Content uploaded by Jeffrey T Parsons
Author content
All content in this area was uploaded by Jeffrey T Parsons on Aug 21, 2017
Content may be subject to copyright.
Correlates of depressive symptoms among
HIV-positive injection drug users: the role of
social support
Y
.
M
IZUNO
,
1
D
.
W
.
P
URCELL
,
1
C
.
D
AWSON-
R
OSE
,
2
J
.
T
.
P
ARSONS
3
&T
HE
SUDIS T
EAM
1
Centers for Disease Control and Prevention, Atlanta,
2
University of California-San Francisco &
3
Hunter College of the City University of New York, USA
Abstract Using cross-sectional data from an ethnically diverse sample of HIV-positive injection
drug users (IDUs), we sought to identify correlates of depressive symptoms. We were particularly
interested in whether perceived social support was associated with depression among HIV-positive
IDUs and whether social support buffered adverse effects of other correlates. Data were collected from a
sample of HIV-positive IDUs recruited from a variety of venues in the New York City and San
Francisco metropolitan areas in the USA. Multiple regression analysis identified four significant
correlates of depressive symptoms. Perceived social support and having a regular place for HIV
medical care were significantly associated with lower levels of depressive symptoms, while history of
mental health problems and non-injection polydrug use were significantly associated with higher levels
of depressive symptoms. Moreover, a significant interaction effect was found between social support
and non-injection polydrug use, indicating that social support buffers the association between non-
injection polydrug use and depression. These results suggest that increasing social support might be a
useful tool for HIV-positive IDUs in reducing depression and the adverse effect of non-injection
polydrug use.
Introduction
Injection drug users (IDUs) constitute 25% of adult and adolescent AIDS cases reported
through December 2001 in the USA. When ‘men who have sex with men and inject drugs’
and the sex partners of IDUs are included, injection drug use has been associated with 35% of
adult and adolescent AIDS cases (CDC, 2001). Research shows that some HIV-positive
IDUs continue to engage in risky sexual and drug using behaviours (Kwiatkowski & Booth,
1998; Rhodes et al ., 1993; Singh et al ., 1993). Thus HIV prevention research on HIV-positive
IDUs should be a priority in the USA.
Address for correspondence: Yuko Mizuno, Prevention Research Branch, Division of HIV/AIDS Prevention*/
Intervention Research and Support, Centers for Disease Control and Prevention, 1600 Clifton Road NE Mail Stop
E37, Atlanta, GA 30333, USA. Tel:
/1 404 639 1925; Fax: /1 404 639 1950; E-mail: ybm2@cdc.gov
AIDS CARE (October 2003), VOL. 15, NO. 5, pp. 689/698
ISSN 0954-0121 print/ISSN 1360-0451 online/03/050689-10 #US Government
DOI: 10.1080/09540120310001595177
Depression is an important problem to be addressed among HIV-positive IDUs. Studies
suggest increased rates of depression among people with HIV (Belkin et al ., 1992; Ciesla &
Roberts, 2001; Dew et al ., 1997). Also, high rates of depressive disorders have been reported
among IDUs (Rabkin et al., 1997). Depression is associated with substantial physical
morbidity and disability, as well as mental suffering (Lyketsos et al ., 1996). Some studies with
drug users indicate significant associations between depression and HIV/AIDS-risk taking
behaviours (Rahav et al., 1998; Hawkins et al ., 1998). Moreover, an association between
psychiatric morbidity and non-adherence to HIV medications has been found among HIV-
positive IDUs (e.g. Ferrando et al., 1996).
A number of studies have examined correlates of depression using samples of HIV-
positive persons (Belkin et al., 1992; Dew et al ., 1997; Ferrando et al ., 1998; Kalichman et al.,
2000; Lyketsos et al ., 1996; Moore et al ., 1999; Serovich et al ., 2000; Siegel et al ., 1997;
Singh et al., 1997; Van Servellen et al., 1998) and IDUs (Knowlton et al., 2000, 2001; Rabkin
et al., 1997). Many of these studies found an inverse association between perceived social
support and depression (Belkin et al ., 1992; Dew et al ., 1997; Kalichman et al., 2000;
Lyketsos et al., 1996; Rabkin et al ., 1997; Serovich et al ., 2000; Siegel et al ., 1997). Some of
these studies included HIV-positive IDUs, but none focused specifically on the relation
between social support and depression in this population.
