Archives of Sexual Behavior

, Volume 36, Issue 6, pp 849–859

Variations in Condom Use by Locale: A Comparison of Mobile Chinese Female Sex Workers in Hong Kong and Mainland China

Authors

    • Centre for Epidemiology and Biostatistics, School of Public Health, Faculty of MedicineThe Chinese University of Hong Kong
    • Centre for Epidemiology and Biostatistics, School of Public Health, Faculty of MedicineThe Chinese University of Hong Kong
  • H. Y. Tsui
    • Centre for Epidemiology and Biostatistics, School of Public Health, Faculty of MedicineThe Chinese University of Hong Kong
  • Shara P. Y. Ho
    • Community Health Organisation for Intervention, Care and Empowerment
Original Paper

DOI: 10.1007/s10508-007-9203-2

Cite this article as:
Lau, J.T.F., Tsui, H.Y. & Ho, S.P.Y. Arch Sex Behav (2007) 36: 849. doi:10.1007/s10508-007-9203-2
  • 117 Views

Abstract

The present study investigated the prevalence of HIV and HIV-related behaviors and perceptions among 336 female sex workers (FSW) traveling from mainland China to work in Hong Kong. None of the participants tested HIV positive. Of the participants, 13.2% and 26.8%, respectively, would have unprotected sex with a client in Hong Kong or in mainland China if they liked the client; 16.7% and 21.7%, respectively, would have unprotected sex with a client in Hong Kong and in mainland China if he paid more. Of the participants, 98.5% used condoms in the last episode of sex work in Hong Kong and 78.1% were consistent condom users in Hong Kong in the last month. Yet, among those who previously worked as FSW in mainland China (n = 141), only 28.6% used condoms consistently in mainland China in the last year; 60% of those who used condoms inconsistently while in mainland China became consistent condom users in Hong Kong. Among those who did not use condoms in mainland China, willingness to have unprotected sex with a client who was liked by the FSW was a strong predictor for consistent condom use in Hong Kong (OR = 18.67, p < .001). Among those who were consistent condom users in Hong Kong, the same variable predicted inconsistent condom use while in mainland China (OR = 14.02, p < .05). Condom use behaviors change when FSW are working in different social contexts and behavioral changes are possible. Education programs should take socio-contextual factors into account.

Keywords

HIV/AIDSSex workersRisk factorsCondomsChina

Introduction

Female sex workers (FSW) are vulnerable in contracting sexually transmitted diseases (STD) and HIV (Ghys, Jenkins, & Pisani, 2001). It is estimated that 4–6 million highly mobile FSW are working in mainland China (The UN Theme Group on HIV/AIDS in China, 2002; Yang et al., 2005b). Mobility increases the difficulty of implementing effective HIV prevention programs for FSW in China. A recent study documented a low level of HIV-related knowledge and prevalent risk behaviors among these FSW (Lau, Tsui, Siah, & Zhang, 2002).

The national estimate of HIV/AIDS cases in mainland China approaches 650,000 (Ministry of Health, People’s Republic of China, UNAIDS, & WHO, 2006). Among all reported HIV cases, 39.3% of the HIV cases could be attributed to injecting drug use, 8.9% could be attributed to heterosexual transmission and such a proportion has been increasing over time (UNCHINA, 2005). The prevalence of STD among FSW in China ranged from 10% to 60% (Joint China Ministry of Health & UN Theme Group on HIV/AIDS in China, 1997). The HIV prevalence of FSW was found to be 1.4% in Guangdong (van den Hoek et al., 2001) and 1.23–6.67% in Yunnan (Lu et al., 2004).

There is a high volume of cross-border activities taking place between Hong Kong and mainland China (Department of Census and Statistics, 2005). Cross-border HIV-related risk behaviors have been prevalent among Hong Kong men (Lau & Tsui, 2003; Lau, Kim, Lau, & Tsui, 2004). A recent study showed that a cohort of male clients of FSW was much more likely to have used condoms when having sex with FSW in Hong Kong than when the same behavior was practiced in mainland China (Lau, Tang, & Tsui, 2003). However, no similar studies have been conducted to compare condom use behaviors among FSW working in Hong Kong and in China.

