Risk Factors of HIV and Other Sexually Transmitted Infections in China: A Systematic Review of Reviews

Background Sexually Transmitted Infections (STIs) are a global challenge. China, once said to have eradicated STIs, is now facing a rapid rise in the prevalence of HIV/STIs. This review of reviews aims to map HIV/STI risk factors among the Chinese population, with the objective of identifying risk factors to inform the formulation of effective prevention strategies. Methods A systematic search using key terms related to HIV/STIs, risk factors and the Chinese population in both English and Chinese databases (PubMed, PsycINFO, the Cochrane Library; Wanfang data, CNKI, VIP and SINOMED) was conducted, and peer-reviewed systematic reviews on the topic from 1991 to 2014 were selected. Identified risk factors were grouped into different level determinants based on the HIV Social Epidemiology Model, and then evaluated and reported based on the PRISMA checklist. Findings Of the twenty-eight reviews included, the majority were focused on well-established, individual level risk factors within key populations, with some highlighting the complexity of interacting factors (e.g., alcohol use and higher income in male migrants). While twenty-two reviews covered individual factors, only ten mentioned social factors and five had contents on structural factors. There are gaps in the evidence on social and structural level impacts of HIV/STIs, such as on stigma, discrimination, health policy, access to care, and illicit drug control policies. Migration and social expectation appear to pose a significant threat in aggravating the HIV/STI situation in China; for example, incarceration patterns indicated a significant risk of HIV/STIs for female sex workers. Conclusions Since international guidelines recommend an integrated and multi-level approach to HIV/STI prevention, a comprehensive approach targeting interventions at all levels along the continuum of care is needed to effectively curtail HIV/STI transmission in China. More research is needed to better understand the impact of socio-political interventions within a Chinese context.


Findings
Of the twenty-eight reviews included, the majority were focused on well-established, individual level risk factors within key populations, with some highlighting the complexity of interacting factors (e.g., alcohol use and higher income in male migrants). While twenty-two reviews covered individual factors, only ten mentioned social factors and five had contents on structural factors. There are gaps in the evidence on social and structural level impacts of HIV/STIs, such as on stigma, discrimination, health policy, access to care, and illicit drug control policies. Migration and social expectation appear to pose a significant threat in aggravating the HIV/STI situation in China; for example, incarceration patterns indicated a significant risk of HIV/STIs for female sex workers.

Introduction
Around the world, over one million people are infected with some kind of Sexually Transmitted Infection (STI) every day, making them a major global health problem [1]. The world's most populous country, China, was said to have eradicated STIs 50 years ago but is now experiencing the most precipitous increase in the prevalence of STIs and facing a huge challenge of identifying how to control HIV and STIs [2,3]. In 2012, 41,929 new cases of HIV/AIDS and about 780,000 People Living with HIV/AIDS (PLHIV) were reported in China [4]. Syphilis and gonorrhoea now represents the third and fifth most common category B infectious diseases in China. In 2012, there were about 1.6 million reported cases of syphilis, gonorrhoea and Hepatitis B in China [5,6]. STIs are known to be associated with many serious health consequences including irreversible neurological problems, prematurity or stillbirth, cervicitis, pelvic inflammatory disease, chronic pelvic pain, infertility, and ectopic pregnancy in women [7][8][9].
A thorough understanding of the risk factors is a cornerstone for designing effective prevention and control interventions [10]. Thus far in China, many reviews have attempted to sum up the risks from various sources in different key populations such as men who have sex with men (MSM), or people who use drugs. Traditional disease prevention intervention often focuses on individual behaviours and overlooks social and contextual factors, some of which are believed to be the fundamental source of disease development [11]. Social epidemiology, which studies "the role of social factors in the aetiology of disease" has been used as a framework for the determinants of HIV, and is used as the underlying conceptual framework for this review [12]. The complexity of HIV/STIs requires a comprehensive socio-epidemiological approach to fully understand the interplay of different risk factors [13,14]. Linking a social approach to epidemiology studies encourages a multidisciplinary approach which goes beyond the traditional individual behaviour method [15][16][17][18].
The aim of this study was to map out the risk factors for HIV/STIs in the Chinese population through a review of systematic reviews. It was anticipated that the overview of risk factors from pooled data could shed light on the priorities and inadequacy of the existing HIV/STIs studies in China and contribute to the fight against HIV/STIs globally.

