China has entered a critical stage of a rapid and widespread increase of human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS). In the past 2 decades, the country has witnessed an alarming increase in HIV, with an annual increase of more than 30% in the number of cases. In recent years, sexual transmission has been the most rapidly growing means of HIV transmission.1 Sexually transmitted cases have increased from 7.2% of the total cases in 2002 to 43.6% by the end of 2005.2 Men who have sex with men (MSM) have now become one of the priority populations for the prevention and control of the HIV pandemic in China. Sociologic and anthropologic studies have indicated that in any given society, the proportion of MSM ranges from 2% to 5% of the adult male population (9.04–22.6 million) regardless of historical, social, or cultural background. In 2005, the Chinese Ministry of Health and United Nation Theme Group on AIDS submitted a joint assessment report, which estimated that HIV infections among MSM accounted for 7.3% of the total 650,000 probable HIV cases in China.3 The 2007 joint assessment report estimated that this proportion had increased to 11.0% of all HIV infections.4 Meta-analyses have shown that HIV prevalence among Chinese MSM has increased sharply, from approximately 1.3% before 2004 to approximately 5.0% in 2009.5,6
With HIV prevalence rapidly increasing among Chinese MSM in recent years, the Chinese government has paid more attention to the MSM population. Since 2005, it has expanded its intervention efforts to MSM, developed nationwide working protocols and guidelines on HIV prevention and control for MSM, and organized national technical workshops on comprehensive HIV prevention interventions for this population. Various programs have been conducted to promote condom use, counseling and testing, peer education, sexually transmitted disease (STD) services, and follow-up outreach, and care for people living with HIV. Third-quarter statistics from 2007 showed that 88,082 MSM were reached by comprehensive HIV-prevention interventions. Some of these have been conducted to prevent the increase of HIV/AIDS, but they have been reported very recently, and thus, to date, there has not been a systematic review of published studies that specifically address Chinese MSM and HIV/AIDS interventions. A critical question is whether these interventions are efficacious in reducing sexual risk behaviors (e.g., increased condom use with sexual partners) and HIV infection; the purpose of this study was to conduct a meta-analysis of these intervention trials.
MATERIALS AND METHODS
Studies investigating HIV infection among MSM in China were identified through article searches in the PsycINFO database, MEDLINE, SpringerLink, and China National Knowledge Infrastructure through October 2, 2011. Various combinations of the terms HIV, AIDS, intervention, efficacy, evaluation, gay, MSM, sex between men, homosexual, and China were used to screen for potentially relevant studies. If the study was duplicated, it was included only once.
All studies identified through the searches described previously were considered for inclusion in our meta-analysis. Inclusion criteria were as follows:
1. Reported the intervention strategy and follow-up time;
2. Measured any one of HIV knowledge, HIV testing, condom use, sexual activity, or HIV and STD incidence; and
3. Used a pre-post design or a quasi-experimental design.
Approximately 67 articles had the potential for inclusion in the meta-analysis; however, the following is an outline of the reasons why it was necessary to exclude some of them:
1. Thirty-four studies (51%) were excluded because they were not intervention studies.
2. Five studies (7%) were excluded because they did not use a pre-post design or quasi-experimental design.
3. Four studies (6%) were excluded because they did not report the baseline sample size or follow-up time.
4. One study (1%) was excluded because it did not include any measures of HIV knowledge, condom use, sexual risk behaviors, HIV testing, or HIV incidence.
5. One study (1%) was excluded because it was based on duplicate data (the duplicate was excluded).
As a result of these exclusions, 22 studies (33%) met all criteria and were included in the meta-analysis. The characteristics of these studies, which were all based on behavioral interventions, are shown in Table 1.
Human immunodeficiency virus knowledge, condom use, sexual risk activity, HIV incidence, STDs incidence, and HIV testing were used as outcomes. Condom use included (1) condom use during the most recent sexual intercourse with a man; (2) condom use during intercourse with a man during the past 6 months; and (3) condom use during intercourse with a woman during the past 6 months. Sexual risk activities included (1) the number of sexual partners during the past 6 months and (2) sex with casual partners during the past 6 months. The effect sizes were calculated for individual studies, and the meta-analysis examined the overall effect; subgroup analyses identified effect modifiers and sources of heterogeneity.
