 82:9: 8:888: :8
ad 2 0:

mvd as md dd dd 3 mm» Am; mmnvv 8:56 938ng 2:208 2:8
de dd 0:

3d dd 5 ed 3m 3 mm; Am; Envy d:_v_:.:: 2.22 58
u. E 3 o:

2.: 3 no 2 m: 8 mg a; $qu a: 2.8

N a ._.o «and me: n $8393, 880 .6 2 ago

 

sang 2. EB:

 

«8.0.-st dais; 50: 28 828.8 3.2% .285 2: 28.52.: .853: 822% :823 8:383: 2: .m. :2 was

 

141

RN 0 m F

 

 

8.2. 2:8

 

9.2%.: SE2 2:3

 

«r
J.

d:.:.:.:.: 5.28.. :96

 

A,
J.

<..< :26

 

o..< :26

 

0...... .o 22$. .8.:._o 22$

 

'IP
an.

52...... 2:8

'P

L J. , <._< 2:3

 

T + . , 0.2 2:3

 

 

 

. + . , 25% :2

.m.m.:m>8 52...; $6850 d:.v.:.:u Em €95.25: .833: 39.2.5 :858 80.5.88: .0 mac d. 3.: 2.6.“.

142

km a 0 ..

 

 

.

T

 

3.2. 2:66

 

. . : r 8.2% SE. 2:66

 

"r
9

9.226 6.69. :96

O

 

<..< :96

'F'
q

 

o..< :96

 

o..< .o 9266. 66.2.6 2:66

J.
J-

 

 

 

T h 6 , 656 2:66

.r . fl , 5:66

.r 4 u t 092.8
T T d. . 9:9“ 92

 

 

 

.6228 26.85 32.2.8.6 6:66.. cam
.96 .66 9.22. 668830 9:226 new .5526,an 66.9.3 89.2.8 5629. 226.8666 .0 #5 .v. .86 9:9“.

143

Many previous studies have found that the association between
childhood physical abuse and drinking and drug problems is present in
females, but absent in males (Simpson 8: Miller, 2002). As we stated in chapter
2 (introduction), from an epidemiological perspective, the nature of the
association in males is actually largely unknown due to the fact that many
previous evidence is based upon subgroups of populations. In this context, in
the next steps, above analysis were restricted in the sub-sample of males.

Robust CPP-drinking associations were found from the current study.

144

 

266.8 626.82. 669:: 6.86.. 656.86 9:96:86 669%
.68 .66 29. 26 .2 6.6 2:226 862:. 2.; 66:22.2. 6 8.6.2.2536 2.. c. 8.266: .6; 268.2 .9 6.6 9:226 862:. 2.; 66:222. 6 8.6.2.886 9..
c. 86:69 6; 29.22 68:62:26... 6.8:: 6.86.. :o. 86:26 268286 «68.2 68.82. 3.98226 6.86.. 26 66.6860 96 .68 .66 :o. 86:66 268.2 .F

 

 

 

N.» 3 tn 2m 0.. 2m to 3 a.a «.6 m... 2m a.a. mm 69. m.
«.2 mm 2: 6.6 26.82. .0 .668
a... 20 ON 6... m... 6.. a.a P... a.a «.6 N... n...” 6... mm 69. 9366.60.26
Nm 3 0.: 8:62:82: 26

2.... 5. Wm N... 3 a.a N.» 3 6..” «.2. 6.. 3 «2.. m. 69.
m... mm 0.: 8822.8 :66
rd v._‘ m.” in o._. mN ed #4. QN m6 @é Nd WON mm mm> mega—0...... 023.6239».
v.2 m2 2: 2.6.86 :96

F... No 5 Nm 3 EN 3 3 3 a... .3 3 ed.” mm 69.
6.2 S. 2: 9.226 6.62. :96
v.9. m.~ 6.2 6.3 E a... 1m. 3 3 m5 2 a... mu. m: 69. 6; mmnv.
ma 2 2: 56.82. 26 266
3m 3 N... «.9 a; o... 6.3 o.~ m... EN 3 3 2m 6 m9. 6:. 80.26.85
m... o. 2: 88268.. 266
6.9. 3 v.3 6.3 E 3 E: m... a... 2: «d m... mm. 3 69. 6:» mmuv.
26.82. 92.6.66...
Gm mm 2: 26.86 266
o... 0.. ca 3 3 o...” N... 3 o.~ m... 3 o.~ om... mo 69. 6:»
gm mo. 2: EU. 9.2%.. 266

m. to I. 6.. .3 a... Qm No me 3,. m... m... 0: mm 69.
N. S E o: 6:. «WV. 2. 266

...o 2.8 mom ._.0 2mm mom ...0 .28 mom ._.0 #8 mo: 0222 6666...... n.n.o

. $1568.. 389. .2682

 

82-52 62-122 as. 98 .362 5 358.8 2.2% 6.86 2.. .6565. 60.2... 68.26 :823 86.868 2:. 3... 9.2

145

 

.2. 8... 69:8.

99.. 2... 66:228. 6 66.28 32.2.: 6. 6.226 62.226 2. .. 266.8 8562 60.2.0 628... .828 22. 2682.66 8.85. ..

 

 

m... 2 8.

m... o... 0.. v... mm 2: o. 6.6 8.82. .o 626
6. : 2 8.

6.6 No «N 0m .N 2: 9:66. 62...... 8822.2. 26
«.0 m 8.

.3 6.0 mm m... E 2: 88226.. :96
«.6 6 8.

N... m... m. a.a mm 2. 68.82. 62.8.66... 2.6.86 :96
8 6 8.

as c... md m9 3 2: 9.226 6.62. :96
.2 2 8.

m... .1. 6.. mm m. 2: 6:. mmuv. 8.82. 26 266
0.0 2 69.

23 3 a.a No 2 2: 6:. mmuv. 8.86 2.82.2.2. 266
v2 2 8.

m... no 6.. Nm 2 2: N 6:. mmuv. 8.8... 92.8866 26.86 266
.8 6 8.

2m o.~ 3. m. .m om 2: 6:. 25. 9.226 .2 :2 266
.2 2 8.

6.. No No 6.3 mm 2: 6:. N—uv. 2. 266
._.o 28 mom 28.6.8; 880 .0 c .8

 

.88... 2 68.2

 

meow—com 68-1.2.5 26:.

660 .665 c. 62:88 9.226 .286 26 856.2... 66.62... 822.5 898.. 8.6.886 2. H .m. 3.... 2.6..

146

 

 

147

.......v .....ov 3.... mm... o. 6.6 8.86 .0 62.0
Edv 35v 3....v 3.9 6666. 60.2.0 88088.. 26
3.9 3.9 36v 36v 88.88.. :96
........v 35v No... E... 68.86 926266;: 26.08 :96
.......v 3.0 3.9 am... 9:28.. 6.62. :96
.......v .......v .......v 36v 6:. mmnv. 8.86 26 266
3.9 36v _......v 36v 6:. mmuv. 8.86 88088.. 266
.2. 80.
..o...v 3.0v 3.9 ..o.ov 8.86 626266.: 26.08 26..
36v .6... mm... mm... 6:. anv. 9:226 266
36v 5... E... 36v 6:. Nan. 2. 266
28.86 80886... 68.86 8.6860
.0328 :0. 6.8:. 6.86.. .0. 8:28.68 6.86.. 68 26 .68
8.8 .__2_o_._....< 8.8 222...? .82. 882.2. .8 a. .82..

 

88-53 .0E-..=2>> 2.0:. 660 250.8880 2.. .0 662-“. .0 m2.6> .. .0. mm... 2.6.—

 

A0. ._” 20¢ magma EQUOE ®>OQM HOH memmmmm mm? amum?mm0500.m .mQGU—m Unwfi 05 fin

b.0qomovso .0 50.6.0.9... .2.... 2.0.8262. 022.00. .0 5.3688000 .m. . é...

Results suggested lack-of-fit for many of the above models. (p values
from the F-test for models for the subpopulation of males also suggested lack-
of-fit, p<0.01.) There are different possible reasons that a model might have
been misspecified, such as missing important covariates, measurement errors,
incorrectly specified distribution of covariates (categorical vs. continuous)
(Begg 8: Lagakos, 1990). Of our major concern is the possibility that parental
drinking/drug problems or mental disturbances and CPP are not independent.
Moreover, “conduct problems” may be correlated with CPP. In this scenario,
the assumption of independence of error terms for logistic regression is
violated. Thus, we used the recursive probit models to (I) explore if there is
endogeneity in above models; (2) estimate coefficient for CPP on drinking
outcomes taking the endogeneity into account.

Table 4.3.1.7 lists results of the Wald test for the rho parameter from
these three models. If the p value is greater than 0.05, little endogeneity is
suggested and vise versa. Firstly possible endogenous variables were
assessed one by one; secondly parental variables were entered together;
finally conduct problems were added. The two equations which are to be

estimated simultaneously are,

CPP= 01+ [51 sex + [52 age + [53 age categories + [54 endogenous variable(s)

+errorl (1)

drinking outcome= 02+Y1 sex + y; age + v3 age categories + Y4 CPP

+error2 (2)

148

 

.8.8.6> 6:08:88 8 8.8.6:. 6866 0.6 8.86 .0880 8.. o 6:06 8.88.. 8083.08 8 80886.:
6.8:: 0.6 86.86 32:86.6 6.866 50: 8.. v 6.00:: 6.8.6:, 6:08:88 8 68396 .0268 8.. m .808 6.8.8:,

6:08:88 8 80886.: 6.8:: 6.866 8.. N 600:: 6.88:, 808:88 8 68.86 32:86.6 6.866 8.. F 682 ..

 

 

F:.:v mm: 8.: mm: 8.: o. 6.6 8.86 .0 680
m:.: _.:.:v 5.: 3.:v 9.: 6:28. 60.5.0 88:88: .8
8.: _.:.:v 3.: 3.: F:.:v 608.088: 68
_.:.:v .0: ~:.: and .9: 86.86 626686.: .668 68
mm: 8.: _.:.:v 8.: F:.:v 966.6 6.28.. 6.6
P:.:v 8.: 8.: 8.: 8.: 6:. mmuv. 8.86 .8 ..66
~:.: 8.: m:.: m 6: _.:.:v 6:. mmuv. 8.86 608.608: ..66
3.? 8o 8.: 8o 8.: .2. 80.

8.86 6266868 26.08 .86
mm: 8.: m:.: 8.: F:.:v 6:. ownv. 96.56 ._:8
mm: 8.: mm: 3.: mm: 6:. «WV. .:. ._66

m 68.2 v 685. m 685. N 68.2 F 68.2

 

89.86 no... .0 .8. 6... .0 88> q

 

88-68 05-122. 8:. 28 2.6882,... :2 08. .62, 52. 820, o. g. E... 63....

 

149

Evidence of endogeneity was found using recursive biprobit
regression model. Parental variables introduced endogeneity to CPP for
outcomes in relation to “dependence”; parental drinking/drug problems also
introduced endogeneity to adverse drinking behavior, such as “early
initiation of MTM drinking” and “history of heavy drinking”; conduct problem
introduced endogeneity to CPP for outcomes in relation to adverse drinking
behavior. Corresponding results for the subpopulation of males can be found
in the appendix table A43. 1 .7.

Table 4.3.1.8 shows estimates for CPP from recursive bivariate probit
models after taking endogeneity into account. Estimates are presented when
endogeneity is suggested from above results. Similar results for the subgroup

Of males can be found in appendix table A4.3.1.8.

One of the disadvantages of probit model is that it is not easy to attach a
Substantive meaning to the coefficient estimate. For example, the coefficient
Of 2.0 means CPP is associated with 2 standard deviation increase in the probit
function for early MTM drinking. Nevertheless, statistically robust and
Positive association between CPP and drinking outcomes were found from
recursive bivariate probit models after taking the endogeneity into
Consideration; when endogeneity is suggested, estimates from recursive
Probit model are greater than those from logistic regression. For all

Statistically robustly correlated error terms (rhoqéO), the correlation is
1”legative, which means unobserved heterogeneity affects CPP and parental

Variables and conduct problem in different ways.

150

 

.00_88> 0000000000 00 00.00.09 .0880 08 8.00.0 .8080 00.. m .0005
.0085? 00080008 00 80858.0 .385 08 05030.0 050.00.05.00 8.8.00 500 00.. 0 .0005 .088. 00080008 00 0830.0 .8080 00..

: .0005 8.8.8., 00080008 00 00085306 .0505 .0880 0.0.. N .0005 .088. 00080008 00 05030.0 00.0.9.0...00 .0880 00.. F .0005. ..

 

 

:.N .:.w :.w .2 .08 8.00.0 .0 .0000
0.: .:.: ...N :.N .:... :.N :.: .5.. ...N 0.80. 80.5.0 00808000 .8
:.: .0... 0N 0.: .3 0N :.: .0.. :.N 00808000 >_-5.mn. .0>0
..N .:.P :.P _..: 0.. :.N 05030.0 02.008.050.000 .0>0
:.N .0... N.N :.N .0... ...N 9.0.5.0 .080; .0>0
:.N ...... :.N :.: 5 :.N .0.. :Nuv. 8.00.0 .8 .80
:.: .2 :.N ...0 .N.N 5: .0.. :Nnv. 8.00.0 00808000 .80
E .0... 0.. .2. RU.
8.00.0 02.008.050.000 .80
:.N .:... :.N :.N .0... 5N :.N .0... :.N .0.. EU. 9.0.5.0 .80
:.N :.: N... :.N N.: E .0.. NWV. ... .80
...o 5:: .000 ._.o 5:: .000 ...o 5:: .000 ._.o .50: .000 ._.o .000
0.0::
m .0005. 0 .0005. : .0005. N .0005. P .0005.

 

.N::N-.::N 65-155. 50.. 0.0: 5:800 05. 5080008 9.0.0. .08 0.0005 .550 020.000. 50.. 000 .0. 0085.3 :. 5:... 0.8»

151

4.3.1.3 Stability of estimates

As described in the Methods chapter, the bootstrap resampling
approach was used to look into the stability of the model-based estimates.
Two estimates are assessed in this step. The first one is the estimate of “CPP-
lifetime occurrence of socially maladaptive drinking” relationship adjusting
for sex, age, and parental drinking or drug problems; the second is the
estimate of “CPP-early onset of socially maladaptive drinking” adjusting for
sex, age, and parental drinking or drug problems. Figure 4.3.1.3 show actual
distributions of estimates under the model just described, based upon 600
bootstrap re-samples of size 1628 from the WMH-mC dataset. As a measure of
central tendency, the mean value for this distribution is 1.1 and the 95% of the
estimates fall within an interval from 0.6 to 1.6 for the first model. The
corresponding OR (and its 95% CI) is 3.1 (1.9, 4.8). For the second model, the
mean value for this distribution is 1.7 and the 95% of the estimates fall within
an interval from 0.8 to 2.6 for the first outcome. The corresponding OR (and its
95% CI) is 5.2 (2.2, 13.1). In summary, this post-estimation data exploration
step helps to confirm that the CPP-alcohol association of primary interest is

non-null, and provides additional evidence against the null.

152

Figure 4.3.1.5 Distribution of coefficients from bootstrap resampling

procedure.

I!)
‘—

10
1

Frequency

 

Frequency
5
l

 

 

Panel 1: lifetime
occurrence of
socially maladaptive
drinking as
outcome. Covariates
include sex, age,
and parental
drinking or drug
problems.

    
  
  
    
  

1
coefficient

Panel 2: lifetime
occurrence of early
onset of socially
maladaptive
drinking as
outcome. Covariates
include sex, age,
and parental
drinking or drug
problems.

  

coefficient

153

4.3.2. Variations of associations between CPP and drinking across
different outcomes

Evidence of variations in the association between CPP and drinking
outcomes was found from both GEE and ALR analysis. Compared with the
CPP-“early try (before teens)” association, stronger association was found in
“early onset of drinking problems” including “early onset of socially
maladaptive problems”, “early onset of alcohol dependence”, and “early
onset of any problem” (table 4.3.2.2). Besides “early try”, stronger
associations were also found when comparing these “early onset” drinking
problems with “early MTM drinking (before 20)”, “ever heavier drinking”,
and “ever socially maladaptive problem” (p<0.05). No other variation was
found across other drinking outcomes (p>0.05).

As we stated in the Method section, ALR is able to take the possible
clustering of outcomes within the primary sampling unit into consideration.
Coefficients yielded from the alternating logistic regression (ALR) model is
slightly less precise compared with those from generalized estimating
equations (GEE). The statistical inferences are identical. The ALR results
suggested that above variations are not due to possible differential

distributions of sex and age across geographic sampling units.

154

 

...... .80. 55. 00.00500 00

0500.00 0... .0. 00.0 .0 0088. 0.000000 0:: 8... 000. 00.0., 0 0. 8.00.0... 50.7000 .0 5.0. 5000.0 0... .0 00.0. 0...

 

 

 

.3 0.0 0.. 00 0.0 0.0 ... 00 0.0 0.0 .0 .0 0:000 880.2008
00.0 .0. 0.. .0 0.0 .0 0.. 00 0.0 00 00 0... 05.00.0050 8000500050
0.... 0.0. 0.. 0.0 0.0 0.0. 0.. 0... 00.0 0... .0 0... 00.00.000.00...
0.... 0.0 0.. 00 0.0 0.0 ... 00 0.0 ..0 00 .0 050020
0>..0000.05 .__0.000 .0>0
00.0 0... 0.. 00 00.0 0.. 0.. 00 00.0 0.0 0.. 00 00......0.0_>00E0>0
.00 0.0. 00 0.0 8.0 0.... 00 0.0 00.0 0.0. 00 0.0 0.009.802.5208
00.0 0.0 00 0.. 3.0 0.00 .0 0.0 00.0 000 00 0.0 5.009.
8.00.0 00808000 .80
.00 0.0. 00 0.0 00.0 0.0. 00 0.0 00.0 0.0. 00 0.. 5.009.800...
02.000085 .._0.000 .80
00.0 00 .... 00 .00 00 0.. 00 00.0 00 0.. 00 5.0.5.5250 5.5.08
8.0 00 0.0 ... 8.0 00 0.0 ... 8...... 00 0.. .0 5.0.0.5500
8.0. 8.0. 8.0.
.0 _0 0.00 00 .0 _0 000 00 .0 _0 .000 00
0:0 08 x00 0.... m...< 0:0 08 x00 0.... 5.0.0 0.0..0>00 0.. 5.5 .wmo

 

NOON-E:N .oE-I5.>>

50.. 0.00. 00500.00 9.0.5.0 08 50550.80 80.0.00 00050.50 805.00 88.00000 0... .0 0..0..0..0> .N.N.:... 0.00.

 

155

ofpe

Simil

To better infer temporality, we restricted the analysis in the subsample
of people who did not initiate drinking (at least 12 drinks in a year) until 16.
Similar pattern was found from results (see appendix tableA 4.3.2.3).

Results from the MIMIC model found that at the same level of riskier
drinking, individuals who suffered from CPP are no more likely to report
“early onset" of drinking problems (p>0.05). Thus, the observed variations

are not due to differential reporting between CPP and non-CPP groups.

156

Iv 0.0.00. 80.50 on... .80
Iv 00808000 .0000.0 >.-5.wo .0>0

Iv 0.0.00. .0255 00808000 .80

 

mmdua .nm.oua
Iv 0.0.00. .8520 00808000 .0>0
. . .3". 0.0"“.
Iv 8.00.0 02.000085 28.00088 4/
Iv 00.0.5.0 5. ..5. .80 00.0.5.0 /i 8500.80.08.00 ”.x
80.0.. \I
Iv 000.808.800.000 \<
00.0". .000

Iv 8.00.0 02.000085 ...0.000 .80
Iv .0 .80

080.80 05-055 as. 0.00 .5050 .000. .0 08... 05:5 .00... 200..

157

Socially maladaptive problems and alcohol dependence problems
include multiple correlated manifestations. Figure 4.3.2.1 depicts the lifetime
occurrence of each of the problem studied. Individuals who suffered from CPP
were more likely to experience each of the drinking problems. According to
GEE and ALR models, the association between CPP and “irresistible desire of
drinking” is stronger than the one between CPP and “drinking despite
physical/ mental problems”, and so is for the association between CPP and
“difficulty cutting down". Furthermore, the association between CPP and
“difficulty cutting down” is also stronger than that between CPP and “drink
more than intended”. No other variation in the strength of associations across

different drinking problems was found.

 

I Figure 4.3.2.2. Lifetime occurrence of drinking problems. Data
from WMH-mC, 2001-2002

I .5

10 . +non-CPP (01392)

I
g 3 , 03—— CPP (n=235)
I

 

 

 

 

 

 

 

 

 

 

 

158

AmodAS ::.500 95.50 2.30%.. no :0hmmo 9.050.. 005.0 5.090. 0. 2.003 0.00.... o: 0.00 .30 Ho... 00.00.30

0..? 000308.65 60:00:08.0 3:008 .0 .0>0. 00.00 0.... .0 .05 00.0wmm00 .0005. 0.2.2 05 5.0.. 0.300%

 

50.0.00: .0 .50. .0000.0 00. .0 00.0. 0. .

 

m... :.: m... :.m :.: :.N :.m :.: :.N 0.00.00.0 0....000 0.0.5
::.: :.:N :.N N.: 8.: :.:N 0N m... 3.: :.:N ..N . :.: 00.02.00 00 020
:m.: m. E N.: :.N 5.: 0.: :.: :.N No: :.:. :.: 5: 0.0.. .0 .000 .005
m:.: :.:. a... :.m ::.: :.N. N.N :.m ::.: N.:. :.N :.m 0.500 .00
::.: :.: o. :.N N.:: N: N. :.N 9...: :.: 0.. :.N 00000.0. 8... 0.05.
: _..: 0:. 0.. m... 8.: :.:. w. :.m ::.: :.:. .N .0 850.05.;
::.: N.N. F... N: ::.: :.: 0.. N: ::.: _.. 3 m. N... 0000.0.0...
::.: :.:. N.N :.: ::.: 0.0.. :.N ...: ::.: .3.. :.: :.: 0.000:
5.: m0: : m.: _.N.: :.:: :._. :.: .N: 5:: N. . :.: 0.8.00.0 00:00
0..: :.:. m... N... :.: _..: m... .00 :.: :.: ..N m... 90.0.00 0000.000...
N.: :.:. m... N... :.: :.N. 0.. N... o...: :.3 m... .0 0.0.00.0 _0.000 0..0000 0.0..:
:0: m. o... ..N ::.: E N. :.N ::.: m: 0.. :.: 0.0.00.0 .068

0...: m. 5.. :.N :0: :.N N. :.: .0.: :.N 0... :.: 0000.0..0.0...._.0.00o000m
.0 60.08 10 .q 5.00:: «.0 .q 6.0%: m0

 

 

II000 08 x00 0..; m..< 0:0 08 x00 0.... mm: 0.08.8 00 0..5 mum.

 

N::N-.::N 65-15...
50.. 0.00 0058.00 000.000 08 8500.000 .00.0...0 00000.50 805.00 88.00000 0... .0 000..08> .:.N.:.0 0.00...

 

159

0000.085. ..._.0.000000m.
Iv .0000.0 00 .800 05.. .0 .000 .00.: <
Iv 000.000 .0 .0>0. 0. 0005000 00 020
I 5.8 50.3 20050

I 000.055

 

Irv
Iv 0088.0.
00.0.... .000".
Iv 0.000 0.000.000.
050 0.0 0:0 "
Iv .0 _ ._ 4/ 0.00000...
Iv 000.000 0000.000... 0000000000 .4 8500.000 80.0.00 H.x
3000.0 \

Iv 050.050 .0.000 0.0000 000.00: \<

Iv 050.050 8.00:

050.050 .0.005._00.0...0
Iv 000000 000.000 0. 000.500

.N::N-_.::N 65-1.23 0.0.. 0.00 0000000000 6000.0 .0 .0005 0.5:... .:.N.:.0 0.0:...

160

4.3.3. The association of CPP and drinking outcomes with respect to the
earliest to the later stages of alcohol involvement

As presented in table 4.3.3.1, in most of the stages of alcohol
involvement, individuals experienced CPP were more likely to be involved in
alcohol compared with their counterparts without such experience.
Endogeneity from parental variables (model 3) was explored. Evidence of
endogeneity (p<0.05 for the test of rho) was found for the first and the last
stage: opportunity to drink alcohol and occurrence of clinical feature of
alcohol dependence in drinkers. The estimate from recursive bivariate probit
model is 1.8 (95% 0.1. =l.5, 2.0) and 2.0 (95% 0.1. =l.0, 2.9), respectively.
Our result highlights the importance of the stage of the occurrence of alcohol
related problems (social maladaption and clinical features of alcohol
dependence) after the initiation of MTM drinking, where the largest relative
risks are found. For example, the occurrence of socially maladaptive
problems in people who experienced CPP is twice of that in people who did

not experience CPP given that they both drink.