Methods
A comprehensive search protocol of peer-reviewed journals was developed. Three English databases (PubMed, PsycINFO, and the Cochrane Library) and four major Chinese databases (Wanfang Data, CNKI, VIP Chinese Journal Database, and SINOMED) were selected. HIV, syphilis and gonorrhoea only became reportable across China since 1991, and thus used as the start of the timeframe of this study   [19].
We included systematic reviews that studied the Chinese population, covering adult HIV/ STI risk factors and including correlates of unsafe behaviour, prevalence and incidence.
Exclusion criteria included: article published in languages other than English or Chinese; articles which were concerned with only congenital STIs; interventional studies; studies which examined non-Chinese populations. Search terms included three aspects: HIV/STIs (including Medical Subject Headings "sexually transmitted diseases"); risk factors relating to the framework (including prevalence, incidence as well as correlates of unsafe behaviours); and Chinese. Both qualitative and quantitative systematic reviews were included. The detailed search strategy can be found in S1 Appendix.
Two independent reviewers screened titles and abstracts of the search results, removing duplicates or ineligible reviews before proceeding to full text screening of eligible papers. Consensus was reached either through discussion or, if necessary, through a third researcher's opinion. The raw data extraction sheet was compiled and modified independently to include basic information about the review based on the Social Epidemiology Framework, namely: the review period; database used; types of STIs; studied population; individual level risk factors (e.g. age), social level risk factors (e.g. networks) and structural factors (e.g. public policy); and limitations. The primary summary measures were the infection risk factors identified. The end point was HIV/STIs. Condom use was a proxy measure for avoiding HIV/STIs. Hepatitis C (HCV) was often included in the reviews as an STI-not only it was more prevalent in gay men with HIV but a lot of HCV patients were presented in the STI clinics. Additionally, HCV and HIV were often researched together. This study followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (the checklist can be found in S2 Appendix). The quality of the studies was assessed according to the validated scale-Assessment of Multiple Systematic Reviews (AMSTAR) ( Table 1) [20,21].

Results
The initial selection process identified 47 out of 425 search results. The initial step was to include both Chinese populations from within and outside of mainland China but very few articles on the latter emerged. A handful of reviews consisting of a mixture of Chinese and other nationalities were excluded due to the lack of clear distinction between risk factors for the two groups. The full-text screening further excluded 19 duplicated reviews leaving a total of 25 English and 3 Chinese systematic reviews for final inclusion (S1 Fig). The reviews reported in the Chinese literature focuse more on HIV and the related risk factors in key populations only (e.g. drug users and MSM) while those in the English literature tended to be more diverse.
Based on the PRISMA checklist, the majority of the reviews identified themselves as a systematic review and/or meta-analysis in the title, with the exception of three [38,40,45]. Most studies presented structured abstracts, clear objectives and rationales but lacked detailed protocol or registration. Inclusion criteria were reported in all but one study [47]. Twenty-five reviews included details of the study selection; three did not [36,38,47]. All reviews mentioned the information sources and provided a full electronic search strategy, although few contacted the study authors to identify additional studies. The majority of the reviews reported the data collection process and items, though more than half failed to include the risk of bias in individual studies. Twenty-six studies stated limitations of the review, but two reviews did not [30,36]. The median number of studies included in each review was 45 with an interquartile range of 40-92.