Different studies used different measures of HIV knowledge, so Cohen d (the difference in treatment and control means divided by the pooled SD) was used as the effect size indicator. All articles reported the results as percentages, which were converted to odds ratios and then to d. For HIV knowledge, condom use, and HIV testing, the effect size of greater than 0 suggests that the intervention was efficacious and less than 0 favored the control group. For HIV incidence, STDs, and sexual activities, the effect size of less than 0 suggests that the intervention was efficacious and greater than 0 favored the control group. We used the random-effect model. Effect sizes were weighted by sample size and are presented with 95% confidence intervals (CI). The Q statistic was used to examine whether significant heterogeneity existed among effect sizes. Effect sizes for the a priori, hypothesized, and categorical moderators were calculated along with 95% CI, and these effect sizes were statistically compared with one another using the QB statistic.29 All analyses were conducted using the Comprehensive Meta-Analysis Software, Version 2.0 developed by a team of experts in the US and the UK.
The studies (k = 22) had a cumulative of N = 11,967 in the baseline/control groups (median n per study = 230) and a cumulative of N = 11,500 in the intervention groups (median n per study = 203). One of the 22 studies (5%) was an MSM college student sample22; another study (5%) was an MSM adolescent sample20; and the remaining (k = 20) were general MSM samples. Two studies (9%) were of quasi-experimental design,10,18 and the remaining (k = 20) were of pre-post design. Five studies (23%) were of closed cohort design,10,14,18,22,26 and the remaining (k = 17) were of open cohort design. The most common intervention type was peer-led, HIV risk reduction education (k = 8). Five studies (23%) used the Voluntary Counseling and Testing (VCT) intervention strategy.7,8,11,17,24 The Chinese government promulgated the HIV VCT implementation plans in 2004, and by 2006, there were 3309 VCT sites in China. Some studies have reported that VCT has been efficacious in HIV prevention. Two of the studies (9%) used the 100% condom use program intervention strategy.14,24 The most common intervention type was group-targeted (k = 16), followed by individually tailored (k = 3)11,21,22 and mixed-type interventions (k = 3).13,17,19 Most of the interventions (k = 19) were not theory based, and the remainder were based on Behavior Change Theory (k = 2),10,28 or the AIDS Risk Reduction Model (k = 1).26 Six of the 22 studies (27%) reported the outcome of STD infection. Five studies reported on incidence of syphilis,8,9,11,13,15 and the remaining (k = 1) study7 did not report on a specific STD. The follow-up times of 8 of the studies (36%) ranged from 3 to 8 months, whereas for 9 studies (41%), follow-up times ranged from 9 to 14 months. Five studies (23%) had follow-up times of 24 months. All studies compared the demographic information between the baseline and intervention groups. One study (5%) reported an age difference between the baseline and intervention groups,15 whereas 2 studies (9%) reported a difference in education between the 2 groups.7,28 One study (5%) reported a difference in occupations between the baseline group and the intervention group,28 whereas 2 studies (9%) reported differences in sexual orientation differences between the 2 groups.12,15 No difference in marriage between the baseline and intervention groups was reported in any of the studies.
Efficacy of Interventions: HIV Knowledge
Although different studies used different measures of HIV knowledge, all studies focused on knowledge of HIV transmission and prevention. We combined these measures as the outcome of “HIV knowledge.” Seventeen studies reported outcomes on HIV knowledge, and the sample size–weighted mean effect size was d = 0.627 (95% CI, 0.460–0.793; Z = 7.38; P < 0.001), which indicates that the interventions are moderately efficacious in increasing HIV knowledge among Chinese MSM (Table 2). There was also evidence of heterogeneity (Q16 = 30.19, P < 0.01) within the studies.
Efficacy of Interventions: Condom Use
Sixteen studies reported outcomes on condom use with men during their most recent intercourse; the sample size–weighted mean effect size was d = 0.417 (95% CI, 0.302–0.532; Z = 7.11; P < 0.001). There was also evidence of heterogeneity (Q15 = 27.23, P < 0.01) within the studies. Sixteen studies reported outcomes of condom use for every session of intercourse among men during the past 6 months (Fig. 1); the sample size–weighted mean effect size was d = 0.394 (95% CI, 0.274–0.514; Z = 6.44; P < 0.001). There was also evidence of heterogeneity (Q15 = 33.38, P < 0.01) within the studies. Seven studies reported outcomes of condom use with women, for every session of intercourse, during the past 6 months; the sample size–weighted mean effect size for this variable was d = 0.605 (95% CI, 0.278–0.932; Z = 3.63; P < 0.001). There was also evidence of heterogeneity (Q6 = 5.436, P < 0.05) within the studies. These results indicate that interventions have moderately increased condom use, with men and women, during intercourse, among Chinese MSM (Table 2).