Hong Kong has a low prevalence of HIV (Advisory Council on AIDS, 2005). It is believed that HIV prevalence in Hong Kong in the future will largely depend on that of Guangdong province in mainland China (Lau & Tsui, 2003). Historically, Hong Kong had been a part of Guangdong. Hong Kong is proximal to a number of other Guangdong cities, which are frequently visited by Hong Kong people. Guangdong ranked fourth among all provinces in China in terms of number of reported HIV cases (UNCHINA, 2005) and the number of HIV cases as well as STD prevalence in Guangdong have been rising sharply (Lai et al., 2005; Zhong, Zheng, & Chen, 2003). A large number of FSW came to work in Hong Kong (Working group on HIV prevention for commercial sex workers and their clients community forum on AIDS Hong Kong Advisory Council on AIDS, 2006). Incentives include better income and the fact that prostitution is illegal in mainland China but not in Hong Kong. Frequency of condom use during sex work may be context dependent. For instance, the keepers of sex establishments play an important role in influencing condom use among FSW (e.g., Morisky et al., 1998; Yang et al., 2005a) and foreign clients were also more likely than local clients to be not using condoms (Wong et al., 2005). Therefore, geographical differentials in condom use among FSW may exist when they work in different places.

The study had several objectives. Firstly, the HIV prevalence of the study population was examined. Secondly, among a cohort of FSW who had worked both in Hong Kong and in mainland China, the frequency of self-reported condom use with their clients in Hong Kong versus in mainland China was compared. It was hypothesized that a geographic differential in self-reported condom use behaviors exists. Furthermore, factors in association with differentials in self-reported condom use with clients in Hong Kong versus in mainland China were investigated. These factors included HIV-related knowledge, perceived susceptibility to contracting HIV or STD, perceived benefits of condom use, and perceived behavioral control over condom use. Factors in association with self-reported condom use with clients in Hong Kong and in mainland China were similarly identified. The Health Belief Model (Rosenstock, 1974; Rosenstock, Strecher, & Becker, 1988) suggests that variables such as perceived susceptibility and perceived benefits to be predictive of the adoption of preventive behaviors whereas the Theory for Planned Behavior suggests that perceived behavioral control to be predictive of the performance of preventive behaviors (Ajzen, 1988).

In data analyses, univariate and multivariate logistic regression analyses were applied separately to those providing sex services in Hong Kong and in mainland China, so as to identify factors associated with self-reported condom use with clients in these two locales. To further identify factors predicting geographic differentials in self-reported condom use, similar logistic regression analyses were applied to a group of FSW who had not been using condoms consistently with clients in mainland China to predict whom of these FSW would change their behavior to become consistent condom users while working in Hong Kong. Similarly, analyses were applied to those who were consistent condom users in Hong Kong, to distinguish those who were not initially consistent condom users while working in mainland China from the others.

Method

Participants

The study population comprised FSW who had engaged in sex work in the last 6 months and came from mainland China to Hong Kong to serve in brothels in a red-light district (Mongkok). A mapping exercise identified about 500 such brothels in the district. A total of 15 brothels were randomly selected and cooperation was sought from the keepers. The majority (over 90%) of the FSW working in these brothels came from mainland China. Trained outreach workers visited the selected brothels and invited eligible FSW to join the study. Out of 379 invitations made, 336 (88.7%) FSW participated in the study.

Among all participants, 42.9%, 46.4%, and 10.7% respectively were of age 20 or less, 21 to 25, and 26 or above; 73.1% had an education level of junior high school or lower; 34.4% had previously worked as FSW in Hong Kong; 8.5%, 19.9%, and 71.6% respectively of them had 5 or less, 6 to 20, and more than 20 clients per week on average in the last month. Those who worked as FSW for at least 1 year were more likely than others to have previously worked in Hong Kong (OR = 3.04, 95% CI: 1.69 to 56.46, p < .001).

Measures

Participants were asked about their age, highest educational attainment, duration of sex work, whether they had previously worked as a FSW in Hong Kong, average number of clients per week in the last month, and whether they had ever had an induced abortion.

Participants were also asked whether they agreed or disagreed with the following statements: (1) a healthy-looking person could transmit HIV to others; (2) once infected with HIV, one could detect the HIV status by an HIV antibody testing within a month after the transmission; (3) one could tell if his/her sex partner had been infected with a STD by mere observation; (4) once infected with an STD, one could not contract the same type of STD in the future, and (5) one could not concurrently contract two types of STD. All but the first statement is considered incorrect. Indicator questions such as these have been used in other HIV-related studies and were often significantly associated with condom use (e.g., Lau et al., 2003; Lau, Feng, Lin, Wang, & Tsui, 2005). In addition, participants were asked to indicate whether they perceived condom use to be efficacious for HIV prevention and STD prevention, as well as to rate the perceived susceptibility for them to contract HIV in the future (“How much chance do you think that you would contract HIV in the future?” and response options included “no chance,” “a little chance,” “quite high chance,” and “very high chance”).

