Int.J. Behav. Med.

DOI 10.1007/s12529-014-9401-y

Impact of an Educational Programme on Reproductive Health

Among Young Migrant Female Workers in Shenzhen, China:
an Intervention Study
Chunyan Zhu & Qingshan Geng & Li Chen &
Hongling Yang & Wei Jiang

# International Society of Behavioral Medicine 2014

Abstract
Background Reproductive health and sexually transmitted diseases (STD) account for a high proportion of health problems
in the rural-to-urban young female migrant workers in China.
Improving these conditions remains highly challenging.
Purpose To developed an educational programme to advance
the reproductive health of the female workers.
Method An intervention study was conducted between July
2010 and April 2011 in Shenzhen. Two commune factories
were selected to participate and provided a control cluster receiving routine local government health services and a second
cluster receiving an educational intervention in addition to the
routine services. The intervention included distribution and free
access to educational study materials. The factory workers
knowledge, attitudes and behaviour in the area of reproductive
health and STD were the main study outcomes.
Results Compared with the control cluster, at the 6-month
follow-up assessment, the intervention cluster had a significantly
C. Zhu
Department of Preventive Medicine, School of Public Health,
Guangzhou Medical University, Guangzhou 510182, Peoples
Republic of China
Q. Geng (*)
Guangdong General Hospital, Guangdong Academy of Medical
Science, Guangzhou 510080, Peoples Republic of China
e-mail: gengqs2010@163.com
L. Chen
Bureau of Health Care, Department of Health of Guangdong
Province, Guangzhou 510060, Peoples Republic of China
H. Yang
Guangzhou Women and Childrens Medical Centre, Guangzhou
Medical University, Guangzhou 510623, Peoples Republic of China
W. Jiang
Department of Internal Medicine, Department of Psychiatry and
Behavioral Sciences, Duke University Medical Center, Durham, NC,
USA

higher proportion of correct answers to queries about human

immunodeficiency virus/acquired immunodeficiency syndrome
(HIV/AIDS) (standardised coefficients of multiple linear regression (B) 0.047; P=0.020) and awareness of places providing
free contraceptives (odds ratio [OR] 2.011, 95 % confidence
interval [CI] 1.6352.472; P<0.001), and a significantly lower
proportion accepting premarital sex (OR 0.492, 95 % CI 0.416
0.582; P<0.001), practising premarital sex (OR 0.539, 95 % CI
0.4780.608; P<0.001) or suffering from gynaecological disorders (OR 0.801, 95 % CI 0.6970.921; P=0.002).
Conclusion A community-based educational intervention
targeting unmarried female migrant workers appears to be
effective in substantially improving their knowledge of reproductive health and their attitudes and behaviour towards
health, and in reducing prevalence of STD.
Keywords Reproductive health . Sexually transmitted
diseases . Female . Community trial . China

Introduction
Reproductive health problems remain the leading cause of ill
health and death for women of childbearing age worldwide.
Impoverished women, especially those living in developing
countries, suffer disproportionately from unintended pregnancies, maternal death and disability, sexually transmitted infections including HIV, gender-based violence and other problems related to their reproductive system and sexual behaviour
[1]. A recent systematic survey in five Chinese provinces
found that over 54 % of urban women interviewed had had
sex prior to marriage, of whom 1232 % became pregnant and
almost all ended with induced abortions [2]. The incidence of
sexually transmitted disease (STD) and other reproductive
health problems in China has increased over the last decade,
at approximately 2030 % per year, particularly among migrant workers [3]. The WHO has recorded 1620 million new

Int.J. Behav. Med.