Social support is an important construct to consider in designing interventions for HIV-
positive populations. Social support is defined as instrumental, emotional or informational
assistance from significant others, and has been found to be one of the major coping resources
for people experiencing stressful life events or chronic strains (Thoits, 1995). A distinction
can be made between perceived social support (the perception that support is available when
needed) and enacted social support (social support that is actually provided), with the former
consistently found to have protective effects on mental health (Thoits, 1995).
Some investigators have studied whether social support has an overall beneficial effect
(main effect model) or if it buffers or interacts with potentially adverse effects of other
variables such as stressful events and physical symptoms (buffering model) (Cohen & Willis,
1985). These studies tested whether the interaction between social support and stress, for
example, was present in a multivariate analysis. Findings from studies of persons with HIV or
AIDS (Siegel et al., 1997), as well as of general populations (see Cohen & Willis, 1985, for a
review), have been somewhat mixed. According to Cohen and Willis (1985) an optimal study
to conduct a comparative test of the main effect and the buffering models would use a large
sample size, instruments with acceptable psychometric properties, and measures of support
and other variables (e.g. stress) that are not highly correlated and, optimally, a longitudinal
design. While the present study did not meet the last criterion, it sufficiently satisfied the other
conditions to warrant the testing of a buffering hypothesis.
This paper sought to identify correlates of depressive symptoms, using cross-sectional
data on an ethnically diverse sample of HIV-positive IDUs recruited from a variety of
community settings in New York and San Francisco metropolitan areas in the USA. Unlike
other previous studies including some HIV-positive IDUs, this study exclusively focused on
HIV-positive IDUs. We were particularly interested in whether perceived social support was
associated with depression among HIV-positive IDUs and whether support buffered adverse
effects of other correlates. Drawing on findings from previous studies, potential correlates
examined were history of mental health problems, substance use, sociodemographic variables,
physical health status, as well as perceived social support. We also examined the associations
with previously untested variables such as receiving HIV medical care and taking HIV
medications as they could logically relieve emotional distress by promoting physical health.
690 Y. MIZUNO ET AL.
Methods
Participants
A sample of 161 HIV-positive IDUs was recruited in the New York City and San Francisco
metropolitan areas. To be included in the study, participants had to be at least 18 years old,
report having injected drugs and having had sex in the past year, self-identify as HIV-positive,
and live in the study area. Approximately equal numbers of male and female IDUs were
recruited from each metropolitan area in locations known to be frequented by IDUs, such as
needle exchanges, HIV health care settings, and other community-based organizations. In the
New York City area, participants were recruited from the Lower East Side of Manhattan and
Jersey City, New Jersey. In the San Francisco area, participants were recruited from East Palo
Alto, San Francisco, and the East Bay (Oakland and Richmond). As formative research
designed to help develop interventions for HIV-positive IDUs, we wanted the sample to be
diverse not only demographically, but also in terms of the primary drug used. Thus, within
each city, a maximum of 60% of participants could be primary heroin users. Because of
missing data on various variables, the final sample size used for multivariate analyses was 137.
Table 1 describes the characteristics of these 137 participants. Preliminary analysis found no
significant demographic differences (gender, age and race) between these 137 participants
and 24 participants who were excluded due to missing data.
Recruitment procedures
The CDC IRB and local human subject review boards approved the study protocol.
Participants were recruited from a variety of community settings through both active outreach
and passive recruiting (posters and flyers). All project materials mentioned that the study was
for HIV-positive IDUs and provided a toll-free number. When cards were handed directly to
someone, the recruiter would say: ‘If this does not apply to you, pass it on to someone else.’ In
this way, persons who took the cards were not identified as HIV-positive by accepting the card.