Participants were also asked (1) whether they could refuse to have sex with clients demanding unprotected sex; (2) whether they would have unprotected sex with clients if they were paid more by the clients; and (3) whether they would have unprotected sex with clients whom they were fond of. These questions were asked twice, firstly with reference to commercial sex work in Hong Kong and secondly with reference to commercial sex work in mainland China.

Further, participants were asked, with respect to vaginal or anal sex, whether they used a condom during the last episode of sex with their client in Hong Kong and the frequency of condom use with clients in Hong Kong in the last month (using a 4-point response scale ranging from never to every time), whether they had previously worked as FSW in mainland China in the last year and, if so, the frequency of condom use with clients in mainland China, whether they had been tested for HIV antibody in the last 6 months, and whether they had contracted a STD in the last 6 months (and, if so, whether they had consulted a physician for the disease).

Procedure

An on-site convenience sampling method was used. Prospective participants were approached while they were expecting business in the brothels. They were briefed that the study was organized by the Chinese University of Hong Kong, that the topic was on HIV prevention in Hong Kong and mainland China, and that it had important policy implications for the HIV prevention work in Hong Kong. They were assured that the participation was voluntary and that data to be obtained would be strictly confidential and anonymous. Verbal consent was obtained before the face-to-face interviews were conducted in a private setting. The interviews were conducted in Mandarin, the national language in China and took about 10 min to complete. It was confirmed that participants had not previously been interviewed. The two interviewers signed a form pledging that they had clearly explained the details of the study to the participants. Ethics approval was obtained from the Survey Research Ethics Committee of the Chinese University of Hong Kong.

Using plastic bottles, unlinked anonymous urine samples were collected from the participants. The samples were kept in temperature below 4°C and were transported to the Department of Health of the Hong Kong government every other day. The GACPAT and Western Blot method (Scientific Committee on AIDS, 2005) was used for HIV antibody testing in an official government laboratory. This laboratory is responsible for relevant testing for the government and operates at an international standard. Participants were not informed about the testing results. Instead, those who requested their testing results were offered a free voluntary counseling and testing service afterwards. Participants were given an incentive of about HK $60 (equivalent to about US $8) and they charged about HK $400 (US $51) for one episode of sex work.

Results

HIV prevalence and HIV-related risk behaviors

None of the participants tested HIV positive. In the last episode of sex work in Hong Kong, 98.5% of the participants used a condom (Table 1). The prevalence of consistent condom use (used every time) with clients in Hong Kong in the last month was 78.1%. Of all participants, 14.9% had been tested for HIV antibody in the last 6 months; the majority of the testees (94%) were, however, tested in mainland China. Further, 29.9% self-reported having contracted a STD in the last 6 months (19% of them did not consult a doctor) and 53.4% had ever had an induced abortion.
Table 1

HIV/STD-related risk behaviors by background characteristics

 

Age group

Duration worked as FSW

All

≤20

≥21

 

<1 year

≥1 year

 

%

%

pa

%

%

pa

%

Used condoms with clients in Hong Kong in the last episode of sex

    No

3.5

0.0

<.01

0.9

1.9

ns

1.5

    Yes

96.5

100.0

 

99.1

98.1

 

98.5

Consistent condom use with clients in Hong Kong in the last month

    No

28.7

16.8

<.01

15.0

25.9

.02

21.9

    Yes (used every time)

71.3

83.2

 

85.0

74.1

 

78.1

Consistent condom use with clients in mainland China in the last yearb

    No

77.1

65.7

ns

70.0

72.0

ns

71.4

    Yes (used every time)

22.9

34.3

 

30.0

28.0

 

28.6

HIV antibody testing in the last 6 months

    No

88.9

82.2

ns

85.1

85.6

ns

85.1

    Yes

11.1

17.8

 

14.9

14.4

 

14.9

Self-reported STD in the last 6 months

    No

65.3

73.7

ns

72.6

68.5

ns

70.1

    Yes

34.7

26.3

 

27.4

31.5

 

29.9

Consulted doctors for STD infectionc

    No

20.4

17.0

ns

20.7

18.2

ns

18.8

    Yes

79.6

83.0

 

