Community mibilzation to reduce drug use quabg ninh, viet nam

RESEARCH AND PRACTICE

Community Mobilization to Reduce Drug Use, Quang Ninh,
Vietnam
Hien Tran Nguyen, MD, PhD, Anh Viet Tran, MSc, Nguyen Binh Nguyen, MD, PhD, Son Hong Nguyen, MD, PhD, Diep Bich Vu, MD, PhD,
Nhu To Nguyen, MD, PhD, Ronald S. Brookmeyer, PhD, and Roger Detels, MD

An epidemic of injection drug use began in
Vietnam in the mid-1990s, concurrent with an
increase in injection drug use in Guangxi,
China.1 Drug abuse has risen threefold in the
last 10 years in Vietnam, involving younger
males, and more recently, females.2 Concurrently, heroin has replaced opium as the
preferred drug.3 The provinces of Vietnam
along the border of China now have among
the highest rates of injection drug use and
HIV among drug users in Vietnam.4 According to national sentinel surveillance data, HIV
prevalence among injection drug users
(IDUs) increased from 10% in 1996 to a peak
of 30% in 2001 to 2002, and then gradually
decreased to 20% in 2008.5 At the time we
conducted our study in Quang Ninh, the
prevalence of HIV among IDUs was 46.7%
in Quang Ninh, compared with 35.7% in
Haiphong and 51.3% in Ho Chi Minh City.6
A variety of strategies have been tried to
prevent drug use in various countries
worldwide, with limited success.7,8 One
approach to prevention of drug use has been
to mobilize the affected communities to take
supportive action.9 Drug users in Asia, unlike

the United States and Europe, tend to remain
a part of their families and their communities.10
This characteristic provides an opportunity to
use the family and the community as an interventional tool to prevent initiation of drug use.11
Following a study to identify the characteristics of drug users in southern Yunnan, China,
Wu et al.12 met with official and unofficial
leaders in 19 villages in Dehong County,
Yunnan, China, to discuss the drug problem
they were experiencing in young men, and to
encourage them to mobilize a community intervention to reduce initiation of drug use. The
meeting was successful in persuading the village leaders that they had to take the initiative
in preventing drug use in their community.
Subsequently, the villagers mounted a broad
intervention program that resulted in a 66%

Objectives. We implemented an intervention to reduce drug use in an urban
commune in northern Vietnam.
Methods. We encouraged the intervention commune to accept responsibility
for developing their own intervention strategies based on a community mobilization model used in southern, rural China. We selected a comparison commune, which had demographic characteristics and a drug history similar to the
intervention commune. The 2-year incidence of new drug users was estimated
retrospectively in the intervention and comparison communes between baseline
(2003) and follow-up (2009).
Results. Increased incidence of new (noninjecting) drug users between 2003
and 2009 in the intervention commune was lower than that in the comparison
commune, and these participants expressed more positive attitudes toward local
authority and people with drug use and HIV/AIDS. Increased condom use during
last intercourse with female sex workers and with female casual partners was
observed in the intervention commune. HIV prevalence and positive opioid tests
decreased more in the intervention commune.
Conclusions. Our results suggested that the community mobilization had

a positive influence in the intervention commune. (Am J Public Health. 2015;105:
189–195. doi:10.2105/AJPH.2014.302101)

decrease in new drug users compared with
matched villages in the same area. The intervention was most effective among those
groups at highest risk of initiating drug use.13
Intervention strategies implemented by the
government and nongovernment organizations have not halted the epidemic of drug
use in Vietnam. Upon learning the results of
the Yunnan community intervention program,
public health leaders in Vietnam requested
that the Hanoi Medical University (HMU)
implement and evaluate a similar community
mobilization strategy to reduce drug use in
Quang Ninh, Vietnam, which was one of the
provinces most affected by the drug epidemic,
with the assistance of researchers from the
University of California, Los Angeles School
of Public Health. In response to the request,
we met with formal and informal leaders
of a commune (Ha Tu) in Ha Long City to
discuss the possibility of their mounting
a community-based intervention to prevent
the initiation of drug use by young men and to
prevent injection use by young men already

January 2015, Vol 105, No. 1 | American Journal of Public Health

using drugs in their commune. We report
the results of that intervention as compared

with a similar commune in the same area.

