Behavioural interventions for HIV positive prevention in developing countries: a systematic review and meta-analysis
Caitlin E Kennedy, Amy M Medley, Michael D Sweat & Kevin R O’Reilly
Volume 88, Number 8, August 2010, 615-623
Table 1. Studies included in systematic review of the literature on behavioural interventions for HIV positive prevention in developing countries
| Author and year | Setting | Population characteristics | Description | Study design |
|---|---|---|---|---|
| Interventions targeting HIV+ and HIV− individuals | ||||
| Allen et al., 1993 |
Kigali, Rwanda | Female paediatric and antenatal care clients Gender: 100% female Agea range: 18–35 |
35-minute educational video and group discussion led by physician and social worker. HIV test results and counselling 3 weeks later. At request of study subjects, male sex partners invited to watch video, attend group sessions and be tested for HIV. Couples could choose to receive test results together. Condoms and spermicide distributed free. | Time series, no comparison arm. Assessments at baseline (n = 1 458), at 12 months (n = 1 254) and at 24 months (n = not reported). Unit of analysis: individual. Participants not randomly selected. |
| Allen et al., 1992 |
Kigali, Rwanda | Female paediatric and antenatal care clients Gender: 100% female Mean age: 29 Age range: 20–40 |
35-minute educational video and group discussion led by physician and social worker. HIV test results and counselling 3 weeks later. At request of study subjects, male sex partners invited to watch video, attend group sessions and be tested for HIV. Couples could choose to receive test results together. Condoms and spermicide distributed free. | Cross-sectional, comparing women who were tested for HIV individually with those tested with their partners. Unit of analysis: individual. Participants not randomly selected. |
| Farquhar et al., 2004 |
Nairobi, Kenya | Women attending antenatal care and their partners Gender: 100% female Mean age (SD): women tested individually: 23.7 (4.4) women whose partners were tested; results received individually: 24.1 (4.6) women whose partners were tested; results received as couple: 23.8 (4.4) |
Group education about HIV transmission, with encouragement to inform male partners about HIV VCT. Return in 1 week, alone or with partner, for optional VCT, counselling on safe sex during pregnancy and on breastfeeding, and free condoms; return in 2 weeks for more counselling: if HIV−, breastfeeding recommended; if HIV+, other infant feeding options given, along with nevirapine for both mother and infant and counselling on its use at delivery. More counselling and optional infant HIV testing offered 3 and 6 months postpartum. Women with HIV-related symptoms treated and referred to local clinics. | Prospective cohort, no comparison arm. Assessments at baseline (group education, n = 2 836), at 1 week (VCT, n = 2 104) and at 2 weeks (follow-up counselling, n = 1 630). In addition, 122 HIV+ women returned 1 week postpartum. Unit of analysis: individual. Participants not randomly selected. |
| King et al., 1995 |
Kigali, Rwanda | HIV+ and HIV− urban women Gender: 100% female Age distribution: 20–24 (4.4%) 25–29 (28.7%) 30–34 (36.7%) 35–39 (25.7%) 40–44 (4.6%) |
15-minute educational video in Kinyarwanda on contraceptive methods and group discussion led by nurse. Oral contraceptives, injectable progestins and Norplant provided free to women enrolled in programme. Other contraceptive methods made available to women and their partners were intrauterine devices, condoms (both before and after intervention), tubal ligation and vasectomy. | Time series, no comparison arm. Assessments at baseline (n = 502) and after intervention (n = 470). Average follow-up time differed by participant and outcome. Unit of analysis: individual. Participants not randomly selected. |
| Machekano et al., 1998 |
Harare, Zimbabwe | Male factory workers Gender: 100% male Age: NR |
Pre-test counselling in factories; subjects encouraged to visit project clinic for HIV test results and counselling 2 weeks after blood draw. Free STD diagnostic and treatment services and condoms also available at project clinic. Video on preventing HIV infection shown at all times in waiting area of project clinic. | Time series, no comparison arm. Assessments at baseline (n = 2 414) and after intervention (n = 2060). Mean follow-up time per subject was 1.2 years. Unit of analysis: individual. Participants not randomly selected. |
| Pickering et al., 1993 |
The Gambia | Female commercial sex workers Gender: 100% female Mean age: 31.9 |