79.3

81.8

 

81.3

Ever had an induced abortion

    No

59.7

36.6

<.001

55.3

41.5

.02

46.6

    Yes

40.3

63.4

 

44.7

58.5

 

53.4

Note: Education level was not associated with any of the above-mentioned variables listed in Table 1. Whether previously worked as FSW in Hong Kong was only associated with whether ever had an induced abortion

ns = non-significant

χ2 test

Among those who reported having worked as FSW in mainland China in the last year

Among those who self-reported having contracted STD in the last 6 months

HIV/STD-related knowledge and perceptions

The distributions of the responses on HIV/STD-related knowledge are summarized in Table 2. Except for two items, HIV/STD-related knowledge was not significantly associated with age, education level, duration of sex work, and previous sex work experience in Hong Kong (Table 2).
Table 2

HIV/STD-related knowledge and perception by background characteristics

 

Age group

Duration worked as FSW

All

≤20

≥21

 

<1 year

≥1 year

 

%

%

pa

%

%

pa

%

Knowledge (appropriate responses)

    % A healthy-looking person could transmit HIV to others (agree)

58.3

71.4

.01

71.9

62.7

ns

65.8

    % One could detect the infection of HIV within one month after the transmission (disagree)

28.5

30.2

ns

30.7

28.1

ns

29.5

    % One could know whether his/her sex partner has infected with STD by observation (disagree)

41.0

46.1

ns

43.0

44.0

ns

43.9

    % One could be immunized from further infection of the same type of STD once contracted an STD (disagree)

64.3

64.1

ns

67.5

62.0

ns

64.2

    % One could not contract 2 types of STD concurrently (disagree)

62.5

57.3

ns

55.3

62.2

ns

59.5

Perception (appropriate responses)

    % Condom use is efficacious for HIV prevention (agree)

58.3

76.6

<.001

64.9

71.0

ns

68.8

    % Condom use is efficacious for STD prevention (agree)

69.4

82.3

<.01

79.8

74.7

ns

76.8

    % Perceived no chance of contracting HIV in the future

62.7

58.6

ns

62.5

58.3

ns

60.4

Perceived control over condom use

    % In Hong Kong, you could refuse sex with clients demanding not using condoms

72.2

77.0

ns

76.3

74.1

ns

74.9

    % In Hong Kong, you would have sex with clients without using condoms if the client paid more

22.2

12.6

.02

11.4

19.9

.05

16.7

    % In Hong Kong, you would have sex with clients without using condoms if you were fond of him

14.0

12.6

ns

8.8

15.3

ns

13.2

    % In mainland China, you could refuse sex with clients demanding not using condoms

71.7

64.4

ns

65.4

68.6

ns

67.6

    % In mainland China, you would have sex with clients without using condoms if the client paid more

29.0

16.1

.01

14.4

25.7

.02

21.7

    % In mainland China, you would have sex with clients without using condoms if you were fond of him

28.3

25.7

ns

19.4

30.5

.04

26.8

Note: Education level was not associated with any of the items listed in Table 2. Whether previously worked as FSW in Hong Kong was only associated with one item (whether one could detect the infection of HIV within one month after the transmission)

ns = non-significant

χ2 test

Of the participants, 68.8% and 76.8%, respectively, believed that condom use was efficacious for HIV prevention and STD prevention; 60.4% believed that there was no chance at all for them to contract HIV in the future. Younger participants were much less likely to perceive condom use to be efficacious for HIV or STD prevention (OR = 0.43 and 0.49, respectively, both p< .01).

Perceived behavioral control over condom use

Of all respondents, 74.9% reported that they could refuse to have sex with a client demanding unprotected sex in Hong Kong versus 67.6% in mainland China (p < .05, McNemar test; Table 2). Similarly, 16.7% versus 21.7%, respectively (p < .05, McNemar test), would have unprotected sex if the client paid more in Hong Kong and in mainland China, and 13.2% versus 26.8%, respectively, would have unprotected sex with clients whom they were fond of in the two respective locations (p < .001, McNemar test).

Those who were aged 20 or less and those who had been working in the sex industry for at least one year were more likely than others to state that they would have unprotected sex with clients who paid more, either in Hong Kong (OR = 1.99, p < .05 and OR = 1.93, p = .05) or in mainland China (OR = 2.13, p < .01 and OR = 2.05, p < .05). Moreover, those who had been working as FSW for at least one year were also much more likely than others to state that they would have unprotected sex with clients whom they were fond of in mainland China (OR = 1.82, p < .05).

