A Predictive Model of HIV Infection Among men who have sex with men (MSM): A Cross-Sectional Study in 6 Cities of China

doi:10.21203/rs.3.rs-4982945/v1

Background

Compared to the overall decrease in infection rates, the prevalence of HIV among key populations like men who have sex with men (MSM) remains high. Understanding the factors influencing HIV in these key populations is crucial. However, currently there are few large-scale studies that can comprehensively incorporate multiple potential influencing factors. Therefore, it is significant to build a predictive model for HIV infection among MSM based on demographic, behavioral, social and psychological characteristics.

Methods

A cross-sectional study of MSM was conducted in six major regions of China. Data were collected through face-to-face interviews. We screen variables through univariate and multivariate analyses, and construct a predictive model through stepwise regression. We evaluated the model through nomograms, ROC(Receiver Operating Characteristic) curves and decision curve analysis (DCA).

Results

Among 2,403 MSM, 199 were HIV positive, resulting in a prevalence rate of 8.3%. MSM aged 35–44, with lower education level, higher smoking frequency, having STDs, assuming the receptive role in same-sex intercourse, using drugs before sex and experiencing depression are more likely to be infected with HIV. The Hosmer-Lemeshow test showed a p-value of 0.3834 for the model. The AUC of the nomogram was 0.783 (95% CI: 0.749–0.816).

Conclusions

HIV infection among the MSM population arises from a combination of demographic, behavioral, social and psychological characteristics. The development of the predictive model aids in formulating targeted and comprehensive intervention strategies. We need to strengthen our focus on HIV infection in MSM and take targeted measures based on the results of predictive model to curb the spread of HIV.

HIV

Men who have sex with men (MSM)

Influencing factor

Predictive model

Nomogram

In recent years, significant progress has been made in the global fight against HIV. According to data released by the Joint United Nations Programme on HIV/AIDS (UNAIDS) in 2023, the number of annual new HIV infections in Asia and the Pacific has steadily declined from 2000 to 2022, with a 14% reduction in new HIV infections since 2010(1). Despite this overall improvement, the median prevalence of HIV in key populations such as men who have sex with men (MSM) remains high. From 2018 to 2022, the median HIV prevalence among MSM in Asia and the Pacific was 4.7%, far exceeding the 0.2% prevalence among adults aged 15–49(1). Given their high risk of infection, MSM has become a key population in the study of HIV epidemiology and prevention strategies.

According to UNAIDS data 2023, in China, MSM have gradually become the highest-risk group for HIV infection(1). Meta-analyses have indicated that the overall HIV prevalence of MSM in China ranges from 2.3–7.4%(2–4). These figures align with the 4.5% HIV prevalence of MSM population in China published by UNAIDS in 2023, suggesting that the HIV epidemic among MSM in China has not been effectively controlled.

Understanding factors related to HIV is crucial for emphasizing vulnerable populations. In several studies and reviews, we have found that a variety of factors related to HIV infection can mainly be categorized into demographic, behavioral, social and psychological characteristics(5–15).

In terms of demographic characteristics, prior researches have reported the effects such as age, education level, marriage, occupation and socioeconomic status. Regarding behavioral characteristics, research has shown that multiple sexual partners(MSP) and unprotected anal intercourse(UAI) among MSM significantly increase the HIV prevalence(8–11). Additionally, research evidence suggests that behaviors such as pre-sex alcohol consumption and pre-sex drug abuse may be associated with a higher risk of HIV infection(12–14). Among social characteristics, the widespread stigma and discrimination against MSM have profound impacts on the psychological health and behavioral choices of MSM(15). Correspondingly, social support plays a crucial positive role in the prevention and treatment of HIV among MSM.

At present, due to the neglect of psychological characteristics, there is limited discussion about the association between these characteristics and HIV infection. As a result, many health intervention measures fail to comprehensively and effectively address the health needs of MSM, limiting our full understanding of this health issue. Therefore, we aim to build a predictive model that integrates psychological characteristics. This approach will allow us to more effectively explore HIV vulnerabilities among MSM.

Study design and sampling

This study is a large-scale national cross-sectional study conducted from November 2023 to March 2024. Survey sites are strategically chosen from six major geographical locations across the country, specifically Urumqi in the Northwest, Shenyang in the Northeast, Xi'an in the Central-Western location, Kunming in the Southwest, Guangzhou in the Southeast and Shanghai in the Eastern coastal area. MSM populations tend to conceal themselves. Therefore, with the assistance of local Centers for Disease Control and Prevention and non-governmental organizations (NGOs), a "snowball" sampling method was implemented at each survey site. Initially, 5–10 MSM "seeds" with strong social influence are recruited. These "seeds" subsequently recruited or recommended other eligible participants.