Sex workers examined, treated for existing STDs and tested for HIV. Given test results and counselled in subsequent visit. All given free condoms and told to return to clinic for treatment or condoms whenever necessary. Free condoms also distributed daily in all bars included in study. | Time series, no comparison group. Assessments at baseline (n = 31), at 1 month (n = 29) and at 2–5 months (n = 31). Unit of analysis: individual. Participants not randomly selected. |
| Roth et al., 2001 |
Kigali, Rwanda | Heterosexual males in cohabitating union Gender: 50% male, 50% female Mean age: males: 39 females: 32 |
Male-focused counselling programme with educational video entitled “Responsibility” and small group discussion. During return visit, trained counsellor gave men their HIV test results in individual counselling session and encouraged them to share them with partners; all female partners had already been tested for HIV. | Before/after, no comparison arm. Assessments at baseline (n = 684 couples) and at 12 months (n = 684 couples). Unit of analysis: individual and couple. Participants not randomly selected. |
| VCT Efficacy Group, 2000 Individuals |
Nairobi, Kenya; Dar es Salaam, United Republic of Tanzania; Port of Spain, Trinidad and Tobago | General population: individuals Gender: 49.2% males, 50.8% females Mean age (SD): intervention males: 28.9 (9.7) intervention females: 28.6 (8.6) control males: 28.1 (9.1) control females: 28.5 (8.8) |
Individuals randomized to HIV VCT or health information. VCT arm: client-centred counselling (personalized risk assessment and risk-reduction plan, role plays and condom use demonstrations); test results available 2 weeks after blood draw. Health information arm: 15-minute video and group discussion about HIV and condom use. | Randomized controlled trial with 1 intervention and 1 control group. Assessments at baseline (n = 3 120), at 7 months (n = 2 550) and at 14 months (n = 2 196). Unit of analysis: individual. Participants not randomly selected. |
| VCT Efficacy Group, 2000 Couples |
Nairobi, Kenya; Dar es Salaam, United Republic of Tanzania; and Port of Spain, Trinidad and Tobago | General population: couples Gender: 50% males, 50% females Mean age (SD): intervention males: 31.5 (8.4) intervention females: 25.9 (6.6) control males: 32.1 (32.1) control females: 26.7 (7.4) |
Couples randomized to HIV VCT or health information. VCT arm: client-centred counselling (personalized risk assessment and risk-reduction plan, role plays and condom use demonstrations); test results available 2 weeks after blood draw. Health information arm: 15-minute video and group discussion about HIV and condom use; VCT offered at first follow-up. | Randomized controlled trial with 1 intervention and 1 control group. Assessments at baseline (n = 1 174), 7-months (n = 1 001) and at 14months (n = 890). Unit of analysis: individual. Participants not randomly selected. |
| Interventions targeting HIV+ individuals only | ||||
| Allen et al., 2003 |
Lusaka, Zambia | HIV serodiscordant couples Gender: 50% male, 50% female Age: NR |
Same-day couples’ VCT service, with free diagnosis and treatment of STDs, condom skills training and free condoms. More counselling provided on request at 3-month intervals and when sexual contacts without protection were reported. | Before/after, no comparison arm. Assessments at baseline (n = 818 couples) and at 12 months (n = 584 couples). Unit of analysis: couple. Participants not randomly selected. |
| Allen et al., 1992 |
Kigali, Rwanda | HIV serodiscordant couples Gender: 50% male, 50% female Age: NR |
Educational video and group discussion led by social worker. Free condoms and spermicides offered. At post-test counselling, project counsellor confidentially distributed HIV test results individually, but couples encouraged to receive them together. | Before/after, no comparison arm. Assessments at baseline (n = 60 couples) and at 12 months (n = 53 couples). Unit of analysis: couple. Participants not randomly selected. |
| Balmer et al., 1994 |
Nairobi, Kenya | HIV+ individuals Gender: NR Age: NR |
Weekly group counselling sessions for 6 months to initiate and sustain behaviour change and provide psychological support. Sessions based on unified HIV/AIDS counselling theory combining behavioural, psychoanalytical and humanistic axioms. | Randomized controlled trial. Assessments at baseline (n = 20) and at 6 months (n = 20). Unit of analysis: individual. Participants not randomly selected. |
| Bunnell et al., 2006 |
Tororo, Uganda | HIV+ adults initiating ART Gender: 25.4% male, 74.6% female Median age: males: 41; females: 37 |