METHODS
Two communes with similar estimated
numbers of drug users, population size,
geography, community political organizations, estimated prevalence of IDUs and
HIV, and national drug prevention and HIV
prevention programs were selected in
Ha Long City (Ha Tu) and Cam Pha City
(Cam Thinh) in northern Vietnam in 2003.
By coin toss, Ha Tu was designated as the
intervention site and Cam Thinh was the
comparison (control) site. At that time, no
other nationwide interventions had been
implemented in Ha Tu. Family Health International (FHI) subsequently initiated an
intervention in Cam Pha City, which includes
Cam Thinh; the comparison area was one
of the communes in that city. The ECHO
intervention model was based on a model

Nguyen et al. | Peer Reviewed | Research and Practice | 189


RESEARCH AND PRACTICE

developed by sociologists at the University of
Connecticut, and subsequently, the peer outreach model was implemented.

The Interventions
The national intervention program operating

in both communes during the study included
media announcements, clean needle distribution programs, AIDS care and support, and
drop-in centers for IDUs.
The theoretical framework used for the intervention in Ha Tu was based on strategies
used in our previous study to reduce the
incidence of new drug users in southern
Yunnan, China. That study used a community
intervention model directed at both at-risk
youths and current drug users, in which community leadership and local residents were
mobilized for social action.13,14 The intervention was based on the behavior change and
self-efficacy model of Bandura’s social learning
theory.15 New behaviors were promoted
through social reinforcement from the community. Persuasive influences included community norms, village leadership, family, and
peers, and supportive attitudes and programs
involving youth, schools, and drug users.
Figure 1 presents the conceptual model for
the intervention program. The model incorporated factors demonstrated to be associated
with uptake of drug and injection use, and
both personal and community factors associated

with self-efficacy. Community involvement
was key in influencing community norms.
Traditional moral principles accepted by the
communes were used to encourage youths
to avoid drugs and to contribute to the wellbeing of the community.16 Social marketing
principles were used to guide the design of the
educational messages.17,18
Initially, a meeting was held in Ha Tu with
key groups, including both commune officials
and semiofficial groups, such as the women’s

and youth associations, to discuss the issue of
drug use by the youths in the commune.
Through these discussions, the participants
recognized their problem and accepted the
idea that they would need to take the primary
responsibility to develop and implement an
intervention program to reduce the incidence
of new drug users and the transition from
noninjecting drug use to injecting drug use.
We then assisted the commune leaders in
developing intervention activities that were
appropriate for the commune, including
monthly meetings at which the commune
subunit leaders reported on the progress
of the intervention in their districts. These
interventions included development of a
didactic school curriculum on drug prevention, school assemblies, informal skits put
on by youths, development of videos and
games, parades to promote nondrug use,
dissemination of drug prevention messages,

Community involvement

Community norms

Education

Family norms

Peer pressure

Smoking

Parental pressure

Drop-in centers

Age

Drug use

Self-efficacy

Youth groups

Youth community
service productivity

FIGURE 1—Conceptual model for factors influencing initiation of drug use: Community
Mobilization to Reduce Drug Use; Quang Ninh, Vietnam; 2003–2009.

190 | Research and Practice | Peer Reviewed | Nguyen et al.

loudspeaker announcements in the residential blocks, bulletins, establishing and staffing
of an intervention center, posters, banners,
and media announcements (including television, radio, and newspapers), visits by youth
to detoxification centers, and messages
from current and former drug users. CDs and
cassette tapes that provided education about
HIV/AIDS and promoted reduction of stigmatization and discrimination against drug users

and persons living with HIV/AIDS were also
distributed to the blocks within the commune.
The intervention program was received with
enthusiasm (as observed by the investigators) by
the commune members who, because they
played a key role in developing the intervention,
felt they had ownership of it.
During the study, FHI also implemented
an enhanced drug intervention in Cam
Thinh, the comparison commune, in 2004
to 2005, which included the founding of
a drop-in center for persons at high risk for
HIV/AIDS, a peer education outreach model,
group events, and media messages emphasizing destigmatization and antidiscrimination. A team of 12 IDUs or former drug users
was also recruited to play the roles of outreach workers and HIV/AIDS educators.