Inclusion and exclusion criteria

The inclusion criteria are as follows: (1)Male and at least 18 years old. (2)Engaged in sexual activity with men in the past six months. (3)Resident of one of the six selected survey sites (primarily residing in the area for the past six months).

The exclusion criteria are as follows: (1)Took less than 300 seconds to complete the questionnaire. (2)Provided incorrect responses to quality control questions. (3)IP address indicates a location outside of the surveyed provinces or cities.

Data collection

All survey personnel received standardized training on survey protocols. Trained surveyors assisted participants in completing the online survey via the Wenjuanxing platform at privacy-protected NGO sites, ensuring data quality and completeness. All participants voluntarily took part in the study. The entire survey took approximately 30 minutes, and each participant received an 80 yuan financial compensation to encourage cooperation.

Variables and measures

The demographic characteristics collected in this study primarily include age, education, marriage, occupation, income, residence and sexual orientation. The behavioral characteristics mainly include smoking, drinking and sexual behavior-related factors. The sexual behavior-related factors mainly include MSP, UAI, sexual frequency, types of sexual behavior, group sex, presexual drinking, presexual drug use, self-reported past sexually transmitted diseases (STD) and recent voluntary counseling and testing for HIV (VCT). In this study, we collected various continuous variables and converted them into categorical variables using the median for binary classification or quantiles for multi-class classification. This approach enhanced the explainability of the research and made the data visualization more intuitive.

Sociological characteristics primarily include social support and social discrimination. Social support refers to the various forms of help and resources that individuals receive from their social networks. It is measured using 12 questions from the Social Support Scale, scored on a Likert 7-point scale, with responses ranging from 1 (strongly disagree) to 7 (strongly agree). Social discrimination refers to physical and non-physical violence experienced due to perceived or known sexual identity(16). Social discrimination is measured using 5 questions from the Discrimination Scale, where a response of "yes" to any of the questions indicates that the individual has experienced social discrimination(16).

Psychological characteristics primarily include self-esteem, depression, anxiety, personality and barriers to healthcare. Self-esteem is measured using 10 questions from the Rosenberg Self-Esteem Scale, scored on a Likert 4-point scale, with answers ranging from 1 (strongly disagree) to 4 (strongly agree)(17). Items 3, 5, 8, 9, and 10 are reverse-scored. Depression is measured using 9 questions from the PHQ-9 Depression Scale, scored on a Likert 4-point scale, with answers ranging from 0 (not at all) to 3 (nearly every day). Anxiety is measured using 7 questions from the GAD-7 Anxiety Scale, scored on a Likert 4-point scale, with answers ranging from 0 (not at all) to 3 (nearly every day). Personality is scored using 4 questions from the brief version of the Chinese-language Big Five Inventory (BFI) personality scale, scored on a Likert 5-point scale, with answers ranging from 1 (strongly disagree) to 5 (strongly agree)(18). Scores of 4–12 indicate a neurotic personality, while scores of 12–20 indicate an agreeable personality. Barriers to healthcare are mainly measured using 12 healthcare-related questions such as "I worry about being discriminated against when seeking medical care", "I worry that my sexual orientation will be exposed" etc(19–20), scored on a Likert 5-point scale, with answers ranging from 1 (strongly disagree) to 5 (strongly agree).

The primary outcome of this study is the status of HIV infection, determined by the self-reported results of the most recent HIV test, coded as a binary variable.

Statistical analysis

The raw data are cleaned and then entered into a database for data analysis. Firstly, descriptive statistics are performed on demographic, behavioral, sociological and psychological characteristics. Secondly, univariate analysis was performed using the chi-square test for statistical inference. All tests are two-sided, and p-values < 0.05 are considered significant. Thirdly, we performed multivariate analysis using the forward LR method for further selection of variables. Finally, we utilized stepwise logistic regression model to explore the relationships between these characteristics and HIV infection. And we assessed the results through nomogram, calibration curve, ROC curve and Decision Curve Analysis (DCA).

Descriptive Analysis and Univariate Analysis

A total of 2,403 people completed the research survey. Among the 2,403 subjects, 199 individuals (8.3%) are HIV positive, and 2,204 individuals (91.7%) are HIV negative. In populations with different HIV infection statuses, there are different distributions of some demographic, behavioral, sociological and psychological characteristics. The results of descriptive analysis and univariate analysis are shown in Table 1.

