We found in the Voluntary HIV-1 Counselling and Testing Efficacy Study, a randomized clinical trial (N = 4293) testing the efficacy of HIV voluntary counselling and testing (VCT) for prevention in three developing countries, that HIV VCT was highly efficacious in reducing sexual risk behavior [1,2]. The study also demonstrated that HIV VCT is cost-effective, especially when directed towards those who are most likely to engage in risk behavior and those infected with HIV-1 , and that those who presented themselves for HIV VCT services are at a higher risk relative to the general population (J. Killewo, C. Furlonge, G. Sangiwa, N.M. Hogan, S.E. Gregorich, T.J. Coates, in preparation) .
Enthusiasm for the widespread promotion of HIV VCT in developing countries is tempered, however, by concerns about the possibility of the negative social consequences of HIV VCT . General concern has been expressed about the level of stigma associated with HIV testing, regardless of the test result, as well as concern about stigma against HIV-seropositive individuals. More specifically, several reports have [6–9] documented a high level of violence against women in Africa, and have suggested a link between violence and HIV infection. There have been calls for additional research on the potential negative consequences of HIV VCT, as well as to examine the potential positive psychosocial outcomes of VCT . To assess the overall impact of HIV VCT, it is important to assess both the occurrence of negative life events that indicate stigma against those receiving HIV VCT, as well as positive life events that indicate positive social outcomes after testing.
The role of disclosure in life event outcomes after HIV VCT has been addressed in the United States and in a few studies in developing countries. Findings in the USA suggested that, despite fears of negative responses, targets of disclosure were generally supportive. High levels of disclosure were reported and negative consequences were rare [10–12]. In developing countries, however, the few studies that have been conducted [5,6,13,14] show lower rates of disclosure and negative outcomes of disclosure for women.
In this paper, we describe the incidence of positive and negative life events among participants in the Voluntary HIV-1 Counselling and Testing Efficacy Study, and explore the impact of treatment group assignment, HIV serostatus and disclosure of serostatus on the incidence of positive and negative life events.
The Voluntary HIV-1 Counselling and Testing Efficacy Study was conducted at three sites: Nairobi, Kenya; Dar es Salaam, Tanzania; and Port of Spain, Trinidad. These sites were chosen to represent a variety of epidemiological and cultural contexts in which voluntary HIV-1 counselling and testing could occur, all within developing countries. All sites were urban; results should be generalized to rural settings with caution. Overall 4293 participants were enrolled. Participants enrolled as couples or as individuals. The intended study population was individuals and couples seeking HIV-related services at each site. Participants were not expected to comprise a random sample of the study site cities. Data were collected from 1995 to 1998. Participants were consented, completed the baseline interview and were then randomly assigned to receive either HIV counselling and testing (HIV VCT) or health information (HI), the comparison intervention. Couples were randomly assigned together so that both couple members always received the same intervention.
The HIV VCT intervention was based on the client-centered HIV counselling model of the United States Centers for Disease Control and Prevention  and the guidelines of the World Health Organization Global Programme on AIDS . This individualized model is also ideal for cultural specificity; at each site counsellors were encouraged to take into account the specific cultural and environmental context of counselling at their site. Participants enrolling as couples were counselled together or individually by their choice; some individual time with the counsellor was given to ensure accurate risk assessment. Couples were informed before HIV testing of the expectation that they would share their test results. Blood was collected and tested by commercial enzyme-linked immunosorbent assay (ELISA). Reactive (positive) samples were confirmed using a second ELISA; if the second ELISA was negative the sample was tested by Western blot. Test results were given individually first, and then the couple was encouraged to share their test results in the counselling session. All participants were instructed in correct condom use, provided with a supply of condoms, and instructed to return for additional condoms as needed. Participants assigned to the HI intervention at baseline watched a locally produced 15 min videotape including HIV/AIDS information and a condom demonstration, and then had their questions answered by a health educator. HI participants were not tested at baseline. The first follow-up was scheduled for 6 months after baseline and was completed an average of 7.3 months after baseline. All participants were offered HIV VCT after the first follow-up survey.
This report focuses on the period of time between baseline and the first follow-up, and reports findings from the first follow-up survey. A detailed description of the Voluntary Counselling and Testing Efficacy Study design and methods has been published previously .