Assessment
In December 2003 and March 2009, all
males aged 15 to 24 years were invited to
participate in anonymous cross-sectional
surveys on drug use and sexual behaviors in
both Ha Tu and Cam Thinh. A meeting was
held in each commune to emphasize to the
young men of the commune the importance
of participating in the survey. Before administration of the interview, verbal informed
consent was obtained. The youths were invited to be interviewed at the village public
house or a school. Youths who did not
come for an interview were followed up by
village leaders and encouraged to participate.
Information about demographic characteristics, HIV/AIDS-related knowledge, attitudes,

sexual behaviors, and substance-use behaviors
were solicited in both surveys. The interview
questions were asked by use of a CD player.11
The participants marked their answers on
blank sheets of paper that contained only
a study identification number and the question
numbers, but not the text of the questions. The

American Journal of Public Health | January 2015, Vol 105, No. 1


RESEARCH AND PRACTICE

participants then marked their answers in the
appropriate spaces. None of the participants
were illiterate. The answers usually involved
numbers or words that would not reveal the
nature of the question. At the completion of the
interviews, the answer sheets, which contained
no identifying information, were placed in
a large box to assure respondents that their
answers could not be attributed to them.
Each participant was then requested to provide a urine specimen that was later tested
for opioids using Quick-Check (ACON Labs,
San Diego, CA). Blood samples were also
collected from all participants and tested for
HIV at the nationally certified HIV laboratory
of Viet Duc Hospital (Hanoi, Vietnam), using
HIV testing strategy III stipulated by the
Vietnam Ministry of Health (for diagnosing

individual HIV cases). According to this strategy,
a sample was considered positive for HIV only
if it was positive with 3 different techniques or
3 different methods of antigen preparation. In our
study, each blood sample was tested for HIV
using the enzyme-linked immunosorbent
assay technique from 3 different test kits
(i.e., 3 different methods of antigen preparation).

Data Analysis
We conducted 2 surveys of young men in
the intervention and comparison communes:
a baseline survey in 2003 and a follow-up
survey in 2009. Data were entered using EPI
INFO version 6.04d (Centers for Disease
Control and Prevention, Atlanta, GA), and
then processed and analyzed using SAS 9.2
(SAS Institute, Cary, NC). Data processing
included cleaning (range and logic checking),
scoring, grouping of variables, and categorization. Because the distribution of demographic
characteristics in the 2 communities at the 2
points in time were not significantly different,
and because the proportions of the mobile
population were low and quite similar, we
assumed that the communes were stable. We
were thus able to retrospectively reconstruct
the 2-year incidence of new drug use by
counting those who reported initiation of any
drug use in the previous 2 years in both the
2003 and 2009 surveys. We calculated

descriptive statistics to describe the changes
in demographic characteristics, drug- and
HIV/AIDS-related knowledge, and attitudes
and behaviors in the period between the 2

surveys in the 2 communes. For comparing
the changes in incidence of drug use initiation
over the time period in the 2 communes, we
calculated the ratio of relative risk (assessment
vs baseline) in the intervention commune to
that in the comparison commune, using the
following formula: ratio of relative risks = r =
RR09/RR03 = (IINT09/ICTR09)/(IINT03/ICTR03).
An r < 1 indicated that the proportionate
increase in incidence over time in the intervention commune was smaller than that in
the comparison commune. The Wald 95%
confidence interval (CI) was calculated for
log r, and then antilogs were taken to obtain
CI for r. The z-test (where Z = log r/SE [log r])
was used for determining the P value for
testing the hypothesis that r = 1.
For prevalence measures (regarding
knowledge, attitudes, behavior, urine test

positivity, and HIV prevalence), we used
a formula analogous to the preceding one
to calculate ratios of relative prevalence
changes, that is: ratio of prevalence ratios =
(PINT09/PCTR09)/(PINT03 /PCTR03 ).

RESULTS
The response rates for the survey of young
men in Ha Tu were 614 of 683 men (89.9%)
in 2003 and 565 of 621 men (91.0%) in
2009. The comparable response rates in
Cam Thinh were 583 of 667 men (87.4%)
in 2003 and 574 of 631 men (91.0%) in
2009. The proportion of married interviewees
was low (approximately 2%) in both communes. The percent unemployed was higher
in Cam Thinh (16.0% vs 10.6%; Table 1).
Levels of education overall were lower in

TABLE 1—Demographic Characteristics of Participants at Baseline (December 2003) in
Intervention and Control Areas: Community Mobilization to Reduce Drug Use; Quang Ninh
Province, Vietnam; 2003–2009
Variable

Intervention Commune, No. (%)

Comparison Commune, No. (%)

370 (60.3)
244 (39.7)

358 (61.4)
225 (38.6)

600 (97.7)

518 (88.9)

14 (2.3)

65 (11.1)

Single

601 (97.9)

572 (98.1)

Married

12 (2.0)

11 (1.9)

1 (0.2)

0 (0.0)

Age, y
15–19
20–24
Ethnic group
Kinh
Others
Marital status

Divorced

Education
£ grade 9

141 (23.1)

100 (17.2)

Grades 10–12

337 (55.2)

343 (59.1)

‡ some college

133 (21.8)

137 (23.6)

Student

347 (56.7)

348 (59.8)

Employed

200 (32.7)

141 (24.2)

Unemployed
Local residency

65 (10.6)

93 (16.0)

Employment status

Yes

455 (74.1)

477 (81.8)

No

159 (25.9)

106 (18.2)

Living alone
Yes

43 (7.0)

17 (2.9)

No

571 (93.0)

566 (97.1)

614 (100.0)

583 (100.0)

Total

Note. Percentages may not add to 100 because of rounding.