The chi-square test results showed that some demographic characteristics, including age, education, marriage and occupation, were associated with HIV infection status. Regarding age, there were differences in HIV prevalence among different age groups (p=0.008), with the highest prevalence of 11.5% in MSM aged 35-44 years. For education, as the educational level increased, the risk of HIV infection decreased (p<0.001), with a prevalence of 15.9% in junior high school or below and 7.2% in college or above. In terms of marital status, the HIV prevalence was significantly higher (15.0%) among the divorced or widowed group compared to other categories (p=0.027). For occupation, physical laborers and freelancers had higher HIV prevalence, respectively 8.5% and 10.1%(p=0.015).

Regarding behavioral characteristics, MSM with a history of STDs had a prevalence of 30.2% (p<0.001), and those who used drugs before sex had a prevalence of 27.5% (p<0.001). Additionally, MSM with everyday smoking (13.9%, p < 0.001), MSP (10.1%, p=0.001), UAI (11.1%, p < 0.001), group sex (10.2%, p=0.032), higher frequency of sex (14.3%, p=0.009), assuming the receptive role in same-sex intercourse (11.6%, p<0.001) and drinking before sex (9.7%, p=0.046) had higher HIV prevalence.

For psychological characteristics, low self-esteem, depression and anxiety were identified as risk factors for HIV infection (p<0.05). The HIV prevalence was 12.7% in MSM with low self-esteem. Compared to other categories, moderate and severe levels of anxiety and depression had significantly higher HIV prevalence.

Multivariate Analysis

We selected 18 factors associated with HIV infection through univariate analysis. Using the forward LR method for multivariate analysis, we refined these variables to eight key factors: age, education, marriage, smoking, sexual pattern, STD, using drugs before sex and depression. Ultimately, we included these variables in the construction of the HIV infection predictive model.

Predictive Model

Based on the preliminary research, we found that using stepwise regression to screen predictive factors can yield the optimal predictive model, with good predictive accuracy and clinical application value(21). Therefore, in this study, we similarly adopted this method for model construction. We employed the full-entry stepwise regression method to include eight variables in the model: age, education, marriage, smoking, sexual pattern, STD, using drugs before sex and depression. The results of the model are shown in Table 2.

Regarding age, MSM aged 35-44 have a higher risk of HIV infection compared to other age groups (OR=1.842, 95% CI 1.147-2.958, p=0.012). Compared to MSM with an education level of junior high school and below, higher educational levels may be associated with a reduced risk of HIV infection (College and above: OR=0.386, 95% CI 0.229-0.650, p<0.001). Sometimes smoking (OR=1.799, 95% CI 1.193-2.713, p=0.005) and everyday smoking (OR=2.225, 95% CI 1.473-3.363, p<0.001) may be related to HIV infection. In terms of sexual patterns, compared to those who only engage as receptive partners, being only the insertive partners (OR=0.364, 95% CI 0.245-0.543, p<0.001) and engaging in both roles (OR=0.613, 95% CI 0.424-0.885, p=0.009) may be associated with a lower risk of HIV infection. Regarding depression, moderate (OR=1.748, 95% CI 1.067-2.864, p=0.027) and moderately severe (OR=2.398, 95% CI 1.414-4.068, p=0.001) levels of depression may be related to HIV infection. In addition, having an STD (OR=5.672, 95% CI 3.868-8.319, p<0.001) and using drugs before sex (OR=3.017, 95% CI 1.575-5.777, p=0.001) may be associated with an increased risk of HIV infection. Meanwhile, we also visualized the OR values of each variable to clarify their impact on HIV infection.

The results of the Hosmer-Lemeshow test indicated that the p-value for the model is 0.3834. Since the p-value is greater than 0.05, it suggests that the model fits the observed data well. We constructed a nomogram based on the model, incorporating the following 10 variables: age, education, marriage, occupation, smoking, STD, MSP, sexual patterns, using drugs before sex and depression. The nomogram is illustrated in Figure 1. The Area Under the Curve (AUC) of the nomogram is 0.783 (95% CI: 0.749-0.816), indicating that the model has good predictive performance, as shown in the Receiver Operating Characteristic (ROC) curve in Figure 2. The nomogram's calibration curve demonstrates the consistency between predicted results and observed outcomes. The results show that the curve aligns well with the diagonal line in the low probability region, with some deviation in the high probability region. Considering that the incidence of HIV infection in the study population typically ranges from 5% to 10%, the model exhibits good consistency. The calibration curve is depicted in Figure 3. Additionally, we evaluated the clinical utility of the predictive model through Decision Curve Analysis (DCA). The DCA results indicate that for HIV infection in the MSM population, the net benefit of using the predictive model is higher than the strategies of treating all or treating none, demonstrating the model's clinical value. The DCA curve is shown in Figure 4. In summary, based on the evaluation results of the model, we can conclude that this HIV predictive model fits well.

