Overall, the level of sexual activity amongst the sample in this study was low. The overall percentage of both women and men on ART who reported sexual activity (38%) is much lower than the proportions found by some studies in developing country contexts see for example [28–32]. For example, the proportion of women in this study reporting sexual activity in the last six months is lower than that found among women in a three-country study of Brazil, South Africa and Uganda  where almost half had sexual intercourse in the last month (46%).
Not surprisingly there was a relationship between sexual intercourse in the past six months and a person’s marital status with married people more likely to have had sexual intercourse. Among those that have never married, were separated/divorced or widowed, 9% reported that they had sexual intercourse with a regular partner, but this is significantly less than those who reported themselves as married or engaged (58%). This finding on the relationship between sexual intercourse and marital status concurs with data from other developing countries with heterosexually driven epidemics. For example, married people with HIV in Kenya were significantly more likely to report sexual intercourse than those who were not . Similarly, in Uganda, compared with HIV-positive people who were widowed, those who identified as married were also significantly more likely to be sexually active [16, 17]. Likewise, a three-country study by Kaida et al.  showed that recent sexual activity for HIV-positive women in Brazil, South Africa and Uganda was significantly associated with being married.
Marital status was similarly important, although not statistically significant, in relation to consistent condom use. Those who reported not being married or engaged reported higher rates of consistent condom use compared with those who were married. HIV-positive people on ART who were married or engaged were more likely to be sexually active with a regular partner but were at the same time less likely to consistently use condoms, irrespective of a partner’s HIV status cf. .
The high level of sexual abstinence by women in this study suggests that although the literature to date has shown little to no control by women over their sexual and reproductive lives, the women in this study, and in particular those who were not married, were not devoid of power to control their sexual life as is indicated elsewhere. And merely because a woman with HIV does not use a condom with her husband does not automatically and unequivocally equate with disempowerment, as suggested by the qualitative interviews. The qualitative data from this study offers some insights into this particular aspect of the behavioural data and involves intangible issues associated with long-term relationships, intimate communication and love. Rather, condom use is negotiated in light of knowledge of a partner’s HIV status, long-term commitment and love. For those women with HIV-negative partners inconsistent condom use appears to be the initiated by their regular male partners. As one of the quotes used in the Results section of this paper illustrates, it appears that their regular partners (husbands) had accepted their HIV status and were not afraid of becoming infected. International research supports this inference, especially research in serodiscordant couples where it has been shown that the HIV-negative partner has been the principal initiator of unsafe sex [33, 34]. One reading of this could be that men had a poor understanding of their risk for HIV acquisition. Another reading of the data is that women were not able to counter their partners’ wishes to discontinue using condoms. A more likely scenario from an detailed reading of that data suggests that this absence of fear was based on elusive aspects of a relationships including acceptance of HIV, understanding and experience of risk and other relationship issues such as duration of the relationship, love and intimacy . Furthermore, what it means to be in a serodiscordant relationship is context-specific and it cannot then be assumed that risk-management and decisions about sexual practices will be and indeed should be the same globally .
Women were not questioned about lifetime sexual or physical violence or about sexual relationship power, so we cannot therefore postulate about their ability to have control in their sexual and reproductive lives more generally, as is possible in other studies cf. . There is however some research in PNG that does link HIV status with lifetime intimate partner violence . Unlike other studies that link HIV risk to intimate partner violence amongst newly diagnosed women with HIV, this study was of women who were already diagnosed with HIV. Furthermore, of the women with HIV-negative partners, all were in relationships that had been initiated following their diagnosis and therefore we cannot comment of the relationship between violence and HIV risk as others have done.
It is true that in PNG women are subjugated to much gender-based violence and have been reported to have minimal to no control over their sexual and reproductive lives [38, 39]. Furthermore, international research has consistently reported that women’s vulnerability to HIV is linked to an absence of power, ability to negotiate sexual relationships and experiences of intimate partner violence cf. [36, 40]. Moreover, it has long been posited that gender-based power differentials between women and men often compromises a woman’s ability to negotiate condoms cf. [41, 42]. Although it may well be possible that women in this study were unable to counter their HIV-negative partners’ desire to discontinue using condoms, this is not what the data suggested. This is an interesting finding and warrants further scrutiny especially in light of the long history of research on gender, HIV and vulnerably that suggests otherwise. It would however appear that it is possible that this data contributes to a recent and increasing recognition within PNG that women can and indeed are exhibiting agency in choosing sexual and lifetime partners , with some becoming involved in polygamous marriages in the desire to obtain status through money and items of modernity such as cars and expensive mobile phones .