Efficacy of Interventions: Sexual Activity
Six studies reported outcomes with MSM having more than 2 sexual partners; the sample size–weighted mean effect size was d = −0.388 (95% CI, −0.644 to −0.132; Z = −2.97, P < 0.01). There was also evidence of heterogeneity (Q5 = 4.431, P < 0.01) within the studies. This indicates that interventions have reduced the number of sexual partners among Chinese MSM (Table 3). Four studies reported outcomes of having sex with casual partners; the sample size–weighted mean effect size was d = −0.139 (95% CI, −0.328 to 0.049; Z = −1.451; P = 0.147). There was also evidence of heterogeneity (Q3 = 2.42, P < 0.01) within the studies. This indicates that interventions have not reduced having sex with casual partners among Chinese MSM.
Efficacy of Interventions: HIV And STD Incidence
Eight studies reported outcomes of the incidence of HIV; the sample size–weighted mean effect size for this variable was d = 0.162 (95% CI, −0.031 to 0.356; Z = 1.643; P = 0.10). The heterogeneity test was significant (Q7 = 5.52, P < 0.01) within the studies. Six studies reported outcomes on STD; the sample size–weighted mean effect size for this variable was d = −0.044 (95% CI, −0.209 to 0.120; Z = −0.528; P = 0.60). The heterogeneity test was significant (Q5 = 7.61, P < 0.01) within the studies. These indicate that interventions have not reduced HIV or STD incidence among Chinese MSM (Table 3).
Efficacy of Interventions: HIV Testing
Eleven studies reported outcomes of HIV testing; the sample size–weighted mean effect size for this variable was d = 0.590 (95% CI, 0.420–0.761; Z = 6.80; P < 0.001), which indicates a positive effect of interventions, as HIV testing has moderately increased among Chinese MSM (Fig. 1). There was also evidence of heterogeneity (Q10 = 18.68, P < 0.01) within the studies.
Participant moderators (including location and age), intervention moderators (including sampling method and intervention strategy), and methodological moderators (including follow-up time and intervention type) were examined for the outcomes of condom use, in accordance with previous meta-analyses (Table 4). Location moderated the effect size of condom use. The effect size of all studies conducted in non–southwest regions was larger than that in southwest regions (QB = 4.45–7.59, df = 1, P < 0.05). Effect sizes based on the respondent-driven sampling (RDS) method were significantly higher than those based on other methods (snowball and MSM venues) for condom use at the last time of intercourse (QB = 12.87, df = 2, P < 0.001), and condom use for men within a 6-month period (QB = 10.14, df = 2, P < 0.01). Effect size based on MSM venues was significantly higher than a snowball sampling of condom use with women during the previous 6 months (QB = 5.09, df = 1, P < 0.05). Interventions were significantly more efficacious when they used the mixed intervention type compared with group-targeted and individually tailored intervention types for condom use with men during the last 6 months (QB = 6.14, df = 2, P < 0.05). Interventions were also significantly more efficacious when they used the popular opinion leader (POL) intervention strategy compared with peer-led education and multiple-intervention strategies for condom use during the last time of intercourse (QB = 11.33, df = 2, P < 0.01) and condom use for men during the last 6 months (QB = 5.00, df = 2, P < 0.05). The moderation of follow-up for condom use was not significant.
The results of the current meta-analysis indicate that interventions have been efficacious in increasing HIV knowledge, HIV testing, and condom use, and in reducing sexual risk activity. However, the efficacies of interventions to reduce HIV and STD incidence were not significant. Most of the studies reported single follow-up, with follow-up periods being less than 12 months. The intervention efficacy of HIV incidence may manifest during a long period, so the rate of HIV and STD incidence has not significantly changed, unlike other outcomes.
The current meta-analysis also examined the efficacy of condom use during intercourse with women. Chinese MSM also tend to have sex with women; according to 1 study, approximately 35% of unmarried and 70% to 80% of married MSM had had sex with women in the last 6 months, and only 16% to 29% of these individuals had consistently used condoms.30 Therefore, condom use with women is important in preventing the spread of HIV/AIDS to women because there is the potential for bridging. In addition, increasing numbers of sexual partners also contribute to HIV/AIDS transmission. It is reported that 79% of MSM had had more than 1 sexual partner during the past 6 months.15 Our results also reflected efficacy for reducing the number of sexual partners among MSM in China.
