Original Research••

Homosexual identity and network overlap predictors of HIV infection among older men from rural China: a case–control study

Abstract

Background The number of newly reported older (≥50 years old) people living with HIV/AIDS in China has been steadily increasing, and heterosexual transmission has been identified as the major route.

Methods In this study, based on biopsychosocial model, a case–control study (74 cases vs 296 controls) was applied to find out predictors (21 potential ones) of HIV infection among older men from rural China. Binary logistic regression with backward selection was applied.

Results Out of 21 potential predictors of HIV infection, 4 were confirmed: receiving HIV-related health education (HRHE) (OR 0.449, 95% CI 0.238 to 0.848), self-reported homosexual identity (16.517 (95% CI 1.073 to 254.182)), visiting female sex workers (FSW) (58.427 (95% CI 7.895 to 432.403)), complete non-overlap of intimate social network and sexual network (2.912 (95% CI 1.372 to 6.177)), partial overlap (3.334 (95% CI 1.536 to 7.236)).

Discussion Findings from this study suggest current comprehensive prevention and control services need to be strengthened to provide effective HRHE including condom use to older men and help them to improve condom use, especially during visiting FSW and among self-reported homosexual older men. Moreover, increasing intimacy in their sexual relationships is necessary.

What is already known on this topic

  • Heterosexual transmission has been identified as the major route of HIV infection among older men in China, and social support which helps to maintain well-being of older men is insufficient in China. Our main concern is whether older men get enough social support from their sexual partners or not.

What this study adds

  • Self-reported homosexual identity (16.517 (95% CI 1.073 to 254.182)), complete non-overlap of intimate social network and sexual network (2.912 (95% CI 1.372 to 6.177)), partial overlap (3.334 (95% CI 1.536 to 7.236)) are found out to be risky factors of HIV infection for older men in China.

How this study might affect research, practice or policy

  • Comprehensive prevention and control services should focus on both on commercial heterosexual transmission and on homosexual transmission at population level. Increasing intimacy in older men’s sexual relationships is as necessary as HIV-related health education.

Introduction

Globally, along with antiretroviral therapy (ART) scale-up, mortality and new cases of HIV decreasing results in increased prevalence of HIV since 2006.1 By December 2020, there are approximately 37.7 million people living with HIV/AIDS (PLHIV) worldwide,2 among which 8.1 million are older (≥50 years old),3 and it is estimated that in some country, the percentage of older male PLHIV will rise to 73% by 2030.4 There are about 1.40 million PLHIV in China,5 the number of newly reported older PLHIV in China has been steadily increasing,6 7 from 32 850 to 66 010 during 2015-2019,8 at a much higher rate than the general population and other age groups,9 and heterosexual transmission has been identified as the major route.6 7 10 There are 486.58 million older population in China in 2020, accounting for 34.52% of the whole population.11 According to a meta-analysis study, during 2010–2018, HIV infection rate in older people is 1.68% in China and 2.13% in west part of China.12 Compared with other age groups, older men have minimal knowledge about HIV/AIDS,13 and condom use among them is extremely low.14 15 Currently, HIV-related health education (HRHE) and behavioural intervention among older people in China mainly adopts the knowledge-attitude-behaviour (KAP) model, and the knowledge and ability reservation of health educators are generally insufficient.10

Older men, whose sexual activities decline obviously after the age of 5016 and who were excluded from the priority of HIV prevention programmes,17–20 were not recognised and listed as a key prevention group until 201721 in China. Due to convention22 and stigma,23 sex has been taken as procreation within a family context in China. Chinese, especially older people, do not talk about sex.22 24 Since the 1980s, when migrating alone, due to a lack of norms and values’ control from original communities,14 22 25–30 sexual attitudes in China have changed to being an individual’s responsibility as long as there is no foreseen negative impact on the well-being of others or the larger society.22 Having sexual behaviour is not only to fulfil physical needs but also emotional needs.14 22 25–30 The closer clients feel with female sex workers (FSW), the less they use condom consistently.14 Moreover, it is estimated that men who have sex with men (MSM) account for 1.73% of men aged 18–64 years in China.31 In 2013, 17.9% of MSM reported that they had sex with both men and women (MSM/W), and living in Chengdu was identified as a risk factor.32

Conventionally, Chinese society is a patriarchy society where men are supposed to be strong and ready to give support to their intimate social networks.33 34 Social support which helps to maintain well-being35–38 for older men in China is insufficient. It is not clear the role of social and sexual network overlap in HIV Infection among older men from rural China. In order to find out the roles of sexual behaviours, sexual identity, health education, and especially social and sexual network overlap in HIV Infection among older men from rural China, a case–control study was conducted.