In countries where viral load testing is part of routine clinical care and management, informed choices about viral suppression and risk of HIV transmission are possible. To date this is not the situation in PNG. And although in other settings there is an awareness of the low risk of sexual transmission while on ART (for example the Swiss Statement) there is no such discourse or health messaging in PNG. In fact, there is evidence in PNG that PLHIV whose partner is also positive are recommended to still use condoms for fear of ‘super-infection’ . Therefore, it is unlikely that such knowledge about viral load and infectivity can account for unprotected sex between sero-discordant partners for whom condoms play a greater role than where both partners are HIV-positive.
Although people living with HIV are increasingly making active choices to have children as a recent qualitative study of prevention of mother-to-child transmission in PNG has found , it is our informed opinion in this study that the low use of condoms was not reflective of a desire on behalf of HIV-positive women to conceive or of HIV-positive men to father a child. In 2008, when this study was conducted, ART was not widely available, and there was little to no discussion of HIV-positive women actively choosing to reproduce (as opposed to testing HIV-positive while pregnant) and prevention of mother to child transmission programs were still in their infancy. No women in the qualitative arm of our study reported wanting to have children after being diagnosed with HIV and neither did any men report wanting to have children. Furthermore, the qualitative data suggests that we can be confident that participants’ low rates of reported sexual activity was not an artefact of social desirability or indeed a discrepancy in what is constituted as sexual intercourse. Rather the qualitative data suggests that in the early stages of ART availability in PNG, many PLHIV were against the notion of both themselves and any other PLHIV from having sex. At the time of this study participants were reporting that reasons for sexual abstinence included a personal and social responsibility not to transmit HIV to another person and a fear of becoming re-infected with HIV. There was no gender difference in reported reasons for sexual abstinence. Additionally, the role of others, particularly health care workers, influencing a person’s beliefs and therefore decisions about sexual behaviour were also deemed important. This is not surprising since adherence to ART was also related to following instruction from the doctor/health care worker . As people stay on treatment and as people live longer it will be important to chart if and how sexual behaviour changes and if the meanings ascribed to sexual intercourse and condom use similarly alter.
Since much of the HIV prevention message in PNG focuses on reducing extra-marital relationships and being faithful to one’s partner, it appears that safe sex within marriages of people living with HIV may need to be emphasised. Condoms have traditionally been seen in PNG as something used outside of a marriage; sex within a marriage is perceived to be safe and this is suggested by the ‘one faithful partner concept’ [47–50]. However in this study of HIV-positive people, those who were married or engaged were proportionally more likely to report having an HIV-positive partner than those who were not, and for those with HIV-positive partners, the need for consistent condom use to prevent transmission is lessened . Though not statistically significant, proportionally less of those with a HIV-positive partner (43%) or of those who did not know the status of their regular partner (38%) consistently used condoms compared with those whose partner was HIV-negative (62%). In other words, consistent condom use was much higher (OR 2.12) among participants whose partners were known to be HIV-negative. Others have similarly identified that the HIV status of one’s partner affects condom use [27, 51].
Almost all of the participants with a regular sexual partner had disclosed their HIV status to them and disclosure was significantly higher amongst those who were married or engaged compared with those who were not. This rate of disclosure to a sexual partner was much higher than that found in other studies, which report disclosure rates to regular partners ranging between 62 - 80% [52–55]. Similarly this data suggests that compared with positive people in other settings who report barriers to disclosure of serostatus cf. , people in PNG experienced fewer obstacles in the disclosure of HIV in intimate sexual relationships.
Consistent use of condoms was not significantly associated with HIV knowledge. Knowledge about the preventative nature of condoms against the sexual transmission of HIV was high (77%), indicating, as others have found, that there was a discrepancy between knowledge and behaviour. Among those that reported vaginal intercourse in the last six months, less than half (46%) reported consistent condom use. However, knowing that ART cannot cure a person of HIV was significantly associated with consistent condom use (OR 3.89). In other words, people with living HIV who were well-informed about their ART were using condoms more consistently. This contrasts with South African data which found that HIV and ART knowledge was not associated with condom use . The association between education level and knowledge that HIV can be transmitted via sexual intercourse, but not via mosquito bites indicates that there could be some confusion about transmission of HIV which may be different from other commonly known blood-borne pathogens such as malaria amongst people who are less educated. This suggests that educational materials targeting this confusion may be beneficial.
Our study found that there was a strong association between having sexual intercourse in the last six months and consuming alcohol. On the other hand, there was a strong association between consistent condom use and consuming alcohol. While other behavioural studies have shown a relationship amongst HIV-positive people between drinking alcohol and high risk sexual behaviour [14, 32, 57, 58] this was not the case in this study of Papua New Guineans on ART.