Factors predicting condom use when having sex with clients in Hong Kong

Three univariate variables were selected by the multivariate model predicting unprotected sex in the last episode of sex work (Table 3). These were perceived efficacy of condom use for HIV prevention, whether they could refuse to have sex with a client demanding unprotected sex while in Hong Kong, and whether they would agree to have unprotected sex with clients whom they were fond of while in Hong Kong.
Table 3

Condom use with clients in Hong Kong (among all participants)

 

Did not use condoms with clients in the last episode of commercial sex

Inconsistent use of condoms with clients in the last month

Row %

ORuniv

ORmult

Row %

ORuniv

ORmult

Age group

    ≤20

28.7

1.00

1.00

    ≥21

   

16.8

0.50*

0.53*

Duration worked as FSW

    <1 year

15.0

1.00

ns

    ≥1 year

   

25.9

1.98*

 

Average number of clients per week in Hong Kong in the last month

    ≤20

14.0

1.00

1.00

    >20

   

25.4

2.10*

2.18*

Condom use is efficacious for HIV prevention

    Disagree/not certain

3.8

1.00

1.00

    Agree

0.4

0.11*

0.10*

   

A healthy-looking person could transmit HIV to others

    Disagree/not certain

28.9

1.00

ns

    Agree

   

18.2

0.55*

 

In Hong Kong, you could refuse sex with clients demanding not using condoms

    No/not certain

4.8

1.00

1.00

    Yes

0.4

0.08*

0.10*

   

In Hong Kong, you would have sex with clients without using condoms if the client paid more

    No/not certain

0.7

1.00

ns

15.5

1.00

1.00

    Yes

5.4

7.81*

 

54.5

6.56***

2.89**

In Hong Kong, you would have sex with clients without using condoms if you were fond of him

    No/not certain

0.3

1.00

1.00

15.7

1.00

1.00

    Yes

9.1

28.70**

23.08**

63.6

9.41***

8.99***

HIV antibody testing in the last 6 months

    No

24.0

1.00

ns

    Yes

   

10.0

0.35*

 

Self-reported STD in the last 6 months

    No

0.4

1.00

ns

15.5

1.00

1.00

    Yes

4.0

9.67*

 

36.0

3.06***

2.85**

Note: Separate univariate logistic regression analyses were conducted to identify factors that were associated with condom use with clients in Hong Kong. Variables that were univariately significant were entered as inputs to a multivariate stepwise logistic regression model. Variables considered included the background characteristics, HIV-related knowledge and perception items, and the three perceived control items related to Hong Kong listed in Table 2, HIV antibody testing behavior, self-reported STD, and ever had an induced abortion

ns = non-significant

ORuni = univariate odds ratio

ORmult = odds ratios obtained from stepwise multivariate logistic regression analysis using univariately significant variables as candidate variables

*p < .05; **p < .01; ***p < .001

Similarly, five univariate variables were retained in the multivariate model predicting inconsistent condom use with clients in the last month: age, having more than 20 clients per week in the last month, whether they would have unprotected sex with clients who were paying more while in Hong Kong, whether they would have unprotected sex with client whom they were fond of while in Hong Kong, and self-reported STD in the last 6 months (Table 3).

Sex work and condom use with clients in mainland china

Of all participants, 42% reported that they had previously been engaged in sex work in mainland China in the last year. The multivariate model showed that the two groups differed significantly in age, duration of sex work, inclination toward not using condoms with clients whom they were fond of while in mainland China, and having ever had an induced abortion (Table 4).
Table 4

Factors predicting sex work in mainland China and condoms use with clients in mainland China (in the last year)

 

Worked as FSW in mainland China

Inconsistent condom use with clients in mainland China

Row %

ORuniv

ORmult

Row %

ORuniv

ORmult

Age group

    ≤20

48.6

1.00

1.00

    ≥21

37.2

0.63*

0.45**

   

Duration worked as FSW

    <1 year

18.4

1.00

1.00

    ≥1 year

54.4

5.28***

4.51***

   

Average number of clients per week in Hong Kong in the last month

    ≤20

28.3

1.00

ns

    >20

46.6

2.21**

    

A healthy-looking person could transmit HIV to others

    Disagree/not certain

51.3

1.00

ns

    Agree

37.3

0.56*

    

