Multi-dimensional factors associated with unprotected anal intercourse with regular partners among Chinese men who have sex with men in Hong Kong: a respondent-driven sampling survey

The human immunodeficiency virus (HIV) epidemic among men who have sex with men (MSM) in China is a public health threat as the HIV prevalence in this population has been high and keeps increasing [1‐3]. According to a national survey in 2011, the average HIV prevalence among MSM in China was 4.9% [4]. In Hong Kong, the HIV prevalence among MSM was 4.08% in 2011 [5]. According to social marketing principles, the target population should not be seen as a homogeneous group and segmentation is required to improve effectiveness of interventions [6]. In this context, previous research has shown that the prevalence of unprotected anal intercourse (UAI) with male regular sex partners (RP) is significantly higher than that with male non-regular sex partners (NRP) among MSM [7]. UAI with RP among MSM contributes to the majority of the new HIV infections in the United States and in Lebanon [8, 9]. Among the 61.4% of the MSM in China having RP, 78.6% also have had sexual intercourse with other male sex partners [10, 11]. The proportion of new HIV infections attributable to sexual intercourse with RP, NRP and commercial sex partners (CSP) among MSM was 40%, 37% and 23% respectively in 2010 in China [10]. The proportion of new infections attributable to RP among MSM was 34% in 2002, and with an increase of 6%, infections attributable to RP among MSM started to prevail in the past decade in China [10]. It seems that HIV prevention in China are mainly venue-based (gay-saunas and bars) and hence focuses mainly on promoting safer sex with NRP [10]. Furthermore, a cohort study targeting MSM has shown that UAI with RP but not UAI with NRP predicts HIV sero-conversion [12]. Hence, the risk of HIV transmission via male RP among MSM is important but may have been overlooked.

Furthermore, our literature search did not locate any study reporting HIV intervention specifically promoting condom use with RP among MSM. To design such programs effectively, we need to identify multi-dimensional risk factors associated with UAI with male RP among MSM. Previous studies have shown that intimacy between heterosexual sex partners is associated with inconsistent condom use [13, 14]. It is also seen that worry about breaching the trust is one of the obstacles hindering condom use between heterosexual regular partners [15]. A few studies have tested similar hypotheses among MSM indicating that trust, emotional connection with the partner, and intimacy were significant predictors of UAI [16‐18]. Cognitive factors are known to be significantly associated with UAI among MSM [19] and HIV/ sexual transmitted infection (STI) risk perception is one of such factors [18‐20]. Very few, if any, studies investigated specific assessment of the risk level of one’s RP among MSM [21]. This is important as monogamy does not seem to be the norm among MSM [22, 23]. Assessment of the perceived risk of contracting HIV via the RP is equally important, as that may cause severe damage to the relationship with the RP, sense of anger, guilt and regret [24]. Previous studies targeting MSM that did not focus specifically on UAI with RP have reported significant contextual factors [19, 25], one of which was the use of alcohol and substances prior to having anal sex with men [26‐28]. Relatively few studies studied the importance of this factor with respect to UAI with RP among MSM [27, 28].

Personality of the MSM is another important but often under-investigated risk factor [29]. Impulsivity usually displays behaviors characterized by little or no forethought, reflection, or consideration of the consequences; it has found to be significantly associated with various types of risk factors such as alcohol abuse [30]. Impulsivity is also likely to be related to HIV risk--particularly in incarcerated substance-abusing youth [31]. A study revealed that the relationship between intensity of methamphetamine use and total unprotected sex was strongest among participants who had higher levels of impulsivity among HIV-positive MSM [32]. According to our knowledge, no study has investigated the association between impulsivity and UAI with RP among MSM.

Respondent driven sampling (RDS) is a relatively new research method designed for collecting data from hard-to-reach populations [33]. It has been used in studies targeting MSM in China [34] and in other countries [35, 36]. It requests participants to refer a fixed number of eligible prospective participants (usually three to five) to join a survey, while similar waves are repeatedly conducted until the required sample size is achieved. Incentives are usually involved when referrals are being made. Theoretically, an equilibrium state would be attained after several waves of recruitment, and unbiased statistical estimates approaching those obtained from random probability sampling would be obtained [33, 35]. The RDS design is hence an improvement over the venue-based sampling method commonly used in HIV research targeting MSM [36, 37].

This study investigated prevalence of UAI specifically with RP among Chinese MSM in Hong Kong, China. Multi-dimensional factors were investigated at different levels, including i) the individual’s background factors (socio-demographic variables such as age, educational level, marital status, and sexual orientation, etc.); ii) contextual factors (use of alcohol and substances prior to having anal intercourse); iii) interpersonal factors (intimacy with the RP, worry that condom use symbolizes mistrust, perceived degree of the RP’s acceptance of condom use, perceived HIV/STI risk level of the RP; and iv) personality factor (impulsivity). It is hypothesized that the aforementioned factors would be associated independently with UAI with RP among MSM.