STD cases occurring every year in China [4]. Younger individuals and migrating workers are at particularly high risk of
STDin Guangzhou, 33.4 % of HIV-infected patients were
migrants [5] and 30.8 % of migrant female workers contracted
a gynaecological disease [6]. However, few of them seek
medical attention, with one study noting that only 18.3 % of
sick individuals had healthcare insurance, 18.9 % believing
medical care was too expensive and 13.5 % reporting they had
no money to see a doctor [6]. Reproductive ill-health and STD
have become a grave public health issue in China and a
tremendous healthcare burden.
Emerging evidence suggests that a substantial reduction in
sexual and reproductive health problems may be achieved by
means of simple and low-cost interventions in primary
healthcare facilities and community settings [712], focusing
on four priority areasidentifying HIV status, promoting
safer sex, optimising links between HIV and sexually transmitted infection services, and dealing with HIV alongside
maternal and infant health [13, 14]. In a study in subSaharan Africa, Madeni et al. [9] reported a significant increase in knowledge and behaviour in the study population
after a 45-min education programme. Al-Iryani et al. reported
that a school-based peer education intervention improved
levels of HIV knowledge about modes of transmission and
prevention, and at the same time decreased levels of stigma
and discrimination in a culturally conservative setting [10].
Another study reported that peer education and conventional
education strategies were effective in improving knowledge
about reproductive health among adolescent girls [11].
Whether and to what extent an education package of
evidence-based interventions at the community level could
improve reproductive health knowledge and reduce reproductive health disease in China is unknown. We therefore undertook an educational project using cluster randomisation to
advance the reproductive health of rural-to-urban unmarried
female migrant workers. We hypothesized that an intervention
targeting such women would produce improvement in their
knowledge of reproductive health, a change of attitude towards premarital sex and a consequent reduction in reproductive health problems.

condensed manufacturing industry or handicraft trades, chiefly undertaking manual work for more than 12 h a day. Boaan
lies at the northwest tip of Shenzhen and is the site of about
1,000 overseas-funded manufacturing units. Socially, workers
are generally confined within each factory and receive certain
community-based health services provided by the local government. Two of these factories, producing motor parts, were
selected to participate in the study (Fig. 1) according to the
following criteria: (1) the factories should have little social
interaction with the outside world, or with each other (there
was in fact a considerable distance between the two); (2) each
factory should have 20,000 or more workers, at least 80 % of
whom were young migrant unmarried females; (3) the factory
administrators should be willing to cooperate in the implementation of the study and (4) community-based health services were provided by the local government for these
workers.
Study Design
The study was conducted between July 2010 and April 2011.
Using a computer-generated randomisation number, one factory was assigned to receive the intervention, the other serving
as the control. For the purpose of cluster randomisation, the
factory was the basic unit for implementing the intervention.
The intervention was not concealed. Because of the distance
(35.5 km) between the factories and the relative social isolation of their work forces, any exchange of knowledge between
the two clusters was minimal.
All study procedures were approved by the ethics review
committees of Guangzhou Medical University, and all participants provided voluntary informed consent. There were some
sensitive questions in the questionnairesexual concepts and
behaviour are sensitive subjects in China, induced abortion
among unmarried females is usually condemned and people
are not willing to talk about these matters with others. As a
result, the participants did not want their names to be used in
the survey. Voluntary verbal informed consent consisted of a
yes response to the question: The survey is about your
knowledge, attitude and behaviour in the field of reproductive
health. It is anonymous, and we will protect your privacy.
Would you be willing to take part in the survey?

Materials and Methods

Intervention
Study Area and Population
Details of the study location have been published previously
[15]. Briefly, two overseas-funded factories in the Baoan
district of Shenzhen were selected for the study. Shenzhen,
the first special economic zone established by the Chinese
central government in 1980, housed 13.22 million migrant
workers in 2010, 81.0 % of the total population. Of those
13.22 million migrants, 47.2 % were female, many working in

The intervention cluster received both the usual local government health services and the intervention package, while the
control cluster received only the government services, and the
duration of the intervention was 6 months. These are delivered
by the community health centre at each factory and are
summarised in Table 1. The intervention package consisted
of free access to educational materials distributed widely to
the female workers, free monthly lectures about reproductive

Int.J. Behav. Med.

Fig. 1 Trial profile

Table 1 Components of healthcare service: study intervention vs. usual care

Components

Intervention

Usual care

Health education

1. Information promoting physical activity

2. Assistance to manage better and/or cope with certain health
conditions, such as influenza, lung tuberculosis or hepatitis
3. Knowledge about reproductive health (healthy sexual activities;
healthy and appropriate use of contraception; STDs and
HIV/AIDS knowledge, effects, prevention, symptoms; appropriate
ways to obtain health care for STD/HIV/AIDS; and menstruation
healthcare)
1. Promoting mental and physical health, such as mental health,
reasonable diet and exercise
2. Disease and injury prevention, such as influenza or workplace
injury prevention
3. Prevention of STD and other reproductive system problems,
such as HIV/AIDS, other STDs and unwanted pregnancy
Postpartum follow-up

1. Information promoting physical activity

2. Assistance to manage better and/or coping
with certain health conditions, such as
influenza, lung tuberculosis or hepatitis

Community preventive

Community health service

1. Promoting mental and physical health,

such as mental health, reasonable diet
and exercise
2. Disease and injury prevention, such as
influenza or workplace injury prevention
Postpartum follow-up

Int.J. Behav. Med.

health, counselling classes and access to contraceptives.