This study utilized a questionnaire survey to collect demographic, behavioral, sociological and psychological information from 2,403 MSM. Among these, 199 individuals were HIV positive, resulting in a prevalence of 8.3%. This rate is higher than the 4.5% HIV prevalence rate reported by UNAIDS for the MSM population in China in 2023(1). The higher prevalence in this study could be attributed to the fact that the survey participants were all from major cities, where the overall HIV prevalence tends to be higher. The findings from this study confirm that the MSM population has a higher HIV prevalence.

Demographic characteristics

Age is a significant risk factor for HIV infection among MSM. As age increases, so does the risk of HIV infection, which aligns with existing research findings(22–23). This could be because older MSM have engaged in high-risk behaviors for a longer period, accumulating more exposure opportunities and thus increasing their risk of HIV infection. In our study, with the increase of age, the incidence of STDs, MSP and the frequency of sexual activities all increased (p < 0.05). These associations support the possibility of the above speculation.

We found that educational attainment is a critical factor influencing HIV infection among MSM. Multiple studies have confirmed that MSM with higher educational levels are more likely to avoid high-risk behaviors (24–25). In this study, lower educational attainment was associated with more frequent smoking behaviors and higher frequencies of sexual activities (p < 0.05). In terms of psychological factors, individuals with lower educational levels exhibited lower social support and higher depression scores (p < 0.05), which may also be related to HIV infection. Individuals with higher educational attainment are often more exposed to reliable information sources, so they tend to have a higher perception of HIV risks and are more likely to take proactive measures to prevent infection.

Behavioral characteristics

The nomogram results indicate that among behavioral characteristics, smoking, STD, being the receptive partner in same-sex behavior and using drugs before sex are all risk factors for HIV infection in the MSM population. Factors such as STDs and using drugs before sex have been reported in multiple studies (8–14). In this study, 9.2% of MSM have STDs and 2.9% use drugs before sex. The OR values for these two characteristics were 5.672 and 3.017, indicating a strong impact of having STDs and using drugs on HIV infection. This result consistent with previous research findings.

Currently, research on the association between smoking and being the receptive partner in same-sex behavior with HIV infection is limited. Our study preliminarily explores the potential risk factors of these behaviors in HIV transmission. Whether smoking increases the risk of HIV infection among MSM has not been fully studied. Its possible mechanisms include immune system suppression and mucosal damage. Being the receptive partner in same-sex behavior is considered a high-risk factor for HIV infection due to the susceptibility to mucosal damage and larger exposure area, increasing the risk of HIV transmission.

Psychological Characteristics

In terms of psychological Characteristics, depression is a significant risk factor for HIV infection in MSM and may influence their behavior and health outcomes through various pathways(26–28). In this study, depression was significantly associated with social discrimination, social support, self-esteem and anxiety (p < 0.01). On the other hand, there was a positive correlation between depression and UAI, group sex and using drugs before sex among MSM (p < 0.05). This result suggests that depression may play a mediating role in the relationship between behavioral, social, psychological characteristics and HIV infection, which requires further research for validation. Currently, research on psychological factors primarily focuses on the stigmatization and discrimination issues faced by MSM, neglecting the impact of depression itself on behavior and HIV infection(29). This study aims to emphasize the influence of depression on HIV infection, enhance mental health support for depressed individuals within the MSM community, and ultimately improve related high-risk behaviors to reduce the prevalence of HIV.

Stratified Analysis

Among various characteristics, demographic characteristics are difficult to change. Therefore, we adopt a stratified prevention strategy for the demographic characteristics of age and education level included in the model.

Our stratified analysis by age revealed that factors such as smoking, using drugs before sex, same-sex orientation, and having an STD have a stronger impact on HIV infection among younger MSM. In contrast, among older MSM, VCT for HIV, depression, personality, social support and social discrimination have a stronger impact on HIV infection. These findings suggest that interventions for younger MSM should focus on related education and behavior change. For older MSM, interventions should enhance healthcare accessibility, provide comprehensive health services and strengthen psychological health and social support.

In the stratified analysis of educational attainment, we found that the behavioral differences are not significant and the main distinctions were observed in sociological and psychological characteristics. In MSM with junior high school education and below, there is no correlation between HIV infection and sociological or psychological characteristics. However, in MSM with senior high school education, depression is identified as a risk factor for HIV infection (p = 0.008). In MSM with a college degree and above, social discrimination, depression, and anxiety are all associated with HIV infection (p < 0.05). This result suggests that individuals with higher educational attainment are more susceptible to social discrimination and stigma. Therefore, we believe that for MSM with lower education levels, prevention efforts should focus on improving basic health knowledge and providing economic support. For those with higher education levels, intervention measures should focus on eliminating discrimination and improving mental health.