Methods

Study site

HIV epidemic information was obtained through China National HIV Sentinel Surveillance System and analysed. Due to a large amount of PLHIV and increasing epidemic, a county-level city (city A) with a population around 810 000 in Chengdu was identified as the study site.39

Study design, participant and procedures

We conducted a case–control study comparing newly diagnosed HIV-infected older men with HIV-uninfected older men.

Eligibility criteria for cases included (1) ≥50 years old, (2) male, (3) living in current address for at least 6 months, (4) newly diagnosed HIV (+) and (5) willing to participate in the study, and sign informed content. We also enrolled controls, those who were HIV (−). Exclusion criteria for both groups included (1) cannot speak clearly and (2) with mental problems and recall disability.

Sample size was calculated according to the formula  Inline Formula , where  Inline Formula  is the value corresponding to the test level, δ is the difference between the overall mean of the two groups and  Inline Formula  is the combined variance of the two groups. The power of a test was taken as 80% and the test level as 0.05. According to our pilot study, the overall mean value of intimacy in the case group was 3.68 with the variance as 1.808, and the control group was 2.71 with 1.440. The sample size of the case group (n1) was calculated as 27, and 108 for control group (n2) at 1: 4 ratios. During 2019–October 2020, 74 cases with 296 controls of the same ages were included. In total, there were 370 respondents in this study.

Eligible older men interested in participation provided informed consent after eligibility was confirmed. An in-depth interview in a separate room with one case was conducted by one skilled staff from local Centers for Disease Control and Prevention (CDC) within 1 month after cases’ HIV diagnosis. Face-to-face interview with a questionnaire was conducted for one control by one well-trained male interviewer in a separate room either at village health centres or activity centres. Local slangs were used to refer sexual behaviours, such as ‘hanging out with Miss(Xiaojie)’ referring ‘visiting FSW’. Both cases and controls’ responses were recorded into same questionnaire. Cases were paid ¥40 (approximately US$7) for participation, and controls got a big bottle of dishwashing liquid (approximately valued US$2) for their participation. Condoms were also provided if needed.

HIV testing

Based on China National Guideline for Detection of HIV/AIDS, Dot immunocolloid gold rapid test (Yingke New Technology) or ELISA was used for HIV antibody test as HIV rapid testing. When HIV (+) results were obtained from HIV rapid testing, 3–5 mL cubital venous blood was sampled and sent to local HIV confirmation laboratory. Immunoblotting (westernblot (WB)) was applied for HIV antibody confirmatory test.

Measures

Based on biopsychosocial model,40 the questionnaire included items on respondents’ demographic characteristics, biological factors,6 7 10 28 30 psychological factors,14 22 25–32 social factors35–37 and health services.13 39 Moreover, overlap of intimate social network and sexual network was measured.

Biological factors

(1)Sexual desire changing after the age of 50? Response options are ‘no sexual need’, ‘decline’, ‘not change’, ‘increase’ and (2) high-risk sexual behaviours: lifetime experiences of visiting FSW, casual sexual behaviour/one night stand, non-commercial non-fixed heterosexual behaviour and homosexual behaviour.

Psychological factors

Sexual identity and sexual attitudes were used to reflect psychological factors.

  1. Sexual identity was assessed by asking ‘how do you identify yourself sexually?’ Response options are heterosexual, homosexual and bisexual.

  2. Sexual attitudes were assessed by asking respondents their attitude towards (people) visiting FSW, towards extramarital heterosexual behaviour and towards homosexual behaviour. Response options of the former two are ‘admire’, ‘it is understandable as a common human behaviour’, ‘undecided’, ‘could not accept’ and ‘strongly objection’. Response options of towards homosexual behaviour are ‘it is understandable and has homosexual behaviour’, ‘it is understandable but not acceptable’, ‘could not accept’, ‘could not accept but be involved in’, ‘strongly objection’ and ‘do not know what homosexual behaviour is’.