Some people were referred into the study by friends who had participated or passed the card
on to them. To participate, people called the toll-free telephone number, were screened for
eligibility, and those who were eligible were scheduled for participation. People were asked to
bring documentation of their HIV serostatus to be able to begin participating in the study at
their first visit. In addition, in San Francisco, a small oral fluid was collected and an Orasure
HIV test (Epitope Inc., Beverton, Oregon, USA) was performed for everyone to confirm their
HIV status.
Interview procedures
Participants completed an in-depth, face-to-face qualitative interview that took approximately
90 minutes, for which they received $20. They then completed a face-to-face quantitative
survey that took approximately 60 minutes and for which they were paid an additional $40.
The qualitative interview and the quantitative survey covered similar HIV-related topics. The
present analysis used data collected by the quantitative survey.
Measures
Dependent variables. Depressive symptoms were measured by the seven-item depression sub-
scale of the Brief Symptom Inventory (BSI; Derogatis & Spencer, 1982). The BSI is a 53-
CORRELATES OF DEPRESSION 691
item, standardized measure of psychological symptoms, problems and complaints. Respon-
dents were given a list of problems (e.g. thoughts of ending life, feeling lonely, feeling blue)
and were asked to indicate how much these problems had bothered them during the past week
including the date of the interview. Responses were scored from 1 (not at all) to 5 (extremely).
The reliability of the BSI sub-scales is well established, with Cronbach’s alphas ranging from
0.70 to 0.88 among counselling centre clients (Broday & Mason, 1991). Cronbach’s alpha for
the seven-item depression sub-scale was 0.85 in our sample.
Independent variables. Perceived social support was measured by a four-item scale. Two items
asked about perceived availability of instrumental support: ‘Is there anybody you could
Ta b l e 1 . Characteristics of participants (n/137)
Variables
Continuous variables MSD
Age (in years) 41.7 6.3
CD4 count 358.9 265.5
Frequency of non-injection drug use in last 30 days 98.2 383.7
Frequency of injection drug use in last 30 days 41.6 75.5
Frequency of drinking in last 30 days 20.5 33.4
Number of non-injection drugs used in last 30 days 1.5 1.4
Number of injection drugs used in last 30 days 1.7 1.1
Categorical variables n %
Gender
Male 74 54.0
Female 63 46.0
Race
White 30 21.9
African American 87 63.5
Hispanic 16 11.7
Other 4 2.9
Annual income
$5000 or more 81 59.1
Less than $5000 56 40.9
Education
High school or more 83 60.6
Less than high school 54 39.4
Work status
Employed 7 5.1
Unemployed 130 94.9
Metropolitan areas of recruitment
New York 72 52.6
San Francisco 65 47.4
History of mental health problems
Yes 75 54.7
No 62 45.3
Having a regular place for HIV medical care
Yes 133 97.1
No 4 2.9
On HIV medications
Yes 90 65.7
No 47 34.3
692 Y. MIZUNO ET AL.
depend on to loan you $10 if you need it?’ ‘Is there anybody you could depend on to take care
of you if you were sick and had to stay in bed?’ The remaining two items asked about
emotional support: ‘Are there people in your life you could talk to if you were sad, nervous or
depressed?’ ‘Are there people in your life you could talk to if you needed advice about a
personal problem?’ Responses were scored from 1 (definitely not) to 5 (definitely yes).
Cronbach’s alpha was 0.83.
Self-reported CD4 count was used to measure HIV-related physical health status. CD4
count was obtained by asking the questions: ‘What was your last CD4 or T-cell count? If you
don’t know the exact number, please give your best guess.’
Whether a respondent had a regular place to go for HIV care was assessed by the
question: ‘Do you have somewhere you go [a clinic or a doctor] regularly for your HIV care?’
Response alternatives were ‘yes’ or ‘no’.
HIV medication taking was measured by the question: ‘Are you currently taking any
medications prescribed by a doctor for HIV/AIDS? (This does not include vitamins or
alternative therapies or drugs such as medical marijuana).’ Again, response alternatives were
‘yes’ or ‘no’.