Gynaecological care, covering examination, diagnosis and
treatment, was provided when needed without charge by the
community healthcare staff who had received gynaecological
training from the study investigators. Before actually
implementing the intervention, the investigators collected
comments from the community healthcare staff, workers
representatives and factory administrators for the purpose of
refining the study procedures to ensure successful
implementation.
Data Collection
Two cross-sectional surveys were conducted to obtain research
data during July and August 2010 before introducing the educational intervention and then 6 months later (MarchApril
2011) in both factories, using the same questionnaires and
methods to guarantee the compatibility of the comparisons.
The data collectors and their supervisors were female students
on undergraduate or masters degree courses at Guangzhou
Medical University, blinded as to the cluster randomisation
and independent of the educational intervention. Standard procedures were established to guide the recruitment, training and
supervision of these data collectors, who received training
before each survey to improve their skills and reduce
intervariation/intravariation in the interview/data collection.
Information on demographics and knowledge levels and on
the attitudes and behaviour of reproductive health services was
obtained via face-to-face interview by the data collectors from
randomly selected samples of all migrant unmarried female
manual workers in both intervention and control factories.
The representative sampling was primarily based on the
distribution of the factory dormitories, which in both factories
are grouped into five areas, east, west, north, south and central.
Randomly, two dormitory buildings were first selected from
each of the five compounds of each factory, then two floors
from each dormitory building, generating a sample of 20 floors
in each factory for each data collection period. Although the
size and the number of beds of each dormitory were identical
and all dormitories were accessible to the study, the number of
unmarried female workers in each varied at times. The final
sample size was therefore different for each factory (Fig. 1).
The primary outcomes were changes in knowledge, attitude and behaviour in the area of reproductive health and
changes in the prevalence of reproductive health problems at
baseline and at the end of the 6-month intervention period.
Gynaecological diseases were identified by the selfreported presence of illnesses via questions like Have you
had any gynaecological disorder, such as menstrual problems,
over the past 6 months?, which were then confirmed by a
medical doctor. Induced abortion was assessed by a self-report
yes response to the question Have you had an induced
abortion in the past 6 months?.

Statistical Analysis
The required sample size was calculated to provide estimates
of the prevalence of reproductive health problems among
unmarried female workers for each cross-sectional survey.
Based on the demographic characteristics of the factories
and their workers annual visits to the community healthcare
centre, we estimated the prevalence of reproductive health
problems among unmarried female workers to be approximately 30.0 %. With 80 % power at a 5 % level of significance, a minimum of 934 participants was required from each
factory for each survey.
Students t tests, chi-square or Wilcoxon tests were used to
test the differences between the intervention and the control
clusters at the baseline and the 6-month assessment in respect
of continuous and categorical measures. Multivariate linear
regression models were used to identify effects of the intervention on knowledge-about-HIV/AIDS scores, and multivariate logistic regression models were used to identify the effects, in both clusters, of the intervention on sexual and
reproductive health indicators over the 6-month intervention
period. The intervention effect was estimated using the adjusted odds ratio (OR) of practice indicators (including knowledge, attitudes, sexual behaviour and reproductive health, and
preventive health services and demand) for the intervention
relative to the comparison, and 95 % confidence interval (CI)
for the risk ratios (ORs) were calculated on a logarithmic scale
using a Taylor series approximation, controlling for age, education level and work duration in the factory.
All data forms underwent scrutiny for logical inconsistencies, skip patterns and missing values. The data was coded
and double-entered into a relational database on EpiData 3.1.
All analyses were carried out at cluster level using SPSS 15.0.
P values of less than 0.05 were considered statistically
significant.

