Skip to main content

Table 1 Behavioural and pyschosocial interventions

From: Equality in sexual health promotion: a systematic review of effective interventions for black and minority ethnic men who have sex with men

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€‰=ā€‰0.004). 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

Freireā€™s 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€‰=ā€‰0.0004)

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€‰=ā€‰0.004);

Statistically significant difference in oral or anal sex with unknown status male partners (pā€‰<ā€‰0.001); 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€‰<ā€‰0.0001, CI 95Ā %, 0.73-1.45); proportion condom-protected intercourse (pā€‰<ā€‰0.0001, CI 95Ā %, 0.87-2.77); unprotected intercourse (pā€‰<ā€‰0.0001, CI 95Ā % 0.68-1.35);multiple sexual partners (pā€‰<ā€‰0.0001, CI 95Ā %, 0.80-1.43); insertive anal intercourse (pā€‰<ā€‰0.0001, CI 95Ā %, 0.81-1.51); receptive anal intercourse (pā€‰<ā€‰0.0001, 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.

Oā€™Donnell

(2014)

Quality Appraisal score: 15

346 Latino MSM

No excuses/Sin buscar excusas

New York, USA

Social Cognitive Theory

Single 45ā€“60 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€‰=ā€‰4.10, Pā€‰<ā€‰0.05); Intervention participants more likely to engage in condom use after intervention (AORā€‰=ā€‰1.69; 95Ā % CI 1.02-2.81, pā€‰<ā€‰0.05).

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 6ā€“12 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€‰=ā€‰5.797; pā€‰<ā€‰.01), also increase in frequency of giving oral sex (Fā€‰=ā€‰3.928; pā€‰<ā€‰.01). They report increase in HIV/AIDS knowledge in Awareness item ā€œsharing clothes and hats not a mode of transmissionā€ (Fā€‰=ā€‰6.671; pā€‰<ā€‰.01). Finally, they report statistically significant differences in acceptance of safer sex (Fā€‰=ā€‰4.811; 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€‰=ā€‰3.03; 95Ā % CIā€‰=ā€‰1.26 ā€“ 7.28); marginal effects on condom use with male partners and with partners with HIV-negative/unknown serostatus (AORā€‰=ā€‰2.64; 95Ā % CIā€‰=ā€‰0.95-7.36) (AORā€‰=ā€‰3.19; 95Ā % CIā€‰=ā€‰0.98-10.4).

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€‰=ā€‰10.84, pā€‰<ā€‰0.05), hepatitis C knowledge (t =12.87, pā€‰<ā€‰0.05); number of sexual partners decreased (t (112)ā€‰=ā€‰4.33, pā€‰=ā€‰0.000); decrease in high risk sexual behaviours in 180Ā days follow-up (tā€‰=ā€‰4.76, pā€‰=ā€‰0.000); increase in specific psychosocial constructs: Self-esteem, coping, Social provisions (all pā€‰<ā€‰0.05) and Collective self-esteem items (Identity; Public; public Latino: all pā€‰<ā€‰0.05)

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, 3ā€“6 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 (25ā€“8 vs. 11ā€“0); 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.