Global Public Health, 2015 Vol. 10, No. 1, 103–118, http://dx.doi.org/10.1080/17441692.2014.972968
Sexual partner mixing and differentials in consistent condom use among men who have sex with men in Maharashtra, India Sucheta Deshpande* and Shalini Bharat Tata Institute of Social Sciences (TISS), Mumbai, India (Received 11 November 2013; accepted 30 August 2014) Sexual partner mixing among men who have sex with men (MSM), based on both gender and partnership status, is an understudied theme in India. Using data from Round 2 of the Integrated Bio-behavioral Survey, this paper reports on gender and partnership status-based sexual mixing and levels of consistent condom use (CCU) among MSM in Maharashtra. A total of 689 MSM were sampled using probabilitybased sampling. Bivariate and regression analyses were carried out on condom use and partnership mixing. Over half (52%) of all MSM reported having only male partners while about one-third (34.5%) reported having partners of both gender. Over 70% of MSM engaged in sex with a mix of casual, regular, commercial and noncommercial partners. MSM with only male partners reported lower CCU as compared to MSM with partners of both genders (47.3% and 62%, respectively, p = 0.11). CCU levels differed significantly by status of sex partner. Overall, MSM having ‘men only’ as partners and those with partners of mixed status have greater risk behaviour in terms of low CCU. HIV prevention interventions need to focus attention on men in ‘exclusively male’ sex partnerships as well as MSM with a mix of casual, regular and commercial partners. Keywords: MSM; sexual mixing; sexual partnerships; condom use; India
Introduction Globally, a concern has been raised for varied rates of HIV infection among men who have sex with men (Baral, Sifakis, Cleghorn, & Beyrer, 2007; Griensven & Wijngaarden, 2010; Griensven, Wijngaarden, Baral, & Grulich, 2009). Comprehensive reviews of HIV burden in MSM worldwide report HIV prevalence among MSM ranging between 3% in the Middle East and North Africa to 25% in the Caribbean (Beyrer, Baral et al., 2012; Beyrer, Sullivan et al., 2012). HIV prevalence among MSM in India is reported between 4.4% and 22% (Brahmam et al., 2008; Hernandez et al., 2006; National AIDS Control Organisation (NACO), 2012; Setia et al., 2008). In the early phase of India’s national AIDS control programme, the focus was on preventing HIV transmission within heterosexual relationships, particularly between female sex workers and their male clients (Dandona, Sisodia, et al., 2005; Halli, Ramesh, O’Neil, Moses, & Blanchard, 2006; Steinbrook, 2007; Thilakavathi et al., 2011). However, there is increasing recognition today of the much greater degree of heterogeneity in HIV transmission dynamics in India (Chandrasekaran et al., 2006). Based on sentinel site data and data from surveys such as the integrated behavioural and *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
104
S. Deshpande and S. Bharat
biological assessment (IBBA) and other studies conducted among MSM (Blanchard et al., 2007; Brahmam et al., 2008; Hernandez et al., 2006), HIV prevalence in this key population is found to be higher than in the general population (4.43% and 0.35%, respectively; NACO, 2012). Globally, two factors are commonly discussed in relation to HIV transmission risk among MSM, namely, multiplicity of sex partners and sexual mixing patterns within ‘same gender’ and between ‘diverse gender’ networks (Doherty, Padian, Marlow, & Aral, 2005; Yuhua et al., 2011). Partner preference plays a vital role in shaping the sexual networks within which sexual mixing takes place. Sexual mixing between ‘unlike’ groups, in which people select partners from other gender groups, and between ‘like’ groups, in which people select partners from their own gender groups, has implications for sexually transmitted infections (STI) (Aral, Adimora, & Fenton, 2008; Doherty, Shiboski, Ellen, Adimora, & Padian, 2006; Hertog, 2007; Yuhua et al., 2011). Compared to Western countries, SouthAsian countries report a multiplicity of sexual identities within the MSM population. In South Asia, sexual identities are culturally derived constructs which cannot be easily captured in terminologies used in the West such as, ‘gay’, ‘transgender’ (TG) or ‘bisexual’ (Bondopadhyay & Haque, 2002; Khan, 1997). For example, men engaged in sex work in the West often identify themselves as ‘heterosexual’; however, such identifications might be different and play an important role in relation to sexual practices in Asian countries (Dowsett, Grierson, & McNally, 2006). Often, MSM classified in Indian studies are a diverse group of men. They are engaged in transactional sex, perform different types of sex roles, i.e. receptive, penetrative, or both, and have male, female and TG partners. Various studies suggest that the sexual networks of MSM in India are usually dense and complex, as they report high levels of both ‘like’ and ‘unlike’ network mixing by having males, TG and females as their sexual partners (Asthana & Oostvogels, 2001; Hernandez et al., 2006; Setia, Sivasubramanian, Anand, Row Kavi, & Jerajani, 2010; Shinde, Setia, Row Kavi, Anand, & Jerajani, 2009). Evidence supporting the dynamics of such complex network mixing among the MSM population in India is growing (Brahmam et al., 2008; Chakrapani, Row Kavi, Ramakrishnan, Rapporpori, & Raghavan, 2002; Khan, 