161

 

 

 

000 00.0.00.0 000.000 .0.00.00 .0. 00.00.00 >._000...000 0 .0000. 000000.00 000 000 .0

.000000.0.0.0 .0.000. .0.00.00
00 .x00 .0. 00.00.00 0 .000... .0

 

 

 

00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0.00 000 00.. 0.00.0000.00.0.00.0
0.0 000 00 0000000000 .0>0
00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0.00 000 00.. 0.00.000 0.
0.00.020 0>..0000.0E
0.00 000 00 0.0.000 .0>0
00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0.00 000 00.. 0.0.0.0 0000.0
0.00 000 00 000.000 .9000 .0>0

00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0.0.. 000 00..
0.00 000 00 00... 00>... 000.000

00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0.00 000 00..
0.00 0000 00 0.0000000 00>... 00.0..

00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0.00 000 00..
0.00 0000 00 3.0000000 .0000.0

.0 $00 000 .o .0000 0.0.0 .o .0000 0.0.0
00.000E 00.00000 0.0000. 00..., 0 000

 

«08-0000 .0E-....2>> 00.

0.0.. 0.00 000.0; 0..; .00000>.0>0. 000.000 000.000.00.000 00.0000 00000000 0003.00 02.0.8000 00 .0 0.0.0.0 0.00.0

 

162

Chapter 5 Discussion

This chapter discusses results presented in chapter 4 with respect to
each study aims in the context of existing knowledge. Pertinent strengths and
limitations of the study will be addressed along with each study aim. The final
section draws conclusions and discusses implications for future studies.

5. l The frequency of beverage alcohol involvement in two
metropolitan cities in China: Beijing and Shanghai
5. l .1 Summary of results

For the population as a whole, the majority of people in Beijing and
Shanghai had the opportunity and had at least tried an alcoholic beverage
prior to the assessment. Less than half of the population had ever consumed
12 drinks in a year. Medium number of drinks is three for drinkers when they
drank the most. Less than one in ten people had a history of heavier drinking
or socially maladaptive drinking. Development of alcohol dependence, as
defined in DSM-IV or ICD-10, seems to be quite rare (<2 %).

Frequency distribution for the age of onset for the first drink of
alcoholic beverages showed a peak during teenage years (13-19). Socially
maladaptive drinking had onset starting in the late teen years; some onsets
were found throughout the 203 and 303.

Less than a third of the population in the two metropolitan cities drank
at least 12 drinks recently (i.e. during the year prior to the assessment).
Medium number of drinks is two during the last 12 months prior to the

assessment for recent drinkers. Less than four percent recently drank heavily.

163

Recent socially maladaptive drinking and alcohol dependence was rare (S
2%).
5.1 .2 Strengths and limitations

To our knowledge, this is the first study describing drinking behavior
and associated problems in the two biggest cities in China: Beijing and
Shanghai. Since the late 19703, alcohol production and consumption have
been increasing steadily concurrent with a booming economy and increased
affluence (WHO, 2003). As such, it is important to have obtained up-to-date
estimates for the frequency of drinking behavior and problems in these two
most advanced economy, metropolitan areas of China. These study estimates

serve as benchmark values for estimating future trends in the extent of

drinking and related problems in contemporary let century China.

Furthermore, as urbanization proceeds, with more and more people
migrating into cities, there may be change in drinking behaviors and
problems in the future for other Chinese cities. As such, one of the strengths of
the current study is that the source population is non-institutional household-
dwelling adults, where the major drinking-related disease burden tends to
occur. Second, although the participation levels are not perfect (~75%),
potential bias in point estimates was minimized by applying probability
weights and post stratification adjustments in estimation. The possible
clustering between individuals, as a result of the multi-stage sampling, was
also taken into account to avoid erroneous statistical inference. These

methodological refinements help promote the likelihood that these study

164

 

results are directly generalizable to the source populations under study. In
contrast to ecological level studies, the measurement was at the individual
level and was based on what individual said about themselves. As compared
to estimates based on clinical populations, this study avoids a possibility that
patients seeking clinical care are just a small portion of the entire set of cases
in population (“the tip of an iceberg”).

Nonetheless, the study does have some counterbalancing limitations.
With respect to the source population, although the study sample covers the
household-dwelling population where most drinking-related disease burden
can be found, some important subgroups were not included in this study (for
instance, people in the military, homeless people, and adolescents). Future
studies in these special populations can be used to supplement to the current
study design. The sampling frame of this study only included people who
were formally registered in a non-agricultural household. Thus the majority of
immigrants, who usually work in Beijing and Shanghai but were not officially
registered in households within these two cities, were not included.
Nonetheless, drinking patterns and problems in immigrants may bear
interesting messages about social-culturally influences on drinking.
With respect to the assessment of drinking, our major concern is that the
drinking-related variables have been measured via a retrospective, self-
report approach. Recall and reporting bias cannot be ruled out, especially
when a long time period has elapsed since the drinking behavior of interest

occurred. For this reason, the study did not make extensive use of the age of

165

onset of drinking variables. Instead, individuals were dichotomized into
“early onset” and “later onset” subgroups to constrain the recall bias,
meanwhile retaining important information. Additionally, the use of a face-to-
face interview approach precluded anonymity of the assessment. The stigma
attached to some drinking outcomes, such as clinical features of alcohol
dependence and socially maladaptive drinking, may have introduced under-
estimations in relation to the occurrence of such variables. Although it has
been shown that self-report can be a reliable and valid method of assessing
drinking behavior, this argument has not yet been confirmed in the Chinese
context (Del Boca 8: Darkes, 2003), especially for sensitive information, such
as legal problems due to drinking. With respect to the assessment of
alcohol dependence, besides concerns about the reliance on self-report,
another notable limitation of study method is “gated” approach as embedded
in the version of CIDI used in this study. That is, respondents lacking a history
of drinking-related socially maladaptive problems or hazard-laden drinking
were not asked about clinical features of alcohol dependence. Evidence from
the NESARC suggests that this “gate” may induce decrements in estimation of
lifetime occurrence and 12 month prevalence of alcohol dependence
(Degenhardt, Bohnert, 8: Anthony, 2007; D. S. Hasin 8: Grant, 2004). Though
the degree of under-estimation is unknown since there has been no
comparison using data from China, our estimates of alcohol dependence may

well serve best as starting point, possibly subject to underestimation.

166

In summary, limitations such as these may have led to under-
estirnations in cumulative occurrence and lZ-month prevalence of drinking-
related problems, especially legal problems and alcohol dependence. To a
lesser extent, there may have been under-estimation of drinking behavior
frequency.

5. l .3 Drinking practices in Beijing and Shanghai

Despite limitations mentioned above, our results are informative in that
this study presetns the first comprehensive description of drinking and
related problems in contemporary metropolitan China based upon
epidemiological samples. First, with respect to experience with alcohol
involvement, our study is the first to observe that over 80% of individuals had
the opportunity to drink alcohol. Along with results from previous studies
conducted in the metropolitan area of Wuhan city in the Hubei province, it
demonstrated that alcohol can be highly accessible in metropolitan cities of
China (Iiafang, Jiachun, Yunxia, Xiaoxia, 8: Ya, 2004; Zhang, Casswell, 8: Cai,
2008). Despite this level of accessibility of alcohol, most people only drank
occasionally and moderately (less than monthly and less than 4 drinks in a
day), which was also found in the Wuhan city study; and the median number
of drinks during a typical drinking day is similar to the Wuhan study as well
(Zhang, Casswell, 8: Cai, 2008). Compared with other countries, people in
these two Chinese metropolitan cities have been found to have a higher level
of lifetime experience with drinking than people in several countries in the

Middle-East and South Asia (e.g. India), and lower than people in America,

167

New Zealand, and European countries (Degenhardt et al., 2008; WHO, 2005).
With respect to the age of first trying alcohol and the initiation of MTM
drinking, our results pointed out peak values during the late teenage years,
with respect to experiences with alcohol. This result is strikingly similar with
other countries participating in the WMHS (Degenhardt et al., 2008).
Adolescence is a developmental period for sustaining brain maturation, and
teen onset of drinking may signed greater susceptibility to developing
drinking problems. The implications for tinting of effective prevention and
intervention programs are obvious (e.g. see O'Brien 8: Anthony, 2005)).
5.1 .4 Drinking problems in Beijing and Shanghai

With respect to drinking-related problems, this study’s finding can be
compared with results from a 1993 study conducted in three provinces in
China, which applied the DSM-III-R criteria. In this comparison, the
occurrence of drinking-related problems in Beijing is observed to be higher
than that in all three outlying provinces, and the occurrence in Shanghai is
intermediate (Wei et al., 1995). This study’s results are not directly
comparable with more recent studies conducted in other areas of China, since
the assessments are distinct from one another (Hao et al., 2004; Zhang,
Casswell, 8: Cai, 2008). Compared with the occurrence of drinking-related
problems in other countries, the occurrence is generally lower than that of
Western countries, slightly lower or equal to that of some other Asian
countries (e.g. Japan and Korea), while higher than that of Nigeria (Gureje,

Lasebikan, Kola, 8: Makanjuola, 2006; Kawakami, Shimizu, Haratani, Iwata, 8:

168

 

Kitamura, 2004; Medina-Mora, Borges, Benjet, Lara, 8: Berglund, 2007; j. T.
Park, Kim, 8: Jhun, 2008). For example, using the same tools of assessment,
surveys in the US and several countries in Europe all showed a lifetime
occurrence well above 10% (CPES, 2001; Rehm, Room, van den Brink, 8:
Jacobi, 2005). One possible explanation is that DSM-IV and ICD—10 criteria are
largely based on clinical observations in Western countries. The sensitivity of
questions might not be as good in the Chinese culture. Nonetheless, the
clinical reappraisal study showed fairly good validity of CIDI in assessing

drinking-related problems (Huang et al., 2008). Thus, the much lower

 

occurrence is not likely to be completely due to the validity of the assessment.

Difference of this type also can be due to socio-cultural and biological
variations across populations. To illustrate with a contextual example,
according to data from the Gender, Alcohol, and Culture: an International
Study (GENACIS), in some European countries, the bar or pub is an especially
common place to drink alcohol; such that males might engage in solitary
drinking (without a companion) on 10-20% of drinking days during a year, as
compared to females on 55-15% of drinking days (Bloomfield et al., 2005). With
respect to China, the 1993 study found that a common reason for drinking is
“to celebrate”. On most drinking occasions, there is a meal and in general,
Chinese drinkers tend to avoid being drunk in front of friends and business
partners (Hao, Chen, 8: Su, 2005; Wei et al., 1995).

As for biological variations, our best example for the Chinese context

involves variations in proteins involved in alcohol metabolizing pathways (e.g.

169

the allele frequencies of genes encoding ADH and ALDH). To illustrate, the
ADH2*1 allele presents in over 95% of many European heritage populations
By comparison, it is found in only 32% of Chinese Han. ALDH2*2 is virtually
absent in the European heritage population and is present in about 20% of
Chinese (Goedde et al., 1992; Y. C. Shen et al., 1997; Thomasson et al., 1991).
These mutations in ADH and ALDH genotypes result in accumulation of
acetaldehyde, which induces the “flushing effect” that functions to punish and
dampen future drinking behavior in many drinkers with the flushing
phenotype. To the extent that culture does not overcome this pharmaco-
genetic process, these mutations may be serving to protect many Chinese
people from excessive drinking. In deed, in many European countries and the
US, people drink alcohol in a higher frequency and larger quantities
(Bloomfield et al., 2005; CPES, 2001). The GENACIS found that considerable
proportions of people drank more than five drinks monthly in selected
European countries (20—50% males, 5-20% females). Genetic variations
mentioned above may play a role behind the higher level of alcohol
consumption among Europeans and US. In summary, the interplay of these
biological and cultural factors results in the lower occurrence of drinking
problems in these two metropolitan cities in China.

Although compared with values observed in metropolitan areas of '
other countries, drinking is not as common in Chinese metropolitan cities,
riskier drinking behavior and drinking problems definitely exists. For

example, 7% people had a history of heavier drinking, which is similar to the

170

results from previous studies (Hao et al., 2004; World Health Organization,
1999), five percent people had a history of socially maladaptive problems
related to drinking, and two percent people have had at least one clinical
feature(s) of alcohol dependence. Given the large population base, these
percentages easily translate into considerable disease burden, not only for
individuals, but for their families and societies.

Some observations about the future may be in order. As compared to
some European countries and the USA, there is less commercial alcohol
beverage market prevention (Degenhardt et al., 2008), and there is still much
latent demand for commercial alcohol in China. With increasing affluence, the
market forces may address this demand, in processes described by Room et
a1. (Room, Schmidt, Rehm, 8: Makela, 2008) and it is very likely that alcohol
consumption will grow. There is some evidence that alcohol production and
consumption in China are climbing (National Bureau of Statistics of China,
2008) and we might project that drinking-related problems will climb as well.
Along with a growing number of automobiles in the cities of China, results
from this study add more reason to call for preventions and harm-reduction
initiatives in China. Looking to the future, the teenage years have been the
key interval for initiation of Chinese drinking, as is true elsewhere
(Degenhardt et al., 2008). Moreover, a large proportion of Chinese drinkers
in this research were found to have had their first socially maladaptive
problem or dependence clinical feature early in that second decade of life.

Hence, to be most affective, prevention and intervention programs should be

171

designed for young people (from teenage to early 20’s) in addition to more
general regulation prevention effort (e.g. control through taxation, as
suggested by the work of Holder, Holder, 2007).

One interesting observation from this study is that a larger proportion
of people in Beijing had drinking-related problems compared with people in
Shanghai. This greater occurrence of drinking problems in Beijing is not
explained by the higher drinking frequency. The 1993 survey of three sites in
China also found significant geographical variations in drinking problems
(Wei et al., 1995). However, due to the lack of assessment of drinking context
and biological factors, as well as the self-report nature of the study instrument,
we are much restricted to any explanation beyond pure observation.

5.2 Subgroup variation with respect to beverage alcohol involvement.
5.2. 1 Summary of results

With respect to time-invariant variables assessed in this dissertation,
sex and age groups, it was found that being male is associated with higher
occurrence of both drinking and drinking-related problems. The strengths of
associations for drinking-related problems are much higher than that in
drinking practices. Compared to people in the oldest age group, those in
younger groups are more likely to be involved in alcohol drinking.
Individuals in the youngest group are especially more likely to be involved
earlier in their life.

With respect to time-variant variables assessed in this dissertation,

marital status, education, and personal income, subgroup variations were

172

found as well. Individuals who were no longer married are more likely to have
a history of heavier drinking, but not recent heavier drinking; individuals with
higher income are more likely to be involved in drinking, but no more likely
to experience drinking-related problems. Individuals who had a job at the
time of assessment were more likely to be involved in drinking, as well as
experience drinking-related problems; however they were no more likely to
initiate drinking early or experience drinking-related problems early.

5.2.2 Strengths and limitations

The main strength of this study is, as stated above, the results are
directly applicable to the source population. Not like studies conducted in
clinical settings, the source population in this study is household dwelling
adults in Beijing and Shanghai. Due to characteristics associated with
treatment seeking behavior, e.g. sex, age, severity of the condition, education,
and so on, it is precluded the generalization of results from clinical studies to
the entire patient population. However, considerable disease burden is from
people who never sought clinical aid. Thus, it is critical to have estimates from
studies which can be generalized to the population where the major disease
burden comes from.

Of major concern of limitations is the “gated” approach in assessing
alcohol dependence, as mentioned above. If the “gate” differentially filtered
individuals in relation to their characteristics, e.g. sex, age, martial status, and
so on, the estimates of subgroup variation in alcohol dependence might have

been biased. Degenhadt et al. compared the patterns of associations with

173

respect to sex and age groups, and did not find appreciable differences in
estimates for alcohol dependence (Degenhardt, Bohnert, 8: Anthony, 2007).
Nevertheless, by comparing percentage loss of cases of alcohol dependence
across different subgroups, Hasin et a1. did show that disproportionally more
cases were missed in females and minority groups (African American, and
Hispanics), although no statistical test was provided. We are not aware of any
such comparisons with respect to marital status, education level, income level,
and employment status; neither in the sample of Chinese. Thus, it remains
unclear if or how much of the estimates for the occurrence of alcohol
dependence is biased because of the “gating” procedure. Nevertheless,
alcohol dependence is only one of the various indicators for “riskier
drinking” in this study. By assessing the pattern of associations between
assessed characteristics and multiple outcomes, we are able to get the profile
of these associations.

One thing merits mention is that under this study aim, our goal is not to
infer causal relationship, but rather to present the differential distribution of
drinking practices and problems in each of the subgroups. By doing this, we
intend to prioritize prevention and intervention programs and to provide
basis for designing effective prevention and intervention programs according
to individual characteristics.

5.2.3 Subgroup variations in drinking outcomes

5.2.3.1. Sex and age

174

 

Consistent with findings from previous studies around the world,
excessive drinking and problems are found in males and younger age groups.
(Degenhardt et al., 2008; D. S. Hasin, Stinson, Ogburn, 8: Grant, 2007; Higuchi,
Parrish, Dufour, Towle, 8: Harford, I994; Kessler et al., 1994; Kim et al., 2008;
Naimi et al., 2003; J. T. Park, Kim, 8:]hun, 2008; Serdula, Brewer, Gillespie,
Denny, 8: Mokdad, 2004; WHO, 2005). This male-female gap in drinking ;,
opportunity only exists in the two older age groups. In the two youngest [
groups, there was no sex-difference in opportunities to drink. Among all

outcomes in this study, opportunity to drink alcohol is the only one that is

 

solely affected by socio-cultural factors. This result may serve as a reflection
of the diminishing stigma attached to female drinking as the reform of the
Chinese society.

Results are consistent with previous studies in China in that males are
more likely to be involved with drinking, and males are much more likely to
have a history of riskier drinking and drinking-related problems (Hao et al.,
2004; Jiafang, Jiachun, Yunxia, Xiaoxia, 8: Ya, 2004; Zhang, Casswell, 8: Cai,
2008; X. Zhou et al., 2006). Moreover, our results found that although the male-
female difference was found in each of the drinking outcomes, it is rather
small in early stages of alcohol involvement (e.g. opportunity, ever tried
alcohol) and increases dramatically in riskier drinking behavior and drinking-
related problems. Drinking-related problems are mainly male phenomenon

in these two metropolitan cities in China. Such robust male-female differences

175

were not likely to be completely accounted for by either differential report or
differential survival.

We found higher alcohol involvement, as well as drinking-related
problems, in younger age groups, which is in agreement with previous
studies conducted in other areas of China (Iiafang, Jiachun, Yunxia, Xiaoxia, 8:
Ya, 2004; Zhang, Casswell, 8: Cai, 2008; H. Zhou et al., 2003). The higher level
of alcohol involvement in younger age groups may well reflect the increasing
alcohol production and consumption in National reports (National Bureau of
Statistics of China, 2008). There are several alternative interpretations of the
observed variations in drinking-related problems across age groups. The first
is that it might be due to differential recall. However, we argue that since
socially maladaptive problems and dependence are quite distinct
experiences during one’s lifetime, it is not likely to be forgotten. Second, it is
also not likely to be due to differential report because the same pattern of
variation was seen in less stigmatized outcomes, such as opportunity to drink
and ever tried alcohol. Previous studies have shown that self-report of
drinking is fairly reliable (Cumming 8: Klineberg, 1994; Lee, Whittemore, 8:
Lung, 1992; S. Liu et al., 1996). One remaining possibility is that drinking-
related problem affects mortality. However, previous studies do not support it.
The study by Hao, et al. found that an individual’s health status is not
associated with drinking behavior (Hao et al., 2004); follow-up studies in the
US failed to show convincing evidence of elevated mortality in heavy drinkers

and people with alcohol dependence (Dawson, 2000; Vaillant, 1996, 2003).

176

One study in Shanghai showed that low to moderate alcohol consumption is
associated with lower mortality in middle-aged Chinese (Yuan, Ross, Gao,
Henderson, 8: Yu, 1997). Furthermore, the same pattern was found in early
stages of alcohol involvement (e.g. opportunity to drink, ever tried alcohol),
which is not very likely to be much associated with mortality. Thus, elevated
mortality in alcohol consumers is not likely to account for the consistent and
robust variations across age groups.

One thing merits attention is that drinking involvement in the youngest
group is more likely to be “right censored”, which means by the time of the
assessment they had not yet experienced relevant drinking outcomes, but
they might later in life. Given that the peak of initiation of MTM drinking is
around 20 years old, the lower occurrence of alcohol involvement in the
youngest group, compared with the adjacent group, is likely to reflect this
“right censorship”. With this in mind, it is striking that the occurrence of
“early MTM drinking” increases with age groups going from older to younger
(5% in the oldest group and almost 30% in the youngest group). Considering
the right censorship, the occurrence of early involvement with drinking in the
youngest group is likely to be higher u the same survey was to be taken later
when they all reached 20 years old. Early initiation of drinking is associated
with higher risk of heavy drinking and alcohol dependence (Caamano-Isorna,
Corral, Parada, 8: Cadaveira, 2008; Dawson, Goldstein, Patricia Chou, June
Ruan, 8: Grant, 2008). Due to the cross-national design of this study, we are not

able to tease age, cohort, and period effect apart beyond the observed

177

 

variations across age groups (Holford, 1991). Nevertheless, these results warn
us of the likely increasing disease burden of alcohol drinking if the young
people are not educated about responsible drinking.
5.2.3.2. Other variables

With respect to drinking practices, our results agree with some
previous studies in the Huaihua area of China and the US national comorbidity
survey replication (NCSR) in that people who are “married”, “working”, and
in “higher income levels” are more likely to be involved in recent drinking,

and education is not extensively associated with recent drinking (Degenhardt,

 

Chiu, Sampson, Kessler, 8: Anthony, 2007; X. Zhou et al., 2006). With respect to I‘
drinking-related problems, our results are quite different from previous
studies in the Wuhan metropolitan area (I iafang, Jiachun, Yunxia, Xiaoxia, 8:
Ya, 2004) in terms of personal income level. In the current study, no
association between income level and drinking problems was found, while
the Wuhan study found higher income was associated with higher likelihood
of drinking problems after adjusting forseveral other covariates (no
unadjusted OR were given). Several alternative reasons may explain the
difference. First, the study population is different. Second, in this study, we
held only sex and age constant. In the Wuhan study, many other covariates,
such as fellow drinking, parental drinking, attitudes about drinking, were
included in the model as well. In the context of exploring subgroup variation,
the inclusion of these variables made results difficult to explain and made

direct comparisons of results impossible. In the context of exploring potential

178

cause, some covariates are more likely to be endogenous rather than
exogenous due to the cross-sectional nature of the study. Thus, the causal
inference for any of the variables is not clear either. Here, we argue that our
results are more relevant in terms of exploring subgroup variation and
providing basis for intervention programs in that we held only sex and age,
which are exogenous to outcomes, constant. By doing this, we seek the
possible subgroup variation, which is not due to the differential composition
in terms of sex and age.

Our results indicate that drinking and related problems are not evenly
distributed in subgroups in the population. Certain subgroups of people
should be priorities for intervention and harm-reduction programs. Special
effort should be made to target at drinking problems in males, younger age
groups. Additional to sex and age, priorities should be given to people who
are working. Policies and regulations on work place drinking may need to be
created and emphasized.

5.3. The association between childhood physical punishment and
drinking and related problems in order to shed light on suspected causes
of drinking-related problems

5.3.1 Summary of results

Robust associations between CPP and negative drinking outcomes are
found. This association is statistically independent of parental drinking
problems, parental mental disturbances, and conduct problems in childhood.

After taking possible endogeneity into account, this association is still robust.

179

 

Variations in association were found across drinking outcomes and across
clinical features of AUD as defined by DSM-IV and ICD-10. Stronger
associations were found between CPP and “early onset of drinking-related
problems” as compared to some other negative drinking outcomes. Stronger
associations were found between CPP and “strong desire” and “difficulty
cutting down” as compared to “drink despite physical/ mental problems". .
According to our results, the most important stage to link CPP to drinking
problems is after the initiation of MTM drinking.

