First author | N | Name | Location | Theory | Description | Design | Primary Outcome | Secondary Outcomes | Findings | Limitations |
---|---|---|---|---|---|---|---|---|---|---|
Adam (2011) Quality Appraisal score: 11 | 40 Spanish speaking gay, bisexual and other MSM | Mano en Mano Latino | Toronto, Canada | NA | An initial day-long session, followed by four two-hour evening sessions. A life-skills and HIV prevention program. | Pre and Post-test | Frequency of protected and unprotected anal intercourse | Loneliness. Self-perceived degree of social connection/self-isolation | Decrease of UAI from pre-test to post-test stat sig. (pā=ā0.004). No difference between scores in loneliness. | Small sample size. No control or experimental group. Self-report (social desirability bias). Different groups thus not a coherent single group. |
Carballo-Dieguez (2005) Quality Appraisal score: 13 | 180 Latinos or of Latin American descent | Latinos Empowering Ourselves (LEO) LGBM (Latin gay and bisexual men) | New York City | Freireās Theory; Empowerment Ideas; researchersā own quantitative and qualitative research | Eight two-hour group sessions, once a week. Basic exercises and focusing on a specific theme. Session themes included oppression, transgression of rules, excuses, substance use, goal setting, the role of pleasure, self-efficacy, and plans for the future. | Randomised Control Trial (RCT) | Unprotected Anal Intercourse (UAI) | Ā | 46Ā % of the intervention participants from baseline to follow-up 1 reported no UAI. For the control group this was 54Ā %. 44Ā % of the intervention participants at follow-up 2 from intervention reported no UAI. 40Ā % from control. No differences between groups. | Selection bias: Participants ready to change |
Assessment effect: Elicitation of critical consciousness comprehensive baseline assessment. | ||||||||||
Convenience sample. Self-report ā social desirability biases. | ||||||||||
Choi (1996) Quality Appraisal score: 17 | 329 Asian and Pacific Islanders (API) | NA | San Fransisco, USA | Health Belief Model; Theory of Reasoned Action; Social Cognitive Theory | A single three-hour group session. | Randomised control trial (RCT) | Number of sexual partners | UAI and AIDS-related knowledge, attitudes and skills | Statistically significant Reductions in number of sexual partners at 3Ā month follow-up (pā=ā0.0004) | Convenience sample Self-report, Social desirability biases |
Increasing Positive ethnic and sexual identity; acknowledging HIV risk behaviours; presenting positive images of H-API. Enhance AIDS knowledge, attitudes to safer sex and sexual negotiations skills. | ||||||||||
Harawa (2013) Quality Appraisal score: 15 | 437 bisexual black men | Men of African American Legacy Empowerment Self (MAALES) | Los Angeles, USA | Theory of Reasoned Action and Planned Behaviour; Empowerment theory; Critical Thinking and Cultural Affirmation model | Six two-hour small-group sessions conducted over three-weeks plus booster sessions at six and 18Ā weeks post intervention. Holistic behavioural change | Randomised Control trial (RCT) | Number of male/female/male to female trans intercourse partners; number of any anal and vaginal intercourse; any unprotected intercourse (UI); any unprotected serodiscordant intercourse; sex while using substances | Ā | Intervention: Reduction in UAI with males (pā=ā.04); reduction in UAI/UVI with females (pā=ā.01); UI with male or females (pā<ā.01); reduction in number of partners both male and female (pā=ā.01); any risky drug use borderline significant (pā=ā.04). No significant differences in sex while using substances. | Convenience sample The focus is mostly on female partners. Self-report and Social desirability biases. |
Hosek (2015) Quality Appraisal score: 13 | 456 Black young men who have sex with men (YBMSM) and 50 young black persons as opinion leaders | Promoting Ovahness through Safer Sex Education (POSSE) | Chicago, USA | Popular opinion leader (OL) model | Opinion leaders trained during four 2Ā h sessions at risk-reduction conversation. Then, out in the community (House Ball Community) to spread knowledge among their peers in relation to sexual health knowledge | Repeated cross-section surveys (five separate circles) | Number of sex partners; Number of condomless anal intercourse (CAI) acts; any CAI with male partners; Sex with partner of unknown serostatus; any oral or anal act with unknown serostatus partners | Frequency of sex under the influence of substances | Decline in mean number of male sex partners (pā=ā0.004); Statistically significant difference in oral or anal sex with unknown status male partners (pā<ā0.001); no statistically significant differences in all other outcomes. | No control group for this study. Convenience sample which might cause generalisability issues. Self-report bias |