Drug use was assessed with a variety of measures. The effect of drug use can be hard to
observe if very little variation exists in this variable, which might be the case among a sample
of IDUs. Because of this possibility we examined frequency of drug use in the last 30 days
(total number of times drugs were used in the last 30 days) and polydrug use in the past 30
days (total number of different drugs used in the past 30 days). Frequency of injection drug
use was calculated by summing up the frequencies of using cocaine, crack, heroin, heroin/
stimulant mix, methadone, amphetamines, steroids, crushed or melted prescription drugs and
other drugs that were injected. For number of injection drugs used, a summary variable
reflecting the total number of different injection drugs used in the past 30 days was created.
Frequency and polydrug use variables were also created for non-injection drugs including
marijuana, crack, powder cocaine, heroin, amphetamines, methadone, nitrate inhalants
(‘poppers’), hallucinogens, PCP, downers and other non-injection drugs.
The extent of alcohol use was measured using a question asking about the frequency of
drinking alcohol in the past three months and a question asking about the number of drinks a
respondent typically had per day. The alcohol use variable was created to reflect the number
of drinks a respondent typically had in the past 30 days.
History of mental health problems was assessed with the following question: ‘Have you
had a significant period in your lifetime, that was not a direct result of alcohol or drug use, in
which you have experienced serious depression (hallucination/serious thoughts about suicide/
attempted suicide/been prescribed psychiatric medications)?’ For each of the questions,
response alternatives were ‘yes’ and ‘no’. These answers were then collapsed to indicate
whether or not a respondent had had any of these mental health problems.
Sociodemographic variables included gender, race, age, annual income, employment,
education and metropolitan area of recruitment. Age was a continuous variable measured in
years. The rest of the sociodemographic variables were categorical variables.
Analytic strategy
Bivariate analyses were conducted to examine the associations between the measure of
depression and its potential correlates by estimating simple linear regression models. Multiple
regression models were then estimated, including variables that were significant in the
bivariate analyses. Finally, interactions between social support and other significant correlates
CORRELATES OF DEPRESSION 693
of depression were tested to examine whether social support had buffering effects. Data were
analyzed using SPSS Version 10.1 (SPSS Inc., Chicago, IL, 60606, USA).
Results
Bivariate analyses
Bivariate analyses showed that history of mental health problems (pB/0.001), increased
frequency of injection drug use in the last 30 days (pB/0.05), and greater number of non-
injection drugs used in the last 30 days (i.e. non-injection polydrug use) (pB/0.05) were
significantly associated with reporting of more depressive symptoms. On the other hand,
increasing age (pB/0.05), male gender (pB/0.05), having a regular place for HIV care (pB/
0.05) and perceiving more social support (pB/0.001) were significantly associated with
reporting of fewer depressive symptoms.
Multivariate analyses
Table 2 presents results of multiple regression analyses. Model 1 includes variables found to
be significant in the bivariate analyses. Controlling for other variables, history of mental health
problems and non-injection polydrug use were associated with reporting of more depressive
symptoms. Perceived social support and having a regular place for HIV care were associated
with reporting of fewer depressive symptoms. Age and gender were no longer significant when
other variables were taken into account. Frequency of injection drug use was marginally
significant (p/0.07). Standardized regression coefficients indicated that the effects of history
of mental health problems, non-injection polydrug use, and perceived social support were
comparable. This model accounted for 29.1% of the variance in depressive symptoms.
Model 2 added the interaction term ‘social support x total number of non-injection drugs
used in the last 30 days’ to Model 1. The interaction was significant and the direction of the
Ta b l e 2 . Multivariate analyses predicting depressive symptoms among HIV-positive IUDs: multiple regression estimates
(n/137)
Variable BSE(B)b
Model 1
Age (in years)
/0.102 0.074 /0.115
Male /0.791 0.902 0.071
History of mental health problems 2.230 0.908 0.200*
Having a regular place for HIV medical care /5.075 2.558 /0.154*
Frequency of injection drug use in last 30 days 0.011 0.006 0.143
Number of non-injection drugs used in last 30 days 0.853 0.332 0.206*
Social support
/0.329 0.116 /0.223*
Model 2
Age (in years) /0.115 0.073 /0.129
Male
/0.652 0.887 /0.058
History of mental health problems 2.247 0.891 0.201*
Having a regular place for HIV medical care /5.789 2.529 /0.175*
Frequency of injection drug use in last 30 days 0.012 0.006 0.156*
Number of non-injection drugs used in last 30 days 4.684 1.616 1.130*
Social support
/0.027 0.169 /0.018
Social support /number of non-injection drugs used in last 30 days /0.236 0.098 /0.956*
Note. R
2
/0.291 for Model 1; R
2
/0.322 for Model 2; *pB/0.05.