In mainland China, you would have sex with clients without using condoms if the client paid more

    No/not certain

40.6

1.00

ns

61.0

1.00

1.00

    Yes

56.5

1.91*

 

97.4

24.30**

11.16*

In mainland China, you would have sex with clients without using condoms if you were fond of him

    No/not certain

37.5

1.00

1.00

55.8

1.00

1.00

    Yes

62.4

2.76***

2.36**

96.2

20.19***

11.89**

Ever had an induced abortion

    No

31.4

1.00

1.00

    Yes

51.4

2.31***

2.42**

   

Self-reported STD in the last 6 months

    No

64.9

1.00

ns

    Yes

   

84.8

3.01*

 

Consistent condom use with clients in the last month in Hong Kong

    No

57.5

1.00

ns

    Yes

38.1

0.45**

    

Note: Separate univariate logistic regression analyses were conducted to identify factors that were associated with previous sex work and condom use with clients in mainland China. Variables that were univariately significant were entered as inputs to a multivariate stepwise logistic regression model. Variables considered included the background characteristics, HIV-related knowledge and perception items, and the three perceived control items related to China listed in Table 1, HIV antibody testing behavior, self-reported STD, and ever had an induced abortion

ns = non-significant

ORuni = univariate odds ratio

ORmult = odds ratios obtained from stepwise multivariate logistic regression analysis using univariately significant variables as candidate variables

*p < .05; **p < .01; ***p < .001

Of the participants having previously engaged in sex work in mainland China, only 28.6% reported consistent condom use in the past year when having sex with their clients in mainland China (Table 1). The multivariate analyses showed that those who would have unprotected sex with clients who paid more for the sex work or with clients whom they were fond of while in mainland China were more likely than others to be inconsistent condom users in mainland China (Table 4).

Variations in frequency of condom use by locale

Of the 141 participants who reported having engaged in sex work both in Hong Kong and in mainland China, 100 (71%) reported inconsistent condom use with clients in the last year in mainland China. Models were fitted to distinguish, among these 100 inconsistent condom users in mainland China, those who used (n = 60, 60%) or did not use (n = 40, 40%) condoms consistently when serving as FSW in Hong Kong in the last month. The results of the univariate analyses showed that among this group of 100 FSW who were inconsistent condom users at sex work in mainland China, those who were amenable to having unprotected sex with clients who paid more in either locale (Hong Kong or mainland China) and those who admitted that they would have unprotected sex with clients whom they were fond of in either locale were statistically significantly more likely than others to be using condoms inconsistently with clients in Hong Kong (Table 5). The multivariate model showed that only the perception that one would have unprotected sex with clients whom they were fond of while in Hong Kong remained statistically significant (Table 5).
Table 5

Factors predicting geographic differential in condom use with clients among those with sex work both in Hong Kong and in mainland China

 

Among inconsistent condom users in mainland China, n = 100

Among consistent condom users in Hong Kong, n = 98

Condom use with clients in Hong Kong

  

Condom use with clients in mainland China

  

Consistent use (n = 60)

Inconsistent use (n = 40)

  

Consistent use (n = 38)

Inconsistent use (n = 60)

  

Row %

Row %

ORuniv

ORmult

Row %

Row %

ORuniv

ORmult

In Hong Kong, you would have sex with clients without using condoms if the client paid more

    No/not certain

70.3

29.7

1.00

 

    Yes

30.8

69.2

5.32**

ns

    

In Hong Kong, you would have sex with clients without using condoms if you were fond of the client

    No/not certain

74.0

26.0

1.00

1.00

    Yes

13.0

87.0

19.00***

18.67***

    

In mainland China, you would have sex with clients without using condoms if the client paid more

    No/not certain

70.5

29.5

1.00

ns

46.3

53.8

1.00

ns

    Yes

42.1

57.9

3.29**

 

5.9

94.1

13.77*

 

In mainland China, you would have sex with clients without using condoms if you were fond of the client

    No/not certain

79.2

20.8

1.00

ns

49.3

50.7

1.00

1.00

    Yes

41.2

58.8

5.43***

 

4.5

95.5

20.45**

14.02*

Note: Separate univariate logistic regression analyses were conducted to identify factors that were associated with condom use with clients in Hong Kong and in mainland China. Variables that were univariately significant were entered as inputs to a multivariate stepwise logistic regression model. Variables considered included the background characteristics, HIV-related knowledge and perception items, and the perceived control over condom use items listed in Table 1, HIV antibody testing behavior, self-reported STD, and ever had an induced abortion

ns = non-significant

ORuni = univariate odds ratio

ORmult = odds ratios obtained from stepwise multivariate logistic regression analysis using univariately significant variables as candidate variables