Strengths and Limitations

The strength of this study lies in collecting comprehensive information on the MSM population and constructing a predictive model that includes demographic, behavioral, and psychological characteristics, exploring the relationship between HIV infection and these variables from different perspectives. We collected basic information from 2,403 MSM, with data sourced from six major regions in China, providing a large sample size with a certain level of representativeness. Additionally, our use of the "snowball" sampling method enabled a hidden and marginalized group like MSM to express themselves more freely and quickly expand the sample size through initial respondents. However, the non-random sampling method may result in a lack of representativeness. Furthermore, we need to supplement our cross-sectional study with longitudinal research to explore the relationship between changes in characteristics and HIV infection.

We summarized the factors influencing HIV infection among key populations like MSM, highlighting the combined effects of demographic, behavioral, social and psychological characteristics. The results from the model and nomogram indicate that aged 35–44, lower education level, higher smoking frequency, having STDs, assuming the receptive role in same-sex intercourse, using drugs before sex and depression make MSM more susceptible to HIV infection. Our study validated several inherent demographic and behavioral characteristics associated with HIV infection. Additionally, we identified a link between smoking and assuming the receptive role in same-sex behavior, as well as the relationship between depression and HIV infection. While sociological factors like social support and discrimination do not have a direct impact on HIV infection, they may indirectly affect the MSM population through their influence on psychological characteristics. We need to strengthen our focus on HIV infection in MSM and take targeted measures based on the results of predictive model to curb the spread of HIV.

AIDS: Acquired immunodeficiency syndrome

HIV: Human Immunodeficiency virus

MSM: men who have sex with men

MSP: multiple sex partnership

NGO: non-governmental organization

BFI: Big Five Inventory

ROC: Receiver Operating Characteristic

AUC: Area Under the Curve

DCA: Decision Curve Analysis

UNAIDS: the Joint United Nations Programme on HIV/AIDS

STD: sexually transmitted disease

UAI: unprotected anal intercourse

VCT: Voluntary Counseling and Testing, for HIV

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement

The recruitment process and study protocol are approved by the Ethics Committee of Shanghai University of Medicine & Health Sciences (2023-MSMMPOX-22-310222197604080237).

Organizations involved in data collection

The authors would like to thank Shanghai Qingai Health Center (Shanghai), Consultation Center of AIDS Aid and Health Service (Liaoning), Trans China (Yunnan), Xinjiang Rong'ai Health Consulting Co., Ltd. (Xinjiang), Xi'an Zhitong AIDS Prevention and Health Center (Shaanxi), and Lingnan Partners Community Support Center (Guangdong) for their contributions to data collection.

Author contributions

MS and YiW: designed the study. FH: provided support for statistical methods.

GX and JZ: collected original data. MZ and YZ: assisted with data entry and organization. CL: conducted statistical analysis of the data. CL and YW: wrote the initial draft of the manuscript. FH, MS, and YiW: revised the initial draft of the manuscript. All authors contributed to the article and approved the submitted version.