2001). A recent study in Bangalore city reported divergent sexual behaviour of MSM with a high percentage of them reporting a mix of women and men (55%), men (45%) and hijra (3%) as their sexual partners (Phillips et al., 2009). A large proportion of MSM in India are married (Brahmam et al., 2008; Dandona, Sisodia, et al., 2005; Phillips et al., 2010) and many are also involved in sex with commercial and casual partners (Brahmam et al., 2008; Setia et al., 2006). In terms of HIV transmission dynamics, the overlap of MSM networks with the general population through marriage, together with the ‘unlike’ sexual mixing patterns, assumes great importance. Sexual practices, networks and identities of MSM in India need to be understood in the context of a largely heteronormative society where there is not only a high level of stigma against same-sex relations, but also where such behaviour is criminalised under section 377 of the Indian Penal code. In addition, social norms that uphold the institution of marriage and place high value on its procreative function make demands of a very different kind on MSM in India, forcing many to marry for social reasons, engage in bisexual behaviour and bear children, without disclosing their sexual orientation to their spouse. The regressive Section 377 was read down by the Delhi High Court in 2009, but the decision was not upheld in the country’s Supreme Court, sparking fresh fears among MSM about societal stigma and possible legal action against them for expressing their sexuality (Mahapatra, 2014; Skanaland, 2009). Stigma and the continued criminalisation of homosexuality compound challenges of HIV prevention work in the country. The sociocultural–legal context framing homosexuality and
Global Public Health
105
available empirical evidence point to the fallacy of treating MSM in India as a homogenous, monolithic group to be reached with standard prevention approaches (Brahmam et al., 2008; Hernandez et al., 2006). But the heterogeneity of the MSM population has not been adequately explored. For example, intermixing based on partnership status such as regular, casual or commercial and that based on gender (male/female/TG) within MSM sexual networks is understudied in India. This paper reports on sexual partner mixing based on both gender and partnership status and levels of condom use among MSM in Maharashtra state. The data are based on the second round of the IBBA survey carried out during 2009–2010. The IBBA surveys were conducted as part of evaluation of Avahan, the India AIDS Initiative intervention programme implemented in six high HIV prevalence states of India. The IBBA surveys were designed to study STI and HIV prevalence among different high-risk populations, including MSM in these states, and were carried out in two rounds in 2007 and 2009 (Indian Council of Medical Research & FHI 360, 2011).
Methods In Maharashtra state, the second round of IBBA was conducted with MSM in Mumbai– Thane and Pune districts in 2009–2010. The target sample size was 400 MSM per district (Saidel et al., 2008). A two-stage cluster sampling design with time location clusters (TLC) was used in the survey. At first, mapping and listing of all existing MSM sites were completed and the sampling frame for TLC was developed. In the first stage, systematic random samples of primary sampling clusters were selected by probability proportional to size. Next, from the selected clusters survey respondents were selected randomly from eligible respondents available during the selected time duration. If the desired sample size was not found in a district, a decision was made to cover the entire population available in that district; this was called a ‘take-all’ approach. In Mumbai– Thane, considered one domain, a probability-based TLC sampling (Chandrasekaran et al., 2008; Saidel et al., 2008) was carried out. In Pune district, since the size of the MSM population was found to be less than 400 in the identified cruising sites, the ‘take-all’ approach was adopted. The recruitment criterion for MSM was, ‘any male, identified at cruising points, 18 years or older, who had any type of sex (oral, manual or penetrative) with another male, paid or unpaid, in the last one month’. The survey was approved by the Ethics committee of the National AIDS Research Institute (NARI), and the screening committee, Health Ministry, Government of India. Community advisory and monitoring boards were set up in the surveyed districts to oversee ethical conduct of the survey. All the participants provided written informed consent at study entry. All those who consented were given a unique identification number. Face-to-face interviews were conducted with MSM by trained staff in the vicinity of MSM cruising points. The key areas of enquiry were: demographic characteristics; sexual partners and sexual practices; condom use; STI and HIV awareness and health seeking behaviour; and information on exposure to the programme (questionnaire posted on www.nari-icmr.res.in). Blood samples were collected for testing HIV-antibodies (anti HSV-2 antibodies on 10% of random sample), and urine samples were tested for Neisseria gonorrhoeae and Chlamydia trachomatis. Participants were offered free medical consultation for STI syndromic treatment. Behavioural and biological information was linked anonymously to safeguard participants’ right to confidentiality.