January 2015, Vol 105, No. 1 | American Journal of Public Health

Nguyen et al. | Peer Reviewed | Research and Practice | 191


RESEARCH AND PRACTICE

Ha Tu (77.0% had more than secondary
school in Ha Tu vs 82.7% in Cam Thinh).
Fewer were local residents in Ha Tu (74.1%)
than Cam Thinh (81.8%) at baseline, and
more respondents lived alone in Ha Tu
(7.0%) than in Cam Thinh (2.9%).
Between 2003 and 2009, the incidence
of new drug users increased more in the
comparison group (from 1.4% to 7.1%) than
in the intervention group (from 2.6% to

7.1%). The ratio of relative risk (assessment
vs baseline) in the intervention commune to
that in the comparison commune was r = 0.6
(95% CI = 0.2, 1.4; Table 2). The impact
of the intervention on the incidence of new
drug users differed by ethnic group and
education level, but not by age. The value
of r was not available for comparison across
categories of marital status and ethnic group
because there were too few participants in
the married and divorced marital status
groups and in others for ethnic groups (Table 2),
resulting in zero values, precluding

calculation of risk ratios. The increases in the
incidence of new drug users (among participants who had never used drugs) were
caused by increases in noninjecting use only
(mainly use of marijuana and amphetaminetype stimulants). Only 10 participants in
Ha Tu and 1 in Cam Thinh reported first
injecting drug use during the 2-year period
before the baseline survey (December
2003); no one reported first injecting drug
use in either commune during the 2-year
period before the assessment survey
(March 2009).
Drug- and HIV/AIDS-related knowledge,
attitudes and behavior, urine test positivity,
and HIV prevalence among participants in
the 2 study sites at the baseline and assessment surveys (in 2003 and 2009, respectively) are presented in Table 3. Knowledge
regarding routes of HIV transmission and

prevention methods remained at approximately 80% at baseline and assessment in
both communes. However, there were

increases in the percentages of participants
who knew that an HIV-infected person might
have a healthy appearance, from 65.6% to
72.9% in Ha Tu and from 62.6% to 72.9%
in Cam Thinh. In the intervention commune,
participants reported greater reductions in
negative attitudes of local authority and
people regarding drug use and HIV/AIDS,
compared with the comparison commune.
In the intervention commune, the percentage
of participants who agreed with the statement “local authority and people currently
see drug users as criminals who need to
be controlled harshly” decreased by 15.7
percentage points (from 31.5% to 15.9%).
In the comparison commune, the percentage
decreased by 9.7 percentage points (from
31.1% to 21.4%). The percentage of participants who agreed with the statement
“distributing clean needles and syringes to
injecting drug users is acceptable to the
local authority and people” increased by
12.7 percentage points (from 50.6% to

TABLE 2—Change in Incidence of New Male Drug Users Before and After Intervention Program in Intervention and Control Areas:
Community Mobilization to Reduce Drug Use; Quang Ninh Province, Vietnam; 2003–2009
Intervention Commune
Incidence Change,
% Points

Assessment Period
3/07–3/09,
% (No.)

Incidence Change,
% Points

Attributable
Incidence Changea

Ratio of
Relative Risksb
(95% CI)

P

7.1 (39/551)

+4.5

1.4 (8/561)

7.1 (39/547)

+5.7

–1.2

0.6 (0.2, 1.4)

.22

2.5 (9/362)

8.0 (24/301)

+5.5

0.9 (3/354)

6.8 (18/265)

+5.9

–0.4

0.4 (0.1, 1.7)

.21

2.7 (6/224)

6.0 (15/250)

+3.3

2.4 (5/207)

7.4 (21/282)

+5.0

–1.7

0.7 (0.2, 2.8)

.68

Assessment Period
3/07–3/09,
% (No.)

2.6 (15/586)

15–19
20–24

Stratifying
Variable
Total

Comparison Commune
Baseline Period
12/01–12/03,
% (No.)

Baseline Period
12/01–12/03,
% (No.)

