The findings from this study have helped to elucidate how interpersonal characteristics of partnerships can influence condom use with non-commercial partners of women in sex work in southern India. Overall consistent condom use (CCU) with non-commercial partners was low and FSWs reported lower CCU with their husband or cohabiting partner than with their non-paying partner. Both FSWs and their non-commercial partners were found to be substantially connected to other types of partners through other sex partnerships. FSWs reported that a considerable proportion of these male partners had other sex partners. These partners included wives, FSWs or other types of female partners – both within and outside their local geographic settings (i.e., district of recruitment). These results highlight the vulnerability of FSWs to both acquisition and transmission of HIV/STIs within complex sexual networks, as well as the integral role of FSWs’ non-commercial partners as bridge populations who may facilitate the transmission of HIV to female partners outside the context of sex work.
The longevity of the sexual partnerships with FSWs’ non-paying partner appears to be particularly important in determining CCU, with a longer relationship duration being associated with lower CCU. A more nuanced understanding of what the duration of the relationship represents (e.g. increased trust, closeness or familiarity; decreased decision-making power or control) and how these can be addressed in HIV/STI prevention programming is needed. Although FSWs in southern India are highly economically vulnerable with few comparably well-paying employment prospects , factors representing the economic stability of the partner (e.g. employment status of the partner, or whether the partner provided economic support) were not significantly associated with CCU. Because the nature of non-commercial relationships is different from commercial relationships, and the economic support, if it exists, is often non-monetary, the decision to use a condom may be more influenced by interpersonal factors related to relationship intimacy (e.g., trust, emotional closeness, power or reproductive desires) than financial dependence. This is supported by studies of non-commercial partnerships of FSWs in other settings [1, 31]. However, economic dependence on the male partner is associated with lower condom use in other settings [23, 32] and studies of transactional sex arrangements have suggested that trade-offs within these relationships occur, such as increasing amounts of transfers of support (in terms of money, goods, gifts) in exchange for risky behaviour that is perceived as valuable to the male partner (such as sex without a condom) [23, 33], even after adjusting for the duration of the relationship . A better understanding of the type and amount of transfers within non-commercial partnerships of FSWs in southern India, both quantitatively and qualitatively could help to better characterize the influence of economic dependence (or co-dependence) on condom use.
While exposure to interventions has been found to be positively associated with increased condom use by FSWs with their clients, including in our setting [4, 12, 34–37], , condom use within non-commercial partnerships has not been a major focus of most interventions and is rarely directly addressed effectively. Condom use within non-commercial partnerships therefore represents an important intervention point, particularly since many non-commercial partners also have other commercial or non-commercial sex partners. Interventions that include a focus on condom use within non-commercial partnerships of FSWs need to go beyond increasing education and access to address issues of intimacy and trust within relationships from the perspectives of women and men, as well as power disparities that favour the male partner [15, 17, 39], . This is particularly true in settings where women have lower status than men and reduced economic opportunities, and traditional social norms frame socio-cultural views of condom use [17, 41], . Interestingly, although education was associated with higher condom use in bivariate analysis in this study, the effects were removed in multivariable analysis. To better understand the complexity of social and environmental factors influencing condom use in non-commercial partnerships and how to develop interventions to increase condom use, conceptual frameworks that have been useful in explaining behaviour could be employed [43–45]. Factors relating to the intention to act, agency and decision-making power (i.e., perception of behavioural control) of women with respect to condom use should be assessed . Furthermore, interventions must be constructed in ways that acknowledge traditional social norms surrounding condom use [17, 41],  and the potential role of gender-based violence [39, 41] and sexual coercion  in lower condom use. Qualitative research methods should be employed to better understand socio-cultural reasons for lower condom use within non-commercial partnerships in this setting.