Funding

Key discipline projects of Shanghai Three-Year Action Plan for Public Health under Grant (GWVI-11.1-29). national health commission of the PRC (Grant No. Z155080000004).Tongren Hospital Talent Launch Project (TR2023rc11, TR2024RC10). Changning District Medical and Health Research Special Project (CNKW2022Y08). The Research Fund of Shanghai Tongren Hospital, ShanghaiJiaotong University School of Medicine (TRYJ2024LC13,TRYJ2022LC10).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Joint United Nations Programme on HIV/AIDS.UNAIDS data 2023. Accessed Oct 31, 2023. https://www.unaids.org
Meng X, Zou H, Beck J, Xu Y, Zhang X, Miao X, Xu F. Trends in HIV prevalence among men who have sex with men in China 2003-09: a systematic review and meta-analysis. Sex Health. 2013 Jul;10(3):211-9.
Zhou Y, Li D, Lu D, Ruan Y, Qi X, Gao G. Prevalence of HIV and syphilis infection among men who have sex with men in China: a meta-analysis. Biomed Res Int. 2014;2014:620431.
Dong MJ, Peng B, Liu ZF, Ye QN, Liu H, Lu XL, Zhang B, Chen JJ. The prevalence of HIV among MSM in China: a large-scale systematic analysis. BMC Infect Dis. 2019 Nov 27;19(1):1000.
Jeffries WL 4th, Greene KM, Paz-Bailey G, McCree DH, Scales L, Dunville R, Whitmore S. Determinants of HIV Incidence Disparities Among Young and Older Men Who Have Sex with Men in the United States. AIDS Behav. 2018 Jul;22(7):2199-2213.
Li QH, Wang JY, Liu SY, Zhang YQ, Li EL, Wang YR, Zhang SL, Zhao WB, Liu SL, Chen XH, Wang FX. Young MSM changed temporal HIV-1 epidemic pattern in Heilongjiang Province, China. Front Microbiol. 2022 Nov 25;13:1028383.
Torres TS, Coelho LE, Konda KA, Vega-Ramirez EH, Elorreaga OA, Diaz-Sosa D, Hoagland B, Pimenta C, Benedetti M, Grinsztejn B, Caceres CF, Veloso VG. Low socioeconomic status is associated with self-reported HIV positive status among young MSM in Brazil and Peru. BMC Infect Dis. 2021 Jul 31;21(1):726.
Guimarães MDC, Kendall C, Magno L, Rocha GM, Knauth DR, Leal AF, Dourado I, Veras MA, Brito AM, Kerr LRFS; Brazilian HIV/MSM Surveillance Group. Comparing HIV risk-related behaviors between 2 RDS national samples of MSM in Brazil, 2009 and 2016. Medicine (Baltimore). 2018 May;97(1S Suppl 1):S62-S68.
Shuper PA, MacLachlan DJ, Joharchi N, Guimond TH, Maxwell J, Adam BD. HIV Risk and Protective Factors in the Context of Alcohol and Substance Use During Pride. AIDS Behav. 2018 Sep;22(9):2797-2806.
Yin L, Zhao Y, Peratikos MB, Song L, Zhang X, Xin R, Sun Z, Xu Y, Zhang L, Hu Y, Hao C, Ruan Y, Shao Y, Vermund SH, Qian HZ. Risk Prediction Score for HIV Infection: Development and Internal Validation with Cross-Sectional Data from Men Who Have Sex with Men in China. AIDS Behav. 2018 Jul;22(7):2267-2276.
Chittamuru D, Icard LD, Jemmott JB 3rd, O'Leary A. Prospective Predictors of Multiple Sexual Partners Among African American Men Who Have Sex with Men. Arch Sex Behav. 2018 Oct;47(7):2081-2090.
Rehm J, Probst C, Shield KD, Shuper PA. Does alcohol use have a causal effect on HIV incidence and disease progression? A review of the literature and a modeling strategy for quantifying the effect. Popul Health Metr. 2017 Feb 10;15(1):4.
Rahimi J, Gholami J, Amin-Esmaeili M, Fotouhi A, Rafiemanesh H, Shadloo B, Rahimi-Movaghar A. HIV prevalence among people who inject drugs (PWID) and related factors in Iran: a systematic review, meta-analysis and trend analysis. Addiction. 2020 Apr;115(4):605-622.