106
S. Deshpande and S. Bharat
Data analyses Sexual mixing among MSM was assessed in terms of two main criteria, namely, sexual mixing based on gender of the partner (male, female, TG or hijra) and partnership status (casual, regular/married and commercial). Participants were categorised according to the gender of the partner and partnership status they reported at the time of interview. Gender-based mixing of partners resulted in the following five categories: (1) ‘Men who have sex with males only’, (2) ‘men who have sex with TGs only’, (3) ‘men who have sex with males and TGs’, (4) ‘men who have sex with males and females’ and (5) ‘men who have sex with males, females and TGs’. For the purposes of this analysis, ‘MSM with mixed partners’, i.e. ‘men who have sex with males and females’ were compared with ‘men who have sex with males only’. The information on types of partners was elicited for the last month. MSM reporting different partnerships, based on their status and regardless of gender, were categorised as one of the following: . MSM with regular partners: defined as MSM reporting regular, i.e. steady partners, irrespective of the gender of the partners. This included regular/steady male and/or female partners, i.e. wife/girlfriend of the participants. . MSM with casual partners: defined as MSM reporting casual or non-regular and non-commercial partners (male and/or female). . MSM with commercial partners: defined as MSM reporting engagement in commercial partnerships. Here commercial partnerships includes paid as well as paying male partners. . MSM with mixed status partners: defined as MSM reporting combinations of partnership statuses; i.e. those who reported engaging in sex with regular as well as casual and commercial partners. Descriptive statistics was used to compare demographic variables and sexual risk behaviour. Pearson’s Chi-squared test with p value < 0.05 was used to assess significance of bivariate relationships between types of partners and partnerships MSM have and their condom use behaviour. MSM categories, namely, having mixed partners versus having only male partners and mixed partnership status versus exclusive partnership status, i.e. MSM who were engaged in only one partnership status (only regular or only casual or only commercial), were compared. These categories were chosen as they had sufficiently large denominators to conduct further analysis. The rest of the categories with smaller denominators, i.e. ‘men who have sex with TGs only’ (n = 76) and ‘men who have sex with males, females and TGs’ (n = 16), were dropped from the analysis. Risk behaviour was assessed in terms of consistent condom use (CCU), taking into consideration eight questions: (1) ‘every time’ condom use with regular male partners; (2) ‘every time’ condom use with casual male partners; (3) ‘every time’ condom use with commercial (paying and paid) male partners; (4) ‘last time’ condom use with regular male partners; (5) ‘last time’ condom use with casual male partners; (6) ‘last time’ commercial (paying and paid) male partners; (7) ‘every time’ condom use with commercial female partners; and (8) ‘last time’ condom use with commercial female partners. All eight questions were combined to get the CCU score among surveyed MSM. Additionally, CCU was computed for all types of male partners of MSM, namely, regular, casual and commercial. The Adjusted Odds Ratio (AOR) and 95% confidence intervals (CIs) were calculated as a measure of association using multiple logistic regressions. In regression analyses, independent variables included were age, education, occupation, marital status, reported
Global Public Health
107
sexual identity, alcohol use, anal sex with regular male partner, number of commercial partners, HIV test, perceived risk of HIV infection, currently having STI symptoms, received any intervention services, whether member of Self Help Group (SHG)/ Community-based Organisation (CBO), and diagnosed HIV serostatus as a biological indicator. These variables were included in the model based on significant association in the univariate regression analyses. CCU was taken as a dependent variable for two categories: MSM who have only male partners and MSM who were in mixed partnerships. The independent variables were adjusted for each other in the models and adjusted odds ratios were calculated at significance level less than 0.05. All statistical calculations were done after adjusting for sampling differences by applying appropriate weights. All analyses were done using STATA (version 12). Results A total of 692 self-reported MSM from Mumbai–Thane and Pune districts participated in the second round of the IBBA survey, and 689 were included for this analysis, as three cases did not fit into the eligibility criteria (Table 1). Of the total sample, more than half (359 or 52%) reported having sex with only male partners in the last month (Table 1). Nearly one-third (238 or 34.5%) reported having sex with both males and females, about one-tenth (76 or 11.1%) with males and TGs; while a small fraction (16 or 2.3%) reported having sex with all three genders. About one-fourth of the sample (27%) reported their sexual engagement in exclusive partnerships such as, ‘only regular/steady partners’, ‘only casual partners’ or ‘only commercial partners’ (Table 1). A large majority of MSM (73%), however, reported mixed status partners, suggesting they had sex concurrently with partners, some of whom were their steady partners, some casual and some paid partners. Table 1. Gender and partnership status-based mixing of partners. Gender-based mixing
n = 689 (%)
Have only male partners Have male and TG partners Have male and female partners Have male, TG and female partners Partnership status-based mixing Exclusive partnership status Have only regular partnership status (M/F) Have only casual partnership status (M/F) Have only commercial partnership status (M/F) Mixed partnership status
359 (52.1) 76 (11.1) 238 (34.5) 16 (2.3) n = 689 (%) 187 (27.1) 95 (13.8) 55 (8) 37 (5.4) 502 (72.9)
Intersection of gender-based and status-based mixing of partners Of 238 MSM with both male and female partners, 52.4% were married and 47.6% were unmarried. Among all MSM, 4% reported sex with paid female partners and 6.7% with casual female partners in the last month. Among MSM reporting sex with only regular/ steady partners (13.8%), a significant proportion reported both gender (male and female) partners (59% p < 0.01). Within this sub-group of MSM, 40% were married and 49.4% of them reported their spouse as their regular female partner. In contrast, almost all MSM who were in casual partnerships had only males and TG as their partners (98.9%
108
S. Deshpande and S. Bharat 0.9
6.2 1.7
100 80
38.2
59
60
4.2
98.9
40
3.3
20
37.6
92 56.7
0 Regular partnership status
Casual partnerhsip status Commercial partnership Mixed partnerhsip status status
Male partners Male and female partners
Figure 1.
