University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

A group-based intervention to increase condom use among HIV serodiscordant couples in India, Thailand, and Uganda
Global Health Sciences Literature Digest
Published June 4, 2007
Journal Article

McGrath JW, Celentano DD, Chard SE, Fullem A, Kamya M, Gangakhedar RR, et al. A group-based intervention to increase condom use among HIV serodiscordant couples in India, Thailand, and Uganda. AIDS Care 2007 Mar;19(3):418-24.


To assess the acceptability and impact of a group-based intervention for sero-discordant couples, designed to increase communication and condom use, and adapted for culturally diverse settings.

Study Design

A non-controlled intervention design, with pre-post comparisons and three months of follow-up.


Pune, India; Chiang Mai, Thailand; Kampala, Uganda


A total of 43 sero-discordant couples were enrolled (15 in India, 14 in Thailand, 14 in Uganda), of whom 40 (91%) completed all study procedures. Index cases were identified from ongoing research projects at STI clinics and VCT sites. Inclusion criteria included 1. Willing to enroll a regular sexual partner of ≥ six months’ duration; 2. Willing to disclose HIV sero-status to the partner; 3. Intent to remain as partners; 4. Willing to attend the intervention sessions and to be interviewed; 5. Willing to report inconsistent condom use; 6. Report no history of domestic violence. Of index cases initially recruited, 89 were ineligible: 18 already used condoms consistently; 16 (18%) were unwilling to recruit a partner; 15 (17%) refused participation in the intervention; and two had been hit by their partners. Of HIV sero-negative partners who were screened, eight were ineligible (none due to refusal). In 27 of 43 (59%) couples, the man was the sero-positive partner. The average age of participants was 34 years, and the average length of relationship was 7.3 yrs. In Thailand, most couples had only completed primary education; in Uganda and India levels of education were higher.


The intervention was based on four sessions of small groups and included "homework" between sessions. The first two sessions were single-sex groups and the last two included couples. Content included providing didactic information on HIV/AIDS and risk-reduction strategies, relapse prevention, impact of disclosure within partnerships, and understanding HIV discordance. Skills training through role-playing and rehearsal using flexible formats, and creating or enhancing motivation for condom use were included. A primary emphasis throughout all of the modules was improving communication about sex.

Primary Outcomes

Primary outcomes included acceptability (willingness to enroll and recruit a partner), feasibility of the intervention (completion of study activities, number of adverse outcomes, satisfaction with intervention activities, and communication skills), social and psychological adverse events. The primary outcomes for measuring impact were reported frequency of unprotected sexual intercourse and condom attitudes. Data were collected through semi-structured interviews administered to participants separately at baseline, one month, and three months post-intervention.


Three out of 43 couples did not complete the study, primarily because of illness in the HIV sero-positive partner. There were no reported psychological or social adverse outcomes. All participants consistently reported liking the intervention and the group format, and wished that it would continue; 80% also wanted more activities without their partners. The authors did not report the proportion who refused to enroll or engage their partner, but presumably this is 31/140 (22%).

Communication skills: At three-month follow-up, 90% of participants reported they could use the communication skills from the intervention, 96% felt more comfortable discussing AIDS with their partner, and 90% were 'very confident' that they could refuse sex if their partner didn't want to use a condom (an increase from 70% at baseline, p=0.001).

Frequency of sexual intercourse: Reported frequency of sexual intercourse in the previous month increased at all sites between baseline and three months. The greatest change occurred among couples in India: at baseline, 73% reported no sex in the previous month compared to 26% at three-month follow-up.

Condom attitudes and use: At baseline, the majority of Thai and Ugandan participants and one-third of Indian participants reported having used a condom with their regular partner; by the first follow-up visit, 100% of couples had used condoms. Despite increased comfort with condoms and improved communication, partners continued to be concerned that transmission might still occur (79% of Thai partners, and 27% of Indian partners).


The authors conclude that, given the inherent tension and risk posed by sero-discordancy, group-based couples' interventions can provide social support and information without fear of stigmatization. The results highlight the feasibility of this intervention and the need for ongoing support for discordant couples.

Quality Rating

Based on the Newcastle-Ottawa quality rating for prospective cohort studies, this study was of adequate quality. The study had several important limitations, primarily due to lack of adequate descriptions of study design and of results. Even though one of the main study objectives was to evaluate willingness to enroll and disclose to a partner, neither the total number of index cases screened nor the sampling strategy were described. Thus, determining the proportion who refused to disclose or participate was not possible, except by calculating these values from numbers given in the text; and these values may not be complete. Although this was a pre-post study design, complete baseline data was not provided. Generalizability of study results is hampered by the small sample size. In addition, HIV-infected participants were recruited from ongoing research projects, so they may have been comfortable with the intervention, and have good communication skills compared to persons recruited from the general population. There is also a possibility that participants provided "socially acceptable" answers about condoms, which were not validated by biological outcomes (HIV transmission), or comparison of responses between each partner in a couple.

In Context

When this study began there were no reported group-based interventions for discordant couples.(1, 2) Previous studies of partner referral for HIV and other STIs have reported low referral rates.(3, 4) Most interventions for sero-discordant couples recruit couples jointly, serving couples who may already have good communication skills. In contrast, this study required HIV-infected persons to reveal their infection to their partner and invite that partner to participate. This approach provides high-risk couples information and support outside of VCT settings.

Programmatic Implications

This study, although somewhat flawed in presentation of results and study design, indicates that group interventions for HIV discordant couples may be useful. This is because groups provide an additional measure of support beyond individual counseling, and may lessen the feelings of stigma and isolation. This study did not address issues of violence regarding disclosure, as they did not enroll anyone who reported physical violence. When social stigma is high, group discussion may not be possible. Participants indicated that individual counseling sessions, not in the presence of their partners, are also needed. When alcohol and substance abuse are prevalent, it may be difficult to implement communication skills. In these circumstances, different intervention approaches would need to be developed.


  1. Kalichman SC, Rompa D, Luke W, Austin J. HIV transmission risk behaviours among HIV-positive persons in serodiscordant relationships. Int J STD AIDS 2002 Oct;13(10):677-82.
  2. Pomeroy, EC, Green DL, Van Laningham L. Couples who care: The effectiveness of a psychoeducational group intervention for HIV serodiscordant couples. Res Soc Work Pract 2002 12(2), 238-52. (No abstract available.)
  3. Machekano R, McFarland W, Hudes E, Bassett MT, Mbizvo MT, Katzenstein D. Correlates of HIV test results seeking and utilization of partner counseling services in a cohort of male factory workers in Zimbabwe. AIDS Behav 2000 4(1), 63-70. (No abstract available.)
  4. Nuwaha F, Faxelid E, Neema S, Eriksson C, Hojer B. Psychosocial determinants for sexual partner referral in Uganda: qualitative results. Int J STD AIDS 2000 Mar;11(3):156-61.