Results
Study Population
The baseline survey recruited 2,980 unmarried female
workers in the intervention factory and 1,060 in the control
factory. The 6-month survey recruited 1,425 in the intervention and 2,139 in the comparison factories (Fig. 1). All sample
sizes exceeded estimates.
Sociodemographics
Table 2 shows the basic demographic characteristics of the
study population. All participants were female single manual
workers, whose ages ranged from 15 to 29, with an average of
20.13.7, most (>95.0 %) being younger than 25. As regards

Int.J. Behav. Med.

Table 2 Key characteristics of participants at baseline and 6-month follow-up surveys

Age (years)

Mean (SD)
Range
Age group (years), % <20
2024
2529
Education level, %
Elementary
Secondary
Tertiary
Work duration in the <1
factory (years), % 1<2
2

Baseline

6-month follow-up

Intervention (N1=2,980) Control (N3=1,060) P

Intervention (N2=1,425) Control (N4=2,139) P

20.0 (2.0)
1629
1,292 (43.4)
1,607 (53.9)
81 (2.7)
2,168 (72.8)
794 (26.6)
18 (0.6)
1,735 (58.2)
562 (18.9)
683 (22.9)

19.3 (1.8)
1527
657 (62.0)
394 (37.2)
9 (0.8)
658 (62.1)
349 (32.9)
53 (5.0)
619 (58.4)
237 (22.4)
204 (19.2)

educational attainment, 70.2 % had received only elementary

education, and 61.2 % were in their first year of employment
in the factories.

Knowledge, Attitudes and Behaviour in the Area

of Reproductive Health
Comparisons of Baseline Information
Between the Intervention and Control Clusters
At the baseline, the intervention clusters knowledge about
reproductive health was less than the controls, except in the
case of the questions STDs are serious and HIV can be
transmitted through contaminated hypodermic needles,
where similar answers were given (Table 3). Overall, the
majority of participants in the intervention (60.9, 73.5 and
67.0 %, respectively) and in the control (64.3, 69.7 and
67.7 %, respectively) thought that induced abortion is serious, STDs are serious and STDs can be prevented.
Moreover, 89.0, 77.8, 82.0 and 77.8 % of intervention participants and 90.9, 79.8, 76.8 and 75.3 % in the control cluster,
respectively, knew HIV/AIDS could be transmitted via infected blood, sexual intercourse, sharing syringes/needles or vertical transmission from mother to child. However, lower rates
in the intervention (37.7 %) and the control (45.3 %) clusters
knew that condoms could prevent HIV transmission during
sexual intercourse, and many had false beliefs about HIV
transmission paths. For instance, 98.1 and 86.9% in the intervention and control clusters, respectively, did not know that
HIV could not be transmitted by a mosquito bite; 73.0 and
71.3 % in the intervention control clusters, respectively, did
not know that HIV could not be transmitted through shaking
hands with or embracing someone who had the condition and
about half (46.2 and 47.3 % in intervention and control

0.001 20.6 (2.2)

1629
479 (33.6)
874 (61.3)
72 (5.1)
0.001 1,025 (71.9)
377 (26.5)
23 (1.6)
0.001 745 (52.3)
376 (26.4)
304 (21.3)

20.8 (2.1)
1729
564 (26.4)
1,468 (68.6)
107 (5.0)
1,357 (63.4)
763 (35.7)
19 (0.9)
1,443 (67.5)
339 (15.8)
357 (16.7)

0.094

0.001

0.008

clusters, respectively) thought HIV could be transmitted

through sharing a meal with an infected person.
Intervention cluster participants, relative to those in the
control cluster, had a significantly higher rate of believing
their peers had had premarital sex and lower rates of ever
having had premarital sex, induced abortion, knowing where
to obtain free contraceptives and free health education (all P
values <0.05) (Table 4).
Comparisons Between Baseline and 6-Month Assessment
Within Each Cluster
In the intervention cluster, relative to the baseline assessment,
all areas of newly acquired knowledge showed improvement
except the STDs are serious question (Table 3). In the control
cluster, knowledge of the following areas increased significantly (all P values <0.05) relative to the baseline: induced
abortion is a serious health problem, HIV can be transmitted
through contaminated hypodermic needles, HIV cannot be
transmitted by a mosquito bite and HIV cannot be transmitted through shaking hands with or embracing an infected
person. However, the rate of correct answers to queries on
STDs are serious, STDs can be prevented, it is possible to
get HIV from having sex without a condom and HIV cannot
be transmitted through sharing a meal with someone who has
the disease was significantly lower relative to the baseline
assessment (all P values <0.05) (Table 3).
Comparison of Findings at 6 Months Comparing Intervention
and Control
At the 6-month point, the knowledge rates of intervention
participants regarding AIDS is an STD, STDs are serious,
STDs can be prevented, it is possible to get HIV from
having sex without a condom and HIV cannot be transmitted

Int.J. Behav. Med.