Male and TG partners Male, TG and female partners
Gender of the partners within partnership status.
p < 0.01). Further, of all MSM who reported sex exclusively with commercial partners, over 90% had only male paid partners (p < 0.01). Of these 90% of MSM who engaged in sex with commercial partners, 12.4% were married at the time of the survey. Among the 73% of MSM who were engaged in sex with a mix of regular, paid and casual partners, more than half (56.7%) had only male partners, while 38.2% reported having both gender partners (male and female; see Figure 1).
Gender-based mixing of partner and condom use Levels of CCU were compared between MSM with only male partners and those with mixed partners (males and females). Overall, CCU was low across partner types; the highest level being 89% with commercial female partners and as low as only 30% with regular female partners. MSM with only male partners reported higher CCU with their regular male partners (73.4%) than with other casual (58.2%) and commercial male partners (60.7%; p < 0.01). However, MSM with exclusively male partners reported overall lower CCU with all types of their male sexual partners as compared to MSM who reported having mixed gender partners (47.3% and 62%, respectively, p = 0.11; Figure 2). Overall, MSM who have ‘male only’ partners as well as those with ‘male and female’ sex partners reported higher CCU with their non-commercial male partners (with regular male 73.4% and 80.8%, respectively; with casual male partners 58.2% and 62%, 100 89
80 80.8 60
73.4 58.2
62
62
60.7
40
47.3 40.5 30
20 0 Regular male partners
Regular female partners
MSM who have male partners
Casual male partners
Commercial male partner
Paid female partners
All male partners*
MSM who have mixed partners (male and female)
Figure 2. CCU among MSM who have only male partners and those with mixed gender partners. Note: *All type of male partners – CCU with only the male partners (of any type).
Global Public Health
109
respectively) compared to their commercial male partners (60.7% and 40.5%, respectively, p = 0.11). In short, among both categories of MSM, CCU was higher with their regular male partners than with casual male partners and lowest with male sex workers. MSM with mixed gender partners reported lowest CCU with their regular female partner (30% with wife or girlfriend), than with their regular male partners (80.8%) and commercial female partners (89%).
Partnership status-based mixing and condom use Unlike gender-based mixing of partners, partner mixing based on relationship status shows a significant difference in levels of CCU with different types of partners across MSM categories (Table 2). Table 2 shows comparatively higher levels of CCU (49–82%) reported by MSM with exclusive partnership types (i.e. either regular, casual or commercial), whereas respondents who engaged in mixed status partnerships reported similar CCU levels (23–88%). Specifically, MSM with mixed status partnerships reported lower levels of CCU with ‘all types of male partners’ (43.2%, p < 0.01). About half (49.4%) of all MSM having only steady or regular partnerships used condoms consistently with their regular female partners (wife/girlfriend), whereas less than one-fourth or 23% of MSM who were involved in mixed partnerships reported CCU with their regular female partners (Table 2). Similarly, 81.8% of MSM in exclusively casual partnerships reported CCU with their casual male partners, while about half (54.4%) of those involved in mixed partnerships (p = 0.02) reported CCU with their casual partners. MSM who were engaged in exclusive commercial partnerships reported higher levels of CCU with their commercial male partners (63.7%) and 100% with female commercial partners. Levels of CCU differed significantly among MSM reporting all types of ‘male only’ partners by their partnership type. For instance, over three-fourths or 75.3% of MSM with ‘male only’ partners reported CCU with their regular male partners, 81.8% with their casual male partners, 63.7% with commercial male partners and 43.2% with mixed male partnerships (p < 0.01; see Table 2).
Factors associated with CCU among MSM reporting partners by gender and partnership status The association of demographic characteristics and sexual behaviour with CCU was explored using multivariate analysis (see Table 3). Factors associated with CCU within the category of MSM who have only male partners were age (26 years and above, AOR = 0.27, CI 0.11–0.70), married status (AOR = 1.24, CI 0.87–1.76) and self-identification as Kothis1 (AOR = 0.14, CI 0.03– 0.55), reporting ‘always anal sex’ with their regular male partner (AOR = 6.85, CI 2.53– 18.52) and currently having an STI symptom (AOR = 0.04, CI 0.00–0.73). Similarly, the determinants of CCU among MSM reporting partners of mixed status were graduate (AOR = 0.22, CI 0.07–0.69), self-identification as Kothis (AOR = 0.30, CI 0.10–0.93), reporting ‘always anal sex’ with regular male partner (AOR = 6.34, CI 2.58– 15.58) and current STI symptom (AOR = 0.12, CI 0.01–1.02).
110 S. Deshpande and S. Bharat
Table 2. CCU among MSM categorised by partnerships based on status.