5.3.2 Strengths and limitations

 

The major contributions of this study are as follows. First, it provides
evidence of a possible causal relationship between childhood physical
punishment and adverse drinking outcomes in the Chinese context. Most of
previous studies are based on Western populations. Nonetheless, CPP is
culturally sensitive. The influence of CPP may vary from society to society.
Our study provided initial basis for future studies of CPP as a potential cause
of drinking-related problems. Second, our estimates are based upon a
community sample. Possible biases in estimation were minimized by efforts to
adjust for clustering and selection probability, as well as non-response
patterns. Thus, results are directly applicable to the household-dwelling
adults living in the two Chinese cities. Third, using a novel statistical method,
we successfully controlled for endogeneity bias and found robust association
after taking the endogeneity into consideration. Fourth, previous studies in

special populations found that the association between CPP or CPA and

180

drinking outcomes only presents in females (Simpson 8: Miller, 2002). This
study found that from a population perspective, the estimated effect of CPP to
negative drinking outcomes is robust in males as well.

The study does suffer from some limitations. Of our major concern is
the assessment of CPP. The assessment of CPP is “one question for multiple

behaviors”. Therefore, we are not able to tease out which behavior(s) account

 

F
for the association. However, these assessed behaviors tend to occur together
as an indication of rearing style of the parents. All these forms of physical .
punishment pose stress throughout the childhood and possibly influence the
devel°Pmem trajectory of the child (D. B. Clark, 2004). One thing merits I?"

attention is that more severe forms of physical punishment, such as biting,
being burnt, being scalded, being hung, etc. were not assessed in this study
due to the potentially emotional upset to the respondent. These severe forms
of physical punishment may or may not present in individuals under study.
Thus, it is possible that estimates were driven by these severe forms of CPA,
but not CPP, per se. more studies are needed to clarify this possibility. Also of
concern is the assessment of parental drinking problems. The assessment of
parental drinking problems is based on the self-report from the respondent
instead of by their parents. No clear definition of “drinking problems” was
given in the question. Thus, the assessment may be subjective and fairly
coarse. Several studies have probed into this question. These studies
suggested that information about parental drinking problems collected from

offspring is pertinent and accurate, and it serves as a valid method of

181

assessing parental problems when the alternative is unfeasible (Prescott et al.,
2005; Rhea, Nagoshi, 8: Wilson, 1993; Sher 8: Descutner, 1986). It is not clear,
however, if this conclusion holds to be true in the Chinese context.
5.3.3 Possible causal inference

Despite these limitations, our findings are of interest. To our knowledge,

this is the first study to show a possible causal relationship between CPP and

 

riskier drinking and drinking problems in the Chinese culture. The results
correspond to some previous findings in US and Canada populations showing

a moderate association between CPA and drinking problems (Holmes 8:

 

Robins, 1987, 1988; Kessler, Davis, 8: Kendler, I997; MacMillan et al., 1999;
MacMillan et al., 2001). Many previous studies found that after controlling for
parental drinking problems, there is no association between CPA and
drinking outcomes (Hughes, Johnson, Wilsnack, 8: Szalacha, 2007; Koss et al.,
2003; Libby et al., 2004; Mullings, Hartley, 8: Marquart, 2004; Young, Hansen,
Gibson, 8: Ryan, 2006), while one small-scaled follow-up study found a robust
association between CPP and binge drinking in 113 African American female
child abuse victims, after holding parental drinking problems constant

(J asinski, Williams, 8: Siegel, 2000). A cross-sectional study also suggested an
association between childhood adversities and AUD holding parental
drinking problems constant; however, no specific estimate was provided for
physical abuse in the study (Carrigan 8: Randall, 2003). Nevertheless, in these
studies, as we described in the background section, the possibility that CPP

might be an endogenous variable of parental drinking problems was not

182

taken into account. The endogenous bias might have resulted in inconsistent
estimates; ignoring that it may have led to erroneous inference. Additionally,
these studies were conducted in special groups of population, such as marine
recruits, lesbians, prisoners, and people attending primary care. Thus, results
from these studies cannot be applied to the general population. Our study
showed from a population perspective that after controlling for parental
drinking problems, the possible causal relationship between CPP and
drinking problem persists. Additionally, we found that the estimated effect

cannot be attributed to parental mental disturbances (depressive mood,

 

anxiety, and suicidal attempt) either.

Due to the cross-sectional and observational nature, this study is more
liable to potential biases, such as differential recall and survival bias, which
makes the inference of causality more difficult. First, it is possible that people
who suffer from drinking problems may search their memory deeper for
experiences of CPP. However, it is not supported by previous studies, which
have shown good to excellent validity and reliability of assessment of CPP
(Bremner, Bolus, 8: Mayer, 2007; Walsh, Macmillan, Trocme, Jamieson, 8:
Boyle, 2008). The finding has been replicated in cocaine dependent
individuals (Kopnisky 8: Hyman, 2002). We further offset this problem by
grouping people who answered “often” and “sometimes” together, and
“rarely" and “never” together. In this way, the CPP variable represents
chronic stress during the childhood, which is reasonably easy to recall.

Second, individuals who suffered from CPP may be less likely to survive,

183

especially from severe forms of abuse (Arias, MacDorman, Strobino, 8: Guyer,
2003). As such, less people with CPP experience will be captured in the cross-
sectional study. This will only bias estimates toward the null since previous

studies have found that a severe form of abuse is highly associated with heavy
drinking and alcohol dependence (Simpson 8: Miller, 2002; Widom, White,

Czaja, 8: Marmorstein, 2007). Thus, our estimates may be an underestimate of F
the real coefficient. Third, reverse causality is a usual concern for observed

associations from retrospective or concurrent studies. It is possible that the

 

occurrence of riskier drinking and related problems could be a manifestation
of a continuity of childhood behavioral problems (including early drinking).
These problems may be the reason for parents to physically punish offspring.
To explore this question, we performed two series of analysis. In the first one,
we restricted our analysis in the subsample of individuals who initiated MTM
drinking after 16 years old. Estimates for CPP remained robust. In the second
series of analysis, we adjusted for the history of childhood conduct problem.
The estimated effect of CPP remains robust as well. Although it cannot be
ruled out that the robustness is due to other unmeasured confounders, our
results provide a solid basis for further inspections.

Summarized from previous studies, Simpson and Miller concluded that
there is evidence of the possible causal relationship between CPA and
drinking problems in females and boys, but not in adult males (Simpson 8:
Miller, 2002). In fact, this issue has been rarely studied in the general adult

male population. The relationship between CPP and negative drinking

184

outcomes may be different in special population, e.g. prisoners, court recruits,
and clinical patients, from that in the general population. Drinking and related
problems are complex human behaviors and conditions with numerous
competing causes. In special populations, individuals who were not exposed
to CPP/CPA may be more likely to experience other possible causes of AUD
(e.g. lack of parental supervision, emotional distress, etc.), compared to the “f
entire population of non-CPP exposed individuals; and eventually develop
drinking problems. Thus, we argue that the association in the majority of adult

the male population is still largely unknown. Our study has provided first-

 

hand evidence of the possible causal relationship between CPP and riskier
drinking and drinking problems in adult males from a population perspective.

Using the novel technique of GEE, we are able to take the inter-
correlation of outcomes into consideration. This technique granted us
statistical efficiency and ensured accurate statistical comparison of estimated
strengths of associations. Results are interesting in that CPP is more strongly
associated with “early onset” of drinking problems, compared with “early
try” and “early MTM drinking”. It suggested that victims of CPP progress
more rapidly than people with no history of CPP. Previous studies have shown
that victims of CPA are more likely to initiate drinking earlier (Bensley,
Spieker, Van Eenwyk, 8: Schoder, 1999; Dube et al., 2006; Hamburger, Leeb,
8: Swahn, 2008; Rothman, DeJong, Palfai, 8: Saitz, 2008). Our results suggested
that above the early involvement of drinking (early try and early MTM

drinking). CPP may pose additional risk to early onset of drinking problems.

185

It is in line with the maturation theory of early onset of substance use
disorders (Blackson 8: Tarter, 1994). According to the maturation theory,
childhood to early adulthood is the key period for brain maturation. Long-
term exposure to environmental stress, e.g. CPP, can push the maturation
process to deviate toward non-normality through behavioral epigenetic
mechanisms. The study by Rothman et al. found that compared with drinkers 'f
with no CPP history, victims of CPP are more likely to use drinking as a tool to
cope (Rothman, DeJong, Palfai, 8: Saitz, 2008). This motive of drinking may get

victims of CPP into unhealthy drinking behavior, e.g. binge drinking, which

 

leads to social interpersonal problems, and alcohol dependence sooner.
Early onset drinking problems is associated with higher level of severity and
worse treatment outcome (1. Brown, Babor, Litt, 8: Kranzler, 1994). Along with
another finding that CPP was more strongly associated with “early onset” of
drinking problems than those with “lifetime occurrence” of drinking
problems, our results highlighted that one focus of prevention strategies may
be placed on fostering coping skills in CPP victims from early on.

Drinking problems in this study covered a broad range of problems,
from drinking-related aggression to neuro-adaptation to compulsive drinking.
Different mechanisms are involved in different clinical features (Koob, 2006; J.
Liu et al., 2006). Previous animal studies have shown that lack of maternal care
or adverse environment leads to changes in the GABA—A receptor, which is
highly involved with the psychoactive effect of ethanol (Caldji, Diorio, 8:

Meaney, 2003), as well as changes in hippocampus and prefrontal cortex

186

 

(Bremner etal., 1997; R. G. Heath, 1972; Teicher, Tomoda, 8: Andersen, 2006).
When these drinking problems were grouped into two categories (socially
maladaptive problems and alcohol dependence) it is not clear which
problems are actually associated with CPP, and furthermore, whether the
magnitudes of associations vary across manifestations. Thus, important
messages may be buried when using one variable to represent different
manifestations. To our knowledge, there has been a vacancy in literature
regarding this issue. Using the GEE technique, we not only generated
manifestation-specific estimates, but also compared the magnitude across
estimates. Heterogeneous associations in manifestations of alcohol
dependence were found. The smallest OR (non-significant) was found in
“drink despite physical/ mental problems”. Compared with the association in
“drink despite physical/mental problems”, (which is a clinical feature of
compulsive drinking) stronger associations were found in “irresistible desire"
and “difficulty in cutting down”, both of which are clinical features of loss of
control over alcohol. One possible explanation is that at the same level of
alcohol dependence, individuals who suffered from CPP may be more likely
to report these two clinical features. However, results from the MIMIC model
did not support this hypothesis: at the same level of problematic drinking,
CPP victims are no more likely to report “irresistible desire” and “difficulty
cutting down” than other problems. We hope the observed heterogeneity of

associations can spur novel hypotheses and promote future studies into this

issue, especially the mechanisms behind loss of control over alcohol.

187

 

With respect to stages of alcohol involvement, our results demonstrated
that the most important stage is arguably the occurrence of drinking problems
after the initiation of MTM drinking rather than the initiation of MTM drinking.
However, it is still a rather broad stage involving many stages of progression.
Due to the cross-sectional design of this study and self-recall of assessment,
the ability of inference of temporality is limited. Our results serve as a call for !'
future studies on this particular stage.
5.4. Future research

First, the observation that people in Beijing had more problems with .

 

drinking deserves further inspection. A study conducted in four Scottish towns
found geographical variations in drinking patterns, and these variations in
drinking patterns were closely associated with alcohol-related crime,
morbidity, and mortality (Plant 8: Pirie, 1979). Given these findings, studies
about drinking culture with respect to drinking context, people’s attitude
about social and excessive drinking may bear important messages for
designing prevention and intervention strategies. Besides, a previous study
showed that the ADH and ALDH genotype and allele frequencies are
differentially distributed across different ethnic groups within China (Y. C.
Shen et al., 1997). There has been no study exploring geographic variations of
these allele frequencies within the Han ethnicity. Here, we argue that since
these genotypes directly affect individual drinking behavior, it is of interest to
integrate these biological measurements into socio-cultural factors to explore

the geographic variations.

188

Second, we found robust subgroup variations with respect to sex, age,
marital status, employment status, and income level. Among these variables,
sex and age are possible determinants of drinking and related problems
because both are exogenous to drinking. Further investigation into these two
variables may provide insights into the etiology and drinking-related
problems. For example, studies in Western countries have shown that the 0—
male-female difference in drinking problems may be due to higher drinking
level in males (Ely, Hardy, Longford, 8: Wadsworth, 1999; Miller, Plant, 8: Plant,

2005). Whether it holds in the Chinese drinking culture will complement

 

current knowledge.

It is interesting that a higher level of personal income is associated with
higher drinking involvement but not drinking problems. The lack of
association between income and drinking problems was also shown in the
NESARC study (D. S. Hasin, Stinson, Ogburn, 8: Grant, 2007). Several possible
explanations can be proposed. For example, the phenomenon may suggest
that people with higher income have more opportunities to drink, e.g.
business—related social activities given that “toasting” has been a popular way
of building relationships in modern China (Cochrane, Chen, Conigrave, 8:
Hao, 2003; Hao, Chen, 8: Su, 2005). It might reflect the infiltration of Western
drinking culture as people in the higher income level started to embrace
drinking into their daily life. It also might suggest that people in the higher

income level are more aware of health benefits of modest drinking. To further

189

probe the reason behind this observation bears useful information for future
education strategies about healthier drinking style.

Third, children from dysfunctional families, including but not limited to
alcoholic families, suffer from various childhood adversities, such as family
tension or parent divorce, witnessing domestic violence, child neglect, and so
on. These clusters of childhood adversities may influence their
neurodevelopment and lead to internalizing or externalizing problems, which
in turn put them into a higher risk of adulthood physical and mental outcomes
(Clemmons, DiLillo, Martinez, DeGue, 8: Jeffcott, 2003; Edwards 8: Gross, 197 6;

Ruchkin, Gilliam, 8: Mayes, 2008; Tremblay et al., 2004). In this context, it is of
interest and importance to tease out the effect of more modifiable factors, e.g.
CPP, from less modifiable factors such as parent divorce or family tension.

Additionally, people with drinking problems vary in severity. Disease
burden is higher in those who suffer from multiple problems than those with
only one problem. Early observations by Tarter, et a]. suggested that drinkers
in higher severity level differ from drinkers in lower severity level in
childhood brain function (Blackson 8: Tarter, 1994). There has been evidence
that CPP is associated with brain function. In this context, it is of interest to
PrObe if CPP is associated with the occurrence of drinking problems only or
With severity as well.

With respect to inferring temporality, longitudinal studies with small

attrition are needed. Nevertheless, observational studies are low in power to

infer causality. Experimental studies are needed to establish the causal

190

relationship between CPP and drinking problems. Previous studies have

found that both genetic and environmental factors influence violent behavior
toward children, including physical abuse (DiLalla 8: Gottesman, 1991; Widom,
1989). In this context, experimental studies are especially of interest to clarify
in what degree physical abuse is modifiable. Our study found stronger
association between CPP and loss of control over drinking. We also found
stronger association between CPP and “early onset of problems”. Survival
analysis from follow-up data will be essential to tease out if the stronger
association is due to earlier onset (longer progression) or due to the influence
of CPP on specific functions of the brain. Last but not the least, the

incorporation of genetic and epigenetic methodology is essential to study the
etiology of drinking-related problems as the developmental process is a

complex network of genetic and environmental factors through multiple

pathways (Blackson 8: Tarter, 1994).

191

Appendix

Questions about drinking

*SUl. The next questions are about your use of alcoholic beverages, including
beer, wine, wine coolers, and hard liquor like vodka, gin or whiskey. How
old were you the very fir_st time you ever drank an alcoholic beverage?

YEARS OLD
(IF VOL): "NEVER" .............. 997 GO TO *SU39
DON’T KNOW ...................... 998
REFUSED .............................. 999

 

* SU2. IF R CAN READ: (RB, PG 17) Please use the table on page 17 in your
booklet as a guide in answering the next questions. How old were you
when you first started drinking at least 12 drinks in a year?

IF R CANNOT READ: When I use the word "drink" in the next questions,
I mean either a glass of wine, a can or bottle of beer, or a shot or jigger of
liquor either alone or in a mixed drink. How old were you when you first
started drinking at least 12 drinks in a year?
IF “ALL MY LIFE” OR “AS LONG AS I CAN REMEMBER,” PROBE:
Was it before your teens?

IF NO/DK, PROBE: Was it before your twenties?

 

YEARS OLD
BEFORE TEENS ................... 12
BEFORE 20$ .......................... 19
NOT BEFORE 203 ................. 20
(IF VOL): "NEVER" .............. 997 GO TO *SU39
DON’T KNOW ...................... 998
REFUSED .............................. 999

 

 

al‘SIJ3. (RB, PG 17) (Look at page 17 in your booklet.) Think about the past 12
months. In the past 12 months, how often did you usually have at least one
drink —- nearly every day, three to four days a week, one to two days a week,
one to three days a month, or less than once a month?

NEARLY EVERY DAY ....... 1
3 - 4 DAYS PER WEEK ........ 2
l - 2 DAYS PER WEEK ........ 3
1- 3 DAYS PER MONTH 4

192

LESS THAN ONCE A MONTH (INCLUDING NEVER

DRINK) .................. 5 GO TO *SU8
DON’T KNOW ...................... 8 GO TO *sus
REFUSED .............................. 9 GO TO *SU8

 

*SU4. (RB, PG 17) (Looking at page 17 in your booklet,) On the days

you drank in the past 12 months, about how many drinks did you usually have

per day?
NUMBER OF DRINKS PER DAY
DON’T KNOW .......... 998
REFUSED .................. 999

*SUS. Was there ever a year in your life when you drank more than you did in the
past 12 months?

YES ......................................... 1 GO TO *SU8
NO ........................................... 5
DON'T KNOW ....................... 8
REFUSED .............................. 9

 

*SU6. INTERVIEWER CHECKPOINT: (SEE *SU3)

*SU3 EQUALS ‘4’ ............................................................ 1
ALL OTHERS .................................................................... 2 GO TO SU12

 

 

*SU7, INTERVIEWER CHECKPOINT: (SEE *SU4)

*SU4 IS EQUALS ‘3’ OR MORE .................................... 1 GO TO SU12
ALL OTHERS .................................................................... 2 GO TO SU39

 

 

*SU8. Think about the years in your life when you drank mgsj. During those
years, how often did you usually have at least one drink — nearly every day,
three to four days a week, one to two days a week, one to three days a
month, or Less than once a month?

NEARLY EVERY DAY ................................................... l
3 - 4 DAYS PER WEEK .................................................... 2
1 - 2 DAYS PER WEEK .................................................... 3
1 - 3 DAYS PER MONTH ................................................ 4

193

LESS THAN ONCE A MONTH ...................................... 5 GO TO *SU39

DON’T KNOW .................................................................. 8 GO TO *SU39
REFUSED .......................................................................... 9 GO TO *SU39

 

*SU9. And on the days you drank during those years, about how many drinks
would you usually have per day?

 

 

NUMBER OF DRINKS PER DAY
DON’T KNOW .......... 998
REFUSED .................. 999
*SUl 0. INTERVIEWER CHECKPOINT: (SEE *SU8)
*SU8 EQUALS ‘4’ ......................................................................... 1
ALL OTHERS ................................................................................ 2 GO TO SU12
*SUI 1. INTERVIEWER CHECKPOINT: (SEE *SU9)
*SU9 EQUALS ‘3’ OR MORE ......................................... 1
ALL OTHERS .................................................................... 2 GO TO *SU39

194

 

.8000 20 003 0.5.020 .0 03.08. H0.0.0.8.. .0000
0.030800. 0.80.0 .o
8.3.... 0.8.... ..o 000000.. 02.0.. 0... .0.. 00.580 .o 00.0080 00.8 :0... 0.08 00.. 0.03 .0N. Dm...

 

0..... 00m 0.000 .5.. 0.0..?
0.8.88.0 8 0.80.... 008.6800 30.. 005000.. 0.0000 .39. 0080000000 ”mmémm >mvc
000.0 0.85.80 5 680008 0 w8.0.0..o
$83.0 6.0.00.0. 0 w...0.. 82.3 0.0.88.0 .0... .8... .00. 0.58 so» 0.2.3 082.03%
8 .o..oo.0 :.o 00803.8 0... .00.... 800.0 0.03 :0.» 80...... 0.... .59. 8 008.. 0.0... 0.03 .0305...

 

.0058. Ev. oz.
00800.. 000... 0...? 080.08.. 000300 .. swoon. 00>0 0.8.0 o. 0.8.800 50.. 0.0. .3. Dm...

 

0N
03m...

OD

UN
03m...

00

UN
03m...
Oh.
CU

A8080 .8 0080... 00.80:. 8.3 080.00... 000000 ”magma. .0va
00.00.5350 .0 0.2.5908 .0080... 58.0.. .50.. 8.3 08030... 08000.0. .0 052.00
.0....0 .o 0.808.030 000000 wcflctp .50.. 8003 0... .00.. 8 08.. 0 .0>0 0.0... 003 .0N. Dm...

 

0.8.5 .00.. 8.3 00.00.08. ”magma. >mv©

0080.. .0 .0 6.0.00 .8 ..oo..o0 .0 00....5.0..o..00. .o v.83 .50.. .....>> 00.00.08.

0.80500... .0>0 was: mafia .o $5.8... .50.. 0.2.3 0.... 8o» 8 08.. 0 .0>0 0.0... 00.5
..0.... .0828... .0 003000.. 00.. 0>00 .008 00.0 080.00... 52.0 0.0 0.8.0.0000 0.0.. 0.: .N. Dw0

 

 

.0.

 

.0.
v8

 

 

.0.
oz

 

c.
00.0

 

 

. .050. 0,0 00 02.. .0. 022.0000 00.5.0 0.00 05.00.00.020 .00. .0000
020.5000 0..: 0000.0... 00 0.00.0... 0 0. .2.”..00052. 00.003082.

 

195

 

00.80...
o. 00

00080.8 80.. 80... 0.08.0.0 0.80.0 ”mm/09.0... >mv.
0000888.
80.. 80... 0.08 .0. 0 .80.... 80.. 80...5 .0 8.80.805 80.. $00.80.. a.a
80.. ..w8o... 80>0 m8.0.8..0 0080.0 80.. 80...5 008.. 0>0.. 80.. 0.0 .0... 3m“.

 

000... 00... 080.80... w8>0.. 80.. ..000. ”mm/DEA. >m.v.
0000... 00... 080.00...
M830... 80... 8000. 0. .880 0 0.00. 80.. 80...5 008.. 0.6.. .0>0 80.. 0R. .00. 3m...

 

00.39.
OH. 00

..05 .00.. .08 ”mm/DE. >mv.

0080.80...

.080..080 .0 000.080 0... 00880.0 00800000.. a082.0. 0v... 080....8..0

000808096 80... 080 w8.0.8..0 .80....5 .805 .0 .8500 .80 n00....0.0

80.. 80...5 008.. 0>0.. .0>0 80.. 0R. 80.50.80: .0808 0... ..0... .080.

.00. 800 080 0080.8. 0.08 0.0 $8.00. 000.. .L ...05 .00.. .08 .08 08..
0800 .0. ...000.0 M83880 .0.... w8.0.8..0 80.0 .0 8500 .80 0.05 0.800.. .80 . 3m»,

 

.0..00.0 ..0 .88080 .090. 0 0.8.0 ”mag... >mv.
0......0 0. 0008 80.. .88080
0... 80 ..w... 0 .0 208...: 0 .0w .0w80. 08 0.800 80.. .0... 08... .0>0 80.. 0.0
.0 ..00...0 80 .0w 0. 3800.0 ..0 .88080 .0m.0. 0 0.8..0 o. 0008 .0>0 80.. 0.n. .00 . Dm«

 

v.8..0 0. 0..000 w80..0 0 ”mag... .55.

$00.0 m.88....80 .0 V.8.... 0. 8805.0 .. 0880.

.0 0.8..0 0 08.0.0. .0.00. ..80.800 80.. .0... .880 o. 0..000 $80 0 ..080 00..

80.... 80.. 80...5 0.... .80.. 8. 08... 0 .0>0 0.0... 00>? 38.0.8.0 .0 008000..
00.. 0>0.. ..08 80.. 080.80... .200 08.00 .8080 0.0 0808008.. 8.08 08C 6:5...

 

 

 

.0.

 

.0.
v...

 

.0.
oz

 

.8
00>

 

.cqu. CF CG 92.4.
.9 wZOmeDO mm—Emm 0.39. QmMmBWZ<ZD .0.... mDOU nm7.O.,.me.DO

 

03.. mmmbmmfi. MO mhmmhomn. M. 0.. U7.0.,..UD~.,..mZ.. mme.>~.m....Z.. .