Individual factors
Individual factors include individual characteristics, socioeconomic position and behaviours.
Individual characteristics.
1. Younger age was a significant risk factor for STIs in four reviews [34,40,47,48], but limited evidence was found for HIV. Among migrant returnees, STIs were higher among 18-30 years (8.7%) than among 31-45 years (2.7%) and 46-60 years (1.5%) [47,50]. Being under 40 was associated with higher rates of HCV infection among entrants of methadone Table 1. Quality assessment of included studies using the Assessment of Multiple Systematic Reviews (AMSTAR).  2. The prevalence of HIV/STIs differs by gender. Three reviews indicated a higher prevalence of STIs for women than men [42,46,47]. One review found that male drug users are oneand-a-half times more likely to be HIV affected in high prevalent areas (defined as >10,000 HIV infected drug users, i.e. Yunnan, Guizhou, Sichuan, Guangxi and Xinjiang) [34]. However, in low prevalence areas, female drug users were twice as likely to be HIV infected, with the odds of male-to-female infection being 0.46 (95%CI 0.27-0.79) [34]. Migrants in cities had 6.70 times (95%CI 6.05-7.41) risk of contracting HIV compared to the general Chinese population, while female migrants were at even higher risk OR = 12.18 (95%CI 11.11-13.35) [42]. Compared to male migrants (4.2%, 95%CI 3.7-4.7%), females had higher STIs prevalence (14.1%, 95%CI 6.4-21.8%) [47,50]. Female drug users had twoto-tenfold higher prevalence of syphilis than male drug users [46]. Differences in gender were not significant among drug users for viral STIs such as HIV/HBV/HCV [34,36,37].
3. Weak associations were found between marriage and HIV/STIs risks among different groups. One study found that unmarried migrants were more likely to engage in commercial sex (OR = 1.  [23]. The number of male sex partners was also related to syphilis infection for MSM [24]. Among the general population, one review reported that the majority of STIs patients had multiple sexual partners within the previous year, with a mean of 6.2 and 7.2 for men and women respectively [47]. Engaging in sex with both genders was associated with 30% increase in HIV infection (OR = 1.30; 95%CI 1.04-1.62), but this correlation was not seen for syphilis [27].
2. Low rates of condom use were observed among general and key populations: only 10% of migrants reported consistent condom use, with nearly 40% having never used a condom [47]. Low rates of consistent condom use were observed among FSW with their stable partners (8-15%) and clients (13-54%) [38]. Similar findings were identified among drug users, with a reported rate of consistent condom use between 0-28%, and a range of 32-100% reported never having used a condom [47]. Low rates of consistent condom use were reported among MSM with regular, non-commercial/casual, and commercial partners with 23

Samples recruited from different networking sites showed different HIV infection rates among MSM
Neighbourhood effects/geographic/physical environment.

Structural factors
Structural factors include: demographic changes; war and militarization; structural violence and discrimination; legal structures, and policy environment. No papers examining the risk of war and militarization, structural violence and discrimination were identified in the selected reviews. Demographic change.
2. Urbanization in China appears to contribute to the spread of HIV/STIs. Pooled estimates of HIV prevalence among migrants returning from urban areas was 0.18% (95%CI 0.12-0.29%), with the odds of HIV infection 3.16 (95%CI 2.06-4.84) times higher than the overall figure [42].
Policy environment.  The lack of structural level studies, such as those on stigma and discrimination, health policy and access to care, and illicit drug control policies, could be explained through multiple reasons. Presently, there is a lack of a clear definition of structural factors. There are no guidelines on how to conduct structural level interventions and limited information on the effectiveness of such interventions [12,61]. However, even with limited evidence, it has been demonstrated that combining structural factors with an individual approach can significantly improve effectiveness of HIV prevention [62]. After the implementation of the National Free ART project in China, the pooled HIV heterosexual infection rate among sero-discordant couples dropped [49]. Internationally, the national 100% Condom Program, introduced in 1991, decreased STIs by 10 folds, and HIV incidence by 5 folds among the young Thai male conscripts from 1991-1993 to 1993-1995 [63].