Age, y

Ethnic group
Kinh

2.6 (15/574)

Others

0 (0/12)

Marital status
Single

2.6 (15/575)

Married

0 (0/10)

Divorced

0 (0/1)

7.0 (38/543)

+4.4

1.6 (8/497)

+12.5

0 (0/64)

7.2 (38/528)

+4.6

1.4 (8/551)

0 (0/17)

0.0

0 (0/10)
NA

12.5 (1/8)

25 (1/4)

+25

7.7 (37/478)

+6.1

–1.7

0.6 (0.2, 1.4)

.07

3.4 (2/58)

+3.4

+9.1

NA

NA

7.4 (38/512)

+6.0

–1.4

0.5 (0.2, 1.4)

.2

3.0 (1/33)

+3.0

–3.0

NA

NA

0 (0/1)

NA

NA

NA

NA

Education level
£ grade 9

3.7 (5/134)

7.9 (5/63)

+4.2

2.1 (2/96)

5.6 (4/72)

+3.5

–0.6

0.8 (0.1, 6.2)

.83

Grades 10–12

2.8 (9/321)

6.5 (21/321)

+3.7

1.2 (4/333)

7.0 (18/258)

+5.8

–1.8

0.4 (0.1, 1.5)

.17

‡ some college

0.8 (1/129)

7.3 (12/164)

+6.5

1.6 (2/129)

7.9 (17/215)

+6.3

+0.2

1.9 (0.2, 22.4)

.37

Note. CI = confidence interval; NA = not available.
a
Attributable incidence change is the incidence change in the intervention group minus incidence change in the control group. Minus signs indicate that the incidence increase in the intervention
group was lower than the incidence increase in the control group (i.e., the intervention helped reduce the risk of starting drug use), and vice versa.
b
A ratio of < 1 indicates that the relative risk (assessment vs baseline) for the intervention group is lower than that for the control group, and vice versa. Some ratios and the corresponding CIs and
P values are left blank because they cannot be calculated when there is at least 1 zero incidence in the same data line.

192 | Research and Practice | Peer Reviewed | Nguyen et al.

American Journal of Public Health | January 2015, Vol 105, No. 1


RESEARCH AND PRACTICE

TABLE 3—Changes in Knowledge, Attitudes, Behavior, Urine Opioid Test Results, and HIV Prevalence in Intervention and Control Areas:
Community Mobilization to Reduce Drug Use; Quang Ninh Province, Vietnam; 2003–2009
Intervention Commune

Variable

Comparison Commune

Baseline Survey Assessment Survey Prevalence Baseline Survey Assessment Survey Prevalence
Ratio of
12/03 (n = 614), 3/09 (n = 565),
Change, 12/03 (n = 583), 3/09 (n = 574),
Change,
Prevalence
% (No.)
% (No.)
% Points
% (No.)
% (No.)
% Points Ratiosa (95% CI)

P

65.6 (386/588)

72.9 (409/561)

7.3

62.6 (356/569)

72.9 (417/572)

10.3

1.0 (0.9, 1.1)

.4

80.9 (473/585)

83.1 (466/561)

2.2

81.5 (463/568)

78.3 (448/572)

–3.2

1.1 (1.0, 1.2)

.1

80.2 (457/570)

80.8 (442/547)

0.6

78.1 (438/561)

78.0 (428/549)

–0.1

1.0 (0.9, 1.1)

.83

31.5 (185/587)

15.9 (89/561)

–15.7

31.1 (177/570)

21.4 (122/571)

–9.7

0.7 (0.5, 1.0)

.04

35.1 (206/587)

15.1 (85/561)

–19.9

27.4 (156/570)

10.7 (61/571)

–16.7

1.1 (0.8, 1.6)

.58

50.6 (297/587)

63.3 (354/559)

12.7

46.3 (263/568)

37.0 (211/571)

–9.4

1.6 (1.3, 1.9)

< .001

59.3 (54/91)

66.1 (80/121)

6.8

58.0 (40/69)

64.0 (87/136)

6.0

1.0 (0.7, 1.4)

.95

100.0 (24/24)

11.5

93.8 (15/16)

92.9 (13/14)

–0.9

1.1 (0.9, 1.4)

.28

74.2 (23/31)

31.3

70.0 (14/20)

76.9 (20/26)

6.9

1.6 (0.8, 3.0)

.16

Knowledge
Knew that an HIV-infected person can have
a healthy appearance
Correctly identified all listed possible routes
of HIV transmission
Correctly identified 3 main measures for prevention
of HIV transmission
Perceived attitude of the local government and people
Drug users are currently seen as criminals who
need to be controlled harshly
HIV-infected people are not allowed to work in
public places
Distributing clean needles and syringes to
injecting drug users is acceptable
Behavior
Used condom in last sexual intercourse with any
female partner in past 12 mo