 

196

.5830 3am wfi3o=£ E s32? 8s 2.633 aim

 

wagers »£ 8.53 038 .8 »n wowsmo

805on Rsowofio ho Rommba 2E3 Mambo 8 93:28 magma» >mM

mwsflsto »n 3.83

BEE so »9 uumsao soon 03% Emu: 85 80305 1253080 so F6533
msotom m on: so» 305 so» 553 M58 8 025:8 S>o so» ED A3 5?.

 

magic Hso» Mo
8233 mofitwom Batons: woosoos 53on .8 as 02% Humémm >m§

 

 

 

 

 

 

 

 

 

s «“3389 van mount.“ mEoom so n383 .353

8:— I magic 5o» mo 8:809 8:253 Batons: Boson: »=mb
so as 03% so» 5:3 $w=2 so 5:9: a mo @028 m 03E 55 so» ED .R: DP,

83 mafia so 2% 2n: ”mmésm 3mm

mom—o widths so.“ 2:: 2::

H on; so» 65 3:83 »o mooobo 2: 80¢ 9:538“ ho waists 08: nose
om 25% so» 553 EOE Ho m»a_o saga mo 30:3 03“: 85 so» BO #2 Dm...

2% Ho: osn wsEsz So» so :38 “so so @on 8 BE ”magma >m¥

H wow oo 8 03a so: 203 so» 85 oosom Ea
wsflsflo oso» so :38 “so so mom 8 BE so» 5:3 8:5 22: 803 $.33?

8 :83 {6% so» 5:3 45.6 ”mmézm >mv~

_ 5%
:23 3.56 05802 was waists “.8.—Sm so» :23 8:5 033 so» ED .93 Dm«

MED? @0235 so» can: 38 o E m»mu 808 ”mmémm >mM

Oh. CU

 

%%§s a» 55 32 Elsi: so

,1on 8% S €53on use owes so» :23 88.: BB 22: 203 .0» Smw

 

 

197

 

 

 

 

 

 

 

 

 

 

 

 

 

8n: .3» For": «6.6 256 «R :56 .55 .33 I. mm“: 260:. 9: aim
+ o
wean: ~»_£:oEuA x56 6% 65m 2
Son: .8» +
mg": .32: 2: x56 :23 >6 6: 9.56 6 :265: .26
+
» :55 A 652 u
is": 2:8: 2:55 2 . _ x :8 c
wmmru: $2: 2: x56 :2; 5:339: 96.5.6 25 fl 82m:
_ :o
«3-: a m » + 26.;
I 3N": $585 E. .8 5:6 .6 3x56 mun x56 :26 .26
2.69: 2: aim _ oz Emu: :.:ww
: a 8» 25:06
v.5": .8» 3a": «2.60:. N? .3: 05 :_ 55.59% m :55 22: .696 .25 :2: $2

Nmu: .mo>

 

of 23...: .2532: «23.622.

.068 Song 2830:: .823 «Sm

8E": 65:2: ~_‘ “mu: 2: :_ 3:262: 96.56 62w I

23": «25:9: E. “an: 2: :_ :2: 22: x56 :23 :3» a 65 :25 I

23": «360:: NF 7.8 2:16 6: 9.5.6 6 6.65: .vnm

oz»
+

258:. :82 a a

 

 

 

 

 

 

 

 

 

 

 

an": 8» a
2.60:. 2: Em Sam“: 23» a :_ 9.5.6 NF x56 .65 £6 _
652
82": mm» a
.262. 2: aim T||||

oz sau==2§§§§z :3

198

 

 

 

FQFPuc
239.. 2: ::.w

 

8m": 2539.: 8:06:83 .363 Sam

 

 

 

%

 

 

 

 

o:

 

23": «mm» 3 8:9. «:6 05 6 >5 65m

 

 

a

2:": .3285 .62 .636

a

2:": .2255 232 £23

a

 

 

 

 

 

 

 

 

 

 

an: .3285 333 2:3... 6% £25

 

mm; a

2.3.": $503th _u_00m .GNFDm

4

2:": .8535... 2:33 .NSm

 

 

 

 

 

 

 

 

 

.mg
6:

199

mm":
.8:

 

 

 

8N": 3:332: .2368: 6 .333: $6.26 :56 3 3568 ..aSm

 

 

mm":
.3:

 

 

 

8a": «9.25.: .2 853:8 e. .26 .55.:

 

 

 

 

man: 25% as: .o .8: .85 a £23

 

 

 

 

 

 

 

SN": 22% so 2 2%.... 9:6

 

 

9

 

 

as": S. .5; .2. 2.. =2; .22: :23

 

 

J

 

 

 

 

«3": «26:35 :2: £66 9.2: 633m

 

 

+

 

 

 

 

«man: .3325 :2: 29.: .325

 

 

 

a

 

 

can: .3385 225 2 :.:: .023.

 

 

+

 

 

man": .9536 «:05? =25 .3: .0: .825

 

 

 

+

 

mama: .6325 .692 :_ :56 .826

 

 

a

 

8w": 6.6% as...» draw

 

 

 

mm":
.8»

200

*SUl.

*SUZ‘

*SU3.

Questionnaire in Chinese

#55er (SU)

Tfififlfi%%¥@fiflfimfififlm%m,m@fl,fim,fifi
mwmfim<mam\flig\fifiM%)ofifi:fimflfifififi

flfifi§k$%?

____$%

(miiifififld): “MXififi” 997 fifii‘hsuzw fiwfi

$951116 ................. 998 "*‘“‘
$Eéfill€$ ............... 999

(mafimxawmm i§1§1iifi$flflfi 14 fimflfiéé’ckcfififiefifii’fi
fiflgTWHflm6¢E§fifio6fifiwflfifi$§¢fifluM$
mfififiTIWWE)N.fifi£i$fi?

 

(mfififikéa$fifi)%fiETflflfl$fifl“#fi”fi¢fi%
Wfi.fifififififi%6fifififi\6%@fi\fi6$5@o§fifi
MF%§$E¢WMUHW3fiE£X$%?

(mfifiifiklfilfié “$Tfiéfifi1¢”§£%“?fififl=flafl7” , iil‘cfl : EE+
gamma?
w%%WAH§“$E”fi%“$flfi”,%M:%E@:+

fiBWWQ?)

 

Em
+3$um ............. m
:+$U% ............. w
:+§uE ............. m
(1211;311wa “Mmfifi‘” 997 fififlhsm fi97fi
Z<iuޤ ................. 998
fififlg ............... wg

666%mbuiifi$£m¢fl¢.fififi%fl¥§ifi¢mml
m<5¢lfii61&),fi%&4&\fififikik\§fiya
m.fiEfi¢Hfi@¢TIm?
R666 ............... 1
6E34w ............. 2

201

8%] 1-2 m ............. 3

fifil- 37k ............. 4

EH"? 1 2x (@331 Tug???) .. 5 fifii‘kwg
Maia ..................... 8 fififihsus
fiéfilg ............... 9 fififijakSUB

*SU4. ifi‘E‘E‘Hfifi‘J ME. ZEi‘if 12 ’I‘Hifi’fifléfifi‘lflfiaqj. ‘Efirfi

 

fiifiz/I‘H M1”

__ mxwwm

$ffli’é ........... 998

ifiéfiflx ......... 999
*SU5.1§’£L1X, 1&2 @“éiifi— immmmzawmx 12 ’I‘FJB’J’EX/ifi
E?

6% ................... 1 fififlfim

19% ................... 5

$fiu‘fa‘ ................. 8

fififlg ............... 9
*SUGA’Jfififififi (fii'SUIB)

*SU3 §§Wfifiwv ............................................... 1

fifi‘fififly ................................................................... 2 fififiksmz
*SU73EififififiR (iii SU4)

*SU4 éfimfifil—sfly“3”§26“3"mtfi$ .................... 1 itfiflrsmz

FEE-fifth, ................................................................... 2 itfiilrsuae

*SU8. I*B~TfiJ;L/LX Mkflifl‘lflflé 5‘43 ZEfiB£3$$J fifi’xfién?
ELEEI/ME [IE—{k 61%] 3—m 6161175: @34— 3: KEEE

”fifiw

fl¥fiiflm ................................. 1

6%] 3—4 75: ................................... 2

6%] 1-2 ‘(fiq ................................... 3

fifil-Mfiz ............................... .....4

fifi¢66m ................................. 5 fififlwmg
fi 97 ii

202

Malia ....................................... 8
fi 97 fi
ifiéfilg ..................................... 9
fi 97 fi
96%. mmmzmawwmimaaza. ififififiiflEE/ymfi?
__ ’B‘iflmfifimfii
Z<inޤ ........... 998
ifiéfilIE$ ......... 999

*SU10. fifififififi (iii*SU8)

fifi§1>kSU39

fififlwmg

 

 

 

 

*&m§§MfiEfiW’ ............................................... 1
mfiam ................................................................... 2 m f? E
*SU12
*suuflfififihfi (iii*SU9)
*su9 fiat)?" .......... 1
Ffifififlfi ..................... 2 mm *SU39 £97fi
£15553? 2 miifikfllfilfimfi’t‘lfilflif‘fis $93] 3‘5
mime,afi*SU1zx§u¢sama.-I§a9mfiam 5 T5?- m g
mamas-9". #mfialarsma (8) (9)
*wm‘TEETM%M%@%%WHfl.E%.@E
E%%~E$E&%%flfi%¥TW%WTfi
E%fi\%fifi1¢fifiimIWfifiE? 1 5 8 9
*fifi:%WT@WIW
$Mh‘fi%%@%%fl5iA\%fi\%Efifl$fi 5 8 9
$$fl®2tlfimfififlfi§flm fi’EfifiiB‘JI‘filgfl £1] 5'] g]
*fifl:5%k~%fi%flfifiifiwmflflo 1 sm wm 8m
2C 0 2c:
$M%‘§§$E%magfi&78w%fizmm~ 1 5 8 9
§Wfl,fimfi@fl

 

 

 

 

 

 

203

 

 

 

1%)

*SU12c\ £3ZPKEBEDEMWMEE Bfii?flfi‘é§i§fl1fi%
B‘J‘T‘ERZT (fiflgfififi‘V‘fi‘ WRAP; EEébkm2§

*fifi]: [flfiflfiflfifi&fi%§fé

 

fiflfi?fi%m%

 

*SU12d\ $i~7fllfifli§7§$$flfl$ififififififififi§

*Eifij: fi‘ggfiflfi TEZHI %

 

 

 

 

 

*SU13‘ filifififififii‘ (iii*SU12l‘5]EX§'J)

Sfifififigfi“? Hg‘étfi ....................
fi—fi‘fiififi‘fl’flgfifi .................

Ffifififlfi ............................................

§J*SU15 INTR02

............. 1 fififilrsuas 39m
............. 2 E§#m-F*su12 . mam?

fl*su15 |NTRO1

............. 3 mafia-Psmz . flfififi

 

*SU15 INT R01

*SU15 INT R02

 

ffim’mfl‘fifi (ééffiifilianE SU12 [£0542
Eifl¢§$2§ “”E: E19), ffifiéfizlfil
'rZEffififlfiiibzfil‘cfliéfiBfimfifi
WEB?

 

 

ifimfkmmbeéflifilfifi SU12 WEE/E
§UEP§$>9 “E” 89» MWWJE
ifififlfiifigrfilflmflfifiém

 

 

in) .................... 1
EM ................... 1
x v ................. 5 SU15b
$5M ................. 5 ill SU15b :21; 8:18U15b
Nana ................ 8 Ell SU15b fafiél‘g """""""""" 9 fl SU15b
ifiéfilé .............. 9§IJSU15b ’ m """"""""
*SUISm mwzméa?
$88 fififlacsme
$8M: .......... . 998 fififihsms
EZEEQQI ....... . 999 itfiilbksme

*SU15b‘ 183%“??? Hi E’L’k‘iflfifi‘iiifiél‘filfl? B‘Jfl/I‘EEEILWW, 3

iii?

204

 

 

 

(Wififikfl?€$¥flfifififififlflfifr,fiflfi:flZEE

 

 

+swmmm?
mfifiifiMfi‘EEW’fiiWfllfi” , 55:19 : EEZ+§I$LX§U
E?)
$@

+E§ufi .................... w

:+2um .................... 19

:+§uE .................... 20 ‘

Z<iuޤ ........................ 998

7&28E388 ...................... 999

*wm‘mfmfl.@fifi—mfiifigflfl¢m~EflfinE%Efizw

fi?fi$—EHW\$fiflfi¢fixfltfl+:Efi‘fi%flfi+: i
’Efikhfi? ;
flf—EEW ............... 1 fififlwms
fih‘ifiEUfi/I‘fi .............. 2 fififlksum
Wtfl+:AH ............ 3 fififlfiMB
Efl+:Afiufi .......... 4
Eini'é.‘ .................... 8 fifiilbrsms
fififlg .................. 9 fififlfims

*wn‘fififiEfim$Wfl?$fi%flfl$%fl¢fififl¢fi,fifigi

 

5%?

$@
Z<iuޤ .................... 998
#Eéfill%$ .................. 999

*wm\fiiflx,fifii%£¢¢$%$flfifi¢flflfififi§flfi?

 

iwfl
Z<inޤ ........... 998
ifiééll%$ ......... 999

205

 

flfifififi:m%§fikflfi¢flfifiifififiifi

 

 

 

 

 

 

 

 

 

 

a E Efl

%,flfi$wwfifl¢&fifiéwfifififlflflfifl f: E 18

79“9”.#flfi¥1*su20 5 8

(Tflflfl%%f$ffiflfifi%fifimgflfl%fl
fie)

*ww18%figéfifi%fifififi%flfl.$@$ 1 5 8
fi,$%fifiw$%%Wfio
*Efl:@fiflflm$flflfifififlfi§

$M%18%E%E%E£%flifififlmmflfi%
Wfifi,fl%fi@%€fimfiufififififi§
%@.@$fififimflfimm@fi? 1 5 8

*fifl:fi%£@fi£%fli%fiflfififlfifififi

DE
$M%1E#&WWW%fi%flmA,Mfififlflfl
fiEfiEfiflofi%@fiwfififim$fi@E 1
fifl.fiflfififimifiofi%fifififiwi fl
@Mfifififi.mmfii‘fifl\%%%fifi 8U 5 8
EE'ffifiél‘IflEE? 19d
*3fl:$%flfifi$fi@
fim%\@%fififi@fl%fiTfiZ$fliflfi%fl
fl? 1 5 8

*flfl:%flflfi%$fl§%fififl$fifl

wmw‘%§E%€W8£flb$@fl,fiE%T 1

fl.fi%.fififi%mw
82181817888188882 5‘“ 5 8

xafl=$8888888,98882x%888 $2
$3
$M%‘%EEW8%flmfififlfifi%flmiflw 1
@Efiflfimgz? fl 5 8
*Efl:%fl%&fi£$fi%%fl§ 6U
19g
fimw‘fi%€€WE}fl@fiW%TflWE%MT? 1 5 8
*fifl:$fi%flfi%fl7
$m%1EE€W@fiEE%@fiME%E.EEM
EEMEEU? 1 5 8

 

 

 

 

*3fl:fi@$@fiflfl§@flfifl$fl

206

 

 

 

 

*SU19h\ Efififlfiéifiifififififl‘fiflfi‘lfl, 58315?!
TngififiliME‘F-ififié‘iqj'tlfififlflfifilflifié, {£583

fi€¥¢WWM%m$%? 1 5 8 9
*3fl:fifi%$%%fi®¢$%flfiMfiflfifi¢
88888

 

8mm‘88888—188888~888,88
E%@fiflfififlifl%fi7§§%fim.w
WWfi%%\I¢1fi§%Efl%A? 1
*fifl:$¥%@fififififl¢7§£%@
$m%\%@flfi%flfl%%%@fiflFE%§Wfi F
8888,88888588888888 ;
H.8fififififi%8? 1 5 8 9
*Efl:%fi%fl%fififi%k§fififififl
8,888888

 

 

 

 

 

 

 

 

 

*SU20\ flifififihfé‘ (iiI*SU19lflEX§TJ)
fiEWW’H’Q‘é‘fi’E 0-2 ’1‘ ......... 1 iifii'PSU39 fi 97 E

Pfifififlh§$ ..................... 2

*W%‘fififi@€@£$EWflF$Wflfl,fiéfiEEfiHflMflfifi~mfi
$fi%flflwm@m$@@?

BIL) ............. 1

MIL) ................ 5 itfiilbksuzsb
$51118; ................ 8 fififlbksuzsb
18 28V; ......... 9 fififiHSUZSb

*SUZSa‘ 5512953115 33 7?; £51 ?

 

2‘ fifiilbksuzs
$8118 ............... 998 fififiksuze
1828.88 ............... 999 fifiillaksuze

$M%\%~mmMfi%$¥mfiF$%Mfiw,fiiflfiififi?
( mifiifixEg‘t‘FflEfiizfi’fii‘fifiam T" , iEI‘EJ‘ 12,482+
EZJLXEFIIE ?
mfifiifiAE’éEWE'fifiWfllfi” , 18181 mar—+3858? )

207

 

$8
—RE§mf ...................... M
:31281 ...................... w
:+$E ...................... m
Z<8ufl8 ........................ 998
fifilg ...................... %9

*w%\8%fi—fifiéfifiWflflfiEfiéWfi?%fi£#Efiw\WfiEfiE
81%tfl+:88.fi%flfl+:Afiu%?

82—188 .............. 1 fififlwms Eh
88E888 .............. 2 fififlwms ;
fitfl+:¢fi ............ 3 fififlwms ‘
flfl+:/\Fj 2113111” .......... 4

$121118 .................... 8 fififlhsm _
fififl§ .................. 9 fififiwms J

 

*wm\fififi#&$%fi%flfififlflfiwMflWE$i$%?
(WfififiAEEETXflfifififiWMflWWP,EM:EE+E?R
mm?
mififikfifiiqfiflwy$flfihfifl:Efl:+§fl@?)

 

 

$8
+E§H% .............. m
:+$ 2115111 ............. 19
:+§ufi ............. 20
Z<8ufié ................. 998
fififlg ............... wg
*wm‘EEux,88£¢E$i8gfi$EQ$m#¢fi$%Mfl?
$8
Z<iuޤ .......... 998
$888Il8$ ........ 999
*ww18fl—$$.8%?8fimm38fi8E3migflfl?
E ............... 1
R8 ............. 2 itfifil 51132
$81118 ........... 8 flififil SU32
1828I8 ......... 9 flfifil SU32

6&9L$Wfifi%$%fiflfifi$:fimfiwfifi$@m$fi”W,8mwfif
EU*SU30

208

fimm $:flfi$W—$$$flfi§flfl$mzAfifiizflflw.fizifi
£51?
( IlflfifiifiAEE‘E'FXfliXfi’Eifi‘fi‘EWI‘fiJ7” , ial‘tfl : Efl+5w>l
311E. ?
mifiifixfiléiWE’fifiWfllifi” , iEI‘n] : Eflz-i-wfim? )

 

 

 

 

$m
+25um .............. m
:+;um ............. m E
:+$UE ............. 20
Z<iuߤ ................. 998
$Eéfi£3€$ ............... 999
*wm‘fiiux.ifififi¢¢$%fi$mfigflfi$mi+fififigm
ma?
$fi
Z<iuޤ ............. 998
iEéfill€$ ........... 999

*SU32‘ Mfiflfi’dfljfflflglfifiififl, 13E %é飢7kflfiifi%ififiifiifi?

fifi
Z<inޤ ............ 998
ifiéfillg$ .......... 999
*SU33‘ §Eififi&%)fiz (firsuaz)
*SU32 gamsfimumafig ................................ 1
Ffifififlb .................................................... 2 fifii‘PSUN

*SU34‘ IMJB LIE. .‘QKHEJ‘EB‘JWIEJafififiéi T 35K?

 

@fifl fimiwm$m: iflfl. Efimz
)2]...3 $m4

Emmi ........... 998

ifiéfilg ......... 999

*SU35\ ¥5fififi%fi : (ilI*SU32 *SU34)
*SU32 59533955379“? ............................................... 1 fifi§J*SU37

 

209

6mm $:fifi$fi~$$mmz%flfl$%EAfi%E£flflw,figkfi
a?
( Mafia-A@§"$Fxmz#"§ta“mem T” , SEI‘fiJ : Efl+z¥m
am ?
mifiifixlfiléfi“$i"ma“$fllfi” , iEI‘cfl : Ear—+35% ? )

 

 

 

$m

+zyuw .............. m

:+;aw ............. m V

:+$ug ............. m

Z<inޤ ................. 998

$E2§Il€$ ............... 999

*wm\€$ux.fifififi¢¢$$fi$flfl$flfl$mEAfifiE£m
ma? k
g

$fl

Z<iuߤ ............. 998

fiéfil‘é ........... 999

*ww\&flfl%$Mfl%flflWE.fi%%£¢&%fifl%fifififl?

 

aw
Z<§ufi§ ............ 998
$E?@Il%$ .......... 999
*SU33‘ filifififififiz (firsusz)
*SU32 §§E§fiflfl1fi£fii§z ................................ 1
Ffifififlt .................................................... 2 fifiilrsuw

*WM\M%HE.@$%m%WM%KfifiT¥R?

 

@fim fifliww$fi= fiml gywm
)5} ...3 gm4
$iui'é; ........... 998
fififl§ ......... 9%
*SU35‘ Xfifififié : (ilI*SU32 *SU34)
*SU32 figfimfififi‘w" ............................................... 1 fififiksuy

209

*SU32 ”éiwfimflw’fifiigfiwsuu gfimfiififlm’fifii
5....2
Ffifififlt ......................................................... 3 flifii‘PSUfl

*SU36\ fefiifigwkwléfimflfifimfiéi 3 A! fifififfifl‘fl‘éfl?

 

ma
Z<iuޤ ............ 998
fifiug .......... %9 r

*SU37~ *Wfififififiii SU26 fi 96 E)

*SU26 Efimfimfi 1‘ 2 312 3 ............................... 1
Ffifififly ................................................. 2 flififirsumen

 

*SU39‘ 5E1)? fififififiii’fl‘f? . fififlifitfi)

Efiéfli‘ififiéfi‘mflifififiifi ......... 1 fififiPSUW
Ffi’fififlfiflg ............................. 2
*SU39.1\ filifififififié‘ (fiiiifikflgflfllfi)

fiifixmmax 01-25 %E .................... 1
Ffifififlyfifi .......................................... 2 ma fififikom aw i 260 fi

 

*wm\$fifififi%fl@fiWfi,Tfimflfl5@ twm‘Efi(@m/
$#&fi%flfi@fimm%fi%o“EM%”E W$)&%,@€£
%:%A%@flfifi&;%A%MflW§W@E 2mmma22$m
%.Wflfim%fiWfifiu%flon%Afififi m2?

fifi%fifi$%%fl$a%EWo gfifi;m;§fi*
<%HE#Tfi$u%M%m>§@$~wfimé 2

Wflfifififiw.flififi£k§fi? “*?fiufi?iu
$fifififizwmsmwx¥mmmsfi,%fifl S'fifl=m4€$
*ama mfimfi§¢mmml

MEBWIQ ?

 

 

 

 

210

 

 

 

 

 

 

 

*SU87a $152 *sussa 7222
gas +33 L161} ............ 12 Mitflflififi...997 $21115 ...... 998

:+5 11.111 ............ 19 “1115‘ ......... 998 flééfilg...999

:7“; L113 ............ 20 $626.12; ...... 999

wmm $$ $M% &fi

+3.5 WU ............ 12 2121621523997 Nana ...... 998
#151: :1; L131; ............ 19 $21115 ......... 998 ifiéélé...999

:1"; WE? ............ 20 flié’filfilg ...... 999

 

 

 

211

 

 

Questions about childhood experiences
*CHl. The next questions are about your childhood background. Did you live with both
of your biological parents up until you were sixteen?

