Status And Influencing Factors Of Club Drug Use Among Men Who Have Sex With Men In Tianjin

Objective: To understand the basic characteristics of new-type drug users among men who have sex with men in Tianjin,and to ascertain the age,ethnicity,education level,working status,income status,sexual orientation,situation of sexual partners and the comparison of general demographic data and sexual behavior between new drug users and non-users in men who have sex with men.To master the drug use of new-type drug users among men who have sex with men in Tianjin.To realize the age of initial use of new drug users,types of new drug use,smoking patterns,presence or absence of mixed use,frequency of use,selection of sexual partners after taking new drugs,and sexual behaviors among men who have sex with men in Tianjin.Random forest and support vector machine(SVM)data mining methods were used to classify the status of MSM according to the characteristics of men who have sex with men,and the performance of the classification prediction was compared with the logistic regression method.It was explored whether the machine learning method was applied to the MSM population with some classification effects.Methods: From July to September 2015,the investigation was conducted by the sexual aid center of Tianjin Center for Disease Control and Prevention.The survey enrolled 222 qualified respondents with a combination of simple random sampling and snowball sampling.The survey was divided into two parts: the on-site questionnaire survey and the laboratory inspection.When the on-site survey was conducted,the survey participants filled in the survey questionnaires themselves,and the professionals collected 5ml of blood samples from the survey subjects and sent them to the laboratory for HIV antibody testing.Using the statistical analysis software SAS 9.4,the basic statistical description and single factor analysis were performed.Using R3.4.3 and SPSS Clementine 12.0 statistical analysis software,logistic regression,random forest,and support vector machine were implemented.Results: A total of 222 men who had sex with men were included in the survey.The average age was 30.09 9.97 years(17-70 years old),of which were mainly unmarried and single people(72.40%),most of which had incomes between 3,000 yuan and 4,999 yuan.Among the participants(45.05%),98(44.14%)received high school education,73(32.88%)received college education or above,and most respondents had jobs(87.78%).121(54.5%)of the respondents used new drugs.The respondents who took drugs were mainly young adults,and those who were older than 40 were less likely to use drugs(OR = 0.152,95% CI: 0.042 ~ 0.555),moreover,their income was more than 5,000 yuan.The survey respondents were more likely to use drugs than those without income(OR = 3.175,95% CI: 1.071 ~ 9.413).Students with junior college or above were more likely to smoke drugs than those who received lower level of education(OR = 5.760,95% CI: 1.082 ~ 30.651).No worker was more likely to use drugs than those who had a job(OR = 0.318,95% CI: 0.108 ~ 0.938);In terms of the number of sexual behaviors in the last three months,individuals with 2-6 times of sex behaviors per week were more likely to take drugs than those who were less than 1 time per week(OR = 2.644,95% CI: 1.246 ~ 5.610).Those having sex with 20 different partner or more in three months were more likely to take drugs than those with only 1 partner(OR = 16.987,95% CI: 2.161 ~ 133.547).Those who had sexual intercourse with commercial sex partners were more likely to smoke drugs than those who had no sexual intercourse with commercial sex partners.(OR=2.648,95%CI: 1.215 ~ 5.772).Those who had sex with temporary sex partners were less likely to take drugs than those who did not have sex with temporary sex partners(OR = 0.536,95% CI: 0.300 ~ 0.960);In terms of the average duration of one-time behavior in the past three months,the average person with a one-time behavior longer than 30 minutes was more likely to take drugs than those with an average one-time behavior less than 10 minutes(OR = 4.400,95% CI: 1.183 ~ 16.367).Of the 121 MSMs who took new drugs,the smoking rate of Rush Popper was 95.87%(116/121),the smoking rate of methamphetamine was 18.18%(22/121),the smoking rate of zero capsule was 9.92%(12/121),the smoking rate of K powder,MDMA,and Magu was less than 5%.Therefore,we further studied the risk factors of high-risk sexual behaviors of Rush Popper abusers and found that the sexual partner pattern was the influencing factor of high-risk sexual behaviors of Rush Popper abusers(P<0.05).Determine whether MSM people used drugs by logistic regression,random forest,and support vector machines.The results showed that support vector machine and random forest were better than logistic regression results in terms of sensitivity,specificity,accuracy,or AUC value.Conclusion: The characteristics of new drug users in the MSM population in Tianjin are youth,singleness,relatively high cultural level,and high income.The relevant departments should strengthen publicity and education work on the new types of drug abuse hazards among these groups.In the future,MSM populations,especially those who use drugs,will have to strengthen the monitoring of new drug use and HIV testing,and reduce the use of new drugs and the new HIV infection rate of MSM.The MSM population in Tianjin has a high rate of new drug use.The abuse rate of Rush Popper is as high as 95% or more.Therefore,relevant departments should focus on the population used by Rush Popper,and increase the supervision of Rush Popper for entertainment venues and online channels to avoid it.Circulation and abuse in the MSM population.According to the characteristics of MSM population,random forests and support vector machines(SVMs)can be used to classify and predict whether or not to use drugs,so as to identify high-risk groups and take targeted measures.