CCU with Regular male partners Regular female partners Casual male partners Commercial male partners Commercial female partners All type of male partnersa
Total (N = 689) %
MSM reporting regular partnerships (n = 95) %
MSM reporting casual partnerships (n = 55) %
MSM reporting commercial partnerships (n = 37) %
MSM reporting mixed partnerships (n = 502) %
p value
74.4 29.8 57.6 53.2 88.6
75.6 49.4 – – –
– – 81.8 – –
– – – 63.7 100
70.4 23.1 54.4 52.6 88.2
0.03* 0.09 0.02* 0.58 0.65
52
75.3
81.8
63.7
43.2
Sexual partner mixing and differentials in consistent condom use among men who have sex with men in Maharashtra, India Sucheta Deshpande* and Shalini Bharat Tata Institute of Social Sciences (TISS), Mumbai, India (Received 11 November 2013; accepted 30 August 2014) Sexual partner mixing among men who have sex with men (MSM), based on both gender and partnership status, is an understudied theme in India. Using data from Round 2 of the Integrated Bio-behavioral Survey, this paper reports on gender and partnership status-based sexual mixing and levels of consistent condom use (CCU) among MSM in Maharashtra. A total of 689 MSM were sampled using probabilitybased sampling. Bivariate and regression analyses were carried out on condom use and partnership mixing. Over half (52%) of all MSM reported having only male partners while about one-third (34.5%) reported having partners of both gender. Over 70% of MSM engaged in sex with a mix of casual, regular, commercial and noncommercial partners. MSM with only male partners reported lower CCU as compared to MSM with partners of both genders (47.3% and 62%, respectively, p = 0.11). CCU levels differed significantly by status of sex partner. Overall, MSM having ‘men only’ as partners and those with partners of mixed status have greater risk behaviour in terms of low CCU. HIV prevention interventions need to focus attention on men in ‘exclusively male’ sex partnerships as well as MSM with a mix of casual, regular and commercial partners. Keywords: MSM; sexual mixing; sexual partnerships; condom use; India
Introduction Globally, a concern has been raised for varied rates of HIV infection among men who have sex with men (Baral, Sifakis, Cleghorn, & Beyrer, 2007; Griensven & Wijngaarden, 2010; Griensven, Wijngaarden, Baral, & Grulich, 2009). Comprehensive reviews of HIV burden in MSM worldwide report HIV prevalence among MSM ranging between 3% in the Middle East and North Africa to 25% in the Caribbean (Beyrer, Baral et al., 2012; Beyrer, Sullivan et al., 2012). HIV prevalence among MSM in India is reported between 4.4% and 22% (Brahmam et al., 2008; Hernandez et al., 2006; National AIDS Control Organisation (NACO), 2012; Setia et al., 2008). In the early phase of India’s national AIDS control programme, the focus was on preventing HIV transmission within heterosexual relationships, particularly between female sex workers and their male clients (Dandona, Sisodia, et al., 2005; Halli, Ramesh, O’Neil, Moses, & Blanchard, 2006; Steinbrook, 2007; Thilakavathi et al., 2011). However, there is increasing recognition today of the much greater degree of heterogeneity in HIV transmission dynamics in India (Chandrasekaran et al., 2006). Based on sentinel site data and data from surveys such as the integrated behavioural and *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
104
S. Deshpande and S. Bharat
biological assessment (IBBA) and other studies conducted among MSM (Blanchard et al., 2007; Brahmam et al., 2008; Hernandez et al., 2006), HIV prevalence in this key population is found to be higher than in the general population (4.43% and 0.35%, respectively; NACO, 2012). Globally, two factors are commonly discussed in relation to HIV transmission risk among MSM, namely, multiplicity of sex partners and sexual mixing patterns within ‘same gender’ and between ‘diverse gender’ networks (Doherty, Padian, Marlow, & Aral, 2005; Yuhua et al., 2011). Partner preference plays a vital role in shaping the sexual networks within which sexual mixing takes place. Sexual mixing between ‘unlike’ groups, in which people select partners from other gender groups, and between ‘like’ groups, in which people select partners from their own gender groups, has implications for sexually transmitted infections (STI) (Aral, Adimora, & Fenton, 2008; Doherty, Shiboski, Ellen, Adimora, & Padian, 2006; Hertog, 2007; Yuhua et al., 2011). Compared to Western countries, SouthAsian countries report a multiplicity of sexual identities within the MSM population. In South Asia, sexual identities are culturally derived constructs which cannot be easily captured in terminologies used in the West such as, ‘gay’, ‘transgender’ (TG) or ‘bisexual’ (Bondopadhyay & Haque, 2002; Khan, 1997). For example, men engaged in sex work in the West often identify themselves as ‘heterosexual’; however, such identifications might be different and play an important role in relation to sexual practices in Asian countries (Dowsett, Grierson, & McNally, 2006). Often, MSM classified in Indian studies are a diverse group of men. They are engaged in transactional sex, perform different types of sex roles, i.e. receptive, penetrative, or both, and have male, female and TG partners. Various studies suggest that the sexual networks of MSM in India are usually dense and complex, as they report high levels of both ‘like’ and ‘unlike’ network mixing by having males, TG and females as their sexual partners (Asthana & Oostvogels, 2001; Hernandez et al., 2006; Setia, Sivasubramanian, Anand, Row Kavi, & Jerajani, 2010; Shinde, Setia, Row Kavi, Anand, & Jerajani, 