Although 7 (32%) of the 22 interventions were conducted in Southwest China, these interventions of condom use were less efficacious than those conducted in other provinces, and MSM in this region have the highest HIV prevalence compared with other regions.5,6 This may be associated with the remarkably high HIV prevalence among MSM in Southwest China in comparison with other regions. One recent meta-analysis revealed that MSM in Southwest China have a higher HIV prevalence, 11.4% (95% CI, 9.6%–13.5%), than other regions, where it ranges from 3.5% to 4.8%.5 Interventions of condom use that are more efficacious should be conducted in Southwest China.
The sampling strategies of current studies have primarily been snowball sampling methods (41%), which might not reflect the heterogeneity of the target populations.1 The results indicate that interventions based on RDS sampling were more efficacious than interventions based on snowball sampling. This may be related to the characteristics of RDS method in that the “seeds” played a crucial role in assisting researchers by leading intervention sessions and maintaining contact with participants throughout the study.26 Future interventions should use more RDS methods.
Most of the studies used single-group pre-post design (91%) or open cohort design (77%). Only 2 studies used a rigorous evaluation methodology (quasi-experimental design). In community-based HIV prevention intervention, it is a challenge to conduct rigorous evaluation methodology.1 The lack of interventions based on the experimental and quasi-experimental designs meant that we could not examine the differences in efficacy between those interventions based on pre-post design and those based on experimental and quasi-experimental designs. Future intervention efforts should use more rigorous evaluation designs.
The present meta-analysis has several limitations. First, this meta-analysis did not examine the efficacy of study designs because of the lack of interventions based on experimental and quasi-experimental designs. By the same token, our meta-analysis did not examine the efficacy of interventions based on theory. Second, although we examined several moderators for condom use, there may be other possible important moderators that were not examined. Other outcomes (e.g., HIV knowledge and HIV testing) were not examined in the moderation analysis. Finally, 1 study sampled participants from 18 cities; thus, some of the data may be repetitive with other articles. Because the study did not report more detailed information on participants, we were unable to identify possibly duplicated data and exclude it from our meta-analysis.
Despite these limitations, our analyses find clear evidence that HIV prevention interventions were efficacious in HIV knowledge, sexual activity, condom use, and HIV testing among Chinese MSM. Interventions for condom use conducted in Southwest China were significantly less efficacious. Interventions for condom use were significantly more efficacious when the RDS method and POL intervention strategy were used. Future interventions need to use more rigorous intervention methodologies.
1. Hong Y, Li XM. HIV/AIDS behavioral interventions in China: A literature review and recommendation for future research. AIDS Behav 2009; 13: 603–613.
2. Chinese Ministry of Health, UNAIDS, World Health Organization. Update on the HIV/STD epidemic and response in China 2005. Beijing, China: China Ministry of Health, 2006.
3. Chinese Ministry of Health, Joint United Nations Programme on HIV/AIDS, World Health Organization. 2005 update on the HIV/AIDS epidemic and response in China. Beijing, China: National Center for AIDS/STD Control and Prevention, China CDC, 2006.
4. State Council AIDS Working Committee Office, UN Theme Group on AIDS in China. Joint assessment of HIV/AIDS prevention, treatment and care in China (2007). Beijing, China: State Council AIDS Working Committee Office, 2007.
5. Chow EPF, Wilson DP, Zhang J, et al.. Human immunodeficiency virus prevalence is increasing among men who have sex with men in China: Findings from a review and meta-analysis. Sex Transm Infect 2011; 38: 845–856.
6. Chow EPF, Wilson DP, Zhang L. HIV and syphilis co-infection increasing among men who have sex with men in China: A systematic review and meta-analysis. PLoS One 2011; 6: e22768.
7. Cao NX, Jiang J, Ge FQ, et al.. Effectiveness of comprehensive interventions for STD/ HIV/ AIDS prevention among male sex workers [in Chinese]. Chin J AIDS STD 2009; 15: 479–428.
8. Ding XB, Feng LG, Xiao Y, et al.. Treatment efficacy of behavior intervention on male homosexual AIDS patients [in Chinese]. J Tropical Med 2010; 10: 323–326.
9. Feng LG, Ding XB, LV F, et al.. Study on the effect of the intervention about AIDS among MSM [in Chinese]. Chongqing, China: Chongqing Preventive Medicine Association Conference, 2009.