' YES ............................ 1 GO TO *CH6
NO .............................. 5
DON’T KNOW ......... 8 GO TO *CH6
REFUSED ................. 9 GO TO *CH6

 

*CH2. Why didn’t you live with your biological parents?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

INTERVIEWER: CIRCLE ALL THAT APPLY. DO NOT READ LIST. F
(IF NEC: Did your biological mother or father die, were they separated or
divorced, or was there some other reason?)
*CH2a. How old were you when (you/ ;
your) (EVENT)? 1
IF VOL “LESS THAN ONE YEAR L;
OLD,” CODE “1.”
YEARS OLD
MOTHER DIED ............................... 1 DON’T KNOW ........ 998
REFUSED ................ 999
YEARS OLD
FATHER DIED ................................. 2 DON’T KNOW ........ 998
REFUSED ................ 999
PARENTS SEPARATED/ DIVORCED YEARS OLD
........................................................... 3 DON’T KNOW 998
REFUSED ................ 999
PARENTS NEVER LIVED TOGETHER
........................................................... 4
YEARS OLD
ADOPTED ........................................ 5 DON’T KNOW ........ 998
REFUSED ................ 999
YEARS OLD
WENT TO BOARDING SCHOOL. 6 DON’T KNOW ........ 998
REFUSED ................ 999
YEARS OLD
FOSTER CARE ................................ 7 DON’T KNOW ........ 998
REFUSED ................ 999
LEFT HOME BEFORE AGE SIXTEEN

 

 

 

212

 

 

 

 

 

 

 

 

 

 

 

 

........................................................... 8 YEARS OLD
DON’T KNOW ........ 998
REFUSED ................ 999

OTHER (SPECIFY) .......................... 9

YEARS OLD

DON’T KNOW ........ 998
REFUSED ................ 999

DON’T KNOW ................................. 98

REFUSED ......................................... 99

*CH6. Up until you were sixteen, were m ever away from home for six months

or longer — either in foster care, with other relatives, in a boarding school, hospital,

juvenile detention center, or elsewhere?
INTERVIEWER: CODE “NO” IF R VOLUNTEERS “RETURNED HOME ON

WEEKENDS” OR OTHER OCCASIONS DURING SIX—MONTH PERIOD.

YES ............................ 1

NO ............................. 5 GO TO *CH8
DON’T KNOW .......... 8 GO TO *CH8
REFUSED .................. 9 GO TO *CH8

*CH6a. Where did you go?

INTERVIEWER: CIRCLE ALL THAT APPLY.

LIVING WITH OTHER RELATIVES ......... 1
BOARDING SCHOOL ................................. 2
HOSPITAL .................................................... 3
JUVENILE DETENTION CENTER ............ 4
FOSTER HOME ............................................ 5
OTHER (SPECIFY) ....................................... 6
DON’T KNOW .............................................. 8
REFUSED ...................................................... 9

*CH6b. How Old were you the first time you went away?

 

.................... YEARS OLD
DON’T KNOW ...................... 998
REFUSED .............................. 999

213

 

*CH6c. Altogether, how many months or years were you away
from home up until you were sixteen?

 

DURATION NUMBER
CIRCLE UNIT OF TIME: MONTHS .1 YEARS 2
DON’T KNOW ...................... 98
REFUSED .............................. 99
*CH8. Who was the head Of your household for most Of your childhood? .fi

INTERVIEWER: IF R SAYS “FATHER”, PROBE: Was that your biological
father, step-father, adoptive father, or someone else?

INTERVIEWER: IF R SAYS “MOTHER”, PROBE. : Was that your
biological mother, step-mother, adoptive mother, or something else?

 

INTERVIEWER: IF R SAYS IT CHANGED, PROBE: Who was the male head
of your household for most Of the time before you turned seventeen?

 

BIOLOGICAL FATHER ........................................... 1

ADOPTIVE FATHER ............................................... 2

STEP FATHER (SPOUSE/ PARTNER OF MOTHER) 3

OTHER MALE (SPECIFY) ...................................... 4

BIOLOGICAL MOTHER ........................................... 5

ADOPTIVE MOTHER ................................................ 6

STEP MOTHER (SPOUSE/ PARTNER OF FATHER) 7

OTHER FEMALE (SPECIFY) ................................... 8

DON’T KNOW .......................................................... 98 GO TO *CH22
REFUSED .................................................................. 99 GO TO *CHZZ

*CH8a. How many years Of school did (he/ she) complete?

 

YEARS
DON’T KNOW .......... 98
REFUSED .................. 99

214

*CH9. How much of your childhood did (male head Of household / female head of
household) either work for pay or work in a family business? Would you say all
Of the time, most, some, a little or not at all?

INTERVIEWER: IF NEC CLARIFY: Work for pay includes self-
employment.

INTERVIEWER: IF R SAYS FATHER WAS A FARMER, CLARIFY: Farming
counts as working in a family business.

ALL ............................ 1 GO TO *CHll

MOST ........................ 2 GO TO *CHll

SOME ........................ 3

A LITTLE .................. 4

NOT AT ALL ........... 5

DON’T KNOW .......... 8

REFUSED .................. 9

*CH9a. ....................................................................... What was the main reason he

was not working for pay during most Of your childhood years?

INTERVIEWER: IF R SAYS: “He/She was self employed,” CLARIFY:
Work for pay includes self-employment. REREAD *CH9.

INTERVIEWER: IF R SAYS, “He/She ways a farmer,” CLARIFY:
Farming counts as working for pay. REREAD *CH9.

IN TERVIEWER: CIRCLE ALL THAT APPLY.

PHYSICAL DISABILITY OR INJURY ................... 1
ALCOHOL OR DRUG ABUSE ............................... 2
MENTAL OR EMOTIONAL DISABILITY ............ 3
TO STAY AT HOME TO RAISE CHILDREN ........ 4
UNABLE TO FIND JOB ........................................... 5
OTHER (SPECIFY) ................................................... 7
DON’T KNOW .......................................................... 8
REFUSED .................................................................. 9

 

 

 

*CHIO. INTERVIEWER CHECKPOINT (SEE *CH9):

*CH9 EQUALS’3’ OR ‘4’ .............................................. 1
ALL OTHERS ................................................................. 2 GO TO *CH22

215

*CH22. Overall, how would you rate (your parents’ relationship/ the relationship of the
people who raised you) while you were growing up — excellent, good, fair, or

poor?

EXCELLENT .................................... 1
GOOD ................................................ 2

FAIR .................................................. 3

POOR ................................................. 4

NO COUPLE (IF VOLUNTEERED) 5
DON’T KNOW .................................. 8
REFUSED .......................................... 9

 

*CH23. How much conflict and tension was there in your household while you were
growing up — a lot, some, a little, or none?

A LOT ........................ 1
SOME ........................ 2
A LITTLE .................. 3
NONE ........................ 4
DON’T KNOW .......... 8
REFUSED .................. 9

 

 

(RB, PG 38) LIST FOR QUESTIONS *CH28 - *CH29

0 PUSHED, GRABBED OR SHOVED
0 THREW SOMETHING
0 SLAPPED, HIT, OR PUNCHED

 

 

 

*CH28. (RB, PG 38) When you were growing up, how Often did someone in your
household do any of the things (on the list on page 38 in your booklet) to ou —
Often, sometimes, rarely, or never?

OFTEN ....................... l
SOMETIMES ............ 2
RARELY .................... 3
NEVER ...................... 4 GO TO *CH29
DON’T KNOW .......... 8 GO TO *CH29
REFUSED .................. 9 GO TO *CH29

216

*CH28a. Who did this to you?
(PROBE: Anyone else?)
INTERVIEWER: CIRCLE ALL THAT APPLY.

BIOLOGICAL FATHER ....... l
ADOPTIVE FATHER ........... 2
STEP FATHER ...................... 3
BIOLOGICAL MOTHER ..... 4
ADOPTIVE MOTHER .......... 5

STEP MOTHER .................... 6
BROTHER/ SISTER ............. 7
OTHER PERSON .................. 8
DON’T KNOW ...................... 98
REFUSED .............................. 99

 

*CH41. During the years you were growing up, did (WOMAN WHO RAISED R)
ever have periods lasting 2 weeks or more where she was sad or depressed most of
the time?

YES ............................ 1

NO .............................. 5 GO TO *CH46
DON’T KNOW .......... 8 GO TO *CH46
REFUSED .................. 9 GO TO *CH46

*CH41a.Was this during all, most, some, or only a little Of your childhood?

ALL ............................ 1
MOST ........................ 2
SOME ........................ 3
A LITTLE .................. 4
DON’T KNOW .......... 8
REFUSED .................. 9

 

 

*CH42. During the time her depression was at its worst, did she also have other
symptoms like low energy, changes in sleep or appetite, and problems with
concentration?

YES ............................ 1
NO .............................. 5 GO TO *CH46

217

DON’T KNOW .......... 8 GO TO *CH46
REFUSED .................. 9 GO TO *CH46

 

*CH46—--- - - - During the time you were growing up, did (WOMAN
WHO RAISED R) ever have periods of a month or more when she was constantly
nervous, edgy, or anxious?

 

 

 

YES ............................ 1

NO .............................. 5 GO TO *CHSI

DON’T KNOW .......... 8 GO TO *CHSl

REFUSED .................. 9 GO TO *CH51 F

*CH46a....-. -- ----Was that during all, most, some, or only a little of :_
your childhood? 5
ALL .......................... 1
MOST ....................... 2 #7
SOME ....................... 3
A LITTLE ................ 4
DON’T KNOW ........ 8
REFUSED ................ 9

*CH47-- - - .......... During the time her nervousness was at its worst, did she

 

also have other symptoms like being restless, irritable, easily tired, and difficulty
falling asleep?

YES ............................ 1

NO .............................. 5 GO TO *CHSI
DON’T KNOW .......... 8 GO TO *CH51
REFUSED .................. 9 GO TO *CHSI

*CH51. Did (WOMAN WHO RAISED R) ever complain about anxiety attacks
where all of a sudden she felt frightened, anxious, or panicky?

YES ............................ 1

NO .............................. 5 GO TO *CH52

DON’T KNOW .......... 8 GO TO *CH52

REFUSED .................. 9 GO TO *CH52

*CH5 1 a. Did she ever comment that during these attacks that her heart was
pounding, or that she was short Of breath, felt ill, or was fearful that she
would die?

218

YES .......................... 1

NO ............................ 5
DON’T KNOW ........ 8
REFUSED ................ 9

 

*CH52. Did (WOMAN WHO RAISED R) ever have a problem with alcohol or drugs?

 

YES ............................ 1
NO .............................. 5 GO TO *CH6]
DON’T KNOW .......... 8 GO TO *CH6]
REFUSED .................. 9 GO TO *CH6]
*CH67. Did (WOMAN WHO RAISED R) ever attempt to commit suicide?
YES ............................ 1
NO .............................. 5
DON’T KNOW .......... 8
REFUSED .................. 9

*CH71. During the years you were growing up, did (MAN WHO RAISED R) ever
have periods lasting 2 weeks or more where he was sad or depressed most Of the

time?

YES ............................ 1

NO .............................. 5 GO TO *CH76
DON’T KNOW .......... 8 GO TO *CH76
REFUSED .................. 9 GO TO *CH76

*CH71a.Was this during all, most, some, or only a little Of your childhood?

ALL ............................ 1
MOST ........................ 2
SOME ........................ 3
A LITTLE .................. 4
DON’T KNOW .......... 8
REFUSED .................. 9

 

 

*CH72. During the time his depression was at its worst, did he also have other

symptoms like low energy, changes in sleep or appetite, and problems with
concentration?

YES ............................ 1
NO .............................. 5 GO TO *CH76

219

 

DON’T KNOW .......... 8 GO TO *CH76
REFUSED .................. 9 GO TO *CH76

 

*CH76 .................................... During the time you were growing up, did (MAN WHO
RAISED R) ever have periods of a month or more when he was constantly
nervous, edgy, or anxious?

 

 

 

YES ............................ 1
NO .............................. 5 GO TO *CH81
DON’T KNOW .......... 8 GO TO *CH8]
REFUSED .................. 9 GO TO *CH8]
*CH76a. ................................. Was that during all, most, some, or only a little of
your childhood?
ALL .......................... 1
MOST ....................... 2
SOME ....................... 3
A LITTLE ................ 4
DON’T KNOW ........ 8
REFUSED ................ 9
*CH77 ............. ...... ---During the time his nervousness was at its worst, did he
also have other symptoms like being restless, irritable, easily tired, and difficulty
falling asleep?
YES ............................ 1
NO .............................. 5 GO TO *CH8]
DON’T KNOW .......... 8 GO TO *CH8]
REFUSED .................. 9 GO TO *CH8]

 

*CH80. Did his nervousness ever interfere a lot with his life or activities?

YES ............................ 1
NO .............................. 5
DON’T KNOW .......... 8
REFUSED .................. 9
*CH8]. Did (MAN WHO

RAISED R) ever complain about anxiety attacks where all Of a sudden he felt
frightened, anxious, or panicky?

YES ............................ l

220

NO .............................. 5 GO TO *CH82
DON’T KNOW .......... 8 GO TO *CH82
REFUSED .................. 9 GO TO *CH82

*CH8la. Did he ever comment that during these attacks that his heart was
pounding, or that he was short Of breath, felt ill, or was fearful that he

would die?

YES .......................... 1
NO ............................ 5
DON’T KNOW ........ 8
REFUSED ................ 9

 

*CH82. Did (MAN WHO RAISED R) ever have a problem with alcohol or drugs?

YES ............................ 1

NO .............................. 5 GO TO *CH91
DON’T KNOW .......... 8 GO TO *CH91
REFUSED .................. 9 GO TO *CH91

 

221

 

 

aficiofi “mm/DEE 3%

we;
#32 an 6.53 @3588 38m Ho «:85. .55 so.» vmv .o _ Q0...

 

28:3. 3.5% ”ESE; Eva

mnemmmfinom :55? 338 aim :33 so.» Ev .300...

 

€853 $5.8m So» 55 ~83 So mavnfim ”mm/DEA >mv¢

wvoEwB 8:28
So» 56 39: He. 55 20:8 :5 Em coao no» vwv ... .ozuos

 

QED no
292 wave 3 H55 HES E :5 ”WEE “mmfiim >8:

~82: 9 war: he 038m @508 .3 ov 8
vOmommsm 0.83 50% mwfifi micv mo So How coco no» vmv. .. .n 35...

 

o—QUR
Oh. CU
fl

A238 85 3 8: 3:3 ”mm/DEE Eva

wov 8 805 v8.83 so» “23 wEov ho mwfifi so» mEZm 85
Bacon x03 8 mo: :8 coao so.» Ev mamas»: 8 Bio a m< 8 ~00...

 

5
..E

 

 

E
M:

 

6
oz

 

3
mm;

 

A. . 359258 S 2.50 m m< ”Umz Ev

 

 

Emu» 03:08 .8 a Box wEBv Ev so»

mmqvv Ba 305 a 322 8 :83 o? .ov 8 SHEEN. BE raov 3:65 mwfifi Sons 0.3 $8526 :8: 2:. A? GA :55 ADD...

 

£539:— 8353 «52.: ...—3:830

222

 

A080: 88% .326 wag ”mmgm Mrmvc

moose ES 208 “gauge 080: 88m 33¢ S: no» Ev 0:90...

 

"90...
OH. 00

982 sea 5.... mans: “mmfiim was

mmxuv .509 H32
3 ..8 $25 .85 E8 080: 80¢ .35 S: .85 so» Ev .2 Q0...

 

Eugen wEmEEE. ”$3me 5%

E533

:o :5me 95:3 Ho .mfiwzqug £8: chmEm $3855

$5305 82 wagoEOm wfiov 3 33¢on $28088 owafinv
bugfionzov 55 so» Ev... 485 wfiuom .3 55 Hofiov EDD...

 

838% 8:8 9 Q 2 2m a wane... ”ESE.“ >va

moms—5v vacuum 828 9 .3 8 DE m Hum 55 so» Ev ... .fiQU...

 

AwEEEn 8 .080: duo a BE wfiufiob “mmémm >m§v

meEEn go 080:
3832 a no .30 E362 928088 85 485 .55 so» Ev .w _QU.._

 

 

 

 

 

 

3? Rs: 8» 038m 83 MERE... ”Egg 5%

2:3 vo>= so» Bacon 55b 05 .8 $5.8m
So.» 80¢ $ch 55o 8 >288 38m $5 so» Ev EDD...

 

 

223

‘ \ I'v- y-Il

 

 

«.808 M80808 .8 $83va88: $888
888 an 18x8 988088 ov 0880 088 8>0 so» Ev...

AGE—U...

 

fiwqfiofia #803 80:30

05 80¢? 083.3 .8 0008 a $5008 8 88 £08 05 .8
0830 .3803 ca? 08088 80¢ 8:008 0888 8: moov
ma ”omz ”5 8808 80¢ E @85sz 3 was 8 0883

68me 8:83 .885 908088 808 88 so» Ev...

.880...

 

aaa—U...
OH. CU

$5.8 82:
9880 8 805 880808 3 8388 8 emu—030m $088

0E: 83.088 8% 02w 9 08088 008m 88 8% Ev...

Mano...

 

580858 8

805 88 v8 :888 a 3 880 380858 8>0 50% 08>». ..

.0035...

 

8888 8 3888 SE 880a .8 E8 988%
8 @883 ova—08 88 800v mEH ”Umz n5 ~809an 8

E 88 van 8880 8 8 880 380838 8>0 8» 083

.vo—QO...

 

A225 8 6:» 6&3 .0288. 82m .83 2383
a BE .8888 8:88 :0 80.803 0 0m: 8>0 so» Ev...

dBQU...

 

mflfi¢ Romans 8 v0>_o>8 80w 80¢o so» Ev...

din—U...

 

 

5

 

v:—

 

Amy
OZ

 

3
ma

 

A. .. 88:08 8 220 m m<

uOmHZ n5

 

 

224

 

 

00.00%. .305 020.500.

0. m .000E ”030.00> 0500000000 >.003 0..; 00.00050. .000.. m. N 0050s 030.00., 0500000000 000000000 0..; 00.00050. .305 0. F 02.0.... ..

 

 

3d cm... 8.0 ~00 mod wmd d. .000 E0305 .0 .0000
odd 8... 2d «ed 3d 2d 0.0.00. .0253 0000000500 >00
and 8d 3d mod add «0... 0000000500 .0>0
9d and add mad Ed and 30305 0>..5000_0E 2.0.000 .0>0
3.0 mud 3d odd ..md dud 500.000 .0300: .0>0
8.5 mmd Ed odd >vd «0.0 $5 mmuv. E0305 >00 >000
00.0 d .d 9d 0 .d mmd de >05 mmuv. E0305 0000000500 >000
8... :5 s... 8.0 0:. 5.0 .0. EU.
E0305 0>_.5000.0E >_.0.000 >000

dmd Nvd mad 5d dvd Ed .05 25. 000.000 >000
mod de dad Rd nod 3d >0; «WV. >.. >000
5:008:00? 0E0305 000.0 E0305 000.0 3.003 000.0 30305

003000., .0.00E +0E0305 .000000 .003 00.00 000000.303 520500.000

3.00.05 020500.000 3.00.05 .0.00E 3.00.05 3.00.05

 

”5000.0 0... .0 .00. 0... .0 0020., 5

 

080-58 .0502; 0.2. 0.05 .800. 5 0.000085 .0. 08. 002. 0.2. $20, 0 ... 030 038

 

0030. 598%?

225

 

 

or 5% 8285 B 520
m.~ dd 5 m.» .3. 3 Same _8_c__o 85288 Em
v.” 66 N.N QM .m._. in fin .e... ed 85338 55
a.” a.a m.~ . 3285 93833668 .26
N.N .2. ~._. N.N ....N flu 9.256 53mm; .96
a.a d... 3. Am; mmnvv Emacs Em 2.8

3. ._‘.N F.» E» munvv Emzoa 8.528% 228

2 .2 3 a;

8qu £285 m>_amum_me-o_8m 25m

Am; mpnvv 9.256 2.8

N.N .2. N; ...N .5.: V... N.N .mé né mN .wd N... Am; N—uvv bu 2.6m
._.0 $8 .38 ._.0 $3 .38 ._.0 $3 .38 ...o $8 #08 ...o $8 .58

3033 82m 8532 82m meocmeamfi 82m msmzeg
savagfimf 5283 5:950 mmamtg 522mm ESE .mEQma maEmccEcu
8.8.52, .555; 5:28

 

88-58 65-12; so: $3 .923 E 2:83 25 €682.» 9%. 5% $82 :85 022:8. .8: a8 .2 £25m m. 3.3 23

 

226

 

 

 

 

 

 

 

n I...

.2865 Eozommo we ::2 6.60.6 9: be m=_m> a.“
so E 2 2 So 3 2 ma .3 3 3 ma 22$._8_=__08§§8>=m
8... a... 3 3 8o 3 I ..m 8o 2 2 a.” 8523.22.
3... 2 2 3 8.0 S 2 3 3... 8 3 2 8.32..

233858.208 55
.2 on 3 5 mg ..m no 2 .3 mm 3 3 8.2.2.3859.
33 EN 3. 2: s3 ..8 3 ...: 33 :.N E. ...: Esauvvefieazmémm
8... 2a 3 3 3... a.a 2 ...... :3 ..2 E 3 gal
5285 882.8% .68
53 man 3. ...: 39v ....8 3 3. 33 NS 3 E. Emmi
5285 «25833668 2.8
one N.N .3 m: .2 3 mo 3 one 3 no 2 Eenvvmééuézémo
2.. no .3 ..N no mo N.N no no @5033...

2 5...... mo 2 a 0.8 mo 2 5.8 .5

83:98 0: 5? $0 mom new .33 5? H5 mam Em xmm £5, m._<

 

5652...... ex... mo... 227mm,...“ 22. .23 "Eng .32.. ._

.mooméoom 66-123 so: 98 559$. 53 2 Sam 9.25:. 8555
on; 288 9.2% 868.8 9.255 new Emsfica 5033a 89.220 53.8 85083 2: B 226:; .2.? 2an

 

227

 

 

 

 

.9
«NM 8... 3.? and 2: no... and «a; Emu Qt Rm? 8» 2356 53285
3 0mm 0: 85288 .96
m~.~ o: 8;. m~.~ 3.? «3 3d am... 3... gm n9 mm» £355 595305
m9 Sm 0: m>_amum_me-o_8w .96
m3 Ed N: 5.. 8.0 w: 84 m3. 3... mdm m2 8.» 29.55 9.9%
m. E Rm 0: 92:2. .98; .96
8.. mad 0: mg 8.0 m: 33 S... «N... 2K om. mg
«.8 5m 0: BE 5% 9:56
o: mad 8.. :.F mad 8.. N: mod 3.. «.5 mom 8»
0mm 52 0: £5530 :95 as?
m: mad 8.. o: 23 2... 8... m3 2... m8 mmm 8»
m3 mam. 0: £55.88 682..
_o$mm mom .0 $8 ”Sn _0 $8 mo:
uEE+uE3+§muoE E_2%qu+omm+xmm :89: $5, : ago
”282: N685

 

88.58 66-12; 2: :5: £8
2ng 5055, EwEm>_o>c_ 9.255 .0 $87. Em EmEfiES _mo_mEQ @8225 :858 833083 9: €03. was

 

228

BIBLIOGRAPHY

Afifi, T. 0., Brownridge,-D. A., Cox, B. J., 8: Sareen, J. (2006). Physical
punishment, childhood abuse and psychiatric disorders. Child Abuse
Negl, 30(10), 1093-1103.

Ager, C. R., Ferrer, H. P., Fillmore, K. M., Golding, J. M., Leino, E. V., 8:
Motoyoshi, M. (1996). Aggregate-level predictors of the prevalence of
selected drinking patterns in multiple studies: a research synthesis
from the Collaborative Alcohol-Related Longitudinal Project. Subst Use
Misuse, 31(11-12), 1503-1523.

Agrawal, A., & Lynskey, M. T. (2008). Are there genetic influences on
addiction: evidence from family, adoption and twin studies. Addiction,
103(7), 1069-1081.

American Psychiatric Association. (1994). Diagnostic and Statistical Manual of
Mental Disorders, Fourth Edition. . Washington, D.C.: the Association.

Anderson, P. (2006). Global use of alcohol, drugs and tobacco. Drug Alcohol
Rev, 25(6), 489-502.

Andrade, 1..., Walters, E. E., Gentil, V., 8: Laurenti, R. (2002). Prevalence of
ICD-10 mental disorders in a catchment area in the city of Sao Paulo,
Brazil. Soc Psychiatry Psychiatr Epidemic], 37 (7), 316-325.

Andreasson, 8., Holder, H. D., Norstrom, T., Osterberg, E., & Rossow, I. (2006).
Estimates of harm associated with changes in Swedish alcohol policy:
results from past and present estimates. Addiction, 101(8), 1096-1105.

Andrews, 0., Henderson, 8., 8: Hall, W. (2001). Prevalence, comorbidity,
disability and service utilisation. Overview of the Australian National
Mental Health Survey. Br 1 Psychiatry, 178, 145-153.