2009). Evidence supporting the dynamics of such complex network mixing among the MSM population in India is growing (Brahmam et al., 2008; Chakrapani, Row Kavi, Ramakrishnan, Rapporpori, & Raghavan, 2002; Khan, 2001). A recent study in Bangalore city reported divergent sexual behaviour of MSM with a high percentage of them reporting a mix of women and men (55%), men (45%) and hijra (3%) as their sexual partners (Phillips et al., 2009). A large proportion of MSM in India are married (Brahmam et al., 2008; Dandona, Sisodia, et al., 2005; Phillips et al., 2010) and many are also involved in sex with commercial and casual partners (Brahmam et al., 2008; Setia et al., 2006). In terms of HIV transmission dynamics, the overlap of MSM networks with the general population through marriage, together with the ‘unlike’ sexual mixing patterns, assumes great importance. Sexual practices, networks and identities of MSM in India need to be understood in the context of a largely heteronormative society where there is not only a high level of stigma against same-sex relations, but also where such behaviour is criminalised under section 377 of the Indian Penal code. In addition, social norms that uphold the institution of marriage and place high value on its procreative function make demands of a very different kind on MSM in India, forcing many to marry for social reasons, engage in bisexual behaviour and bear children, without disclosing their sexual orientation to their spouse. The regressive Section 377 was read down by the Delhi High Court in 2009, but the decision was not upheld in the country’s Supreme Court, sparking fresh fears among MSM about societal stigma and possible legal action against them for expressing their sexuality (Mahapatra, 2014; Skanaland, 2009). Stigma and the continued criminalisation of homosexuality compound challenges of HIV prevention work in the country. The sociocultural–legal context framing homosexuality and
Global Public Health
105
available empirical evidence point to the fallacy of treating MSM in India as a homogenous, monolithic group to be reached with standard prevention approaches (Brahmam et al., 2008; Hernandez et al., 2006). But the heterogeneity of the MSM population has not been adequately explored. For example, intermixing based on partnership status such as regular, casual or commercial and that based on gender (male/female/TG) within MSM sexual networks is understudied in India. This paper reports on sexual partner mixing based on both gender and partnership status and levels of condom use among MSM in Maharashtra state. The data are based on the second round of the IBBA survey carried out during 2009–2010. The IBBA surveys were conducted as part of evaluation of Avahan, the India AIDS Initiative intervention programme implemented in six high HIV prevalence states of India. The IBBA surveys were designed to study STI and HIV prevalence among different high-risk populations, including MSM in these states, and were carried out in two rounds in 2007 and 2009 (Indian Council of Medical Research & FHI 360, 2011).
Methods In Maharashtra state, the second round of IBBA was conducted with MSM in Mumbai– Thane and Pune districts in 2009–2010. The target sample size was 400 MSM per district (Saidel et al., 2008). A two-stage cluster sampling design with time location clusters (TLC) was used in the survey. At first, mapping and listing of all existing MSM sites were completed and the sampling frame for TLC was developed. In the first stage, systematic random samples of primary sampling clusters were selected by probability proportional to size. Next, from the selected clusters survey respondents were selected randomly from eligible respondents available during the selected time duration. If the desired sample size was not found in a district, a decision was made to cover the entire population available in that district; this was called a ‘take-all’ approach. In Mumbai– Thane, considered one domain, a probability-based TLC sampling (Chandrasekaran et al., 2008; Saidel et al., 2008) was carried out. In Pune district, since the size of the MSM population was found to be less than 400 in the identified cruising sites, the ‘take-all’ approach was adopted. The recruitment criterion for MSM was, ‘any male, identified at cruising points, 18 years or older, who had any type of sex (oral, manual or penetrative) with another male, paid or unpaid, in the last one month’. The survey was approved by the Ethics committee of the National AIDS Research Institute (NARI), and the screening committee, Health Ministry, Government of India. Community advisory and monitoring boards were set up in the surveyed districts to oversee ethical conduct of the survey. All the participants provided written informed consent at study entry. All those who consented were given a unique identification number. Face-to-face interviews were conducted with MSM by trained staff in the vicinity of MSM cruising points. The key areas of enquiry were: demographic characteristics; sexual partners and sexual practices; condom use; STI and HIV awareness and health seeking behaviour; and information on exposure to the programme (questionnaire posted on www.nari-icmr.res.in). Blood samples were collected for testing HIV-antibodies (anti HSV-2 antibodies on 10% of random sample), and urine samples were tested for Neisseria gonorrhoeae and Chlamydia trachomatis. Participants were offered free medical consultation for STI syndromic treatment. Behavioural and biological information was linked anonymously to safeguard participants’ right to confidentiality.