Used condom in last intercourse with a female sex

88.5 (23/26)

worker in past 12 mo
Used condom in last intercourse with a casual female 42.9 (9/21)
sex partner in past 12 mo
Reported current use of any drug

1.0 (6/614)

1.2 (7/565)

0.3

1.2 (7/583)

1.2 (7/574)

0

1.2 (0.3, 5.6)

.77

Reported current heroin use
Reported current injecting use

0.7 (4/614)
0.7 (4/614)

0 (0/565)
0.4 (2/565)

–0.7
–0.3

1.2 (7/583)
1.0 (6/582)

0.2 (1/574)
0 (0/574)

–1.0
–1.0

NA
NA

NA
NA

Injected drugs at least once in past month

1.1 (7/614)

0 (0/565)

–1.1

0.9 (5/583)

0 (0/574)

–0.9

NA

NA

–1.4

2.9 (17/583)

1.9 (11/574)

–1.0

0.7 (0.2, 2.2)

.5

–2.3

2.2 (13/583)

0.3 (2/574)

–1.9

NA

NA

Urine test opioid-positive

2.4 (15/614)

HIV-positive

2.3 (14/614)

1.1 (6/564)
0 (0/564)

Note. CI = confidence interval; NA = not available.
a
A ratio of < 1 indicates that the prevalence ratio (assessment vs baseline) for the intervention group is lower than that for the control group, and vice versa. Some ratios and the corresponding CIs
and P values are left blank because they cannot be calculated when there is at least 1 zero prevalence in the same data line.

63.3%) in the intervention commune,
but decreased by 9.4 percentage points
(from 46.3% to 37.0%) in the comparison
commune. These relative changes over
time were statistically different between the
intervention and comparison communes,
and reflected a more positive attitude
among young men in Ha Tu following the
intervention. Although measures did not
reach statistical significance individually,

most measures of knowledge, attitudes,
and behavior showed more improvement
in Ha Tu.

Regarding HIV-related behaviors, the level
of condom use during last sexual intercourse
with a female sex worker increased from
88.5% to 100% in the intervention commune,
but slightly decreased (from 93% to 92.9%)
in the comparison commune. The level of
condom use during last sexual intercourse
with a female casual partner increased from
42.9% to 74.2% in the intervention commune,
and from 70.0% to 76.9% in the comparison
commune. Prevalence of reported injecting
drug use at least once in the past month was
low in both communes at baseline (only 7 of

January 2015, Vol 105, No. 1 | American Journal of Public Health

614 participants reported using an injection
drug in the past month in Ha Tu, compared
with 5 of 583 participants in Cam Thinh), and
reduced to zero in both communes in 2009.
In Ha Tu, urine opioid positivity among
participants decreased from 2.4% to 1.0%
(–1.4 percentage points), whereas in Cam
Thinh, it decreased from 2.9% to 1.9% (–1.0
percentage points). Analogously, HIV prevalence (determined by laboratory serum testing)
decreased from 2.3% to 0% in Ha Tu, whereas

it decreased from 2.2% to 0.3% (–1.9 percentage points) in Cam Thinh.

Nguyen et al. | Peer Reviewed | Research and Practice | 193


RESEARCH AND PRACTICE

DISCUSSION
Studies of community and outreach-based
interventions have reported differing results.
Singh reported that a community-based outreach program that used a multipronged
approach directed at IDUs in Manipur, India,
was successful, although there was no comparison group, and the outcome variables
were primarily process variables (number of
interventions implemented, condoms distributed, and education workshops held).19
Coyle et al., in a review of published papers
on outreach programs in the United States,
concluded that
outreach-based interventions have been effective
in reaching out-of-treatment IDUs, providing the
means for behavioral changes and inducing
behavior change in the desired direction.20(p20)

