Development And Assessing The Performance Of A Predictive Risk Model Of HIV Infected Among Students Men Who Have Sex With Men In China

Objective: Although the global AIDS epidemic has stabilized,new diagnoses among men who have sex with men(sometimes referred to as MSM)are increasing rapidly.HIV infected has received widespread attention in recent years among students MSM,who are one of the subgroups among MSM.On average,8 college MSM students in China are infected with HIV every day.Meanwhile,students MSM are a concealed group with a large number.Contrast to repress physiological and psychological requirements at ordinary times,students MSM are strongly desired to release themselves in a more extreme pattern.Unprotected anal intercourse,multiple sexual partners,usage of recreational drugs and sexually transmitted infection are common among students MSM,however,the rate of HIV counseling and testing services are low among this group.The poor risk perception of HIV infected show that nearly half of students MSM believe that they have low or no risk of HIV infected,which may be one of the primary causes of HIV infection among this group.Therefore,an accurate and easily used predictive risk model of HIV infected among students MSM is urgent to develop,which might not only help to strengthen risk perception of HIV infection,but also facilitate HIV test.However,there is no international or domestic predictive risk model of HIV infected developed for students MSM.We have two aims in this study as the follows: 1.To develop a predictive risk model of HIV infected among students MSM;2.To assess the performance of a predictive risk model of HIV infected among students MSM,based on a cross-sectional study in Shenyang.Methods: This study includes students MSM enrolled in two studies,one is a multiple sites cross-sectional study from 2012 to 2013,the other one is in a cross-sectional study in Shenyang between February 2009 and December 2014.All individuals have written informed consent.Social demographics,recent sexual behaviors,history of drug use,and information of sexual transmitted infection are included in this study.Laboratory tests include HIV screening and confirmation,syphilis screening and confirmation and HSV-2 antibody testing.Univariate and multivariate logistic regression was used to develop model;stepwise procedure was used to select variables.Hosmer–Lemeshow test was used to estimate the Model calibration;C-statistics test was used to assess the model discrimination.The model scoring system is due to the odds ratio(OR)of each variables.Diagnostic Odds Ratio(DOR)and receiver operating characteristic(ROC)analysis were used to assess the model scoring system.Bootstrap resampling technology was used to model validation.The database of students MSM for model test was enrolled in a cross-sectional study in Shenyang between January 2015 and December 2017.All individuals have written informed consent,male,self-reported have sex with man(oral sex/anal sex)in the last 6 months,students,older than 18 years old.Model test was used receiver operating characteristic(ROC)analysis.SAS 9.2 and SPSS 22.0 were used for analysis,all analysis is two-side test,p<0.05.Results: 1.A total of 742 students MSM were enrolled in this study,and HIV prevalence was 5.9%(44/742).According to chi-square analysis,compared with who was HIV non-infected,HIV-infected students MSM had the following characteristics: the infection rate of sexually transmitted diseases was higher,the larger proportion of students MSM provided commercial sexual services,the larger proportion of clubdrug use in the past six months,and proportion of receptive or both in the most recent intercourse of MSM was larger.2.According to univariate and multivariate logistic regression,the variables and score in predictive risk model of HIV infected among students MSM were sexual transmitted infection(OR = 5.4,95% CI = 2.5~11.4,P <0.01)(5 points);clubdrug use in the past six months(OR = 2.2,95% CI = 1.2~4.2,P = 0.02)(2 points);the most recent intercourse of MSM was receptive or both(OR = 2.3,95% CI = 1.0~5.0,P = 0.04)(2 points).The AUC of C-statistics is 0.71,which shows the model has moderate discriminative power.3.The AUC of the model scoring system was 0.71(95% CI=0.63~0.80).The diagnostic odds ratio(DOR)was greater than 1.All the above results show that the scoring system can discriminate the risk of each variables effectively.4.The frequency of the three variables in the final model appeared in 1000 Bootstrap samples were 94.3%,69.3% and 51.3% respectively,all of which were greater than 50%,indicating that all three variables had good credibility.5.A total of 155 MSM students were enrolled in Shenyang between August 2016 and December 2017.The HIV infection rate was 6.45%(10/155).The AUC of the model test was 0.72(95% CI=0.58~0.86),indicating that the model has moderate discriminative power when applied to other students MSM.Conclusion: As a simple and convenient method,the predictive risk model of HIV infected among students MSM can help students MSM strengthen their risk perception of HIV infection and maybe help facilitate HIV test.In addition,in the future research work,it is still necessary to collect more information about students MSM to further model test and improve the performance of the model.