Anthony, C. B. (1995). Dramshop liability laws. In D. H. Jaffe (Ed.),
Encyclopedia of drugs and alcohol. New York, Ny: Macmillian Library
Reference.

Anthony, J. C., 8: Petronis, K. R. (1989). Epidemiologic evidence on suspected
causal associations between cocaine use and psychiatric disturbances.
Washington D.C.: US. government printing office.

Anthony, J. C., 8: Van Etten, M. L. (1998). Epidemiology And Its Rubrics. In A.

Bellack & M. Hersen (Eds.), Comprehensive Clinical Psychology. (pp.
255-380). Oxford, UK: Elsevier Science Publications.

229

 

 

Anthony, J. C., Warner, L. A., 8: Kessler, R. C. (1994). Comparative
epidemiology of dependence on tobacco, alcohol, controlled
substances, and inhalants: Basic findings from the National Comorbidity
Survey. Experimental and Clinical Psychopharmacology, 2, 244-268.

APA. (1952). Diagnostic and Statistical Manual of Mental Disorders, First Edition.
Washington, D.C.: American Psychiatric Association.

APA. (1968). Diagnostic and Statistical Manual of Mental Disorders, Second
Edition. . Washington, DC.

APA. (1994). Diagnostic and Statistical Manual of Mental Disorders, Fourth
Edition. . Washington, D.C.: the Association.

Archer, K. J., 8: Lemeshow, S. (2006). Goodness—of-fit test for a logistic
regression model fitted using survey sample data. The Stata journal,
6(1), 97-105. if
Archer, K. J., Lemeshow, S., 8: Hosmer, D. W. (2007). Goodness—of-fit tests for D
logistic regression models when data are collected using a complex
sampling design. Computational Statistics and Data Analysis 51(200'1),
4450 - 4464.

 

Arias, E., MacDorman, M. P., Strobino, D. M., 8: Guyer, B. (2003). Annual
summary of vital statistics--2002. Pediatrics, 112(6 Pt 1), 1215-1230.

Ary, D. V., Tildesley, E., Hops, H., 8: Andrews, J. (1993). The influence of
parent, sibling, and peer modeling and attitudes on adolescent use of
alcohol. Int ] Addict, 28(9), 853-880.

Avery, G. (2005). Endogeneity in logistic regression models. Emerg Infect Dis,
11(3), 503-504; author reply 504-505.

Bahlmann, M., Preuss, U. W., 8: Soyka, M. (2002). Chronological relationship
between antisocial personality disorder and alcohol dependence. Eur

Addict Res, 8(4), 195-200.

Bahr, S. J., Marcos, A. C., 8: Maughan, S. L. (1995). Family, educational and
peer influences on the alcohol use of female and male adolescents. ]

Stud Alcohol, 56(4), 457-469.
Barnes, G. M., Welte, J. W., 8: Dintcheff, B. (1992). Alcohol misuse among

college students and other young adults: findings from a general
population study in New York State. Int ] Addict, 27(8), 917-934.

230

Barnow, S., Schuckit, M. A., Lucht, M., John, U., 8: Freyberger, H. J. (2002). The
importance of a positive family history of alcoholism, parental rejection
and emotional warmth, behavioral problems and peer substance use
for alcohol problems in teenagers: a path analysis. [Stud Alcohol, 63(3),
305-31 5.

Barrish, H. H., Saunders, M., 8: Wolf, M. M. (1969). Good behavior game:
effects of individual contingencies for group consequences on
disruptive behavior in a classroom. [App] BehavAnal, 2(2), 119-124.

Barros, M. B., Botega, N. J., Dalgalarrondo, P., Marin-Leon, L., 8: de Oliveira, H. P
B. (2007). Prevalence of alcohol abuse and associated factors in a
population-based study. Rev Saude Publica, 41(4), 502-509.

Bastian, H. (2000). A consumer trip into the world of the DALY calculations: an
Alice-in-Wonderland experience. Reprod Health Matters, 8(15), 1 13-116.

 

...-l
Bazzano, L. A., On, D., Reynolds, K., Wu, K., Chen, C. S., Duan, K., et a1. (2007). D
Alcohol consumption and risk for stroke among Chinese men. Ann
Neurol.

Begg, M. D., 8: Lagakos, S. (1990). On the consequences of model
misspecification in logistic regression. Environ Health Perspect, 87, 69-

75.

Bensley, L. S., Spieker, S. J., Van Eenwyk, J., 8: Schoder, J. (1999). Self-
reported abuse history and adolescent problem behaviors. II. Alcohol
and drug use. [Adolesc Health, 24(3), 173-180.

Bewick, B. M., Trusler, K., Mulhern, B., Barkham, M., 8: Hill, A. J. (2008). The
feasibility and effectiveness of a web-based personalised feedback and
social norms alcohol intervention in UK university students: a
randomised control trial. Addict Behav, 33(9), 1192-1 198.

Blackson, T. C., 8: Tarter, R. E. (1994). Individual, family, and peer affiliation
factors predisposing to early-age onset of alcohol and drug use. Alcohol
Clin Exp Res, 18(4), 813-821.

Bloomfield, K., Allamani, A., Beck, P., Bergmark, K. H., Csemy, L., Eisenbach-
Stangl, I., et al. (2005). Gender, Culture and Alcohol Problems: A Multi-
national Study (Project Report). Berlin: Institute for Medical Informatics,
Biometrics 8: Epidemiology, Charité Universitatsmedizin Berlin.

Bobashev, G. V., 8: Anthony, J. C. (2000). Use of alternating logistic regression
in studies of drug-use clustering. Subst Use Misuse, 35(6-8), 1051-1073.

231

 

Boehm, S. L., 2nd, Ponomarev, 1., Jennings, A. W., Whiting, P. J., Rosahl, T. W.,
Garrett, E. M., et a1. (2004). gamma-Aminobutyric acid A receptor

subunit mutant mice: new perspectives on alcohol actions. Biochem
Pharmacol, 68(8), 1581-1602.

Bolton, J. M., Robinson, J., 8: Sareen, J. (2008). Self-medication of mood
disorders with alcohol and drugs in the National Epidemiologic Survey
on Alcohol and Related Conditions. ] Affect Disord.

Borges, G., Mondragon, L., Medina-Mora, M. E., Orozco, R., Zambrano, J., 8:
Cherpitel, C. (2005). A case-control study of alcohol and substance use
disorders as risk factors for non-fatal injury. Alcohol Alcohol, 40(4), 257-
262.

Bowirrat, A., 8: Oscar-Berman, M. (2005). Relationship between dopaminergic
neurotransmission, alcoholism, and Reward Deficiency syndrome. Am]
Med Genet B Neuropsychiatr Genet, 132(1), 29-37.

Bremner, J. D., Bolus, R., 8: Mayer, E. A. (2007). Psychometric properties of the
Early Trauma Inventory-Self Report. I Nerv Ment Dis, 1 95(3), 21 1-218.

Bremner, J. D., Randall, P., Vermetten, E., Staib, L., Bronen, R. A., Mazure, C.,
et a1. (1997). Magnetic resonance imaging-based measurement of
hippocampal volume in posttraumatic stress disorder related to
childhood physical and sexual abuse-a preliminary report. Biol
Psychiatry, 41(1), 23-32.

Brems, C., Johnson, M. E., Neal, D., 8: Preemon, M. (2004). Childhood abuse
history and substance use among men and women receiving
detoxification services. Am [Drug Alcohol Abuse, 30(4), 799-821.

Briere, J., 8: Elliott, D. M. (2003). Prevalence and psychological sequelae of
self-reported childhood physical and sexual abuse in a general
population sample of men and women. Child Abuse Neg], 27 (10), 1205-
1222.

Briscoe, J., Akin, J., 8: Guilkey, D. (1990). People are not passive acceptors of
threats to health: endogeneity and its consequences. Int ] Epidemiol,
19(1), 147-153.

Brown, E. C., Catalano, R. P., Fleming, C. B., Haggerty, K. P., 8: Abbott, R. D.
(2005). Adolescent substance use outcomes in the Raising Healthy
Children project: a two-part latent grth curve analysis. ] Consult Clin
Psychol, 73(4), 699-710.

232

 

Brown, G. R., 8: Anderson, B. (1991). Psychiatric morbidity in adult inpatients
with childhood histories of sexual and physical abuse. Am ] Psychiatry,
l48(1), 55-61.

Brown, J., Babor, T. P., Litt, M. D., 8: Kranzler, H. R. (1994). The type A/type B
distinction. Subtyping alcoholics according to indicators of
vulnerability and severity. Ann N YAcad Sci, 708, 23-33.

Bucholz, K. K., Heath, A. C., Reich, T., Hesselbrock, V. M., Kramer, J. R.,
Nurnberger, J. 1., Jr., et a1. (1996). Can we subtype alcoholism? A latent
class analysis of data from relatives of alcoholics in a multicenter family 3
study of alcoholism. Alcohol Clin Exp Res, 20(8), 1462-1471. r?

Caamano-Isorna, P., Corral, M., Parada, M., 8: Cadaveira, F. (2008). Factors 1‘
associated with risky consumption and heavy episodic drinking among
Spanish university students. ] Stud Alcohol Drugs, 69(2), 308-312.

 

Caldji, C., Diorio, J., 8: Meaney, M. J. (2003). Variations in maternal care alter F
GABA(A) receptor subunit expression in brain regions associated with
fear. Neuropsychopharmacology, 28(11), 1950-1959.

Cameron, C. A., 8: Trivedi, P. K. (2009). Microeconometrics using stata.
College station, TX: A stata press pulication, StataCorp LP.

Carey, V. J., Zeger, S. L., 8: Diggle, P. (1993). Modelling multivariate binary
data with alternating logistic regressions. Biometrika, 80(3), 517-526.

Carrigan, M. H., 8: Randall, C. L. (2003). Self-medication in social phobia: a
review of the alcohol literature. Addict Behav, 28(2), 269-284.

Caspi, A., 8: Moffitt, T. E. (2006). Gene-environment interactions in psychiatry:
joining forces with neuroscience. Nat Rev Neurosci, 7(7), 583-590.

Chen, C. C., Lu, R. B., Chen, Y. C., Wang, M. P., Chang, Y. C., Li, T. K., et a1.
(1999). Interaction between the functional polymorphisms of the
alcohol-metabolism genes in protection against alcoholism. Am ] Hum
Genet, 65(3), 795-807.

Chen, C. Y., 8: Anthony, J. C. (2003). Possible age-associated bias in reporting
of clinical features of drug dependence: epidemiological evidence on

adolescent-onset marijuana use. Addiction, 98(1), 71-82.

Chen, C. Y., Dormitzer, C. M., Gutierrez, U., Vittetoe, K., Gonzalez, G. B., 8:
Anthony, J. C. (2004). The adolescent behavioral repertoire as a context

233

for drug exposure: behavioral autarcesis at play. Addiction, 99(7), 897-
906.

Chen, J., Dunne, M. P., 8: Han, P. (2006). Child sexual abuse in Henan province,
China: associations with sadness, suicidality, and risk behaviors among
adolescent girls. [Adolesc Health, 38(5), 544-549.

Chou, S. P., 8: Pickering, R. P. (1992). Early onset of drinking as a risk factor
for lifetime alcohol-related problems. Br ] Addict, 87(8), 1199-1204.

Clark, D. B. (2004). The natural history of adolescent alcohol use disorders.
Addiction, 99 Suppl 2, 5-22.

Clark, D. B., Lesnick, L., 8: Hegedus, A. M. (1997). Traumas and other adverse
life events in adolescents with alcohol abuse and dependence. [Am
Acad Child Adolesc Psychiatiy, 36(12), 1744-1751 .

Clark, W. B., 8: Hilton, M. E. (1991). Alcohol in America: Drinla'ng Practices and
Problems. Albany, NY: SUNY Press.

Clemmons, J. C., DiLillo, D., Martinez, 1. G., DeGue, S., 8:Jeffcott, M. (2003).
Co-occurring forms of child maltreatment and adult adjustment
reported by Latina college students. Child Abuse Neg], 27(7), 751-767.

Cochrane, J., Chen, H., Conigrave, K. M., 8: Hao, W. (2003). Alcohol use in
China. Alcohol Alcohol, 38(6), 537-542.

Coombs, R. H., Paulson, M. J., 8: Richardson, M. A. (1991). Peer vs. parental
influence in substance use among hispanic and anglo children and
adolescents. ] Youth Adolesc, 20, 73-88.

Couzigou, P., Coutelle, C., Fleury, B., 8: Iron, A. (1994). Alcohol and aldehyde
dehydrogenase genotypes, alcoholism and alcohol related disease.
Alcohol Alcohol Supp], 2, 21-27.

CPES, C. P. E. S. (2001). NATIONAL COMORBIDITY SURVEY REPLICATION
(NCS-R): CPES (Collaborative Psychiatric Epidemiology Surveys).

Crabbe, J. C., Phillips, T. J., Harris, R. A., Arends, M. A., 8: Koob, G. F. (2006).
Alcohol-related genes: contributions from studies with genetically
engineered mice. Addict Bio], ll (3-4), 195-269.

Crum, R. M., Chan, Y. P., Chen, L. S., Storr, C. L., 8: Anthony, J. C. (2005).
Incidence rates for alcohol dependence among adults: prospective

234

 

data from the Baltimore Epidemiologic Catchment Area Follow-Up
Survey, 1981-1996. ] Stud Alcohol, 66(6) , 795-805.

Crum, R. M., Helzer, J. E., 8: Anthony, J. C. (1993). Level of education and
alcohol abuse and dependence in adulthood: a further inquiry. Am ]
Public Health, 83(6), 830-837.

Crum, R. M., Juon, H. S., Green, K. M., Robertson, J., Fothergill, K., 8:
Ensminger, M. (2006). Educational achievement and early school
behavior as predictors of alcohol-use disorders: 35-year follow-up of
the Woodlawn Study. 1 Stud Alcohol, 67(1), 75-85.

Crum, R. M., Storr, C. L., 8: Anthony, J. C. (2005). Are educational aspirations
associated with the risk of alcohol use and alcohol use-related
problems among adolescents? Subst Use Misuse, 40(2), 151-169.

Cumming, R. G., 8: Klineberg, R. J. (1994). A study of the reproducibility of
long-term recall in the elderly. Epidemiology, 5(1), 116-1 19.

Dai, K., Thavundayil, J., Santella, S., 8: Gianoulakis, C. (2007). Response of the
HPA-axis to alcohol and stress as a function of alcohol dependence and
family history of alcoholism. Psychoneuroendocrinology, 32(3), 293—305.

Dawson, D. A. (2000). Alcohol consumption, alcohol dependence, and all-
cause mortality. Alcohol Clin Exp Res, 24(1), 72-81.

Dawson, D. A., Goldstein, R. B., Patricia Chou, S., June Ruan, W., 8: Grant, B. F.
(2008). Age at First Drink and the First Incidence of Adult-Onset DSM-N
Alcohol Use Disorders. Alcohol Clin Exp Res.

Degenhardt, L., Bohnert, K. M., 8: Anthony, J. C. (2007). Case ascertainment of
alcohol dependence in general population surveys: 'gated' versus
'ungated' approaches. Int] Methods Psychiatr Res, 16(3), 1 1 l-123.

Degenhardt, L., Chiu, W. T., Sampson, N., Kessler, R. C., 8: Anthony, J. C.
(2007). Epidemiological patterns of extra-medical drug use in the
United States: evidence from the National Comorbidity Survey
Replication, 2001-2003. Drug Alcohol Depend, 90(2-3), 210-223.

Degenhardt, L., Chiu, W. T., Sampson, N., Kessler, R. C., Anthony, J. C.,
Angermeyer, M., et a1. (2008). Toward a global view of alcohol, tobacco,
cannabis, and cocaine use: findings from the WHO World Mental Health
Surveys. PLoS Med, 5(7), e141.

235

 

Del Boca, F. K., 8: Darkes, J. (2003). The validity of self-reports of alcohol
consumption: state of the science and challenges for research.
Addiction, 98 Supp12, 1-12.

Demyttenaere, K., Bruffaerts, R., Posada-Villa, J., Gasquet, 1., Kovess, V.,
Lepine, J. P., et a1. (2004). Prevalence, severity, and unmet need for
treatment of mental disorders in the World Health Organization World
Mental Health Surveys. lama, 29] (21), 2581-2590.

Desousa, G., Murphy, S., Roberts, C., 8: Anderson, L. (2008). School policies
and binge drinking behaviours of school-aged children in Wales—a
multilevel analysis. Health Educ Res, 23(2), 259-271.

DeWit, D. J., Adlaf, E. M., Offord, D. R., 8: Ogborne, A. C. (2000). Age at first
alcohol use: a risk factor for the development of alcohol disorders. Am ]
Psychiatry, 157(5), 745-750.

Dick, D. M., 8: Bierut, L. J. (2006). The genetics of alcohol dependence. Curr
PsychiattyRep, 8(2), 151-157.

Dick, D. M., Rose, R. J., Viken, R. J., Kaprio, J., 8: Koskenvuo, M. (2001).
Exploring gene-environment interactions: socioregional moderation of
alcohol use. [Abnorm Psycho], 110(4), 625-632.

DiLalla, L. F., 8: Gottesman, II. (1991). Biological and genetic contributors to
violence--Widom's untold tale. Psychol Bull, 109(1), 125-129; discussion
130-122.

Donovan, J. E., 8: Molina, B. S. (2008). Children's introduction to alcohol use:
sips and tastes. Alcohol Clin Exp Res, 32(1), 108-119.

Downs, W. R., Capshew, T., 8: Rindels, B. (2004). Relationships between adult
women's alcohol problems and their childhood experiences of parental
violence and psychological aggression. [Stud Alcohol, 65(3), 336-344.

Dube, S. R., Anda, R. F., Felitti, V. J., Chapman, D. P., Williamson, D. F., 8: Giles,
W. H. (2001). Childhood abuse, household dysfunction, and the risk of

attempted suicide throughout the life span: findings from the Adverse
Childhood Experiences Study. lama, 286(24), 3089-3096.

Dube, S. R., Anda, R. F., Felitti, V. J., Croft, J. B., Edwards, V. J., 8: Giles, W. H.
(2001). Growing up with parental alcohol abuse: exposure to childhood
abuse, neglect, and household dysfunction. Child Abuse Neg], 25(12),

1 627- l 640.

236

 

Dube, S. R., Miller, J. W., Brown, D. W., Giles, W. H., Felitti, V. J., Dong, M., et
a1. (2006). Adverse childhood experiences and the association with
ever using alcohol and initiating alcohol use during adolescence. [
Adolesc Health, 38(4), 444 e44l-410.

Duncan, G. J., Boisjoly, J., Kremer, M., Levy, D. M., 8: Eccles, J. (2005). Peer
effects in drug use and sex among college students. [Abnorm Child
Psycho], 33(3), 375-385.

Duncan, S. C., Duncan, T. E., 8: Strycker, L. A. (2006). Alcohol use from ages 9
to 16: A cohort-sequential latent growth model. Drug Alcohol Depend,
81(1), 71-81.

Edenberg, H. J., Dick, D. M., Xuei, X., Tian, H., Almasy, L., Bauer, L. O., et a1.
(2004). Variations in GABRAZ, encoding the alpha 2 subunit of the
GABA(A) receptor, are associated with alcohol dependence and with
brain oscillations. Am [Hum Genet, 74(4), 705- 714.

Edenberg, H. J., Xuei, X., Chen, H. J., Tian, H., Wetherill, L. F., Dick, D. M., et
a1. (2006). Association of alcohol dehydrogenase genes with alcohol
dependence: a comprehensive analysis. Hum Mol Genet, 15(9), 1539-
1549.

Edwards, G., 8: Gross, M. M. (1976). Alcohol dependence: provisional
description of a clinical syndrome. Br Med], 1 (6017), 1058-1061 .

Ehlers, C. L., Gilder, D. A., Wall, T. L., Phillips, E., Feiler, H., 8: Wilhelmsen, K.
C. (2004). Genomic screen for loci associated with alcohol dependence
in Mission Indians. Am [Med Genet B Neuropsychiatr Genet, 1293(1),
1 10-1 15.

Elkins, 1. J., McGue, M., 8: Iacono, W. G. (2007). Prospective effects of
attention-deficit/hyperactivity disorder, conduct disorder, and sex on
adolescent substance use and abuse. Arch Gen Psychiatry, 64(10), 1145-
1152.

Ely, M., Hardy, R., Longford, N. T., 8: Wadsworth, M. E. (1999). Gender
differences in the relationship between alcohol consumption and drink
problems are largely accounted for by body water. Alcohol Alcohol,
34(6), 894-902.

Engle, R. F., Hendry, D. F., 8: Richard, J. F. (1983). Exogeneity. Econometrica
51 277-304.

237

 

Feighner, J. P., Robins, E., Guze, S. B., Woodruff, R. A., Jr., Winokur, G., 8:
Munoz, R. (1972). Diagnostic criteria for use in psychiatric research.
Arch Gen Psychiatry, 26(1), 57-63.

Felitti, V. J., Anda, R. F., Nordenberg, D., Williamson, D. F., Spitz, A. M.,
Edwards, V., et a1. (1998). Relationship of childhood abuse and
household dysfunction to many of the leading causes of death in adults.
The Adverse Childhood Experiences (ACE) Study. Am [Prev Med, 14(4),
245-258.

Fillmore, K. M., Golding, J. M., Kniep, S., Leino, E. V., Shoemaker, C., Ager, C.
R., et a1. (1995). Gender differences for the risk of alcohol-related
problems in multiple national contexts. Recent DevAlcohol, 12, 409-439.

Fillmore, K. M., Golding, J. M., Leino, E. V., Ager, C. R., 8: Ferrer, H. P. (1994).
Societal-level predictors of groups' drinking patterns: a research
synthesis from the Collaborative Alcohol-Related Longitudinal Project.
Am [Public Health, 84(2), 247-253.

Flanders, W. D. (2006). On the relationship of sufficient component cause
models with potential outcome (counterfactual) models. Eur [ Epidemic],
2] (12), 847-853.

Fromme, K., 8: Corbin, W. (2004). Prevention of heavy drinking and
associated negative consequences among mandated and voluntary
college students. [ Consult Clin Psychol, 72(6), 1038-1049.

Gable, R. S. (2004). Comparison of acute lethal toxicity of commonly abused
psychoactive substances. Addiction, 99(6), 686-696.

Galaif, E. R., Stein, J. A., Newcomb, M. D., 8: Bernstein, D. P. (2001). Gender
differences in the prediction of problem alcohol use in adulthood:
exploring the influence of family factors and childhood maltreatment. [
Stud Alcohol, 62(4), 486-493.

Gallo, J. J., Anthony, J. C., 8: Muthen, B. O. (1994). Age differences in the
symptoms of depression: a latent trait analysis. [ Geronto], 49(6), P251-
264.

Giancola, P. R., 8: Moss, H. B. (1998). Executive cognitive functioning in
alcohol use disorders. Recent DevAlcohol, 14, 227-251.

Glindemann, K. E., Ehrhart, 1. J., Drake, E. A., 8: Geller, E. S. (2007). Reducing

excessive alcohol consumption at university fraternity parties: a cost-
effective incentive/reward intervention. Addict Behav, 32(1), 39-48.

238

 

 

l“

 

Goedde, H. W., Agarwal, D. P., Fritze, G., Meier-Tackmann, D., Singh, S.,
Beckmann, G., et a1. (1992). Distribution of ADHZ and ALDHZ genotypes
in different populations. Hum Genet, 88(3), 344-346.

GOLDBERG, L. (1966). Behavioral and Physiological Effects of Alcohol on Man.
Psychosom Med, 28, 570-595.

Gordis, L. (2004). Epidemiology (3rd ed.). Philadelphia: Elsevier Saunders.

Gorenstein, E. E. (1987). Cognitive-perceptual deficit in an alcoholism
spectrum disorder. [ Stud Alcohol, 48(4), 310-318.

Gorwood, P., Batel, P., Gouya, L., Courtois, F., Feingold, J., 8: Ades, J. (2000).
Reappraisal of the association between the DRDZ gene, alcoholism and
addiction. Eur Psychiatry, 15(2), 90-96.

 

Grant, B. F., 8: Harford, T. C. (1995). Comorbidity between DSM-IV alcohol use __
disorders and major depression: results of a national survey. Drug F
Alcohol Depend, 39(3), 197-206.

Grant, B. F., Harford, T. C., Muthen, B. O., Yi, H. Y., Hasin, D. S., 8: Stinson, F. S.
(2007). DSM-N alcohol dependence and abuse: further evidence of
validity in the general population. Drug Alcohol Depend, 86(2-3), 154-
166.