Structured face-to-face interviews were used at baseline and at the first follow-up to assess HIV risk behavior in the 2 months before the interview. The baseline interview was completed before randomization; at each follow-up the interview was conducted before receiving services. Couples were interviewed individually, and it was stressed that their responses would not be revealed to their partners. The baseline survey assessed demographic information, including age, relationship status, income, level of schooling, and religion. Interviewers needed to be blinded to the baseline serostatus of participants during the follow-up interview; therefore, several series of questions assessing life events and disclosure of HIV serostatus were deferred from the interview and ascertained later by a counsellor. The treatment group (VCT or HI) refers to the treatment group the participant was randomly assigned to at baseline. Enrollment status refers to whether the participant was enrolled as an individual or as a member of a couple.
Life events were assessed by asking a series of questions about the occurrence of specific events since the participant enrolled in the study. Positive life events included strengthening of sexual relationships and increased emotional support from family, peers, health providers or employers. Negative life events included break-up of a marriage, break-up of a sexual relationship, physical abuse by a sexual partner, neglect by family, being disowned by family, being estranged by peers, and being discriminated against by health professionals or employers. The proportion of participants reporting each life event was calculated after eliminating those who said that life events were not applicable to them (e.g. some participants were not married so could not experience the break-up of a marriage); there was a ‘not applicable’ option for each life event question. Three negative life events were not included in the multivariate analysis because they occurred too infrequently: being estranged by peers (1% overall); being discriminated against by health professionals (1%); and being discriminated against by employers (1%).
Disclosure was assessed by a series of questions about the disclosure of serostatus to their spouses, sexual partners, parents, children, brothers, sisters, other relatives, friends, landlord, neighbors, religious leaders, community leaders, physicians and employers. Again, respondents were able to indicate ‘not applicable’ if they did not have such a person to disclose to. For analyses examining the relationship of disclosure with the occurrence of positive and negative life events, we included only a composite variable indicating disclosure to any spouse or sexual partner.
First we show the demographic characteristics of the entire baseline sample (N = 4293) separately for those who enrolled as individuals and those who enrolled as members of a couple. Next, a series of logistic regression models describe the relationship between positive and negative life events and (i) treatment group, (ii) serostatus, and (iii) the disclosure of serostatus. When appropriate, logistic regression models accounted for intracouple response correlation via GEE estimation .
The effect of treatment group (HIV VCT versus HI) on the occurrence of positive and negative life events was assessed using a series of logistic regression models. Models were fit for each life event, separately for individuals and couples. Explanatory variables included treatment group assignment, site of recruitment and respondent age. These analyses included the entire baseline sample.
The effect of baseline HIV serostatus on subsequent life events was also assessed using a series of logistic regression models. Models were fit separately for individuals and couples. Explanatory variables included baseline serostatus, site of recruitment and respondent age. For individuals, serostatus was simply each respondent's baseline test result. For couple members the serostatus variable indicated one of four types of couple serostatus: concordant HIV positive, concordant HIV negative, serodiscordant with HIV-positive man and serodiscordant with HIV-positive woman. These analyses included everyone in the baseline sample who was assessed at follow up.
The effect of serostatus disclosure to one or more sexual partners on life events was assessed with a series of logistic regression models. Models were fit separately for individual men, individual women, male couple members, and female couple members. Explanatory variables included serostatus disclosure, baseline serostatus, disclosure by serostatus interaction, site of recruitment and respondent age. As a result of sparse data, individual serostatus was used in these models instead of couple serostatus. These analyses included individuals and couple members who were assigned to HIV VCT at baseline, tested for HIV, received their test result, participated in follow-up and who reported having a sexual partner to disclose to.
Table 1 presents the demographic characteristics of the 3120 participants who enrolled as individuals and the 1173 who enrolled as members of couples. We found no significant differences between the 4293 participants who enrolled at baseline and the 3551 (82.7%) who returned at follow-up . By design, study participants were evenly divided between men and women and between the three study sites in Kenya, Tanzania and Trinidad. Couples were targeted for recruitment, and almost 30% of participants enrolled as couples. The median age of participants was 27 years (interquartile range 22–33 years).
Throughout the analyses there was a consistent effect of site. Participants at the study site in Tanzania were more likely to report each of the positive life events than participants at the other sites. There were no significant effects of site for the occurrence of negative life events. There was also a consistent overall effect of age; older respondents were less likely than younger respondents to report positive and negative life events.