The sampled MSM with RP were at high risk of HIV/STD transmission. A previous study showed that UAI with RP but not that with NRP predicted HIV seroconversion [12]. In this study, prevalence of UAI with RP was high (45.8%) and was comparable to that reported in a study conducted in Beijing in 2011 (49.7%) [45], while it was lower than that obtained from studies conducted in Argentina (58.0%) and Kazakhstan (69.0%) [35, 46]. Interventions focusing on UAI with RP among MSM are hence greatly warranted. Attention should also be given to sex with NRP among those with RP, as about 30% of the participants had had sex with NRP in the last six months and UAI was only involved. To socio-ecological model reminds us that we need to look at factors associated with UAI with RP among MSM at inter-personal, individual and structural levels [47].

Interpersonal factors need to be considered. Unlike results obtained from studies targeting heterosexual couples [13, 14], quality of relationship was not associated with UAI with RP in this study. The majority of the participants reported a good relationship with their RP. Other factors on relationship between the MSM and his RP (e.g. trust) were however, important. Many participants worried that condom use with the RP would be interpreted as mistrust. In this study and previous studies conducted in homosexuals and heterosexual populations [13, 14, 16‐18], the concern over mistrust was significantly associated with condom use during sexual intercourse. The data suggested that mistrust was common as one third of the participants believed that their RP was at high risk of HIV infection in the future or would pass HIV infection to them. The beliefs may be a reflection of the high prevalence of participants having both RP and NRP. In the context of mistrust but a good relationship, discussion about condom use might be avoided by the couple to minimize tension. A vicious cycle of perceived high risk of HIV transmission, mistrust, avoidance, and then perceived high risk might be created, which needs to be broken with improvement in communication skills through interventions. Furthermore, the absence of legal marriage to the same sex might further hinder MSM to form their families and remain monogamous. Structural factors therefore also need to be considered to improve the situation. Couple-based HIV interventions are warranted and have shown to be effective [48‐50]. Difficulties include worry about separation, sexual abuse and sorry that such information would be a shock to the partner or would harm the partner [51].

Contextual individual-level factors such as psychoactive substance use prior to anal intercourse should be considered. In this study, the prevalence was about 15%. Similar to the results of other studies [25, 28], the variable was significantly associated with UAI. In China, there is no substance use prevention program targeting MSM. Such programs are different from those targeting the general population as sexual orientation and risks need to be considered. Prevention and risk reduction programs are warranted to fill up the service gap. In this study, alcohol use was not significantly associated with UAI among MSM. The result was similar to that of a recent conducted among Hong Kong MSM [7]. However, mixed findings of the associations between alcohol/drug use and UAI among MSM [7, 46, 51] have been reported.

Impulsivity is another important individual-level factor. Impulsivity is defined as acting without first considering the possible consequences is another important individual-level factor [32]. In this study and others, impulsivity refers to relatively stable personality and its measure was not specific to sex-related behaviors. It is potentially related to sexual adventurism, which is usually defined as sexual risk taking and sexual sensational seeking [52‐54]. UAI is therefore related to sexual adventurism and sensational seeking [52, 54]. Although previous studies have reported a significant association between higher impulsivity and unprotected sex among adult young women, high school students [51, 55], drug users [56] and substance-abusing youth [31], this is the first study finding out that impulsivity was significantly associated with UAI during sex with RP among MSM. A study has shown greater intensity of methamphetamine use and higher levels of impulsivity predicted more unprotected sex among HIV-positive MSM, which suggests that targeting impulsivity in interventions may help reduce sexual risk behaviors [32]. Some interventions have been successful to instill stronger self-control on alcohol abuse and drug addiction [57, 58] and such components may need to be added to HIV interventions targeting this population of MSM.

The study has several limitations. The sample size was modest; some non-significant associations may be subjected to inadequate power. We used community based RDS method to increase representativeness. However, we were still unable to control for non-response bias. Non-random recruitment of peers by RDS method could influence the estimates in unknown ways [33]. The development of statistical analysis methods for RDS data was not readily available and risk factors analysis in the study was unweighted because the RDSAT does not entertain multivariate analyses. Data were self-reported and may involve reporting bias due to social desirability, while the survey was anonymous. Furthermore, the study did not involve HIV antibody testing and it is a limitation that we did not ask about HIV sero-status of the participants and their male regular sex partners. One of the reasons was that such data might not be reliable due to strong social desirability and may hence strongly be subjected to reporting bias. The HIV prevalence was about 4% in the study population, HIV testing rate (last six months) was rather low (about 33.5%) and we do not have data about disclosure of HIV status to male regular sex partners. We do not know about the prevalence of sero-discordant couples among our participants but we contend that the number of such cases would be small. Some of the scales were constructed for this study and had not been validated as there are few studies specifically focused on UAI with RP among MSM.

In sum, the participants were at high risk of HIV/STI transmission as they have relatively high prevalence of having NRP and low prevalence of condom use with both RP and NRP. Participants in general, perceived that their RP might be at high risk of HIV/STI transmission. Issues arising from the relationship between such MSM and their RP, including feeling of mistrust associated with condom use, need to be addressed. Unlike the case of heterosexual risk factors, intimacy between the couple was however, not a significant factor. The issue on impulsivity has seldom been discussed in HIV prevention targeting MSM. It was a significant factor and further studies on interventions to increase self-control are warranted. Overall, HIV interventions aiming at reduction of UAI with RP among MSM need to consider multi-dimensional and socio-ecological factors at structural, interpersonal and individual levels. Lastly, the study used RDS method for data collection and hence to reduce sampling bias.