Table 3 Change in level of knowledge about reproductive health by study arm at baseline and 6-month follow-up
Intervention

Control

Baseline
(N1=2,980)

6-month follow-up
(N2=1,425)

Ever heard of emergency contraception (%)

25.8

31.9b

<0.001

30.2

29.3

0.096

Induced abortion is serious (%)

AIDS is an STD (%)
STDs are serious (%)
STDs can be prevented (%)
Modes of HIV/AIDS transmission
HIV can be transmitted through a transfusion
of infected blood (%)
HIV can be transmitted through sex with someone
who is HIV-infected (%)
HIV can be transmitted through contaminated
hypodermic needles (%)
HIV can be spread by mother-to-child
transmission (%)
It is possible to get HIV from having sex
without a condom (%)
HIV cannot be transmitted by mosquito
bite (%)
HIV cannot be transmitted through sharing
a meal with someone who has HIV (%)
HIV cannot be transmitted through shaking
hands with or embracing someone who has HIV (%)

60.9
42.5
73.5
67.0

65.9b
52.8b
71.9
69.7

0.048
0.009
0.018
0.699

64.3
47.2
69.7
67.7

68.1a
46.7
60.1b
34.0b

0.176
<0.001
<0.001
<0.001

89.0

87.4

0.093

90.9

87.9

0.642

77.8

80.5a

0.169

79.8

81.2

0.595

82.0

83.2

<0.001

76.8

81.4b

0.178

77.8

77.3

0.096

75.3

75.7

0.272

37.7

43.8b

<0.001

45.3

34.5b

<0.001

10.9

23.0b

0.049

13.1

22.9b

0.939

46.2

48.1

0.541

47.3

42.3b

0.001

27.0

32.9b

0.307

28.7

34.5b

0.312

Comparison of baseline and 6-month values of intervention and control arms, P<0.05

Comparison of baseline and 6-month values of intervention and control arms, P<0.01

Baseline
(N3=1,060)

6-month follow-up
(N4=2,139)

Table 4 Change in attitudes about premarital sex, sexual behaviour and reproductive health, and preventive health services and needs, by study arm at
baseline and at 6-month follow-up
Intervention

Control

Baseline
(N1=2,980)

6-month follow-up
(N2=1,425)

Baseline
(N3=1,060)

6-month follow-up
(N4=2,139)

Premarital sex is acceptable (%)

Premarital sex is common among peers (%)
Ever had premarital sex (%)
Using contraception when having sexa (%)

3.3
69.1
7.4
71.6

4.1
68.4
10.9c
76.9

0.554
<0.001
0.001
0.280

2.9
57.9
10.8
78.1

16.7c
69.5c
31.3c
83.0

<0.001
0.487
<0.001
0.077

Ever had induced abortiona (%)

Ever had gynaecological disease (%)
Ever been treated for gynaecological diseaseb (%)
Think information about reproductive health
should be provided (%)
Know where free contraceptives are provided (%)
Know where free health education is
provided (%)

13.1
7.7
89.4
79.3

12.8
8.7
95.6
80.8

0.003
0.083
0.265
0.745

26.3
9.3
94.4
79.8

22.5
13.3c
96.3
95.6c

0.007
<0.001
0.781
<0.001

<0.001
<0.001

10.9
10.8

2.3c
7.0c

<0.001
0.750

5.5
3.5

7.4d
6.7c

Answered only by those who answered yes to ever had premarital sex

Answered only by those who answered yes to ever had gynaecological disease

Comparison of baseline and 6-month values of intervention and control arms, P<0.01

Comparison of baseline and 6-month values of intervention and control arms, P<0.05

Int.J. Behav. Med.