ART delivered weekly at home, with referral as needed for free medical and psychological care. Behavioural intervention with group education on ART at enrolment, testing of cohabitating partners through home-based family VCT and counselling on risk reduction strategies to protect HIV− partners or those with unknown HIV status. Free condoms provided on request. | Before/after, no comparison arm. Assessments at baseline (n = 926) and at 6 months (n = 815). Unit of analysis: individual, couple and aggregated cohort. Participants not randomly selected. |
| Da Silveira et al., 2006 |
Pelotas, Brazil | HIV+ women attending HIV outpatient clinic Gender: 100% female Age distribution: 15–19 (5.6%); 20–29 (41.5%); 30–39 (29.4%); 40–49 (17.6%) ≥ 50 (5.9%) |
Educational programme with four educational modules on HIV infection, general health-care measures and condom use delivered by physician during routine medical consultation. Use of flipcharts with graphic displays to reinforce intervention’s main contents. Patients could request as many free condoms as they wished from clinic pharmacy after consultation. | Non-randomized trial with 1 intervention group and 1 control group. Assessments at baseline (n = 340), 30 days (n = 332) and at 60 days (n = 335). Unit of analysis: individual. Participants not randomly selected. |
| Jones et al., 2006 |
Lusaka, Zambia | Sexually active HIV+ women Gender: 100% women Mean age: 29 |
Three-session group educational and skills-building intervention with videos and role playing to illustrate sexual barrier products, risk reduction and sexual negotiation strategies. At each visit, participants given1-month supply of male and female condoms and vaginal lubricants and screened and treated for STDs and vaginal infections. | Randomized controlled trial. Assessments at baseline (n = 240), at 6 months (n = 233) and at 12 months (n = 166). Unit of analysis: individual. Participants not randomly selected. |
| Kamenga et al., 1991 |
Kinshasa, Democratic Republic of the Congo | HIV serodiscordant couples Gender: 50% male, 50% female Mean age: M+, F− couples: males: 39.8; females: 32.0 M−, F+ couples: males: 37.9; females: 30.7 |
HIV serodiscordant couples invited to special clinic where serostatus confirmed. Each couple was subsequently informed of HIV test results, first individually by counsellor of the same sex and then together as couple by both members of the counselling team. Couples then counselled about STDs, HIV infection and condom use; followed monthly at counselling centre and given condoms and a sexual activity calendar at each visit. | Time series, no comparison arm. Assessments at baseline (n = 168 couples), at 6 months (n = 149 couples) and at 18 months (n = 140 couples). Unit of analysis: couple. Participants not randomly selected. |
| MacNeil et al., 1999 |
United Republic of Tanzania | Sexually active HIV+ adults Gender: 34% male, 66% female Mean age: 31.9 |
At least once a month, enhanced care and support consisting of ongoing counselling of HIV+ person on prevention and problem-solving, education of other family members, condom provision and, when necessary, referral for treatment. | Randomized control trial comparing enhanced programme with standard health services. Assessments at baseline (n = 154) and at 6 months (n = 144). Unit of analysis: individual. Participants not randomly selected. |
| Ryder et al., 2000 |
Kinshasa, Democratic Republic of the Congo | HIV serodiscordant couples Gender: 50% male, 50% female Mean age: M+, F− couples: males: 40.7; females: 25.4 M−, F+ couples: males: 38.7; females: 31.0 |
Couples tested for HIV and informed of results first individually by counsellor of the same sex and then together as couple; counselled on safe sex, condom use and HIV-associated risks, and followed by monthly counselling sessions to identify difficulties faced and develop corrective strategies. | Time series, no comparison arm. Assessments at baseline (n = 178 couples), at 18 months (n = 167 couples) and at 30 months (n = 139 couples). Unit of analysis: individual and couple. Participants not randomly selected. |
| Yang et al., 2001 |
Fuyang City, China | HIV+ individuals and their spouses Gender: 50% male, 50% female Age range: 20–49 |
Couples educated at home on HIV and condom use every other month for 12 months. Emphasis on importance of consistent condom use; free condoms distributed. | Before/after, no comparison arm. Assessments at baseline (n = 90 couples) and at 12 months (n = 84 couples). Unit of analysis: individual and couple. Participants not randomly selected. |
AIDS, acquired immunodeficiency syndrome; ART, antiretroviral therapy; HIV, human immunodeficiency virus; NR, not reported; SD, standard deviation; STD, sexually-transmitted disease; VCT, voluntary counselling and testing.
a Age throughout the table is expressed in years.