106
S. Deshpande and S. Bharat
Data analyses Sexual mixing among MSM was assessed in terms of two main criteria, namely, sexual mixing based on gender of the partner (male, female, TG or hijra) and partnership status (casual, regular/married and commercial). Participants were categorised according to the gender of the partner and partnership status they reported at the time of interview. Gender-based mixing of partners resulted in the following five categories: (1) ‘Men who have sex with males only’, (2) ‘men who have sex with TGs only’, (3) ‘men who have sex with males and TGs’, (4) ‘men who have sex with males and females’ and (5) ‘men who have sex with males, females and TGs’. For the purposes of this analysis, ‘MSM with mixed partners’, i.e. ‘men who have sex with males and females’ were compared with ‘men who have sex with males only’. The information on types of partners was elicited for the last month. MSM reporting different partnerships, based on their status and regardless of gender, were categorised as one of the following: . MSM with regular partners: defined as MSM reporting regular, i.e. steady partners, irrespective of the gender of the partners. This included regular/steady male and/or female partners, i.e. wife/girlfriend of the participants. . MSM with casual partners: defined as MSM reporting casual or non-regular and non-commercial partners (male and/or female). . MSM with commercial partners: defined as MSM reporting engagement in commercial partnerships. Here commercial partnerships includes paid as well as paying male partners. . MSM with mixed status partners: defined as MSM reporting combinations of partnership statuses; i.e. those who reported engaging in sex with regular as well as casual and commercial partners. Descriptive statistics was used to compare demographic variables and sexual risk behaviour. Pearson’s Chi-squared test with p value < 0.05 was used to assess significance of bivariate relationships between types of partners and partnerships MSM have and their condom use behaviour. MSM categories, namely, having mixed partners versus having only male partners and mixed partnership status versus exclusive partnership status, i.e. MSM who were engaged in only one partnership status (only regular or only casual or only commercial), were compared. These categories were chosen as they had sufficiently large denominators to conduct further analysis. The rest of the categories with smaller denominators, i.e. ‘men who have sex with TGs only’ (n = 76) and ‘men who have sex with males, females and TGs’ (n = 16), were dropped from the analysis. Risk behaviour was assessed in terms of consistent condom use (CCU), taking into consideration eight questions: (1) ‘every time’ condom use with regular male partners; (2) ‘every time’ condom use with casual male partners; (3) ‘every time’ condom use with commercial (paying and paid) male partners; (4) ‘last time’ condom use with regular male partners; (5) ‘last time’ condom use with casual male partners; (6) ‘last time’ commercial (paying and paid) male partners; (7) ‘every time’ condom use with commercial female partners; and (8) ‘last time’ condom use with commercial female partners. All eight questions were combined to get the CCU score among surveyed MSM. Additionally, CCU was computed for all types of male partners of MSM, namely, regular, casual and commercial. The Adjusted Odds Ratio (AOR) and 95% confidence intervals (CIs) were calculated as a measure of association using multiple logistic regressions. In regression analyses, independent variables included were age, education, occupation, marital status, reported
Global Public Health
107
sexual identity, alcohol use, anal sex with regular male partner, number of commercial partners, HIV test, perceived risk of HIV infection, currently having STI symptoms, received any intervention services, whether member of Self Help Group (SHG)/ Community-based Organisation (CBO), and diagnosed HIV serostatus as a biological indicator. These variables were included in the model based on significant association in the univariate regression analyses. CCU was taken as a dependent variable for two categories: MSM who have only male partners and MSM who were in mixed partnerships. The independent variables were adjusted for each other in the models and adjusted odds ratios were calculated at significance level less than 0.05. All statistical calculations were done after adjusting for sampling differences by applying appropriate weights. All analyses were done using STATA (version 12). Results A total of 692 self-reported MSM from Mumbai–Thane and Pune districts participated in the second round of the IBBA survey, and 689 were included for this analysis, as three cases did not fit into the eligibility criteria (Table 1). Of the total sample, more than half (359 or 52%) reported having sex with only male partners in the last month (Table 1). Nearly one-third (238 or 34.5%) reported having sex with both males and females, about one-tenth (76 or 11.1%) with males and TGs; while a small fraction (16 or 2.3%) reported having sex with all three genders. About one-fourth of the sample (27%) reported their sexual engagement in exclusive partnerships such as, ‘only regular/steady partners’, ‘only casual partners’ or ‘only commercial partners’ (Table 1). A large majority of MSM (73%), however, reported mixed status partners, suggesting they had sex concurrently with partners, some of whom were their steady partners, some casual and some paid partners. Table 1. Gender and partnership status-based mixing of partners. Gender-based mixing
n = 689 (%)
Have only male partners Have male and TG partners Have male and female partners Have male, TG and female partners Partnership status-based mixing Exclusive partnership status Have only regular partnership status (M/F) Have only casual partnership status (M/F) Have only commercial partnership status (M/F) Mixed partnership status
359 (52.1) 76 (11.1) 238 (34.5) 16 (2.3) n = 689 (%) 187 (27.1) 95 (13.8) 55 (8) 37 (5.4) 502 (72.9)
Intersection of gender-based and status-based mixing of partners Of 238 MSM with both male and female partners, 52.4% were married and 47.6% were unmarried. Among all MSM, 4% reported sex with paid female partners and 6.7% with casual female partners in the last month. Among MSM reporting sex with only regular/ steady partners (13.8%), a significant proportion reported both gender (male and female) partners (59% p < 0.01). Within this sub-group of MSM, 40% were married and 49.4% of them reported their spouse as their regular female partner. In contrast, almost all MSM who were in casual partnerships had only males and TG as their partners (98.9%
108
S. Deshpande and S. Bharat 0.9
6.2 1.7
100 80
38.2
59
60
4.2
98.9
40
3.3
20
37.6
92 56.7
0 Regular partnership status
Casual partnerhsip status Commercial partnership Mixed partnerhsip status status
Male partners Male and female partners
Figure 1.