Interventions designed for clients of FSWs as well as FSWs have been observed to contribute to declines in STI prevalence and increases in condom use [47, 48]. However, interventions designed for non-commercial male partners are uncommon, despite the often significant role of non-commercial male partners in determining if condoms are used within these partnerships [1, 49]. In India, this is in part because such partners are often hidden or difficult to access . Husbands are particularly inaccessible, since many are not aware of their partner’s involvement in sex work, while other non-commercial partners may be actively involved in the management of their partner’s sex work . Fewer husbands than main cohabiting partners in our study were aware of their partner’s involvement in sex work (23% versus 40%, p=0.01). It may be more effective to design male-focused interventions specifically for other non-commercial partners of FSWs, in this and other study settings. There is some evidence for the success of male-focused interventions which address social norms and gender-based violence in increasing condom use, including in India (the Sonagachi Project) and South Africa (Stepping Stones) [50, 51]. Targeting younger males may be particularly effective in terms of changing social norms surrounding sexual behaviour for men and women in India. A systematic review of six couples-focused behavioural interventions (outside the context of sex work) found that involvement in the programs was associated with reduced unprotected sex , suggesting that these types of interventions could be beneficial and should be explored. In some areas targeted by the Avahan AIDS Initiative in southern India, an increasing focus of the program has been on increasing safety within non-commercial partnerships of FSWs. Avahan’s primary focus in terms of increasing condom use, however, has been for condom use within commercial partnerships. In this study, condom use with non-commercial partners was statistically significantly different in multivariable analysis according to the district where women were recruited into the study. In particular, condom use was higher in Mysore district compared to the other districts. These differences may be due to differences in program implementation or timing of the intervention implemented in each district. Avahan has been the only intervention in Mysore district, but was not the first (though it is now the only) intervention in other districts. Notably, though, a recent study of three districts in Karnataka state (Mysore, Belgaum and Bellary) found that condom use at last sex with non-commercial (non-marital) partners of FSWs has increased significantly (24.4% to 55.9%) over six years of the intervention , although this was not the case earlier in the intervention . The largest increases were observed in Belgaum and Mysore. These results indicate that it may take longer for interventions targeted toward FSWs to have an effect on condom use within non-commercial compared to commercial partnerships, but that positive results can be observed. Additional research should be conducted to better understand the social and structural factors that operate within each district on a macro level to influence condom use, and the intersecting relationship between these factors and intervention impact. Individual-level variables such as those explored in this study may not necessarily capture these influences. Lower condom use with non-commercial relative to commercial partners persists and sustained interventions targeted toward increasing condom use within non-commercial partnerships should continue to be explored and developed. The use of HIV prevention methods that women have more control over (i.e., microbicides, female condom) should be investigated in this study population and within non-commercial partnerships.
In our study, CCU was almost two-fold higher with FSWs’ husband or cohabiting partner when this partner knew the FSW was in sex work. Condom use could be higher in these relationships because of an increased awareness of the risks incurred by these women by the male partner. This could also be due to greater exposure and involvement in HIV/STI programming designed for FSWs, or increased access by HIV/STI programs to male partners. Sex work occupational stigma, which influences women to hide their sex work status from their partners and families, has increasingly been postulated as a major barrier to health access for FSWs [54, 55]. These results provide support for sex work being recognized as a more legitimate occupation, where women do not feel obligated to hide their work from their non-commercial partners. However, disclosure of sex work needs to be understood within the context of local socio-cultural views and social norms regarding women’s status and sex work, to help ensure that women’s safety is not compromised by disclosure.
Sentinel surveillance and observational studies suggest that HIV and STIs have decreased among FSWs in Karnataka state since Avahan was introduced [12, 56], and mathematical modelling has indicated that the increase in condom use among FSWs with their clients after the intervention was introduced is consistent with decreasing HIV epidemiological trends over 2-3 rounds of survey data collection [57, 58]. Although commercial sex partnerships of FSWs play a large role in the spread of HIV [7, 59], the lack of information on non-commercial partners makes it difficult to assess their role in the spread of HIV through local sexual networks, and their overall potential contribution to HIV epidemics. However, since condom use remains low in non-commercial partnerships in this setting, to better inform the contribution of non-commercial partnerships to the spread of HIV, future empirical studies should collect information on the sexual behaviour of non-commercial partners with other partners (e.g., numbers and types of other partners, frequency of sexual contacts, condom use) and the presence of HIV infection among non-commercial partners of FSWs. Since almost all information on non-commercial partners has been collected second-hand from FSWs rather than from the perspective of their male partners, studies should be conducted with men.
There are several limitations to this study. The study is based in four districts in Karnataka state, southern India, and may not be generalizable to other regions in India. It is based on self-reported data from cross-sectional surveys, and self-reported data may be subject to social desirability bias . However, our sample size was large, particularly for a marginalized and hidden population of FSWs, and the cluster sampling design was aimed to make the sample as representative as possible. Reported condom use was substantially lower with non-commercial rather than commercial partnerships, indicating that women may have been comfortable reporting higher-risk behaviour with these partners. At the same time, it may be more socially acceptable for women to report lower condom use with non-commercial partners, since women as well as men may associate condom use with infidelity or reduced trust. We were unable to control for fertility desires of respondents, which may affect levels of condom use with non-commercial sex partners . However, since the majority of respondents reported using some kind of birth control for family planning purposes, this indicates that most women were not planning on becoming pregnant. Finally, developing questionnaires grounded in theoretical frameworks previously used in similar populations and settings could be helpful in explaining the reasons for condom use .