Hsiang E, Gyamerah A, Baguso G, Jain J, McFarland W, Wilson EC, Santos GM. Prevalence and correlates of substance use and associations with HIV-related outcomes among trans women in the San Francisco Bay Area. BMC Infect Dis. 2022 Nov 26;22(1):886.
Sun S, Pachankis JE, Li X, Operario D. Addressing Minority Stress and Mental Health among Men Who Have Sex with Men (MSM) in China. Curr HIV/AIDS Rep. 2020 Feb;17(1):35-62.
Denney MR, Pichon LC, Brantley ML. Violence, Discrimination, Psychological Distress, and HIV Vulnerability Among Men Who Have Sex With Men in Memphis, Tennessee. Am J Mens Health. 2023 Mar-Apr;17(2):15579883231163727.
Rosenberg M. Society and the Adolescent Self-Image.
Carciofo R, Yang J, Song N, Du F, Zhang K. Psychometric Evaluation of Chinese-Language 44-Item and 10-Item Big Five Personality Inventories, Including Correlations with Chronotype, Mindfulness and Mind Wandering. PLoS One. 2016 Feb 26;11(2):e0149963.
Wagner AC, Hart TA, McShane KE, Margolese S, Girard TA. Health care provider attitudes and beliefs about people living with HIV: Initial validation of the Health Care Provider HIV/AIDS Stigma Scale (HPASS). AIDS Behav. 2014 Dec;18(12):2397-408.
Kupferschmidt K. Why monkeypox is mostly hitting men who have sex with men. Science. 2022 Jun 24;376(6600):1364-1365.
Liu S, Xia D, Wang Y, Xu H, Xu L, Yuan D, Liang A, Chang R, Wang R, Liu Y, Chen H, Hu F, Cai Y, Wang Y. Predicting the risk of HIV infection among internal migrant MSM in China: An optimal model based on three variable selection methods. Front Public Health. 2022 Oct 25;10:1015699.
Qiao YC, Xu Y, Jiang DX, Wang X, Wang F, Yang J, Wei YS. Epidemiological analyses of regional and age differences of HIV/AIDS prevalence in China, 2004-2016. Int J Infect Dis. 2019 Apr;81:215-220.
Cao Z, Xu J, Zhang H, Song D, She M, Wang J, Wang J, Fan Y, Sun Y, Zhang C. Risk factors for syphilis among married men who have sex with men in China. Sex Transm Dis. 2014 Feb;41(2):98-102.
Jiao K, Wei R, Li H, Chow EPF, Piqueiras E, Lewis T, Xu Z, Ren C, Ma W. HIV testing frequency and associated factors among five key populations in ten cities of China: a cross-sectional study. BMC Infect Dis. 2022 Feb 28;22(1):195.
Johansson K, Persson KI, Deogan C, El-Khatib Z. Factors associated with condom use and HIV testing among young men who have sex with men: a cross-sectional survey in a random online sample in Sweden. Sex Transm Infect. 2018 Sep;94(6):427-433.
Liu Z, Chen Y, Yao T, Zhang T, Song D, Liu Y, Yu M, Xu J, Li Z, Yang J, Cui Z, Li C, Ma J. Factors related to HIV testing frequency in MSM based on the 2011-2018 survey in Tianjin, China: a hint for risk reduction strategy. BMC Public Health. 2021 Oct 20;21(1):1900.
Yang Z, Li S, Zhang R, Na J, Li Y, Mu H, Yu L, Liu L, Sun W, Pan G, Yan L. Current heterosexual marriage is associated with significantly decreased levels of anxiety symptoms among Chinese men who have sex with men. BMC Psychiatry. 2020 Apr 6;20(1):151.
Ninnoni JP, Nsatimba F, Agyemang SO, Commey IT, Bennin L, Agyare E, Gyimah L, Senya K, Baddoo NA, Obiri-Yeboah D. An exploratory qualitative study of the psychological effects of HIV diagnosis; the need for early involvement of mental health professionals to improve linkage to care. BMC Public Health. 2023 Dec 15;23(1):2518.
Krueger EA, Holloway IW, Lightfoot M, Lin A, Hammack PL, Meyer IH. Psychological Distress, Felt Stigma, and HIV Prevention in a National Probability Sample of Sexual Minority Men. LGBT Health. 2020 May/Jun;7(4):190-197.