Discussion
Our review identified that among the structural risks, the incarceration pattern reveals a significant risk of HIV/STIs for FSW. Internationally, it is found that compared with the general public, the prevalence of HIV/STIs in the prison population is two-to-ten times higher, due to prevalence of unprotected sexual intercourse and intravenous drug use [64]. The objective of the re-education of sex workers, according to the Chinese policy, is to guide FSW away from "bad habits" [60]. However, evidence suggests the re-education policies are exposing FSW to increased risk of HIV/STIs, thus, there is need to re-consider the current policies [65]. Further, modelling suggests that decriminalization of sex workers could avert up to 33%-46% of HIV infections [66]. In China, more than 40 scholars and lawyers advocated for abolishment of detention education to the National People's Congress and the Chinese Political Consultative Conference in 2014 [67].
Most key populations (e.g. PWID, LDTD and FSW) were included in 1995 in the national surveillance system in China. Yet MSM was only added in 2002 [68], despite MSM studies accounting for more than one third of all reviews explored in this study. This is due to rapid increase of HIV infection in MSM in recent years in China from 1.77% (95%CI 1.26-2.57) in 2000 to 5.98% (95%CI 4. 43-8.18) in 2010 [44]. Aligning with official reports, indicates that although HIV prevalence in key populations are stabilizing or decreasing, HIV infection among the estimated 3.1-6.3 millions of MSM have increased dramatically [4, 23-26, 28, 31, 69]. It has been identified that condom-less anal intercourse, multiple partners, migration and sex work were all found to pose significant risks for HIV infection among MSM. HIV prevention needs to take into consideration context-specific strategies. Engaging in sex with both male and female partners was associated with 30% increase in HIV infection, which presents a particular challenge in China given that at least one-third of Chinese MSM have wives or female partners to fulfil social expectations and pressures to marry and have children [27,70]. In addition, a recent review found that although behavioural interventions alone can increase consistent condom use, it had little impact on HIV or syphilis infection [71], suggesting a need to incorporate behavioural, biomedical, social and structural dimensions [62].
This multi-level approach to HIV/STIs prevention and treatment is becoming increasingly important internationally [62]. The World Health Organization (WHO) recognizes the concurrent nature of vulnerabilities for key populations, thus recommends integrated service be provided [72]. For example, intravenous drug use is a significant individual risk factor for drug-using FSW to HIV/STIs infection [4,34,37,40,53,55,58]. A more holistic approach for an intervention would add opiate substitution treatment and needle and syringe programmes, to the traditional behavioural approaches [73]. WHO calls for screening, diagnosis and treatment of STIs to be offered routinely to key populations, "as part of comprehensive HIV prevention and care" [72]. Along with other international studies, the finding that co-infection is a risk factor for contracting HIV/STIs reinforces WHO guidelines suggests a need to scale up provider-initiated and voluntary testing services as well as counselling in STIs clinics in China [74,75].
There are several limitations to this review. First, as most reviews contained data from cross-sectional studies, risk factors cannot infer causality. The pooled data nonetheless reinforces the associations of the risk factors and infections as the data is gathered systematically from all identified reviews. Secondly, as the review covers diverse sampling methods, population groups and STIs over a large time span, the overall picture might not reflect the real situation at a particular point in time. However, the review by Cai et al. concluded that no significant time trends were identified between 1980 and 2012, studying high-risk sexual behaviours through 174 observational studies [76]. Thirdly, most studies included were conducted in urban cities and might not represent all of China. Given the vast number of floating populations in China and the limited health facilities in rural areas, the rural situation of HIV/ STIs may be equivalent if not worse than urban cities, thus demanding further investigation.

Conclusion
It is a critical moment for China in terms of HIV/STIs control and prevention. A comprehensive picture of the risk factors, as presented in this review can help in effective planning of HIV/STIs prevention strategies. However, reviews on HIV/STIs risks among the Chinese populations are limited to individual factors. Though important, social and structural risk factors are desperately lacking, as highlighted by international guidelines and research for HIV/STIs prevention. A comprehensive approach targeting interventions at all levels, along the continuum of care is needed to effectively curtail HIV/STIs transmission in China. Our study recommends that more research is needed on the impact of socio-political interventions within a Chinese context.