694 Y. MIZUNO ET AL.
coefficient was negative, meaning that non-injection polydrug use was significantly associated
with an increase in depression but this effect decreased in persons who perceived more social
support. In Model 2, history of mental health problems, having a regular place for HIV care
and frequency of injection drug use were also significant. The addition of the interaction term
significantly improved the fit of Model 2 compared with Model 1 (DR
2
/0.031, pB/0.05).
Model 2 accounted for 32.2% of the variance in depressive symptoms. Other interaction
terms (e.g. history of mental health x social support) were also tested, but were not significant.
Discussion
This study makes an important contribution as it identifies several factors that might help
improve the mental health of HIV-positive IDUs. By exclusively focusing on IDUs who are
HIV-positive, the study extends prior research findings on the inverse association between
social support and depression (e.g. Belkin et al ., 1992) to HIV-positive IDUs. We also found a
pattern indicating that social support buffers the adverse effect of non-injection polydrug use.
These results point to the importance of incorporating a social support component to mental
health services or interventions that are designed for HIV-positive IDUs, particularly if they
also use a variety of non-injection drugs, which is common in this population. The potential
effect of receiving HIV medical care has previously received little attention, but the present
study found that having a regular place for HIV care is associated with fewer depressive
symptoms. This finding suggests the importance of linking HIV-positive IDUs to the system
of health care.
Because our sample consists of HIV-positive IDUs, we used a variety of drug use
measures and found that non-injection polydrug use in the past 30 days was significantly
associated with depressive symptoms. This finding was contrary to that of a prior study
(Rabkin et al., 1997) that found no association between polydrug use and depression. These
different findings may be due to different measures of polydrug use employed among studies.
Our analysis also showed some support for the association between frequency of injection
drug use in the past 30 days and depressive symptoms.
Preliminary analyses found that the two drug use variables (i.e. the number of non-
injection drugs used and frequency of injection) were not significantly correlated with each
other (r/0.14, p/0.10), suggesting that they might be measuring different aspects of drug
use. For example, frequency of injection drug use might primarily measure the level of
physical addiction to injection drugs while non-injection polydrug use might be a broader
indicator of an IDUs’ involvement in drugs and a life style that is heavily influenced by drug
use. Research has found associations between polydrug use and risk-taking behaviours such as
getting into fights and engaging in unprotected sex (Feigelman et al ., 1998). It could be that
the lives of polydrug users are likely to be chaotic and unstable. It is not clear, however, why
the number of injection drugs was not significantly associated with depressive symptoms in
the present analysis. Also, questions still remain as to what combinations of drugs are
associated with depression and why. These topics are important to pursue in future research.
The present paper shows some support for the buffering model suggesting that the
adverse effect of non-injection polydrug use would be smaller among those who perceive more
social support. One hypothesis to explain this interaction effect may be that HIV-positive
IDUs who use a variety of non-injection drugs have such chaotic lifestyles that whatever the
support perceived to be available might mean a lot to them. The present study, however, did
not provide relevant data to test this hypothesis. Perceived social support did not buffer the
adverse effects of history of mental health and frequent injection drug use. Another topic of
CORRELATES OF DEPRESSION 695
future studies may be to explore why social support buffers the negative effects of certain
factors while not buffering the effects of others.