Table 5 Comparison of sexual and reproductive health indicators by study cluster at the 6-month point
Intervention
(%, N2=1,425)

Control
(%, N4=2,139)

Adjusted odds ratio

(intervention to control)a (95 % CI)

Agreed with premarital sex

Ever had premarital sex
Ever had induced abortion
Ever had gynaecological disease

4.1
10.9
12.8
8.7

16.7
31.3
22.5
13.3

<0.001
<0.001
0.468
0.002

0.492 (0.4160.582)
0.539 (0.4780.608)
0.873 (0.6061.259)
0.801 (0.6970.921)

Think information about reproductive health should be provided

Know where free contraceptives are provided
Proportion (%) of correct answers to questions on
HIV/AIDS knowledge (meanSD)c

80.8
7.4
59.524.1

95.6
2.3
57.622.8

<0.001
<0.001
0.020

0.476 (0.4110.550)
2.011 (1.6352.472)
0.047b

Adjusted for age, education level and work duration in the factory

Standardised coefficients (B) of intervention to control

The eight questions on modes of HIV/AIDS transmission included the following: (1) HIV can be transmitted through a transfusion of infected blood, (2)
HIV can be transmitted through sex with someone who is HIV-infected, (3) HIV can be transmitted through contaminated hypodermic needles, (4) HIV
can be spread by mother-to-child transmission, (5) It is possible to get HIV from having sex without a condom, (6) HIV cannot be transmitted by
mosquito bite, (6) HIV cannot be transmitted through sharing a meal with someone who has HIV and (7) HIV cannot be transmitted through shaking
hands with or embracing someone who has HIV

through sharing a meal with an infected person are all significantly greater than those of the control cluster (all P values
<0.05) (Table 3). The rate of change (between baseline and 6month point) of the intervention participants who gave correct
answers to the HIV/AIDS questions was significantly higher
than that of control cluster participants (3.5 vs. 1.1 %; P<0.05;
data not shown).
At the 6-month point, the intervention cluster had significantly lower rates of believing premarital sex was acceptable,
admitting having had premarital sex and having undergone an
induced abortion or suffered from a gynaecological disease
over the past 6 months, than the control cluster. Furthermore,
the intervention cluster showed a significant increase in their
knowledge of where to obtain free contraceptives and health
education, while this area of knowledge in the control cluster
was significantly reduced compared with the baseline (all P
values <0.05) (Table 4).
After adjusting for age, education level and work duration
in the factory at the 6-month point, the proportion of correct
answers to questions about HIV/AIDS (B 0.047; P=0.020)
and where to obtain free contraceptive (OR 2.011, 95 % CI
1.6352.472; P<0.001) was significantly higher in the intervention cluster than in the control (Table 5). A pronounced
reduction was noted, in the intervention compared with the
control cluster, in permitting premarital sex (OR 0.492, 95 %
CI 0.4160.582; P<0.001), experience of premarital sex (OR
0.539, 95 % CI 0.4780.608; P<0.001), gynaecological disease (OR 0.801, 95 % CI 0.6970.921; P=0.002) and belief
that reproductive health information was needed (OR 0.476,
95 % CI 0.4110.550; P<0.001), and there was no significant
difference in ever had induced abortion between intervention
cluster and the control (Table 5).

Discussion
The study demonstrated that implementation of an educational
programme aimed at improving the knowledge, attitude and
behaviour in the area of reproductive health among unmarried
female migrant workers, via cluster randomisation, was associated with significant improvements in participants knowledge of reproductive health, HIV/AIDS and premarital sexual
behaviour. The interventions were based on findings from
formative research and the input of community healthcare
providers to address the fundamental needs of unmarried
female workers and implemented with the active participation
of the community administrative and healthcare staff. The
study highlights the importance of understanding the existing
sociocultural context when translating scientific evidence into
effective and sustainable reproductive health strategies at the
community level, involving community healthcare providers
in decisions affecting implementation of the study, as well as
building the communitys capacity for sustained action.
The findings of the study are consistent with the results of
previous research in countries other than China. One review
evaluated 37 interventions from 28 studies in Latin American
and Caribbean nations, and noted that school- or communitybased interventions that included about 3 h of STD/HIV/
AIDS instruction and behavioural change significantly increased both knowledge and condom use, though the effects
varied widelythe interventions produced more condom use
when focused on high-risk individuals via distributing condoms to them and explicitly addressing sociocultural factors
[16]. OGrady and colleagues reported that a single-session
peer-led safer sex intervention, based on the informationmotivation-behavioural skills theoretical model, increased