Green, A. H. (1988). Child maltreatment and its victims. A comparison of
physical and sexual abuse. Psychiatr Clin North Am, 11(4), 591-610.

Greenland, S., 8: Brumback, B. (2002). An overview of relations among causal
modelling methods. Int [ Epidemic], 31(5), 1030-1037.

Grimm, D. (2008). Public health. Staggering toward a global strategy on
alcohol abuse. Science, 320(5878), 862-863.

Gu, D., Wildman, R. P., Wu, 11., Reynolds, K., Huang, J., Chen, C. S., et al.
(2007). Incidence and predictors of hypertension over 8 years among
Chinese men and women. [Hypertens, 25(3), 517-523.

Gureje, O., Lasebikan, V. O., Kola, L., 8: Makanjuola, V. A. (2006). Lifetime and
12-month prevalence of mental disorders in the Nigerian Survey of
Mental Health and Well-Being. Br] Psychiatry, 188, 465-471.

Hamburger, M. E., Leeb, R. T., 8: Swahn, M. H. (2008). Childhood
maltreatment and early alcohol use among high-risk adolescents. [Stud
Alcohol Drugs, 69(2), 291-295.

239

 

 

Hao, W., Chen, H., 8: Su, Z. (2005). China: alcohol today. Addiction, 100(6),
737-741.

Hao, W., Su, Z., Lin, B., Zhang, K., Yang, H., Chen, S., et a1. (2004). Drinking

and drinking patterns and health status in the general population of five

areas of China. Alcohol Alcohol, 39(1), 43-52.

Harada, S., Agarwal, D. P., Goedde, H. W., Tagaki, S., 8: lshikawa, B. (1982).

Possible protective role against alcoholism for aldehyde
dehydrogenase isozyme deficiency in Japan. Lancet, 2(8302), 827.

Harburg, E., DiFranceisco, W., Webster, D. W., Gleiberman, L., 8: Schork, A.
(1990). Familial transmission of alcohol use: 11. Imitation of and aversion

to parent drinking (1960) by adult offspring (l977)-Tecumseh,
Michigan. [ Stud Alcohol, 51 (3), 245-256.

Harford, T. C., Grant, B. F., Yi, H. Y., 8: Chen, C. M. (2005). Patterns of DSM-1V
alcohol abuse and dependence criteria among adolescents and adults:

results from the 2001 National Household Survey on Drug Abuse.
Alcohol Clin Exp Res, 29(5), 810-828.

Harford, T. C., 8: Muthen, B. O. (2001). The dimensionality of alcohol abuse

and dependence: a multivariate analysis of DSM-IV symptom items in
the National Longitudinal Survey of Youth. [ Stud Alcohol, 62(2), 150-157.

Harford, T. C., 8: Parker, D. A. (1994). Antisocial behavior, family history, and

alcohol dependence symptoms. Alcohol Clin Exp Res, 18(2), 265-268.

Hasin, D., 8: Paykin, A. (1999). Alcohol dependence and abuse diagnoses:

concurrent validity in a nationally representative sample. Alcohol Clin

Exp Res, 23(1), 144—150.

Hasin, D. S., Grant, B., 8: Endicott, J. (1990). The natural history of alcohol
abuse: implications for definitions of alcohol use disorders. Am [
Psychiatry, 147(11), 1537-1541.

Hasin, D. S., 8: Grant, B. F. (2004). The co-occurrence of DSM-IV alcohol abuse
in DSM-IV alcohol dependence: results of the National Epidemiologic
Survey on Alcohol and Related Conditions on heterogeneity that differ

by population subgroup. Arch Gen Psychiatry, 61(9), 891-896.

Hasin, D. S., Schuckit, M. A., Martin, C. S., Grant, B. F., Bucholz, K. K., 8: Helzer,

J. E. (2003). The validity of DSM-IV alcohol dependence: what do we

240

@I
_— _ ._'_..-JIJ_._

 

 

know and what do we need to know? Alcohol Clin Exp Res, 27(2), 244-
252.

Hasin, D. S., Stinson, F. S., Ogburn, E., 8: Grant, B. F. (2007). Prevalence,
correlates, disability, and comorbidity of DSM-IV alcohol abuse and
dependence in the United States: results from the National
Epidemiologic Survey on Alcohol and Related Conditions. Arch Gen
Psychiatry, 64(7), 830-842.

Hasin, D. S., Van Rossem, R., McCloud, S., 8: Endicott, J. (1997). Differentiating
DSM-IV alcohol dependence and abuse by course: community heavy p...
drinkers. [SubstAbuse, 9, 127-135. 3

Hawkins, J. D., Catalano, R. F., 8: Miller, J. Y. (1992). Risk and protective factors
for alcohol and other drug problems in adolescence and early
adulthood: implications for substance abuse prevention. Psychol Bull,
112(1), 64-105.

'—

 

Heath, D. B. (1995). Alcohol: history of drinking. In]. H. Jaffe (Ed.),
Encyclopedia of drugs and alcohol
New York, NY: Macmillan Library Reference.

Heath, R. G. (1972). Electroencephalographic studies in isolation-raised
monkeys with behavioral impairment. Dis Nerv Syst, 33(3), 157-163.

Heckman, J. (1978). Dummy Endogenous Variables in a Simultaneous
Equation System. .Econcmetrica, 46, 931-959.

Helzer JE, B. A., McEvoy LT. (1991). Alcohol abuse and dependence. In R. L.
DA (Ed.), PSychiatric Disorders in America. The Epidemiologic
CatchmentArea Study. (pp. 81-115). New York: The Free Press.

Higuchi, S., Matsushita, S., Maesato, H., 8: Osaki, Y. (2007). Japan: alcohol
today. Addiction, 102(12), 1849—1862.

Higuchi, S., Parrish, K. M., Dufour, M. C., Towle, L. H., 8: Harford, T. C. (1994).
Relationship between age and drinking patterns and drinking
problems among Japanese, Japanese-Americans, and Caucasians.
Alcohol Clin Exp Res, 18(2), 305-310.

Hingson, R. W., Heeren, T., 8: Winter, M. R. (2006). Age at drinking onset and

alcohol dependence: age at onset, duration, and severity. Arch Pediatr
Adolesc Med, 160(7), 739-746.

241

Hishimoto, A., Liu, 0. R., Drgon, T., Pletnikova, O., Walther, D., Zhu, X. G., et
a1. (2007). Neurexin 3 polymorphisms are associated with alcohol
dependence and altered expression of specific isoforms. Hum Mol
Genet, 1 6(23), 2880-2891 .

Holder, H. D. (2007). What we learn from a reduction in the retail alcohol
prices: lessons from Finland. Addiction, 102(3), 346-347.

Holder, H. D., Gruenewald, P. J., Ponicki, W. R., Treno, A. J., Grube, J. W., Saltz,
R. F., et al. (2000). Effect of community-based interventions on high-risk
drinking and alcohol-related injuries. [ama, 284(18), 2341-2347. r“

Holford, T. R. (1991). Understanding the effects of age, period, and cohort on
incidence and mortality rates. Annu Rev Public Health, 12, 425-457.

Holmes, S. J., 8: Robins, L. N. (1987). The influence of childhood disciplinary
experience on the development of alcoholism and depression. [ Child F, ~

 

Psycho] Psychiatry, 28(3), 399-415.

Holmes, S. J., 8: Robins, L. N. (1988). The role of parental disciplinary practices
in the development of depression and alcoholism. Psychiatry, 51(1), 24-
36.

Hooper, M., Hardy, K., Handyside, A., Hunter, S., 8: Monk, M. (1987). HPRT-
deficient (Lesch-Nyhan) mouse embryos derived from germline
colonization by cultured cells. Nature, 326(6110), 292-295.

Horwitz, A. V., Widom, C. S., McLaughlin, J., 8: White, H. R. (2001). The impact
of childhood abuse and neglect on adult mental health: a prospective
study. [Health Soc Behav, 42(2), 184-201.

Hosmer, D. W., 8: Lemeshow, S. (1980). Goodness-of-fit tests for the multiple
logistic regression model. Communications in Statistics: 771ecry and
Methods, Part A, 9, 1043-1069.

Huang, G., Zhang, Y., Momartin, S., Cao, Y., 8: Zhao, L. (2006). Prevalence and
characteristics of trauma and posttraumatic stress disorder in female
prisoners in China. Compr Psychiatry, 47(1), 20—29.

Huang , Y. 0., Liu, Z. L., Zhang, M. Y., Shen , Y. C., Tsang , A., He , Y. L., et a1.
, (2008). Mental disorders and service use in China. . In Ronald C Kessler
8: T. B. Ustiin (Eds.), World Health Organization World Mental Health
Survey Series Volume 1 - Patterns of mental illness in the WMH Surveys
(Vol. 1). New York: Cambridge University Press.

242

Hughes, T. L., Johnson, T. P., Wilsnack, S. C., 8: Szalacha, L. A. (2007).
Childhood risk factors for alcohol abuse and psychological distress
among adult lesbians. Child Abuse Neg], 31(7), 769-789.

Hume, D. (1977). An enquiry concerning human understanding. Indianapolis:
Hackett.

Jackson, K. M., 8: Sher, K. J. (2003). Alcohol use disorders and psychological
distress: a prospective state-trait analysis. [Abnorm Psycho], 112(4),
599-6 13.

Jasinski, J. L., Williams, L. M., 8: Siegel, J. (2000). Childhood physical and
sexual abuse as risk factors for heavy drinking among African-
Arnerican women: a prospective study. Child Abuse Neg], 24(8), 1061-
107 l .

Jiafang, Z., Jiachun, W., Yunxia, L., Xiaoxia, 0., 8: Ya, F. (2004). Alcohol abuse
in a metropolitan city in China: a study of the prevalence and risk
factors. Addiction, 99(9), 1103-1 110.

Johnson, B. A., Cloninger, C. R., Roache, J. D., Bordnick, P. S., 8: Ruiz, P. (2000).
Age of onset as a discriminator between alcoholic subtypes in a
treatment-seeking outpatient population. Am [ Addict, 9(1), 17-27.

Johnson, R. C., Nagoshi, C. T., Schwitters, S. Y., Bowman, K. S., Ahern, F. M., 8:
Wilson, J. R. (1984). Further investigation of racial/ ethnic differences
and of familial resemblances in flushing in response to alcohol. Behav
Genet, 14(3), 171-178.

Katz, J., Carey, V. J., Zeger, S. L., 8: Sommer, A. (1993). Estimation of design
effects and diarrhea clustering within households and villages. Am [
Epidemic], 138(11), 994-1006.

Kawakami, N., Shimizu, H., Haratani, T., Iwata, N., 8: Kitamura, T. (2004).
Lifetime and 6-month prevalence of DSM-III-R psychiatric disorders in
an urban community in Japan. Psychiatry Res, 121(3), 293-301.

Kellam, S. G., Brown, C. H., Poduska, J. M., Ialongo, N. S., Wang, W., Toyinbo,
P., et a1. (2008). Effects of a universal classroom behavior management
program in first and second grades on young adult behavioral,
psychiatric, and social outcomes. Drug Alcohol Depend, 95 Suppl 1, S5-
328.

Kendler, K. S., Kuo, P. H., Webb, B. T., Kalsi, G., Neale, M. C., Sullivan, P. F., et
a1. (2006). A joint genomewide linkage analysis of symptoms of alcohol

243

 

dependence and conduct disorder. Alcohol Clin Exp Res, 30(12), 1972-
1977.

Kendler, K. S., Myers, J., 8: Prescott, C. A. (2007). Specificity of genetic and
environmental risk factors for symptoms of cannabis, cocaine, alcohol,
caffeine, and nicotine dependence. Arch Gen Psychiatry, 64(11), 1313-
1320.

Kendler, K. S., Prescott, C. A., Neale, M. C., 8: Pedersen, N. L. (1997).
Temperance board registration for alcohol abuse in a national sample
of Swedish male twins, born 1902 to 1949. Arch Gen Psychiatry, 54(2),
178-184.

Kessler, R. C., Davis, C. G., 8: Kendler, K. S. (1997). Childhood adversity and
adult psychiatric disorder in the US National Comorbidity Survey.
Psychol Med, 27(5), 1101-1119.

Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes, M.,
Eshleman, S., et a1. (1994). Lifetime and lZ-month prevalence of DSM-
IlI-R psychiatric disorders in the United States. Results from the National
Comorbidity Survey. Arch Gen Psychiatry, 51(1), 8-19.

Kessler, R. C., 8: Ustun, T. B. (2004). The World Mental Health (WMH) Survey
Initiative Version of the World Health Organization (WHO) Composite
International Diagnostic Interview (CIDI). Int [ Methods Psychiatr Res,
13(2), 93-121.

Keyes, K. M., Grant, B. F., 8: Hasin, D. S. (2007). Evidence for a closing gender
gap in alcohol use, abuse, and dependence in the United States
population. Drug Alcohol Depend.

Keyes, K. M., 8: Hasin, D. S. (2008). Socio-economic status and problem
alcohol use: the positive relationship between income and the DSM-IV
alcohol abuse diagnosis. Addiction, 103(7), 1120-1130.

Kim, J. H., Lee, S., Chow, J., Lau, J., Tsang, A., Choi, J., et a1. (2008). Prevalence
and the Factors Associated with Binge Drinking, Alcohol Abuse, and
Alcohol Dependence: A Population-Based Study of Chinese Adults in
Hong Kong. Alcohol Alcohol.

King, K. M., 8: Chassin, L. (2007). A prospective study of the effects of age of

initiation of alcohol and drug use on young adult substance
dependence. [ Stud Alcohol Drugs, 68(2), 256-265.

244

 

Kloner, R. A., 8: Rezkalla, S. H. (2007). To drink or not to drink? That is the
question. Circulation, 116(1 1), 1306-1317.

Koob, G. F. (2006). The neurobiology of addiction: a neuroadaptational view
relevant for diagnosis. Addiction, 10] Suppl 1, 23-30.

Kopnisky, K. L., 8: Hyman, S. E. (2002). Neurcpsychcpharmacclcgy: The Fifth
Generation of Progress (1 ed.). Philadelphia: Lippincott Williams 8:
Wilkins.

Koss, M. P., Yuan, N. P., Dightman, D., Prince, R. J., Polacca, M., Sanderson, B.,
et al. (2003). Adverse childhood exposures and alcohol dependence
among seven Native American tribes. Am [Prev Med, 25(3), 238-244.

Koutakis, N ., Stattin, H., 8: Kerr, M. (2008). Reducing youth alcohol drinking
through a parent-targeted intervention: the Orebro Prevention Program.
Addiction, 103(10), 1629-1637.

 

Kuehn, M. R., Bradley, A., Robertson, E. J., 8: Evans, M. J. (1987). A potential
animal model for Lesch-Nyhan syndrome through introduction of HPRT
mutations into mice. Nature, 326(6110), 295-298.

Kunitz, S. J., Levy, J. E., McCloskey, J., 8: Gabriel, K. R. (1998). Alcohol
dependence and domestic violence as sequelae of abuse and conduct
disorder in childhood. Child Abuse Neg], 22(1 1), 1079-1091.

Kuo, P. H., Aggen, S. H., Prescott, C. A., Kendler, K. S., 8: Neale, M. C. (2008).
Using a factor mixture modeling approach in alcohol dependence in a
general population sample. Drug Alcohol Depend, 98(1-2), 105-114.

Lee, M. M., Whittemore, A. S., 8: Lung, D. L. (1992). Reliability of recalled
physical activity, cigarette smoking, and alcohol consumption. Ann
Epidemic], 2(5), 705-714.

Lewis, M. J. (1996). Alcohol reinforcement and neuropharmacological
therapeutics. Alcohol Alcohol Suppl, 1, 17-25.

Libby, A. M., Orton, H. D., Novins, D. K., Spicer, P., Buchwald, D., Beals, J., et
al. (2004). Childhood physical and sexual abuse and subsequent
alcohol and drug use disorders in two American-Indian tribes. [ Stud
Alcohol, 65(1), 74-83.

Lin, Y., Kikuchi, S., Tamakoshi, A., Wakai, K., Kawamura, T., Iso, H., et a1.

(2005). Alcohol consumption and mortality among middle-aged and
elderly Japanese men and women. Ann Epidemic], 15(8), 590-597.

245

Liu, J., Lewohl, J. M., Harris, R. A., Iyer, V. R., Dodd, P. R., Randall, P. K., et a1.
(2006). Patterns of gene expression in the frontal cortex discriminate
alcoholic from nonalcoholic individuals. Neuropsychopharmacology,
31(7), 1574-1582.

Liu, S., Serdula, M. K., Byers, T., Williamson, D. F., Mokdad, A. H., 8: Flanders,
W. D. (1996). Reliability of alcohol intake as recalled from 10 years in
the past.Am [Epidemic], 143(2), 177-186.

Looby, A. (2008). Childhood attention deficit hyperactivity disorder and the
development of substance use disorders: valid concern or '-
exaggeration? Addict Behav, 33(3), 451-463.

Lopez, A. D., Mathers, C. D., Ezzati, M., Jamison, D. T., 8: Murray, C. J. (2006).
Global and regional burden of disease and risk factors, 2001:
systematic analysis of population health data. Lancet, 367 (9524), 1747- i

 

1757.

Luo, Y., Parish, W. L., 8: Laumann, E. O. (2008). A population-based study of
childhood sexual contact in China: Prevalence and long-term
consequences. Child Abuse Neg], 32(7), 721-731 .

MacMillan, H. L., Boyle, M. H., Wong, M. Y., Duku, E. K., Fleming, J. E., 8:
Walsh, C. A. (1999). Slapping and spanking in childhood and its
association with lifetime prevalence of psychiatric disorders in a
general population sample. Cmaj, 161(7), 805-809.

MacMillan, H. L., Fleming, J. E., Streiner, D. L., Lin, E., Boyle, M. H., Jamieson,
E., et a1. (2001). Childhood abuse and lifetime psychopathology in a
community sample. Am [ Psychiatry, 158(11), 1878-1883.

Mathers, C., 8: Ayuso-Mateos, J. L. (2000). Global burden of alcohol use
disorders in the Year 2000: summary of methods and data sources.
Geneva: WHO.

Mathers, C. D., Bernard, C., Iburg, K., Inoue, M., Ma, F. D., Shibuya, K., et al.
(2003). The Global Burden of Disease in 2002: data sources, methods and
results. Geneva: World Health Organization

Mathers, C. D., 8: Loncar, D. (2006). Projections of global mortality and burden
of disease from 2002 to 2030. PLcS Med, 3(11), e442.

McBride, 0., Teesson, M., Slade, T., Hasin, D., Degenhardt, L., 8: Baillie, A.
(2008). Further evidence of differences in substance use and

246

dependence between Australia and the United States. Drug Alcohol
Depend.

McGovern, P. E., Zhang, J., Tang, J., Zhang, Z., Hall, G. R., Moreau, R. A., et a1.
(2004). Fermented beverages of pre- and proto-historic China. Proc
Natl Acad Sci U S A, 101 (51), 17593-17598.

Medina-Mora, M. E., Borges, G., Benjet, C., Lara, C., 8: Berglund, P. (2007).
Psychiatric disorders in Mexico: lifetime prevalence in a nationally
representative sample. Br] Psychiatry, 190, 521-528.

Merikangas, K. R. (1990). The genetic epidemiology of alcoholism. Psychol
Med, 20(1), 11-22.

Merikangas, K. R., Leckman, J. F., Prusoff, B. A., Pauls, D. L., 8: Weissman, M. M.
(1985). Familial transmission of depression and alcoholism. Arch Gen
Psychiatry, 42(4), 367-372.

Merikangas, K. R., Risch, N. J., 8: Weissman, M. M. (1994). Comorbidity and
co-transmission of alcoholism, anxiety and depression. Psycho] Med,
24(1), 69-80.

Michalak, L., Trocki, K., 8: Bond, J. (2007). Religion and alcohol in the US.
National Alcohol Survey: how important is religion for abstention and
drinking? Drug Alcohol Depend, 87 (2-3), 268-280.

Mihic, S. J., Ye, 0., Wick, M. J., Koltchine, V. V., Krasowski, M. D., Finn, S. E.,
et al. (1997). Sites of alcohol and volatile anaesthetic action on GABA(A)
and glycine receptors. Nature, 389(6649), 385-389.

Miller, P., Plant, M., 8: Plant, M. (2005). Spreading out or concentrating weekly
consumption: alcohol problems and other consequences within a UK
population sample. Alcohol Alcohol, 40(5), 461-468.

Molina, B. S., Pelham, W. E., Gnagy, E. M., Thompson, A. L., 8: Marshal, M. P.
(2007). Attention—deficit/hyperactivity disorder risk for heavy drinking
and alcohol use disorder is age specific. Alcohol Clin Exp Res, 31(4),
643-654.

Moss, H. B., Chen, C. M., 8: Yi, H. Y. (2008). DSM-IV criteria endorsement
patterns in alcohol dependence: relationship to severity. Alcohol Clin
Exp Res, 32(2), 306-313.

Moss, H. B., 8: Kirisci, L. (1995). Aggressivity in adolescent alcohol abusers:
relationship with conduct disorder. Alcohol Clin Exp Res, 19(3), 642-646.

247

 

 

 

Mullings, J. L., Hartley, D. J., 8: Marquart, J. W. (2004). Exploring the
relationship between alcohol use, childhood maltreatment, and
treatment needs among female prisoners. Subst Use Misuse, 39(2), 277-
305.

Muthen, B. O., Hasin, D., 8: Wisnicki, K. S. (1993). Factor analysis of ICD-IO
symptom items in the 1988 National Health Interview Survey on Alcohol
Dependence. Addiction, 88(8), 107 l- 1077 .

Muthen, B. O., 8: Muthen, L. K. (2007). Mplus user's guide (4.1 ed.). Los Angeles:
Muthén 8: Muthén. Fr:

Myers, M. G., Brown, S. A., 8: Mott, M. A. (1995). Preadolescent conduct
disorder behaviors predict relapse and progression of addiction for
adolescent alcohol and drug abusers. Alcohol Clin Exp Res, 19(6), 1528-
1536.

 

Naimi, T. S., Brewer, R. D., Mokdad, A., Denny, C., Serdula, M. K., 8: Marks, J. S. i]
(2003). Binge drinking among US adults. [ama, 289(1), 70—75.

National Bureau of Statistics of China. (2008). Yearbook of Chinese statistics.
Retrieved. from.

Nelson, C. B., Rehm, J., Ustun, T. B., Grant, B., 8: Chatterji, S. (1999). Factor
structures for DSM-IV substance disorder criteria endorsed by alcohol,
cannabis, cocaine and opiate users: results from the WHO reliability
and validity study. Addiction, 94(6), 843-855.

Newman, I. (2002). Cultural aspects of drinking patterns and alcohol controls
in China. The Globe(l).

Newton, P. M., 8: Messing, R. O. (2006). Intracellular signaling pathways that
regulate behavioral responses to ethanol. Pharmaco] Ther, 109(1-2),
227-237.

NIAAA. (2004). NIAAA Council Approves Definition of Binge Drinking (V 01. 3).
Bethesda: National Institute on Alcohol Abuse and Alcoholism.

O'Brien, M. S., 8: Anthony, J. C. (2005). Risk of becoming cocaine dependent:

epidemiological estimates for the United States, 2000-2001.
Neurcpsychcpharmacclcgy, 30(5), 1006-1018.

248

O'Grady, K. E., Arria, A. M., Fitzelle, D. M., 8: Wish, E. D. (2008). Heavy
Drinking and Polydrug Use among College Students. [Drug Issues,
38(2), 445-466.

Oscar-Berman, M., 8: Marinkovic, K. (2007). Alcohol: effects on
neurobehavicral functions and the brain. Neurcpsychcl Rev, 17(3), 239-
257.

Parascandola, M., 8: Weed, D. L. (2001). Causation in epidemiology. [
Epidemic] Community Health, 55(12), 905-912.

Park, J. T., Kim, B. G., 8: Jhun, H. J. (2008). Alcohol consumption and the CAGE
questionnaire in Korean adults: results from the Second Korea National
Health and Nutrition Examination Survey. [ Korean Med Sci, 23(2), 199-
206.

Park, J. Y., Huang, Y. H., Nagoshi, C. T., Yuen, S., Johnson, R. C., Ching, C. A.,
et a1. (1984). The flushing response to alcohol use among Koreans and
Taiwanese. [Stud Alcohol, 45(6), 481-485.