*p < .05; **p < .01; ***p < .001

Similar analyses were conducted to identify factors distinguishing, among FSW who were consistent condom users at sex work while in Hong Kong (n = 98), those who were consistent (n = 38, 38.8%) or inconsistent (n = 60, 61.2%) condom users at sex work in mainland China. The results of univariate analysis showed that consistent condom users in Hong Kong who were willing to have unprotected sex with clients who paid more or whom they were fond of while in mainland China were much more likely than other consistent condom users in Hong Kong to also be inconsistent condom users in mainland China (Table 5). The multivariate model showed that only the perception item related to willingness to have unprotected sex with clients whom they were fond of while in mainland China remained statistically significant (Table 5).

Discussion

 No HIV positive cases were detected. Surveillance studies documented low prevalence of HIV of 0.4% among FSW in Shenzhen, Guangdong only (Xu, 2005). The moderate sample size of this study may, therefore, be unable to detect the low prevalence. The sampled participants, however, were at considerable HIV-related risk as indicated by the high prevalence of self-reported having contracted a STD in the last 6 months and induced abortion. Prevention and surveillance are most needed for places where HIV prevalence is still low but risk behaviors are prevalent (WHO/UNAIDS, 2000). A service gap thereby exists as only a minority of the sampled FSW had had an HIV antibody test in the last 6 months.

Similar to other relevant studies (e.g., Lau et al., 2002), our participants had low levels of educational attainment, HIV-related knowledge, and perceived susceptibility. In this study, a higher percentage of participants perceived they could not refuse to have sex with clients demanding unprotected sex while in mainland China than in Hong Kong. This differential may be related to the setting of the sex-work establishment. This variable on perceived behavioral control was predictive of condom use in the last episode of commercial sex in Hong Kong. Moreover, about 20% reported that they would have unprotected sex with clients who paid more for the service while in Hong Kong and while in mainland China.

While only 13.2% of the participants reported that they would engage in unprotected sex if they were fond of the client while working in Hong Kong, 26.8% would do so in mainland China. Reasons for the difference were not obtained from this study. It is, however, speculated that even they were fond of a client in Hong Kong; the chance of developing a long-term relationship was very low. The chance for developing a similar relationship might however be higher in mainland China and such may affect condom use behaviors. Furthermore, those FSW working in brothels in Hong Kong would not go out with their Hong Kong clients whereas it is common for FSW to go to dinner, karaoke, and other places with their clients in mainland China. Being transient in Hong Kong, there is hence a much lower chance for developing intimacy with their clients in Hong Kong, as compared to the case in mainland China.

Based on data obtained from the same cohort, a differential in the frequency of condom use by locale was apparent. The aforementioned contextual differences (such as “fond of clients” and “willingness to have unprotected sex with clients who are paying more”) may partially account for the observed geographic differential with regard to using condoms more frequently with clients in Hong Kong than in mainland China. Such conjecture is also supported by the results of the logistic regression analyses (Table 5). In the literature, relationship intimacy (e.g., feeling of affection and trust) with clients has, in fact, been associated with condom use behaviors (Kerrigan et al., 2003). HIV workers need to be aware of these issues when designing their intervention programs.

Behavioral change is not impossible and may occur within a short period of time, especially when the social context changes. This also corroborates the results obtained from local studies conducted on clients of FSW (Lau et al., 2003; Lau, Tsui, & Ho, 2005), showing similar differentials by locales. As migrant sex workers and commercial sex activities among travelers have been common (Hawkes et al., 1994; Lau & Tsui, 2003; Matteelli et al., 2003; van Haastrecht et al., 1993), the observed geographic differentials in condom use among FSW and/or their clients may also occur in other places. To our knowledge, this is the first study of this type.

 Interestingly, another study showed that male FSW clients’ perception on the difference in the degree of insistence on using condoms among FSW in Hong Kong and FSW in mainland China was also a significant predictor of the observed geographic differential in condom use behaviors among clients of FSW (Lau et al., 2005). More studies on the role of perceived behavioral control over condom use under different circumstances are warranted.