There are a number of limitations in the study that should be noted. First, our sample
may not be representative of the population because (a) obtaining a random sample of HIV-
positive IDUs is difficult, (b) we recruited in two large urban areas, and (c) because we set
quotas on heroin users. Second, all data were collected by interviewer-administered
questionnaire, potentially leading to socially desired responding (Turner et al., 1998), which
might result in underestimates of substance use and sexual risk practices.
Another limitation is that the data used here were cross-sectional and thus we were
unable to establish causal relationships; we cannot ascertain whether polydrug use preceded
depressive symptoms. Drug use generally has been viewed as a maladaptive coping strategy
that reduces the ability to develop better coping skills among people who are under stress and
thus makes them more susceptible to psychological distress (Singh et al ., 1997). It is possible,
however, that depression also causes people to use drugs in the first place as a form of self-
medication or leads people to use more drugs over time. A similar claim can be made for the
association between not having a regular place for HIV care and higher levels of depressive
symptoms; here, it could be that depression could lead to not getting to medical
appointments. In addition, we would be more assured of our testing of buffering versus
main effect models if a longitudinal design had been used (Cohen & Willis, 1985).
Finally, a number of measurement issues need to be discussed. The measure of
depressive symptoms consisted of the seven items from the BSI. Although the administration
of this instrument was relatively simple, the measure does not have a clear cut-off point that
indicates probable clinical depression. Some researchers are concerned that there may be an
overlap between the measures of emotional distress and somatic symptoms of HIV (e.g.
Knowlton et al., 2000); thus what is intended to be a measure of depression, for example,
might actually be measuring symptoms of HIV associated with disease progression. Our
bivariate analysis found no significant association between depressive symptoms and CD4
count, a potential measure to indicate HIV-related physical health status. CD4 count,
however, was self-reported and thus might not have been accurate, although some studies
(e.g. Cunningham et al ., 1997) found that self-reported CD4 count and CD4 count obtained
from medical records were highly correlated. Our study lacked a measure of physical
functioning or HIV-related symptoms that could be a better indicator of HIV-related physical
health status.
Despite these limitations, the present findings have important intervention implications.
First, we have identified groups of HIV-positive IDUs who may be more likely to report
depressive symptoms: those with a history of mental health problems, those who were also
non-injection polydrug users and, to some extent, those who were frequent injectors. They
may significantly benefit from mental health services, drug treatment programmes, and HIV
prevention programmes tailored to their specific needs. In addition, perceived social support
was found to have a protective effect, and appeared to buffer the adverse effect of non-
injection polydrug use.
However, more studies that use elaborate measures of social support may be needed to
explore specifically what types of social support are most effective for this population. For
example, social support can have various components like instrumental (e.g. help with
money), emotional or informational (Thoits, 1995). In order to develop effective interven-
tions, researchers and programme developers may need to ask whether these components are
equally important to the needs of HIV-positive IDUs or whether one component is more
important. It is also possible that the effects of social support vary depending on the type of
people from whom support is perceived to be available, e.g. family members versus friends
696 Y. MIZUNO ET AL.
who also use drugs. Future research should look into these questions to better address the
mental health needs of HIV-positive IDUs.
Acknowledgements
The Seropositive Urban Injectors Study (SUDIS) was funded by cooperative agreements of
the Centers for Disease Control and Prevention with New Jersey City University (U62/
CCU213605) and the University of California, San Francisco (U62/CCU913557). The
following colleagues contributed to the development, design and implementation of SUDIS:
Thomas Borkowski, Kimberly Boyd, Cynthia Gomez, Kelly Knight, Beatrice Krauss, Paula
Lum, and Richard Wolitski. We also acknowledge the many other project staff who helped
with various aspects of SUDIS.
References
B
ELKIN,
G
.
S
.
,F
LEISHMAN,
J
.
A
.
,S
TEIN,
M
.
D
.
,P
IETTE,
J
.
&M
OR,
V
.
(1992). Physical symptoms and depressive
symptoms among individuals with HIV infection. Psychosomatics ,33 ,416/427.
B
RODAY,
S
.
F
.
&M
ASON,
J
.
L
.
(1991). Internal consistency of the Brief Symptom Inventory for counseling-center
clients. Psychological Reports ,68 , 94.