TABLE 1 Characteristics of the MSMs.

Characteristics		Total （n=2403）	HIV Status		Univariate analysis
Characteristics		Total （n=2403）	positive（n=199）	negative（n=2204）	X2	p value
Age	18-24	564	52（9.2%）	512（90.8%）	11.704	0.008
	25-34	1228	81（6.6%）	1147（93.4%）
	35-44	459	53（11.5%）	406（88.5%）
	≥45	152	13（8.6%）	139（91.4%）
Education	Junior high school and below	157	25（15.9%）	132（84.1%）	17.526	＜0.001
	High school	358	38（10.6%）	320（89.4%）
	College degree and above	1888	136（7.2%）	1752（92.8%）
Marriage	Unmarried	2035	163（8.0%）	1872（92.0%）	7.233	0.027
	Married	255	19（7.5%）	236（92.5%）
	Divorced or widowed	113	17（15.0%）	96（85.0%）
Occupation	Mental labor	386	23（6.0%）	363（94.0%）	10.43	0.015
	Physical labor	774	66（8.5%）	708（91.5%）
	Student	320	17（5.3%）	303（94.7%）
	Freelance	923	93（10.1%）	830（89.9%）
Monthly Income（￥）	<3000	502	40（8.0%）	462（92.0%）	1.438	0.697
	3001-6000	851	78（9.2%）	773（90.8%）
	6001-12000	802	61（7.6%）	741（92.4%）
	>12001	248	20（8.1%）	228（91.9%）
Residence	Local	936	75（8.0%）	861（92.0%）	2.856	0.24
	Short term<5year	870	65（7.5%）	805（92.5%）
	Long term>5year	597	59（9.9%）	538（90.1%）
Sexual orientation	Heterosexual orientation	516	36（7.0%）	480（93.0%）	1.472	0.225
Sexual orientation	Homosexual orientation	1887	163（8.6%）	1724（91.4%）	1.472	0.225
Smoking	Never	1464	90（6.1%）	1374（93.9%）	30.844	＜0.001
	Sometimes	392	48（12.2%）	344（87.8%）
	Always	224	16（7.1%）	208（92.9%）
	Everyday	323	45（13.9%）	278（86.1%）
Drinking	Never	1122	80（7.1%）	1042（92.9%）	6.474	0.091
	Sometimes	1067	94（8.8%）	973（91.2%）
	Always	197	22（11.2%）	175（88.8%）
	Everyday	17	3（17.6%）	14（82.4%）
VCT	No	457	39（8.5%）	418（91.5%）	0.047	0.828
VCT	Yes	1946	160（8.2%）	1786（91.8%）	0.047	0.828
STD	No	2182	135（6.2%）	2047（93.8%）	137.009	＜0.001
STD	Yes	221	64（30.0%）	157（70.0%）	137.009	＜0.001
MSP	No	1206	78（6.5%）	1128（93.5%）	10.485	0.001
MSP	Yes	1197	121（10.1%）	1076（89.9%）	10.485	0.001
UAI	No	1395	87（6.2%）	1308（93.8%）	18.306	＜0.001
UAI	Yes	1008	112（11.1%）	896（88.9%）	18.306	＜0.001
Sexual frequency	Four or more times per week	91	13（14.3%）	78（85.7%）	11.691	0.009
	Once to three times a week	419	47（11.2%）	372（88.8%）
	Once to three times a month	982	76（7.7%）	906（92.3%）
	Less than once a month	911	63（6.9%）	848（93.1%）
Sexual pattern	0	704	82（11.6%）	622（88.4%）	22.283	＜0.001
	1	921	48（5.2%）	873（94.8%）
	all	778	69（8.9%）	709（91.1%）
Group sex	No	1727	130（7.5%）	1597（92.5%）	4.593	0.032
Group sex	Yes	676	69（10.2%）	607（89.8%）	4.593	0.032
Drinking before sex	No	1510	112（7.4%）	1398（92.6%）	3.994	0.046
Drinking before sex	Yes	893	87（9.7%）	806（90.3%）	3.994	0.046
Using drugs before sex	No	2334	180（7.7%）	2154（92.3%）	34.676	＜0.001
Using drugs before sex	Yes	69	19（27.5%）	50（72.5%）	34.676	＜0.001
Self-esteem	Low	511	65（12.7%）	446（87.3%）	16.836	＜0.001
Self-esteem	High	1892	134（7.1%）	1758（92.9%）	16.836	＜0.001
Social support	Low	190	27（14.2%）	163（85.8%）	10.042	0.007
	Medium	1165	86（7.4%）	1079（92.6%）
	High	1048	86（8.2%）	962（91.8%）
Discrimination	No	1070	68（6.4%）	1002（93.6%）	9.422	0.002
Discrimination	Yes	1333	131（9.8%）	1202（90.2%）	9.422	0.002
Personality	Neuroticism	1024	90（8.8%）	934（91.2%）	0.606	0.436
Personality	Agreeableness	1379	109（7.9%）	1270（92.1%）	0.606	0.436
Barriers to healthcare	Low	662	49（7.4%）	613（92.6%）	0.931	0.335
Barriers to healthcare	High	1741	150（8.6%）	1591（91.4%）	0.931	0.335
GAD-7	No	1393	91（6.5%）	1302（93.5%）	26.465	＜0.001
	Mild	576	48（8.3%）	528（91.7%）
	Moderate	290	45（15.5%）	245（84.5%）
	Severe	144	15（10.4%）	129（89.6%）
PHQ-9	No	1197	69（5.8%）	1128（94.2%）	36.023	＜0.001
	Mild	665	57（8.6%）	608（91.4%）
	Moderate	260	30（11.5%）	230（88.5%）
	Moderately severe	166	30（18.1%）	136（81.9%）
	Severe	115	13（11.3%）	102（88.7%）

TABLE2 Stepwise regression model

No competing interests reported.

A Predictive Model of HIV Infection Among men who have sex with men (MSM): A Cross-Sectional Study in 6 Cities of China

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Background

METHODS

Study design and sampling

Inclusion and exclusion criteria

Data collection

Variables and measures

Statistical analysis

Results

Discussion

Demographic characteristics

Behavioral characteristics

Psychological Characteristics

Stratified Analysis

Strengths and Limitations

Conclusion

Abbreviations

Declarations

References

Tables

Additional Declarations

Status:

Version 1