10. Gao MY, Wang S. Participatory communication and HIV/AIDS prevention in a Chinese marginalized (MSM) population. AIDS Care 2007; 19: 799–810.
11. Li JJ, Huan XP, Yan HJ, et al.. Effectiveness of behavioral intervention combined with voluntary counseling and testing on high risk behaviors among MSM [in Chinese]. Chin Prev Med 2011; 12: 666–669.
12. Liu F, Liu HH, Zhou C, et al.. Effect of integrated intervention on HIV/AIDS of MSM in a certain district of Chongqing [in Chinese]. Pract Prev Med 2008; 15: 1055–1058.
13. Ma F, Shen LT, Su CH, et al.. Effect evaluation of HIV/ AIDS integrated intervention of MSM in Xiamen City [in Chinese]. Chin J Dis Control Prev 2010; 14: 726–728.
14. Meng XY, Shao Y. Evaluation on the effect of AIDS related high risk behavior intervention among MSM in Pulandian City [in Chinese]. Prev Med Trib 2010; 16: 334–335.
15. Nong QX, Xu YF, Lin XQ, et al.. Evaluation on HIV/A IDS intervention among men have sex with men [in Chinese]. J Prev Med Inf 2010; 26: 200–203.
16. Wang Y, Zhang HB, Li ZJ, et al.. Effect evaluation on behavior intervention of AIDS/STI prevention and treatment in MSM group [in Chinese]. Pract Prev Med 2009; 16: 654–657.
17. Wang L, Liu ZF, Cao WH. Effect evaluation of testing initiated HIV/AIDS comprehensive interventions among men who have sex with men [in Chinese]. Chin Prim Health Care 2011; 25: 94–95.
18. Wang SG, Zhang SK. Effects of HIV education intervention among MSM [in Chinese]. Xinjiang Soci Sci 2004; 12: 59–66.
19. Wang Y, Zhang HB, Zhang GG, et al.. Effectiveness of AIDS related health education and behavioral [in Chinese]. J Prev Med Inf 2008; 24: 962–967.
20. Wang M, Wu CX. Cognitive status of AIDS and the effects of health intervention among MSM in Wuhan city [in Chinese]. J Pub Health Prev Med 2009; 20: 43–46.
21. Wang F, Huang YL. Analysis on effect of AIDS related high risk behavior intervention among MSM in Yingtan city [in Chinese]. Prev Med Trib 2011; 17: 518–520.
22. Wu GF, He FM, Heng SY. Intervention effects evaluation on aids-related know ledge and high-risk behaviors among 244 MSM in colleges [in Chinese]. Anhui J Prev Med 2011; 17: 86–88.
23. Xiu CZ, Liu MH, Li XF, et al.. Study of AIDS knowledge, behavior, intervention and the status of infection among MSM [in Chinese]. Prev Med Trib 2009; 15: 97–103.
24. Xu J, Zhou W, Zhou DJ, et al.. Evaluation of integrative behavioral intervention on AIDS among MSM in Wuhan [in Chinese]. Chin J Health Educ 2009; 25: 736–738.
25. Yao LH, Luo W, Wang RQ, et al.. Evaluation of intervention on HIV knowledge among MSM [in Chinese]. Sci Technol Innov Her 2010; 8: 247–249.
26. Zhang HB, Wu ZY, Zhang YJ, et al.. A pilot intervention to increase condom use and HIV testing and counseling among men who have sex with men in Anhui, China. J Acquir Immune Defic Syndr 2010; 53: S88–S92.
27. Zhang ZK, Wen XQ, Chen W, et al.. Evaluation of intervention on AIDS related high risk behaviors among men having sex with men in Guilin [in Chinese]. J Prev Med Inf 2010; 26: 784–787.
28. Zeng G, Xiao Y, Xu P, et al.. Evaluation of effect of community-based HIV/AIDS interventions among men who have sex with men in eighteen cities, China [in Chinese]. Chin J Prev Med 2009; 43: 977–980.
29. Noar SM, Black HG, Pierce LB. Efficacy of computer technology-based HIV prevention interventions: A meta-analysis. AIDS 2009; 23: 107–115.
30. Zhang B, Li X, Chu Q, et al.. A survey of HIV/AIDS related behaviors among 2250 MSM in nine major cities of China [in Chinese]. Chin J AIDS STD 2008; 14: 541–547.