Jemmott (2014) Quality Appraisal score: 16 | 503 African American MSM | Being responsible for Ourselves (BRO) | Philadelphia, USA | Social Cognitive Theory; the reasoned-action approach; qualitative research (formative research with black MSM). | A three session one-on-one HIV/STI risk reduction designed to increase consistent condom use vs. three one-on-one health promotion intervention. | Randomised Controlled Trial (RCT) ā cross-sectional study. | Condom-protected intercourse acts | Proportion of condom-protected intercourse acts; unprotected sexual intercourse; multiple sexual partners; IAI; RAI. Theoretical constructs: condom use hedonistic outcome expenctancy; condom use prevention outcome expectancy; condom use self-evaluative outcome expectancy; condom use availability self-efficacy; condom use negotiation self-efficacy; condom use technical skills efficacy; condom use impulse control self-efficacy; HIV risk-reduction knowledge; condom use knowledge. | Increase in consistent condom use at 90Ā days post intervention follow-up (pā<ā0.0001, CI 95Ā %, 0.73-1.45); proportion condom-protected intercourse (pā<ā0.0001, CI 95Ā %, 0.87-2.77); unprotected intercourse (pā<ā0.0001, CI 95Ā % 0.68-1.35);multiple sexual partners (pā<ā0.0001, CI 95Ā %, 0.80-1.43); insertive anal intercourse (pā<ā0.0001, CI 95Ā %, 0.81-1.51); receptive anal intercourse (pā<ā0.0001, CI 95Ā %, 0.57-1.13). There were no differences between the arms of the intervention. | One-on-one approach instead of group intervention which might have affected the results. Black populations more socially close-knit; an aspect important in relationships related negotiations. Self-report/social desirability bias. Generalisability limited as participants not randomly selected. |
Baseline, IPT, 6-months follow-up; 12-months follow-up. | ||||||||||
OāDonnell (2014) Quality Appraisal score: 15 | 346 Latino MSM | No excuses/Sin buscar excusas | New York, USA | Social Cognitive Theory | Single 45ā60 min session intervention (either English or Spanish) vs. non-attention control condition. Three-month follow-up | Randomised control trial/Randomised field trial | 1) total number of unprotected anal intercourse (UAI) acts with last two male partners; 2) condom use at last intercourse with a male partner; 3) self-report of an HIV test during the 3-month follow up window. | Ā | Decline in mean number of UAI (59Ā % vs. 39Ā %, Fā=ā4.10, Pā<ā0.05); Intervention participants more likely to engage in condom use after intervention (AORā=ā1.69; 95Ā % CI 1.02-2.81, pā<ā0.05). | Short-term impacts (3-month follow-up). Self-report bias. Convenience sample. |
Stein (2015) Quality Appraisal score: 14 | 337 young MSM of colour (Black/African-American and Latino/Hispanic males). | Community based organisation behavioural (CBO) outcomes of Many Men, Many Voices (CBOP-3MV) Project | New York, Tampa, New Orleans, USA | Ā | Small group-level intervention facilitated by peers in groups of 6ā12 clients. | Repeated measures design with no control group | 1) Sex partners: Number and prevalence of sex partners; number of serodiscordant/unknown status partners; | Ā | Significant decreases in all outcomes, both Sex partners and sex events related. | No control group for the specific study. Internal validity is not strong. Self-report bias. The three different CBOs offered the intervention in slightly different ways. |
2) Sex events: Number and prevalence of sex events without a condom; number of sex events with serodiscordant/unknown status partners; number and prevalence of sex events without a condom withĀ another (male); number of sex events without a condom under the influence of substances. | ||||||||||
Somerville (2006) Quality Appraisal score: 13 | 766 Young Latino MSM and 37 Young Latino Promotores (YLP) | Young Latino Promotores (YLP) | Vista, California; McAllen, Texas, USA | Popular Opinion Leader model ā Theory of diffusion | Uses social networks to deliver HIV prevention messages. Identification and enlistment of popular persons within the community. Training them in disseminating prevention and risk reduction messages related to HIV/STIs. Supporting and reinforcing successful waves of POL in order to create a culture of normative change. | Repeated cross-section surveys | 1) HIV Risk behaviour: frequency of receptive anal sex with condoms; frequency of receiving and giving oral sex without using condoms; | Ā | Most of the items observed were already high. Increase in the use of condoms in receptive anal sex (Fā=ā5.797; pā<ā.01), also increase in frequency of giving oral sex (Fā=ā3.928; pā<ā.01). They report increase in HIV/AIDS knowledge in Awareness item āsharing clothes and hats not a mode of transmissionā (Fā=ā6.671; pā<ā.01). Finally, they report statistically significant differences in acceptance of safer sex (Fā=ā4.811; pā<ā.05). | No control group which impacts integral validity. Self-report bias and social desirability biases. Convenience sample. Cross section surveys with different participants each time and not the same participants. |
3) HIV/AIDS knowledge 2) Social norms: acceptance of safer sex; | ||||||||||
The theory of diffusion which OL is based recommends 15Ā % of the population should be trained as OL in order for change to occur. This is not the case for this particular study. | ||||||||||