By contrast, Giesbert and Haydon, who
reviewed community-based interventions at
the local level, concluded that “many but not
all interventions have modest or equivocal
impact.”9(p633)
It should be noted that our study, which
was originally meant to be an evaluation of

a community intervention in comparison
with a community with no intervention, was
a comparison of 2 different interventions in
2 communes. In Ha Tu, a “bottom-up” approach
was implemented, which focused on at-risk
youths and current drug users. The key factor
in the intervention was, in our opinion, assisting
the community to recognize that they had
a problem with their young people becoming
drug users and, secondly, assisting the community to recognize that they needed to take
the initiative in designing and implementing
the intervention program. Thus, the key
elements of the intervention strategy in Ha Tu
were designed by the commune, giving them
“ownership” of the program.
After this study began, FHI implemented
some drug and HIV/AIDS intervention
activities in the comparison commune
(Cam Thinh, as part of Cam Pha City). The
activities were mainly directed at current
drug users, but some activities (such as media
messages emphasizing destigmatization and
antidiscrimination and recruitment of 12
drug users or former drug users to play the
role of outreach workers and HIV/AIDS

educators) might have had the effect of
improving the knowledge and attitudes and
reducing high-risk behaviors and HIV transmission among people in the comparison
commune as a whole. Low response rate is

a common problem for many studies of
sensitive topics. However, the response rates
in our study were relatively high—88.9% in
Cam Thinh and 87.4% in Ha Tu for the
baseline survey, and 91% in both for the
assessment survey. Because drug use is
a stigmatized behavior, nonresponders were
probably more likely to be drug users and
HIV-infected persons.
Significantly greater improvement in
perceived attitudes of local authority and
people toward drug users and persons living
with HIV/AIDS were observed in Ha Tu, the
intervention commune, than in Cam Thinh.
Because stigma is a major barrier to testing,
this result suggested that the intervention
strategy used in Ha Tu helped get people
tested for HIV, an essential element of
any HIV intervention program, by improving acceptability and willingness of
drug users to accept testing and seek
assistance. 21---22
We did not objectively verify reported
injecting status. Because of the strong repeated
messages against drug injecting during the
study in both communes, it was likely that
the zero incidence of injecting during the
assessment period (2007---2009) reflected
bias. The greater decline in HIV prevalence
in Ha Tu, however, suggested that there was
greater decline in injecting in the Ha Tu than

in the comparison area.
Our results indicated that positive results
from both urine testing for opioids and blood
testing for HIV declined, somewhat more
in the intervention area. The reported incidence of any drug use increased in both
areas, although the rate of incidence increase
was lower in the intervention area. The
reported prevalence of injecting drugs at
least once in the past month was lower in
both areas at the second survey, but the rate
of decline was greater in the comparison
area. The greater decline in the prevalence of
positive urine testing results and prevalence
of positive HIV results in Ha Tu suggested
a greater impact of the intervention in Ha Tu,
despite the lack of statistical significance

194 | Research and Practice | Peer Reviewed | Nguyen et al.

for each element alone. The data also suggested that new drug users were more likely
to not inject at the time of the second survey
than at the baseline survey.
The results of this community intervention
study paralleled those observed by Wu et al.
in rural China.10 The common factor was the
design of the intervention by the community,
and thus, their ownership of the intervention.
The details of the intervention, however, differed, reflecting the differences in the cultures
of the 2 countries and the urban (Vietnam)
versus rural (China) settings of the intervention.

Therefore, it was not the details of the intervention strategy that were important in our
opinion, but the local ownership of the intervention. It was this strategy of mobilizing
the community to recognize and take responsibility for designing and implementing
the intervention that should be adopted by
other localities with significant problems
with drug use among their youths. Health
officials are, therefore, likely to achieve better
success at preventing drug use by working with
affected communities to develop interventions appropriate for those communities and
involving the commune leaders in designing
the intervention. j

About the Authors
Roger Detels and Ronald S. Brookmeyer are with the
Fielding School of Public Health, University of CaliforniaLos Angeles. Hien Tran Nguyen, Nguyen Binh Nguyen, and
Diep Bich Vu are with the National Institute of Hygiene and
Epidemiology, Hanoi, Vietnam. Anh Viet Tran and Son
Hong Nguyen are with the Hanoi Medical University,
Hanoi. Nhu To Nguyen is with the Family Health International 360, Hanoi.
Correspondence should be sent to Roger Detels,
Department of Epidemiology, UCLA Fielding School of Public
Health, 71-269 CHS, Box 951772, Los Angeles, CA
90095-1772 (e-mail: ). Reprints can be
ordered at by clicking the “Reprints” link.
This article was accepted May 13, 2014.

Contributors
H. T. Nguyen implemented the study and participated in
analysis and writing the article. A. V. Tran assisted with
study implementation and data management. N. B.

Nguyen performed data analysis and took part in writing
the article. S. H. Nguyen assisted with study implementation. D. B. Vu took part in data management and analysis.
N. T. Nguyen took part in data management and analysis.
R. S. Brookmeyer provided the statistical design and contributed to writing the article. R. Detels took part in study
design, implementation, analysis, and writing the article.