Social factors included (1) whom did they live with; (2) daily entertainment: visiting tea-house with friends, playing mahjong, exercises, watching televison (TV), tourism, having party with friends, dancing at squares and other.

Health services provision was measured by asking whether they got HRHE before the investigation.

Overlap of intimate social network and sexual network

Intimate social network was measured by asking respondents ‘please nominate your current most intimate five social network member, starting from the most intimate one’, characteristic information was obtained. Sexual network was measured by ‘please nominate your most intimate ten sexual partners during the past 3 years, starting from the most intimate one’, detail information was also obtained.

Intimate social network members and sexual network members were cross-checked. If a partner in social network and a partner in a sexual network conform to the following characteristics, the partner was regarded as the same person: (1) type of relationship, (2) gender, (3) age, (4) nationality, (5) marital status, (6) educational level, (7) profession and (8) history of migrant workers.

When the number of intimate social network members of the respondents is equal to or less than that of sexual network members, the sexual network members were taken as the reference. If all the intimate social network members of the respondents were their sexual network members, it was regarded as ‘completely overlap’ between his intimate social network and sexual network. If some of the intimate social network members were their sexual network members, it was regarded as ‘partially overlap’. If the intimate social network members were not his sexual network members, they are regarded as ‘complete non-overlap’.

When the number of intimate social network members of the respondents is more than that of sexual network members, the members of the intimate social network were taken as a reference. If some of the members of the sexual network are members of his intimate social network, it is regarded as ‘partial overlap’. If all the members of the sexual network are not members of his intimate social network, it is regarded as ‘complete non-overlap’.

Data analysis

Frequencies for nominal variables, mean and SD for interval variables were assessed. T-test and χ2 tests/Fisher’s exact test as univariate analyses were used to examine the relationships between HIV infection and independent variables. Binary logistic regression with backward selection was applied to examine factors associated with HIV infection with p<0.05 in univariate analyses. Adjusted OR (AOR) and 95% CIs were calculated. Factors with AOR greater than 1 were categorised as risk factors and less than 1 as protective factors.

Results

Demographic characteristics, social factors and health education

Most respondents were at the age of 60–69 (40.54%), primary school educated (50.81%), married (81.08%), held rural residency (96.22%), latest worked as a farm worker/migrant worker (81.35%), migrated more than 1 year ago (45.95%), currently had monthly income less than ¥1000 (47.30%) and spent less than ¥500 for entertainment (74.59%), lived with their spouses and other family members (parents or children%) (48.11%), took visiting tea-house with friends (68.92%) and watching TV (63.51%) as daily entertainment. And more than half (59.48%) received HRHE before the investigation.

Five demographic variables were associated with HIV infection: household residency, marital status, migration, monthly income, monthly expense for entertainment. Four social factor variables were associated with HIV infection: living with, and visiting tea-house with friends, exercising and watching TV as daily entertainment. HRHE was also associated with HIV infection. In the multivariate model, only HRHE (OR 0.449, 95% CI 0.238 to 0.848) was retained. Details are shown in table 1.

Table 1
|
Demographic characteristics, social factors and health education associated with HIV infection (n=370)

Sexual desire change, sexual identity, attitude and high-risk sexual behaviours

Most of the respondents’ sexual desire declined after the age of 50 years, and self-reported as heterosexual, held undecided attitude or could not accept towards visiting FSW. 37.57% of respondents did not know what homosexual behaviour is. 29.46% visited FSW, 4.86% practised casual sexual behaviour and 1.35% reported practising homosexual behaviour.

Sexual identity, attitude towards visiting FSW, attitude towards visiting FSW, attitude towards extramarital heterosexual behaviour and attitude towards homosexual behaviour, visiting FSW, casual sexual behaviour, non-commercial non-fixed behaviour and homosexual behaviour were associated with HIV infection. In the multivariate model, sexual identity (16.517 (95% CI 1.073 to 254.182)), visiting FSW (58.427 (95% CI 7.895 to 432.403)) and homosexual behaviour (16.517 (95% CI 1.073 to 254.182)) were retained. Details are shown in table 1.