There are other factors that have not been included in the study but may be responsible to explain the observed geographic differentials. First, the client-worker power structures of the two locales may be different. Female sex workers coming to work in Hong Kong may be somehow empowered by their interaction with some local sex workers or by some local non-governmental organizations (NGO) that are serving these FSW. In fact, there are a few NGO conducting outreach education for our study population. Men in mainland China may also be more dominant than men in Hong Kong, where education for gender equality has better roots. Second, different norms for condom use may exist. To our knowledge, FSW were told by the establishments’ keepers in Hong Kong that they could refuse a client if he demanded unprotected sex. The influences of keepers were shown to affect condom use behaviors among FSW (Morisky et al., 1998; Yang et al., 2005a). In Hong Kong, but not in mainland China, brothels are always “guarded” by male keepers. FSW working in Hong Kong, therefore, may feel more protected. Health education and HIV prevention campaigns have also been more available in Hong Kong than in mainland China. Institutional differences may also be related to the level of condom use. An income effect may also exist, as the service fee is higher in Hong Kong and, on average, the majority of the studied FSW had more than 20 clients a week in Hong Kong. The higher income in Hong Kong allows FSW to be more able to turn down Hong Kong clients demanding unprotected sex. As an exploratory study, these variables were not investigated and future studies are warranted.

The study had several limitations. First, only one particular group of mobile FSW was studied and generalization to other FSW groups should be made with caution. Further studies are required to see if a general trend exists among mobile FSW in different countries. Second, data were self-reported and a reporting bias may exist. However, for that bias to affect the conclusions, responses involving inconsistent condom use needs to be more socially desirable in Hong Kong than in mainland China. There does not seem to be an obvious reason for such a differential in social desirability to exist; the data were also obtained anonymously. Self-reported STD data may also have inflated the estimated prevalence of STD. However, as health services seeking behaviors have not been ideal in this group, clinical data would under-estimate the prevalence by even a larger extent.

The duration for determining the status of inconsistent condom users in Hong Kong and in mainland China differed. The time frame was “the last month” in Hong Kong but “the last year” in mainland China. There were reasons behind this discrepancy. The studied FSW usually stayed in Hong Kong for a relatively short period of time (say 1–2 months). Therefore, the time frame for condom use in Hong Kong had to be quite short (in this study, last month). On the other hand, some of the studied FSW might have worked in Hong Kong for several months and a short time frame (say the last month) for condom use in mainland China would make the question non-applicable to these FSW (therefore, one year was used as the time frame). The response categories of the two situations were, however, the same (never, seldom, occasionally, and always). The contrast in the prevalence of consistent condom use in Hong Kong versus mainland China (78.1% versus 28.6%) was, in fact, very large. Further, 98.5% of the participants had used a condom in the last episode of commercial sex encounter in Hong Kong. It is, therefore, reasonable to claim that geographic differentials for condom use behaviors do exist.

On-site random sampling was also not feasible and convenience samples were obtained. Although some selection bias may, therefore, exist, the analyses using data of the same cohort should be less influenced by the selection bias. The ordering of the questions on condom use in Hong Kong and in mainland China was the same for all participants. Order effects may, therefore, be present though we believe that such effects should be minimal. Measurements of the work environment were also not given. It is impossible to rate the work environment in Chinese accurately as multiple and diversified establishments may have been included. Observation by researchers is not feasible and reporting by FSW may not be reliable.

In summary, geographic differentials in frequency of condom use with clients have been detected for this group of mobile FSW. Perceived control of condom use, especially the willingness to have unprotected sex with clients whom were liked by the FSW, significantly predicted frequency of condom use in Hong Kong and in mainland China, as well as the afore-mentioned geographic differentials. Further studies are warranted to confirm the generalizability of the observed differences. Behavioral changes were detected and HIV intervention and research work need to take sociocultural contexts into account, instead of using an over-simplified medical/health approach.

Acknowledgements

The authors would like to thank all participants of the study and Dr. W. L. Lim and Dr. K. M. Ho of the Department of Health, Hong Kong Special Administrative Region for their help and advice on the study. Thanks are extended to Dr. Jean H. Kim of the Chinese University of Hong Kong for proofreading the manuscript. The study was supported by the Hong Kong Council for the AIDS Trust Fund. An earlier version of this article was presented at the National University of Singapore Centennial Conference August 1–3, 2005, Singapore and the presentation was sponsored by the National University of Singapore.

Copyright information

© Springer Science+Business Media, LLC 2007