Male and TG partners Male, TG and female partners
Gender of the partners within partnership status.
p < 0.01). Further, of all MSM who reported sex exclusively with commercial partners, over 90% had only male paid partners (p < 0.01). Of these 90% of MSM who engaged in sex with commercial partners, 12.4% were married at the time of the survey. Among the 73% of MSM who were engaged in sex with a mix of regular, paid and casual partners, more than half (56.7%) had only male partners, while 38.2% reported having both gender partners (male and female; see Figure 1).
Gender-based mixing of partner and condom use Levels of CCU were compared between MSM with only male partners and those with mixed partners (males and females). Overall, CCU was low across partner types; the highest level being 89% with commercial female partners and as low as only 30% with regular female partners. MSM with only male partners reported higher CCU with their regular male partners (73.4%) than with other casual (58.2%) and commercial male partners (60.7%; p < 0.01). However, MSM with exclusively male partners reported overall lower CCU with all types of their male sexual partners as compared to MSM who reported having mixed gender partners (47.3% and 62%, respectively, p = 0.11; Figure 2). Overall, MSM who have ‘male only’ partners as well as those with ‘male and female’ sex partners reported higher CCU with their non-commercial male partners (with regular male 73.4% and 80.8%, respectively; with casual male partners 58.2% and 62%, 100 89
80 80.8 60
73.4 58.2
62
62
60.7
40
47.3 40.5 30
20 0 Regular male partners
Regular female partners
MSM who have male partners
Casual male partners
Commercial male partner
Paid female partners
All male partners*
MSM who have mixed partners (male and female)
Figure 2. CCU among MSM who have only male partners and those with mixed gender partners. Note: *All type of male partners – CCU with only the male partners (of any type).
Global Public Health
109
respectively) compared to their commercial male partners (60.7% and 40.5%, respectively, p = 0.11). In short, among both categories of MSM, CCU was higher with their regular male partners than with casual male partners and lowest with male sex workers. MSM with mixed gender partners reported lowest CCU with their regular female partner (30% with wife or girlfriend), than with their regular male partners (80.8%) and commercial female partners (89%).
Partnership status-based mixing and condom use Unlike gender-based mixing of partners, partner mixing based on relationship status shows a significant difference in levels of CCU with different types of partners across MSM categories (Table 2). Table 2 shows comparatively higher levels of CCU (49–82%) reported by MSM with exclusive partnership types (i.e. either regular, casual or commercial), whereas respondents who engaged in mixed status partnerships reported similar CCU levels (23–88%). Specifically, MSM with mixed status partnerships reported lower levels of CCU with ‘all types of male partners’ (43.2%, p < 0.01). About half (49.4%) of all MSM having only steady or regular partnerships used condoms consistently with their regular female partners (wife/girlfriend), whereas less than one-fourth or 23% of MSM who were involved in mixed partnerships reported CCU with their regular female partners (Table 2). Similarly, 81.8% of MSM in exclusively casual partnerships reported CCU with their casual male partners, while about half (54.4%) of those involved in mixed partnerships (p = 0.02) reported CCU with their casual partners. MSM who were engaged in exclusive commercial partnerships reported higher levels of CCU with their commercial male partners (63.7%) and 100% with female commercial partners. Levels of CCU differed significantly among MSM reporting all types of ‘male only’ partners by their partnership type. For instance, over three-fourths or 75.3% of MSM with ‘male only’ partners reported CCU with their regular male partners, 81.8% with their casual male partners, 63.7% with commercial male partners and 43.2% with mixed male partnerships (p < 0.01; see Table 2).
Factors associated with CCU among MSM reporting partners by gender and partnership status The association of demographic characteristics and sexual behaviour with CCU was explored using multivariate analysis (see Table 3). Factors associated with CCU within the category of MSM who have only male partners were age (26 years and above, AOR = 0.27, CI 0.11–0.70), married status (AOR = 1.24, CI 0.87–1.76) and self-identification as Kothis1 (AOR = 0.14, CI 0.03– 0.55), reporting ‘always anal sex’ with their regular male partner (AOR = 6.85, CI 2.53– 18.52) and currently having an STI symptom (AOR = 0.04, CI 0.00–0.73). Similarly, the determinants of CCU among MSM reporting partners of mixed status were graduate (AOR = 0.22, CI 0.07–0.69), self-identification as Kothis (AOR = 0.30, CI 0.10–0.93), reporting ‘always anal sex’ with regular male partner (AOR = 6.34, CI 2.58– 15.58) and current STI symptom (AOR = 0.12, CI 0.01–1.02).
110 S. Deshpande and S. Bharat
Table 2. CCU among MSM categorised by partnerships based on status.
CCU with Regular male partners Regular female partners Casual male partners Commercial male partners Commercial female partners All type of male partnersa
Total (N = 689) %
MSM reporting regular partnerships (n = 95) %
MSM reporting casual partnerships (n = 55) %
MSM reporting commercial partnerships (n = 37) %
MSM reporting mixed partnerships (n = 502) %
p value
74.4 29.8 57.6 53.2 88.6
75.6 49.4 – – –
– – 81.8 – –
– – – 63.7 100
70.4 23.1 54.4 52.6 88.2
0.03* 0.09 0.02* 0.58 0.65
52
75.3
81.8
63.7
43.2