Int.J. Behav. Med.

college students information about HIV and other STDs,

motivation to practise preventive behaviour and skills to implement safer sexual behaviour (e.g. keeping condoms available) [17].
Certain foreign studies noted that knowledge and attitude
were most likely to change, while changes in sexual behaviour
(such as sexual debut and condom use) were less likely to
occur [18, 19]. There has been one previous study in China
aimed at testing the feasibility of implementing a workplacebased intervention to promote contraceptive use among unmarried female migrants in a privately owned factory, using a
quasi-experimental before-and-after design [20]. The results
of the study were somewhat contradictorywhile 90 % of the
598 participating women at baseline considered using contraceptives if having sex before marriage, only 78 % (N=62/79)
reported having used them when having sex over the 3-month
study period. Faced with these findings, the authors concluded
that implementing a complex intervention with a hard-toreach population through a factory in China, using a quasiexperimental design, was not at all easy.
Since the present intervention aimed to disseminate information to alter attitudes and behaviour in a certain population,
randomly assigning the intervention at the individual level
would not guarantee concealment and cluster randomisation
at the factory level was therefore chosen. Although a clusterrandomised controlled trial is considered the most appropriate
design for studies of this nature, it is not without methodological limitations and biases [2123]. Firstly, a potential limitation of the present study was that the changes at the endpoints
relied on retrospective recall. The standardised and unified
data collection methods used at both factories, at baseline and
6 months, were to eliminate the differences in potential recall
lapses across the study arms and over time. Secondly, certain
interventions were delivered by the community health centre
staff in the factory as part of the programme implementation.
Such an approach helped the local staff to build rapport with
the workers and probably enhanced the effectiveness of the
programme, but individual variations and possible inconsistency remained concerns. However, utilising independent data
collectors to gather study information, rigorous monitoring,
data quality assurance and random checks on implementation
are believed to have kept the potential bias to a minimum.
Thirdly, the issue of contamination and diffusion between
intervention and control clusters might be a concern. There
were no transfers or migrations of female workers between
intervention and control factories during the trial, though the
possibility could not be entirely dismissed of an exchange or
diffusion of information between the two factories through
community medical workers and female workers. Fourthly,
because of the high mobility of the workers (more than 50 %
had been employed for less than a year at the time of the
study), most of the study population did not receive any
intervention for 6 months, thus possibly minimising its

effectiveness. Furthermore, the mobility of the workers made

for a high rate of loss to the follow-up, and 88.5 % of the
population at the 6-month point had not participated in the
baseline assessment. Two cross-sectional surveys were carried
out to obtain information at baseline and 6 months in the same
communities, and conservative analytical methods were used
as recommended for analysis of cluster-randomised controlled
trials. Fifthly, health education serving both intervention and
control clusters were strengthened during the trial, the control
population receiving basic reproductive healthcare through
the existing health education programme.
Despite these limitations and the short intervention period,
the trial provides encouraging results, suggesting that a
community- or workplace-based educational intervention programme to promote reproductive health can lead to improvement in knowledge, attitudes and behaviour and thus healthier
practice among unmarried migrant female workers and in
certain other communities in China. Further, the results suggest that educational interventions to modify other popular
unhealthy attitudes and behaviour may also be effective in
communities with condensed populations and similar cultural
backgrounds.
Acknowledgments We wish to thank Dr. Feiyue Chai of Shajing
General Hospital in Shenzhen City and Dr. Xianhua Fu of Guangzhou
Medical University for their kind assistance.
Informed Consent All procedures followed were in accordance with
the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of
1975, as revised in 2000. Informed consent was obtained from all patients
for being included in the study.
Conflict of Interest All authors declare that they have no conflict of
interest.