 

Peng, G. S., Chen, Y. C., Tsao, T. P., Wang, M. F., 8: Yin, S. J. (2007).
Pharmacokinetic and pharmacodynamic basis for partial protection
against alcoholism in Asians, heterozygous for the variant ALDH2*2
gene allele. Pharmaccgenet Gencrnics, 17(10), 845-855.

Perry, C. L., Grant, M., Ernberg, G., Florenzano, R. U., Langdon, M. C., Myeni,
A. D., et a1. (1989). WHO Collaborative Study on Alcohol Education and
Young People: outcomes of a four-country pilot study. Int [ Addict,
24(12), 1145-1171.

Petraitis, J., Flay, B. R., 8: Miller, T. O. (1995). Reviewing theories of adolescent
substance use: organizing pieces in the puzzle. Psychol Bull, 117(1), 67-
86.

Plant, A. M., 8: Pirie, F. (1979). Self-reported alcohol consumption and alcohol-
related problems: A study in four Scottish towns. Social Psychiatry and
Psychiatric Epidemiology, 14(2).

Poduska, J. M., Kellam, S. G., Wang, W., Brown, C. H., Ialongo, N. S., 8:
Toyinbo, P. (2008). Impact of the Good Behavior Game, a universal
classroom-based behavior intervention, on young adult service use for
problems with emotions, behavior, or drugs or alcohol. Drug Alcohol
Depend, 95 Suppl 1, $29-44.

249

Pollock, V. E., Briere, J., Schneider, L., Knop, J., Mednick, S. A., 8: Goodwin, D.
W. (1990). Childhood antecedents of antisocial behavior: parental
alcoholism and physical abusiveness. Am [ Psychiatry, 147(10), 1290-
1293.

Power, 0., Rodgers, B., 8: Hope, S. (1999). Heavy alcohol consumption and
marital status: disentangling the relationship in a national study of
young adults. Addiction, 94(10), 1477-1487.

Prescott, C. A., Aggen, S. H., 8: Kendler, K. S. (2000). Sex-specific genetic
influences on the comorbidity of alcoholism and major depression in a
population-based sample of US twins. Arch Gen Psychiatry, 57(8), 803-
81 1.

Prescott, C. A., 8: Kendler, K. S. (2001). Associations between marital status
and alcohol consumption in a longitudinal study of female twins. [Stud
Alcohol, 62(5), 589-604.

Prescott, C. A., Sullivan, P. F., Myers, J. M., Patterson, D. G., Devitt, M.,
Halberstadt, L. J., et a1. (2005). The Irish Affected Sib Pair Study of
Alcohol Dependence: study methodology and validation of diagnosis
by interview and family history. Alcohol Clin Exp Res, 29(3), 417-429.

Proudfoot, H., Baillie, A. J., 8: Teesson, M. (2006). The structure of alcohol
dependence in the community. Drug Alcohol Depend, 81(1), 21-26.

Quine, S., 8: Stephenson, J. A. (1990). Predicting smoking and drinking
intentions and behavior of pre-adolescents: The influence of parents,
siblings, and peers. .Fam Sys Med. , 8, 191-200.

Ouitkin, F. M., Rifl<in, A., Kaplan, J., 8: Klein, D. F. (1972). Phobic anxiety
syndrome complicated by drug dependence and addiction. A treatable
form of drug abuse. Arch Gen Psychiatry, 27(2), 159-162.

Racz, 1., Bilkei-Gorzo, A., Toth, 2. E., Michel, K., Palkovits, M., 8: Zimmer, A.
(2003). A critical role for the cannabinoid CB1 receptors in alcohol
dependence and stress-stimulated ethanol drinking. [Neurcsci, 23(6),
2453-2458.

Radouco-Thomas, S., Garcin, F., Laperriere, A., Marquis, P. A., Lambert, J.,
Denver, J., et al. (1979). Genetic epidemiology and the prevention of
functional mental disorders and alcoholism: family study and biological
predictors. Prcg Neurcpsychcpharmaccl, 3(1-3), 165-189.

250

 

 

 

Rahav, G., Wilsnack, R., Bloomfield, K., Gmel, G., 8: Kuntsche, S. (2006). The
influence of societal level factors on men's and women's alcohol
consumption and alcohol problems. Alcohol Alcohol Suppl, 41(1), i47-55.

Reed, P. L., Anthony, J. C., 8: Breslau, N. (2007). Incidence of drug problems in
young adults exposed to trauma and posttraumatic stress disorder: do
early life experiences and predispositions matter? Arch Gen Psychiatry,
64(12), 1435-1442.

Regier, D. A., Farmer, M. E., Rae, D. S., Locke, B. Z., Keith, S. J., Judd, L. L., et
a1. (1990). Comorbidity of mental disorders with alcohol and other drug
abuse. Results from the Epidemiologic Catchment Area (ECA) Study.
[ama, 264(19), 2511-2518.

Regier, D. A., Farmer, M. E., Rae, D. S., Myers, J. K., Kramer, M., Robins, L. N.,
et a1. (1993). One-month prevalence of mental disorders in the United
States and sociodemographic characteristics: the Epidemiologic
Catchment Area study. Acta Psychiatr Scand, 88(1), 35-47.

Rehm, J., Room, R., van den Brink, W., 8: Jacobi, F. (2005). Alcohol use
disorders in EU countries and Norway: an overview of the
epidemiology. Eur Neurcpsychcpharmaccl, 15(4), 377-388.

Rehm, J., Taylor, B., 8: Patra, J. (2006). Volume of alcohol consumption,
patterns of drinking and burden of disease in the European region 2002.
Addiction, 101(8), 1086-1095.

Reidpath, D. D., Allotey, P. A., Kouame, A., 8: Cummins, R. A. (2003).
Measuring health in a vacuum: examining the disability weight of the
DALY. Health Policy Plan, 1 8(4), 351-356.

Rhea, S. A., Nagoshi, C. T., 8: Wilson, J. R. (1993). Reliability of sibling reports
on parental drinking behaviors. [ Stud Alcohol, 54(1), 80-84.

Riggs, S., Alario, A. J., 8: McHorney, C. (1990). Health risk behaviors and
attempted suicide in adolescents who report prior maltreatment. [
Pediatr, 1 16(5), 815-821.

Robins, L. N., Locke, B. Z., 8: Regier, D. A. (1991). An overview of Psychiatric
disorders in America. New York: The Free Press.

Robinson, J., Sareen, J., Cox, B. J., 8: Bolton, J. (2009). Self-medication of

anxiety disorders with alcohol and drugs: Results from a nationally
representative sample. [Anxiety Discrd, 23(1), 38-45.

251

Room, R. (2006). Taking account of cultural and societal influences on
substance use diagnoses and criteria. Addiction, 101 Suppl 1, 31-39.

Room, R., Babor, T., 8: Rehm, J. (2005). Alcohol and public health. Lancet,
365(9458), 519-530.

Room, R., Schmidt, L., Rehm, J., 8: Makela, P. (2008). International regulation of
alcohol. ij, 337, a2364.

Rose, G. (2001). Sick individuals and sick populations. Int [ Epidemic], 30(3),
427-432; discussion 433-424.

Ross, C. A., Keyes, B. B., Xiao, 2., Yan, H., Wang, Z., Zou, Z., et a1. (2005).
Childhood physical and sexual abuse in China. [ Child Sex Abus, 14(4),
1 15-126.

Rothman, E. F., DeJong, W., Palfai, T., 8: Saitz, R. (2008). Relationship of age of
first drink to alcohol-related consequences among college students
with unhealthy alcohol use. SubstAbus, 29(1), 33-41.

Rounsaville, B. J., Spitzer, R. L., 8: Williams, J. B. (1986). Proposed changes in
DSM-III substance use disorders: description and rationale. Am [
Psychiatry, 143(4), 463-468.

Rowe, M., Bellamy, C., Baranoski, M., Wieland, M., O'Connell, M. J., Benedict,
P., et al. (2007). A peer-support, group intervention to reduce
substance use and criminality among persons with severe mental
illness. Psychiatr Serv, 58(7), 955-961 .

Ruchkin, V., Gilliam, W. S., 8: Mayes, L. (2008). Developmental pathway
modeling in considering behavior problems in young Russian children.
Child Psychiatry Hum Dev, 39(1), 49-66.

Ruixing, Y., Jiaqiang, D., Dezhai, Y., Weixiong, L., Shangling, P., Jinzhen, W.,
et a1. (2006). Effects of demographic characteristics, health-related
behaviors and lifestyle factors on the prevalence of hypertension for
the middle-aged and elderly in the Guangxi Hei Yi Zhuang and Han
populations. Kidney Blood Press Res, 29(5), 312-320.

Saha, T. D., Chou, S. P., 8: Grant, B. F. (2006). Toward an alcohol use disorder
continuum using item response theory: results from the National
Epidemiologic Survey on Alcohol and Related Conditions. Psychol Med,
36(7), 931-941.

Schinke, S. P., Schwinn, T. M., Di Noia, J., 8: Cole, K. C. (2004). Reducing the
risks of alcohol use among urban youth: three-year effects of a

252

computer-based intervention with and without parent involvement. [
Stud Alcohol, 65(4), 443-449.

Schuckit, M. A. (1985). The clinical implications of primary diagnostic groups
among alcoholics. Arch Gen Psychiatry, 42(11), 1043-1049.

Schuckit, M. A. (2009a). Alcohol-use disorders. Lancet, 373(9662), 492-501.

Schuckit, M. A. (2009b). An overview of genetic influences in alcoholism. [
SubstAbuse Treat, 36(1), S5-l4.

Schuckit, M. A., Daeppen, J. B., Tipp, J. E., Hesselbrock, M., 8: Bucholz, K. K.
(1998). The clinical course of alcohol-related problems in alcohol
dependent and nonalcohol dependent drinking women and men. [Stud
Alcohol, 59(5), 581-590.

Schuckit, M. A., 8: Smith, T. L. (2006). An evaluation of the level of response to
alcohol, externalizing symptoms, and depressive symptoms as
predictors of alcoholism. [ Stud Alcohol, 67(2), 215-227.

Schuckit, M. A., Smith, T. L., Danko, G. P., Anderson, K. G., Brown, S. A.,
Kuperman, S., et a1. (2005). Evaluation of a level of response to alcohol-
based structural equation model in adolescents. [ Stud Alcohol, 66(2),

1 74-184.

Schuckit, M. A., Smith, T. L., Danko, G. P., Bucholz, K. K., Reich, T., 8: Bierut, L.
(2001). Five-year clinical course associated with DSM-1V alcohol abuse
or dependence in a large group of men and women. Am [ Psychiatry,
158(7), 1084-1090.

Schuckit, M. A., Smith, T. L., 8: Kalmijn, J. (2004). The search for genes
contributing to the low level of response to alcohol: patterns of findings
across studies. Alcohol Clin Exp Res, 28(10), 1449-1458.

Schutz, C. G. (1995). Drug use in Asia. In D. H. Jaffe (Ed.), Encyclopedia of
drugs and alcohol. New York, NY: Macmillian Library Reference.

Serdula, M. K., Brewer, R. D., Gillespie, C., Denny, C. H., 8: Mokdad, A. (2004).
Trends in alcohol use and binge drinking, l985-l999: results of a multi-
state survey. Am [Prev Med, 26(4), 294-298.

Shen, M., Chai, J., Yang, B., Huang, 8., 8: Yan, J. (2003). llre World Mental

Health Survey in China: An Overview of Design and Field Procedures.
Beijing: Research Center for Contemporary China, Peking University.

253

 

Shen, Y. C., Fan, J. H., Edenberg, H. J., Li, T. K., Cui, Y. H., Wang, Y. F., et al.
(1997). Polymorphism of ADH and ALDH genes among four ethnic
groups in China and effects upon the risk for alcoholism. Alcohol Clin
Exp Res, 21(7), 1272-1277.

Shen, Y. C., Zhang, M. Y., Huang, Y. 0., He, Y. L., Liu, Z. R., Cheng, H., et a1.
(2006). 'I'welve-month prevalence, severity, and unmet need for
treatment of mental disorders in metropolitan China. Psycho] Med, 36(2),
257-267.

Sher, K. J., 8: Descutner, C. (1986). Reports of paternal alcoholism: reliability
across siblings. Addict Behav, 11(1), 25-30.

Sher, K. J., Gershuny, B. 8., Peterson, L., 8: Raskin, G. (1997). The role of
childhood stressors in the intergenerational transmission of alcohol use
disorders. [ Stud Alcohol, 58(4), 414-427.

Simpson, T. L., 8: Miller, W. R. (2002). Concomitance between childhood
sexual and physical abuse and substance use problems. A review. Clin
Psycho] Rev, 22(1), 27-77.

Sloan, F. A., Reilly, B. A., 8: Schenzler, C. (1994). Effects of prices, civil and
criminal sanctions, and law enforcement on alcohol-related mortality. [
Stud Alcohol, 55(4), 454-465.

Smith, G. W., 8: Shevlin, M. (2008). Patterns of alcohol consumption and
related behaviour in Great Britain: a latent class analysis of the alcohol
use disorder identification test (AUDIT). Alcohol Alcohol, 43(5), 590-594.

Solomon, R. L., 8: Corbit, J. D. (1974). An opponent-process theory of
motivation. 1. Temporal dynamics of affect. Psychol Rev, 81(2), 119-145.

Soueif, M. 1., Yunis, F. A., 8: Taha, H. S. (1986). Extent and patterns of drug
abuse and its associated factors in Egypt. Bull Narc, 38(1-2), 113-120.

Stabenau, J. R. (1984). Implications of family history of alcoholism, antisocial
personality, and sex differences in alcohol dependence. Am [ Psychiatry,
141(10), 1178-1182.

Stafstrom, M., Ostergren, P. O., Larsson, S., Lindgren, B., 8: Lundborg, P.
(2006). A community action programme for reducing harmful drinking
behaviour among adolescents: the l'relleborg Project. Addiction, 101(6),
813-823.

254

Stallings, M. C., Cherny, S. 5., Young, S. E., Miles, D. R., Hewitt, J. K., 8: Fulker,
D. W. (1997). The familial aggregation of depressive symptoms,
antisocial behavior, and alcohol abuse. Am [Med Genet, 74(2), 183-191.

StataCorp LP. (2005). Stata Manual, . College Station, Texas: A Stata Press
Publication.

Swett, C., Jr., Cohen, C., Surrey, J., Compaine, A., 8: Chavez, R. (1991). High
rates of alcohol use and history of physical and sexual abuse among
women outpatients. Am [Drug Alcohol Abuse, 17(1), 49-60.

Teicher, M. H., Tomoda, A., 8: Andersen, S. L. (2006). Neurobiological
consequences of early stress and childhood maltreatment: are results
from human and animal studies comparable? Ann N YAcad Sci, 1071,
3 13-323.

Tevyaw, T. O., Borsari, B., Colby, S. M., 8: Monti, P. M. (2007). Peer
enhancement of a brief motivational intervention with mandated
college students. Psycho] Addict Behav, 21(1), 114-1 19.

Thomasson, H. R., Edenberg, H. J., Crabb, D. W., Mai, X. L., Jerome, R. E., Li, T.
K., et a1. (1991). Alcohol and aldehyde dehydrogenase genotypes and
alcoholism in Chinese men. Am [Hum Genet, 48(4), 677-681.

Tremblay, R. E., Nagin, D. 3., Seguin, J. R., chcolillo, M., Zelazo, P. D., Boivin,
M., et al. (2004). Physical aggression during early childhood:
trajectories and predictors. Pediatrics, 114(1), e43-50.

Treno, A. J., Gruenewald, P. J., Lee, J. P., 8: Remer, L. G. (2007). The
Sacramento Neighborhood Alcohol Prevention Project: outcomes from
a community prevention trial. [ Stud Alcohol Drugs, 68(2), 197-207.

Trent, L., Stander, V., Thomsen, C., 8: Merrill, L. (2007). Alcohol abuse among
US. Navy recruits who were maltreated in childhood. Alcohol Alcohol,
42(4), 370-375.

Trim, R. S., Leuthe, E., 8: Chassin, L. (2006). Sibling influence on alcohol use in
a young adult, high-risk sample. [ Stud Alcohol, 67(3), 391-398.

Turner, J., Perkins, H. W., 8: Bauerle, J. (2008). Declining negative
consequences related to alcohol misuse among students exposed to a
social norms marketing intervention on a college campus. [Am Coll

Health, 57(1), 85-94.

255

U.S.EPA. (1994) . CHEMCAL SUlVHVIARYFOR ACETALDEHYDE. Retrieved 02-19-
2008. from http://www.epa.gov/chemfact/s_acetal.txt.

Uhl, G. R., Drgon, T., Johnson, C., Fatusin, O. 0., Liu, 0. R., Contoreggi, C., et
al. (2008). "Higher order" addiction molecular genetics: convergent
data from genome-wide association in humans and mice. Biochem
Pharmacol, 75(1), 98-111.

United States Department of Health, E. a. W. (1964). Report of the advisory
committee to the Surgeon General. Retrieved. from.

Vaillant, G. E. (1996). A long-term follow-up of male alcohol abuse. Arch Gen
Psychiatry, 53(3), 243-249.

Vaillant, G. E. (2003). A 60-year follow-up of alcoholic men. Addiction, 98(8),
1 043- 1 05 1 .

Vaillant, G. E., 8: Milofsky, E. S. (1982). Natural history of male alcoholism. IV.
Paths to recovery. Arch Gen Psychiatry, 39(2), 127-133.

van de Wiel, A. (2004). Diabetes mellitus and alcohol. Diabetes Metab Res Rev,
20(4), 263-267.

van Lier, P. A., Huizink, A., 8: Crijnen, A. (2008). Impact of a preventive
intervention targeting childhood disruptive behavior problems on
tobacco and alcohol initiation from age 10 to 13 years. Drug Alcohol
Depend.

Vicente, B., Kohn, R., Rioseco, P., Saldivia, S., Levav, 1., 8: Torres, S. (2006).
Lifetime and lZ-month prevalence of DSM-III-R disorders in the Chile
psychiatric prevalence study. Am [ Psychiatry, 163(8), 1362-1370.

Walsh, C. A., Macmillan, H. L., Trocme, N., Jamieson, E., 8: Boyle, M. H. (2008).
Measurement of victimization in adolescence: Development and
validation of the Childhood Experiences of Violence Questionnaire.
Child Abuse Neg].

Warner, L. A., 8: White, H. R. (2003). Longitudinal effects of age at onset and
first drinking situations on problem drinking. Subst Use Misuse, 38(14),
1 983-20 16.

Watson, C. G., Hancock, M., Gearhart, L. P., Malovrh, P., Mendez, C., 8: Raden,

M. (1997). A comparison of the symptoms associated with early and late
onset alcohol dependence. ] Nerv Ment Dis, 185(8), 507-509.

256

Wei, H., Derson, Y., Xiao, S., Li, L., 8: Zhang, Y. (1999). Alcohol consumption
and alcohol-related problems: Chinese experience from six area
samples, 1994. Addiction, 94(10), 1467-1476.

Wei, H., Young, D., Lingjiang, L., Shuiyuan, K., Jian, T., Hanshu, S., et a1. (1995).
Psychoactive substance use in three sites in China: gender differences
and related factors. Addiction, 90(11), 1503-1515.

Werner, D. F., Blednov, Y. A., Ariwodola, O. J., Silberman, Y., Logan, E., Berry,
R. B., et a1. (2006). Knockin mice with ethanol-insensitive alphal-
containing gamma-aminobutyric acid type A receptors display
selective alterations in behavioral responses to ethanol. [Pharmacol

Exp Ther, 319(1), 219-227.

WHO. (1967). Manual of the International Statistical Classification of Diseases,
Injuries, and Causes of Death, Eighth Revision. . Geneva: World Health i
Organization,.

WHO. (1992). International Statistical Classification of Diseases and Related
Health Problems (10 ed.). Geneva: World Health Organization.

WHO. (2003). Global Alcohol Database (GAD) Country Profile-China: WHO.

WHO. (2004a). The global burden of disease: 2004 update Geneva: WHO.

WHO. (2004b). WHO Global Status Report on Alcohol 2004. In W. H.
Organization (Ed.). Geneva: World Health Organization. Department of
Mental Health and Substance Abuse.

WHO. (2005). Alcohol, Gender and Drinla'ng Problems: Perspectives from Low
and Middle Income Countries. Geneva: WHO.

WHO (Cartographer). (2009). Burden of disease attributable to: Alcohol [map].

WHO expert committee on addiction-producing drugs. (1964). Technical
report series. No.273. . Geneva: WHO.

Widom, C. S. (1989). The cycle of violence. Science, 244(4901), 160-166.

Widom, C. S., 8: Hiller-Sturmhofel, S. (2001). Alcohol abuse as a risk factor for
and consequence of child abuse. Alcohol Res Health, 25(1), 52-57.

Widom, C. S., White, H. R., Czaja, S. J., 8: Marmorstein, N. R. (2007). Long-term

effects of child abuse and neglect on alcohol use and excessive
drinking in middle adulthood. [ Stud Alcohol Drugs, 68(3), 317-326.

257

Wightman, R. M., 8: Robinson, D. L. (2002). Transient changes in mesolimbic
dopamine and their association with 'reward'. [ Neurcchem, 82(4), 721-
735.

Williams, D. M. (1998 ). Alcohol indulgence in a Mongolian community of
China. Bulletin of Ccncemed Asian Scholars, 30(1), 13-22.

Wilsnack, R. W., Vogeltanz, N. D., Wilsnack, S. C., Harris, T. R., Ahlstrom, S.,
Bondy, S., et a1. (2000). Gender differences in alcohol consumption and
adverse drinking consequences: cross-cultural patterns. Addiction,
95(2), 251-265.

Windle, M., Windle, R. C., Scheidt, D. M., 8: Miller, G. B. (1995). Physical and
sexual abuse and associated mental disorders among alcoholic
inpatients.Am [ Psychiatry, 152(9), 1322-1328.

Wood, M. D., Read, J. P., Mitchell, R. E., 8: Brand, N. H. (2004). Do parents still
matter? Parent and peer influences on alcohol involvement among
recent high school graduates. Psycho] Addict Behav, 18(1), 19-30.

Wooldridge, J. M. (2002). Econometric Analysis of Cross Section and Panel Data.
London, England: The MIT Press.

World Health Organization. (1977). Manual of the International Statistical
Classification of Diseases, Injuries, and Causes of Death, Nineth Revision. .
Geneva: World Health Organization,.

World Health Organization. (1999). WHO Global NCD InfoBase. WHO Monica
Project: World Health Organization. Geneva. .

Wu, N. S., Lu, Y., Sterling, S., 8: Weisner, C. (2004). Family environment
factors and substance abuse severity in an HMO adolescent treatment
population. Clin Pediatr(Phi1a), 43(4), 323-333.

Young, S. Y., Hansen, C. J., Gibson, R. L., 8: Ryan, M. A. (2006). Risky alcohol
use, age at onset of drinking, and adverse childhood experiences in
young men entering the US Marine Corps. Arch PediatrAdclesc Med,
160(12),1207-1214.

Yuan, J. M., Ross, R. K., Gao, Y. T., Henderson, B. E., 8: Yu, M. C. (1997). Follow

up study of moderate alcohol intake and mortality among middle aged
men in Shanghai, China. ij, 314(7073), 18-23.

258

 

 

Zhang, J., Casswell, S., 8: Cai, H. (2008). Increased drinking in a metropolitan
city in China: a study of alcohol consumption patterns and changes.
Addiction, 1 03(3), 416-423.

Zhang, J., 8: Yu, K. F. (1998). What's the relative risk? A method of correcting
the odds ratio in cohort studies of common outcomes. [ama, 280(19),
1 690-1 69 1 .

Zhao, C., Liu, Z., Zhao, D., Liu, Y., Liang, J., Tang, Y., et a1. (2004). Drug abuse
in China. Ann N YAcad Sci, 1025, 439-445.

Zhou, H., Deng, J., Li, J., Wang, Y., Zhang, M., 8: He, H. (2003). Study of the
relationship between cigarette smoking, alcohol drinking and cognitive
impairment among elderly people in China. Age Ageing, 32(2), 205-210.

Zhou, 11., Su, Z., Deng, H., Xiang, K., Chen, H., 8: Hao, W. (2006). A
comparative survey on alcohol and tobacco use in urban and rural
populations in the Huaihua District of Hunan Province, China. Alcohol,
39(2), 87-96.

Zlotnick, C., Johnson, D. M., Stout, R. L., Zywiak, W. H., Johnson, J. E., 8:

Schneider, R. J. (2006). Childhood abuse and intake severity in alcohol
disorder patients. [ Trauma Stress, 19(6), 949-959.

259