Acknowledgments
This study was funded by National Institutes of Health/
Fogarty International Center grant D43 TW005795.

American Journal of Public Health | January 2015, Vol 105, No. 1


RESEARCH AND PRACTICE

Human Participant Protection
The study was approved by the ethical institutional
review boards of both Hanoi Medical University and the
University of California, Los Angeles.

References
1. Des Jarlais DC, Kling R, Hammett TM, et al.
Reducing HIV infection among new injecting drug users
in the China-Vietnam Cross Border Project. AIDS.
2007;21(suppl 8):S109---S114.
2. Tran TN, Detels R, Long HT, Phung LV, Lan HP.
HIV infection and risk characteristics among female sex
workers in Hanoi, Vietnam. J Acquir Immune Defic Syndr.
2005;39(5):581---586.
3. Nguyen VT, Scannapieco M. Drug abuse in Vietnam:

a critical review of the literature and implications for
future research. Addiction. 2008;103(4):535---543.

18. Kotler P, ed. Marketing for Nonprofit Organizations.
2nd ed. Englewood Cliffs, NJ: Prentice Hall; 1975.
19. Singh N. AIDS outreach program targeting injecting
drug user community intervention program. Int Conf
AIDS. 1996;11:503 (abstract 1419).
20. Coyle SL, Needle RH, Normand J. Outreach-based
HIV prevention for injecting drug users: a review of
published outcome data. Public Health Rep. 1998;113
(suppl 1):19---30.
21. Kalichman SC, Simbayi LC. HIV testing attitudes,
AIDS stigma, and voluntary HIV counselling and testing
in a black township in Cape Town, South Africa. Sex
Transm Infect. 2003;79(6):442---447.
22. Sullivan SG, Xu J, Feng Y, et al. Stigmatizing attitudes
and behaviors toward PLHA in rural China. AIDS Care.
2010;22(1):104---111.

4. UNAIDS, World Health Organization. AIDS Epidemic
Update. Geneva, Switzerland: World Health Organization; 2007.
5. Ministry of Health. Vietnam. National Report on the
HIV/AIDS Situation, Prevention and Control Activities in
2011: Direction and Major Tasks in 2012. Hanoi,
Vietnam: Ministry of Health; 2012.
6. Ministry of Health. Vietnam: HIV/AIDS Control
Review. Hanoi, Vietnam: Ministry of Health; 2008.
7. Mathers BM, Degenhardt L, Phillips B, et al. Global
epidemiology of injecting drug use and HIV among

people who inject drugs: a systematic review. Lancet.
2008;372(9651):1733---1745.
8. Panda S, Sharma M. Needle syringe acquisition
and HIV prevention among injecting drug users:
a treatise on the “good” and “not so good” public
health practices in South Asia. Subst Use Misuse.
2006;41(6-7):953---977.
9. Giesbrecht N, Haydon E. Community-based interventions and alcohol, tobacco and other drugs: foci,
outcomes and implications. Drug Alcohol Rev. 2006;
25(6):633---646.
10. Wu Z, Detels R, Zhang J, et al. Risk factors for
intravenous drug use and sharing equipment among
young male drug users in Longchuan County, southwest
China. AIDS. 1996;10(9):1017---1024.
11. Detels R. HIV/AIDS in Asia: introduction. AIDS
Educ Prev. 2004;16(suppl A):1---6.
12. Wu Z, Zhang J, Detels R, et al. Characteristics of
risk-taking behaviors, HIV/AIDS knowledge and risk
perception among young males in southwest China. AIDS
Educ Prev. 1997;9(2):147---160.
13. Wu Z, Detels R, Zhang J, Li V, Li J. Communitybased trial to prevent drug use in Yunnan, China. Am J
Public Health. 2002;92(12):1952---1957.
14. Rothman J. Three Models of Community Organization
Practice, Their Mixing and Phasing. Strategies in Community
Organization. 3rd ed. Itasca, IL: F. F. Peacock Publishers
Inc.; 1979:86---102.
15. Bandura A. Social Learning Theory. Englewood
Cliffs, NJ: Prentice Hall; 1977.
16. Gossop M, Grant M. Preventing and Controlling Drug
Abuse. Geneva, Switzerland: World Health Organization;

1990.
17. Kotler P, Zaltman G. Social marketing: an approach
to planned social change. J Mark. 1971;35(3):3---12.

January 2015, Vol 105, No. 1 | American Journal of Public Health

Nguyen et al. | Peer Reviewed | Research and Practice | 195