References
1. UNFPA. Improving reproductive health. 2013. http://www.unfpa.
org/rh/. Accessed Dec. 22, 2013.
2. Qian X, Tang SL, Garner P. Unintended pregnancy and induced
abortion among unmarried women in China: a systematic review.
BMC Health Serv Res. 2004. doi:10.1186/1472-6963-4-1.
3. Gong XD. STD epidemiology and control. Bull for STI Prev and
Ctrl. 2011;25:913.
4. Wu SM. Status and trends of STIs epidemic in China. Acta Med Sin.
2006;19(5):103941.
5. Zou XP. Investigation of reproductive health status and service needs
among unmarried young men in Changshu. Zhonghua Nan Ke Xue.
2002;8(5):3529.
6. Yin Q, Gong QP, Mao JS, Zong ZH, Yu LL, Xu P. The living
conditions of floating women in Nanjing city. J Nanjing Colg Pop
Pgm Mgmt. 2008;24(3):412.
7. Speizer IS, Magnani RJ, Colvin CE. The effectiveness of adolescent
reproductive health interventions in developing countries: a review of
the evidence. J Adolesc Health. 2003;33(5):32448.

Int.J. Behav. Med.

8. Hawks SR, Madanat HN, Merrill RM, Goudy MB, Miyagawa T. A
cross-cultural comparison of health promoting behaviors among
college students. IEJHE. 2002;5:8492.
9. Madeni F, Horiuchi S, Iida M. Evaluation of a reproductive health
awareness program for adolescence in urban Tanzaniaa quasiexperimental pre-test post-test research. Reprod Health. 2011;8:21.
doi:10.1186/1742-4755-8-21.
10. Al-Iryani B, Basaleem H, Al-Sakkaf K, Crutzen R, Kok G, van den
Borne B. Evaluation of a school-based HIV prevention intervention
among Yemeni adolescents. BMC Public Health. 2011;11:279. doi:
10.1186/1471-2458-11-279.
11. Parwej S, Kumar R, Walia I, Aggarwal AK. Reproductive health
education intervention trial. Indian J Pediatr. 2005;72(4):28791.
12. de Anda D. Replication of an intensive educational intervention for
youth pregnancy and STI prevention: the GIG. Child Adolesc Soc
Work J. 2008;25:5569.
13. Sarah ML. Make or break? 2010: a pivotal year for scaling up RH/
HIV integration and accelerating progress towards MDGS 5 and 6.
2010. http://www.aidsalliance.org/includes/document/RH_
REPORT.pdf. Accessed 21 Jan 2014.
14. Dickinson C, Attawell K, Drucea N. Progress on scaling up integrated services for sexual and reproductive health and HIV. Bull WHO.
2009;87:84651.
15. Zhu CY, Wang JJ, Fu XH, Zhou ZH, Zhao J, Wang CX. Correlates of
quality of life in China rural-to-urban female migrate workers. Qual
Life Res. 2012;21(3):49503.

16. Huedo-Medina TB, Boynton MH, Warren MR, Lacroix JM, Carey
MP, Johnson BT. Efficacy of HIV prevention interventions in Latin
American and Caribbean nations, 19952008: a meta-analysis. AIDS
Behav. 2010;14:123751.
17. OGrady MA, Wilson K, Harman JJ. Preliminary findings from a
brief, peer-led safer sex intervention for college students living in
residence halls. J Prim Prev. 2009;30:71631.
18. Gallant M, Maticka-Tydale E. School-base HIV prevention
programmes for African youth. Soc Sci Med. 2004;58:1337
51.
19. Paul-Ebohimhen VA, Poobalan A, van Teijlingen ER. A systematic
review of school-based sexual health interventions to prevent STI/
HIV in sub-Saharan Africa. BMC Public Health. 2008;8:4. doi:10.
1186/1471-2458-8-4. 10.
20. Qian X, Smith H, Huang W, Zhang J, Huang Y, Garner P. Promoting
contraceptive use among unmarried female migrants in one factory in
Shanghai: a pilot workplace intervention. BMC Health Serv Res.
2007;7:17. doi:10.1186/1472-6963-7-77.
21. Puffer S, Torgerson DJ, Watson J. Cluster randomized controlled
trials. J Eval Clin Pract. 2005;11:47983.
22. Hahn S, Puffer S, Torgerson DJ, Watson J. Methodological bias in
cluster randomised trials. BMC Med Res Methodol. 2005;5:10. doi:
10.1186/1471-2288-5-10.
23. Lewis SC, Warlow CP. How to spot bias and other potential problems
in randomised controlled trials. J Neurol Neurosurg Psychiatry.
2004;75:1817.