C
ENTERS FOR
D
ISEASE
C
ONTROL AND
P
REVENTION
(2001). HIV/AIDS Surveillance Report ,13 (2), 14.
C
IESLA,
J
.
A
.
&R
OBERTS,
J
.
E
.
(2001). Meta-analysis of the relationship between HIV infection and risk for depressive
disorders. American Journal of Psychiatr y ,158 , 725/730.
C
OHEN,
S
.
&W
ILLIS,
T
.
(1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin ,98 , 310 /
357.
C
UNNINGHAM,
W
.
E
.
,R
ANA,
H
.
M
.
,S
HAPIRO,
M
.
F
.
&H
AYS,
R
.
D
.
(1997). Reliability and validity of self-report CD4
counts in persons hospitalized with HIV disease. Journal of Clinical Epidemiology ,50 , 829 /835.
D
EROGATIS,
L
.
R
.
&S
PENCER,
P
.
M
.
(1982). The Brief Symptom Inventory (BSI): Administration, scoring, and procedure
manual /1. Baltimore, MD: John Wiley.
D
EW,
M
.
A
.
,B
ECKER,
J
.
T
.
,S
ANCHEZ,
J
.
,C
ALDARARO,
R
.
,L
OPEZ,
O
.
L
.
,W
ESS,
J
.
,D
ORST,
S
.
K
.
&B
ANKS,
G
.
(1997).
Prevalence and predictors of depressive, anxiety and substance use disorders in HIV-infected and uninfected men: a
longitudinal evaluation. Psychological Medicine ,27 ,395/409.
F
EIGELMAN,
W
.
,G
ORMAN,
B
.
S
.
&L
EE,
J
.
A
.
(1998). Binge drinkers, illicit drug users, and polydrug users: an
epidemiological study of American collegians. Jour nal of Alcohol & Dr ug Education ,44 ,47/69.
F
ERRANDO,
S
.
,G
OGGIN,
K
.
,S
EWELL,
M
.
,E
VANS,
S
.
,F
ISHMAN,
B
.
&R
ABKIN,
J
.
(1998). Substance use disorders in
gay/bisexual men with HIV and AIDS. American Journal of Addictions ,7,51/60.
F
ERRANDO,
S
.
J
.
,W
ALL,
T
.
L
.
,B
ATKI,
S
.
L
.
&S
ORENSEN,
J
.
L
.
(1996). Psychiatric morbidity, illicit drug use and
adherence to Zidovudine (AZT) among injection drug users with HIV disease. American Journal of Drug Alcohol
Abuse ,22 , 475 /487.
H
AWKINS,
W
.
,L
ATKIN,
C
.
,H
AWKINS,
M
.
&C
HOWDURY,
D
.
(1998). Depressive symptoms and HIV-risk behaviors in
inner-city users of drug injections. Psychological Reports ,82 , 137 /138.
K
ALICHMAN,
S
.
C
.
,H
ECKMAN,
T
.
,K
OCHMAN,
A
.
K
.
,S
IKKEMA,
K
.
&B
ERGHOLTE,
J
.
(2000). Depression and thoughts
of suicide among middle-aged and older persons living with HIV-AIDS. Psychiatric Services ,51 ,903/907.
K
NOWLTON,
A
.
R
.
,L
ATKIN,
C
.
A
.
,C
HUNG,
S
.
,H
OOVER,
D
.
R
.
,E
NSMINGER,
M
.
&C
ELENTANO,
D
.
D
.
(2000). HIV and
depressive symptoms among urban injection drug users. AIDS & Behaviour ,4, 353 /360.
K
NOWLTON,
A
.
R
.
,L
ATKIN,
C
.
A
.
,S
CHROEDER,
J
.
R
.
,H
OOVER,
D
.
R
.
,E
NSMINGER,
M
.
&C
ELENTANO,
D
.
D
.
(2001).
Longitudinal predictors of depressive symptoms among low-income injection drug users. AIDS Care ,13 ,549/559.
K
WIATKOWSKI,
C
.
F
.
&B
OOTH,
R
.