The overlap of intimate social network and sexual network

52.70% of the case group’s intimate social network and sexual network were complete non-overlap, more than the control group (18.58%) (p<0.05). In the multivariate model, the overlap of intimate social network and sexual network was retained (complete non-overlap: 2.912 (95% CI 1.372 to 6.177); Partial overlap: 3.334 (95% CI 1.536 to 7.236)). Details are shown in table 1.

Discussion

Both cases and controls in this study come from rural communities and represent older male population in China. At population level, our study confirms the function of commercial heterosexual behaviour as known6 7 10 and of self-reported homosexual identity31 32 in HIV transmission independently among older men from rural China. Instead of condom, intrauterine device for women has been main contraception method among married couples in China during 1979–2012, some older men never use condom in their whole life,24 38 condom use during visiting FSW is extremely low14 which put them at the risk of HIV infection. Multivariate model shows that homosexual identity is a risk factor, but not homosexual behaviour. Older men in Chengdu live in an environment where homosexual behaviours are tolerated.32 4.05% of cases self-reported as homosexual and 6.76% reported having sex with men. In this study, we did not measure the frequency of homosexual behaviour. Self-reported homosexual identity implicitly reflects the frequency of homosexual behaviour. The frequency of homosexual behaviour among homosexual identified MSM is more than among heterosexual identified MSM.28 The more frequently they have homosexual behaviour, the more possibility of HIV infection happens. Our previous study finds out that HIV-1 subtype CRF 55_01B which was first found out among MSM in China exists among older men at our study site,41 which provide molecular evidence for the role of homosexual behaviours in HIV transmission among older men. This study gives epidemiological evidence for the function of homosexual identity in HIV transmission among older men at population level. Moreover, older MSM/W play role of ‘bridge population’ as transmitting HIV infection from MSM to their heterosexual partners.28 30

As expected, receiving HRHE is a protective factor13 38 39 for avoiding HIV infection. During 2006–2017, awareness rate of HIV/AIDS core knowledge among the older population in China was 51.0%,42 far from the target in 2017 as 90%.21 Even though comprehensive prevention and control has been targeting older male population since 2017,21 during 2019–2020, only 59.48% of older men got HRHE in this study. Health services provision is not effective and sufficient to reach older men to achieve the ‘90-90-90 targets’, namely 90% of PLHIV knowing their status, 90% of those diagnosed receiving sustained ART, and 90% of those receiving ART achieving viral suppression by 2020.43 During 2016–2020, 92.9% of PLHIV in China received ART, and 96.1% of them achieve viral suppression,44 better than the global level,2 but the goal of knowing their status has not been achieved.44

Overlap of intimate social network and sexual network reflect that older men get social support from their sexual partners which helps to maintain their well-being.35–37 Conventionally, Chinese men are supposed to be a ‘giver’ rather than a ‘receiver’.33 34 Older men get little support from intimate social network, even from family members.45 In this study, 52.70% of cases’ intimate social network and sexual network were complete non-overlap, which means they did not receive social support from their sexual networks within the past 3 years. They are not close to their sexual partner, sexual behaviours are mainly commercial heterosexual behaviour (94.59%). At the same time, nearly 80% of controls have close relationships with their sexual partners, and overlap of intimate social network and sexual network help them avoid getting HIV infected. It is worth mentioning that we find out that emotional support is protective meanwhile material support only is risky for becoming HIV-positive among older men.38

Therefore, in order to prevent HIV transmission among older men, current comprehensive prevention and control services21 need to be strengthen to provide effective HRHE to access older men effectively, and to help older men to improve condom use, especially visiting FSW and among self-reported homosexual older men. In a conventional patriarchy society, Chinese society needs to reconsider older men as vulnerable ‘older people’ first, then ‘men’. As ‘older people’, increasing intimacy in their sexual relationships is necessary to help them both avoid HIV infection and maintain well-being.35–37

Limitations

In the current study, self-report behaviour information should be noted. In order to reduce report bias, our interviewers were well trained, interviews were conducted in separate rooms and local slang was used.

Moreover, in-depth interviews were conducted by skilled medical staff members from county A CDC for case group, but by well-trained young male interviewers for control group. In order to minimise informational bias due to the different types of interviewers between two groups, interviewers were trained by the same trainers, only well-trained interviewers conducted the in-depth interview and interviewers followed the same procedure.

Overview

Article metrics
Altmetric data not available for this article.
Dimensionsopen-url