Tobin (2013) Quality Appraisal score: 16 | 147 African American MSM | Unity in diversity (UND) | Baltimore, USA | Information-motivation-behaviour model (IBM); social network theory (SNT); social cognitive theory (SCT) ā self-efficacy construct. | Six group sessions group sessions as the intervention arm and a single session as control condition. 3-month follow-up | Randomised control trial (RCT) | Number of partners; condom use; sex while drunk or high. | Ā | Increased odds of reporting fewer male partners (AORā=ā3.03; 95Ā % CIā=ā1.26 ā 7.28); marginal effects on condom use with male partners and with partners with HIV-negative/unknown serostatus (AORā=ā2.64; 95Ā % CIā=ā0.95-7.36) (AORā=ā3.19; 95Ā % CIā=ā0.98-10.4). | Short term results. Convenience sample which might impact generalisability. Self-report and social desirability biases. Potential contamination between conditions. |
Vega (2011) Quality Appraisal score: 13 | 113 Latino gay men | SOMOS: We are | New York, USA | Social Identity Theory | Five group sessions exploring concepts of sexual, ethnic and cultural identity; coming out process; body image and sexual acts. Ten intervention cycles. | Pre- and post test | HIV/AIDS and hepatitis C knowledge; number of sexual partners, different types of partners, high risk sexual encounters; psychosocial constructs (coping, self-efficacy, internalized homophobia, self-esteem, sources for social support and collective self-esteem). | Ā | Increase in HIV/AIDS and hepatitis C knowledge from baseline to 90Ā days follow-up (tā=ā10.84, pā<ā0.05), hepatitis C knowledge (t =12.87, pā<ā0.05); number of sexual partners decreased (t (112)ā=ā4.33, pā=ā0.000); decrease in high risk sexual behaviours in 180Ā days follow-up (tā=ā4.76, pā=ā0.000); increase in specific psychosocial constructs: Self-esteem, coping, Social provisions (all pā<ā0.05) and Collective self-esteem items (Identity; Public; public Latino: all pā<ā0.05) | No control group which might increase the self-report bias and create problems of integral validity. Self-report and self-desirability biases. Sample size not randomised but rather specifically focused on Latino populations in New York ā issues of generalisability |
Baseline, Follow-up 90Ā days, follow-up 180Ā days | ||||||||||
Williams (2013) Quality Appraisal score: 15 | 88 HIV+ African American MSMW with history of childhood sexual abuse. | Enhanced Sexual health Intervention for Men (ES-HIM) | Los Angeles, USA | Cognitive-behavioural approach and ecological framework | Six two-hour small group sessions administered over a period of three weeks. A stress-focused sexual risk reduction intervention condition (ES-HIM) vs. a general health promotion intervention condition (HP). | Randomised control trial (RCT) | Sexual risk behaviours (unprotected anal and vaginal sex); number of sexual partners; psychological symptoms; stress biomarkers (urinary cortisol and catecholamines and neopterin (indicator of HIV progression). | Ā | Both groups reduced unprotected anal insertive sex (time Pā<ā.01). Both reduced vaginal sex (time Pā<ā.01). Both groups reduced number of sexual partners which was sustained at the 6-months follow-up as well as time Pā<ā.001). Reductions in depression and PTSD (time Pā<ā.01). Overall, intervention arm showed no significant advantage over control group. | Participants despite randomisation differences at baseline in CSA severity, psychological symptoms and biomarkers. Major intervention group effect (ES-HIM lower biomarker composite scores than HP at baseline). Failure in randomisation that could not be corrected due to small sample size. Convenience sample which might impact generalisability. Self-report and social desirability biases. |
Assessment baseline, 3ā6 month follow-ups. | ||||||||||
Young (2013) Quality Appraisal score: 13 | 112 African American and Latino MSM | HOPE (Harnessing Online Peer Education) UCLA | Los Angeles, USA | The community-peer leader model. Similar to the opinion leader model. | Twelve-week HIV prevention intervention designed to use peer-led Facebook groups to diffuse HIV information to increase testing among black and Latino MSM. Home-based HIV testing kit and completed questionnaire at baseline and 12-week follow-up. Intervention: received peer-delivered information on HIV prevention. Control: Peer-delivered information on general health information | A hybrid design of RCT with diffusion approach | Behaviour change: requesting a home-based kit, returning the kit and following-up for test results | Number of sexual partners, observed and self-reported communication using the social networking community. | More intervention participants requested and followed-up with a kit (25ā8 vs. 11ā0); high participation and engagement rates in both groups; number of sexual partners decreased in both groups. | The statistical data are not present for the primary outcome as the numbers are low, small sample. |
Self-report bias. Convenience sample. |