(1998). HIV-seropositive drug users and unprotected sex. AIDS and Behavior ,2,
151 /160.
L
YKETSOS,
C
.
G
.
,H
OOVER,
D
.
R
.
,G
UCCIONE,
M
.
,D
EW,
M
.
A
.
,W
ESCH,
J
.
E
.
,B
ING,
E
.
G
.
&T
REISMAN,
G
.
J
.
(1996).
Changes in depressive symptoms as AIDS develops. American Journal of Psychiatry ,153 , 1430 /1437.
M
OORE,
J
.
,S
CHUMAN,
P
.
,S
CHOENBAUM,
E
.
,B
OLAND,
B
.
,S
OLOMON,
L
.
&S
MITH,
D
.
(1999). Severe adverse life
events and depressive symptoms among women with, or at risk for, HIV infection in four cities in the United States
of America. AIDS ,13 , 2459/2468.
CORRELATES OF DEPRESSION 697
R
ABKIN,
J
.
G
.
,J
OHNSON,
J
.
,L
IN,
S
-
H
.
,L
IPSITZ,
J
.
D
.
,R
EMIEN,
R
.
H
.
,W
ILLIAMS,
J
.
B
.
W
.
&G
ORMAN,
J
.
M
.
(1997).
Psychopathology in male and female HIV-positive and negative injecting drug users: longitudinal course over 3
years. AIDS ,11 ,507/515.
R
AHAV,
M
.
,N
UTTBROCK,
L
.
,R
IVERA,
J
.
J
.
&L
INK,
B
.
(1998). HIV infection risks among homeless, mentally ill,
chemical misusing men. Substance Use & Misuse ,33 , 1407 /1426.
R
HODES,
T
.
J
.
,D
ONOGHOE,
M
.
C
.
,H
UNTER,
G
.
&S
TIMSON,
G
.
(1993). Continued risk behavior among HIV positive
drug injectors in London: implications for inter vention. Addiction ,88 ,153/156.
S
EROVICH,
J
.
M
.
,B
RUCKER,
P
.
S
.
&K
IMBERLY,
J
.
A
.
(2000). Barriers to social support for persons living with HIV/AIDS.
AIDS Care ,12 , 651 /662.
S
IEGEL,
K
.
,K
ARUS,
D
.
&R
AVEIS,
V
.
(1997). Correlates of changes in depressive symptomatology among gay men with
AIDS. Health Psychology,16 ,230/238.
S
INGH,
B
.
K
.
,K
OMAN,
J
.
J
.
,C
ATAN,
V
.
,S
OUPLY,
K
.
,B
IRKEL,
R
.
&G
OLASZEWSKI,
T
.
(1993). Sexual risk behavior among
injection drug-using human immunodeficiency virus positive clients. International Journal of the Addictions ,28 ,
735 /747.
S
INGH,
N
.
,S
QUIER,
C
.
,S
IVEK,
C
.
,W
AGENER,
M
.
M
.
&Y
U,
V
.
(1997). Psychological stress and depression in older
patients with intravenous drug use and human immunodeficiency virus infection: implications for intervention.
International Journal of STD & AIDS ,8,251/255.
T
HOITS
, P.A. (1995). Stress, coping, and social suppor t processes: where are we? What next? Journal of Health and
Social Behavior , extra issue, 53
/79.
T
URNER,
C
.
F
.
,K
U,
L, R
OGERS,
S
.
M
.
,L
INDBERG,
L
.
D
.
,P
LECK,
J
.
H
.
&S
ONENSTEIN,
F
.
L
.
(1998). Adolescent sexual
behavior, drug use, and violence: increased reporting with computer survey technology. Science ,280 ,867/73.
V
AN
S
ERVELLEN,
G
.
,S
ARNA,
L
.
,N
YAMATHI,
A
.
,P
ADILLA,
G
.
,B
RECHT,
M
.-
L
.
&J
ABLONSKI,
K
.
J
.
(1998). Emotional
distress in women with symptomatic HIV disease. Issues in Mental Health Nursing ,19 ,173/189.
698 Y. MIZUNO ET AL.