Effect of treatment group on the occurrence of life events
Table 2 presents the overall occurrence of each positive and negative life event by treatment group (HIV VCT versus HI) among those who returned at the first follow-up. Positive life events were relatively common, ranging from strengthening of a sexual relationship (42% overall) to increased emotional support from employers (12% overall). Negative life events were rare. The most common negative life event was the break-up of a sexual relationship (27% overall). The break-up of a marriage was reported by 5% of participants, and physical abuse by a sexual partner was reported by 4.5% of participants. Other negative life events were reported by 0–4% of participants.
There were no significant differences between treatment groups in the occurrence of positive life events and few significant differences in the proportion reporting negative life events. Couple members assigned to HIV VCT at baseline were significantly more likely than couple members assigned to HI to report being neglected or disowned by their families (3 versus 1%, P < 0.01). Among those who enrolled as individuals, there was a trend suggesting that those who were assigned to HI at baseline were more likely to report being neglected or disowned by their family (4 versus 2% of those who received HIV VCT at baseline, P < 0.06).
Effect of HIV serostatus on the occurrence of life events
At baseline 14.6% of participants who were assigned to HIV VCT and tested were found to be HIV seropositive (13.6% of individual and 17.6% of couple members).
Table 3 shows the occurrence of life events by serostatus for individuals and for couples. HIV-seronegative individuals were more likely to report the positive life events strengthening of a sexual relationship (43 versus 20%, P < 0.0001) and increased emotional support from peers (28 versus 18%, P < 0.01). HIV-seropositive individuals were more likely to report the positive life event of increased emotional support from health professionals (41 versus 23%, P < 0.05), and more likely to report the negative life events break-up of a marriage (9 versus 3%, P < 0.05) and being neglected or disowned by their family (5 versus 2%, P < 0.05).
Participants in seroconcordant negative couples were most likely (56%) and participants in serodiscordant female-positive couples were least likely (18%) to report the strengthening of a sexual relationship (P < 0.005). Participants in serodiscordant female-positive couples also tended to be the most likely to report the break-up of a sexual relationship (P < 0.08). Models predicting three of the negative life events, however, could not be estimated as a result of sparse data. For these life events (break-up of a marriage, physical abuse and being neglected or disowned by family), we collapsed couple serostatus to two categories: seroconcordant and serodiscordant (not tabled). There were no significant differences between serodiscordant and seroconcordant couples in the occurrence of physical abuse or being neglected or disowned by their family. For the break-up of a marriage, however, the difference approached significance, with 13% of serodiscordant and 4% of seroconcordant couples reporting the break-up of a marriage (P < 0.09).
Effect of serostatus disclosure on the occurrence of life events
Of the 1463 participants who were tested at baseline and returned for the first follow-up, 76% disclosed their serostatus to a sexual partner (70% of those who enrolled as individuals and 91% of those who enrolled as members of a couple). Overall, participants who tested HIV seropositive at baseline were less likely than participants who tested negative at baseline to disclose their serostatus to a sexual partner (52 versus 79%, P < 0.0001).
Table 4 shows the impact of serostatus disclosure on life events for men and for women who enrolled as individuals. Among men, both HIV-seropositive and HIV-seronegative men who disclosed were more likely to report the strengthening of a sexual relationship compared with men who did not disclose. Among women, however, HIV-seronegative women who disclosed tended to be more likely to report the strengthening of a sexual relationship (P < 0.07); but there was no difference among HIV-seropositive women. There were trends suggesting that HIV-seropositive women who did disclose and HIV-seronegative women who did not disclose were the most likely to report the break-up of a sexual relationship. Twenty-five per cent of HIV-seronegative women who did not disclose and 26% of seropositive women who did disclose reported the break-up of a sexual relationship compared with 14% of HIV-seronegative women who did disclose and 19% of HIV-seropositive women who did not disclose (P < 0.07).
Table 5 shows the impact of serostatus disclosure for participants who enrolled as members of couples. All effects were either not significant or the data were too sparse to estimate the impact of serostatus disclosure; this analysis included all participants who enrolled as members of couples, were tested at baseline, received their test results and returned at follow-up (188 men and 188 women).
Individuals who were assigned to HIV VCT at baseline were not significantly more likely to experience negative life events compared with individuals who enrolled in the study and were assigned to HI at baseline. Our findings suggest that low levels of stigma were associated with HIV testing and counselling. The only trend that even approached significance was in the percentage of individuals who reported being neglected or disowned by their families (4 versus 3%), although the incidence was very low. This suggests that just having HIV VCT was not in itself stigmatizing; it did not result in an increased occurrence of negative life events compared with a group of participants comparable in demographic characteristics, HIV risk behavior and motivation to receive HIV VCT. For couples there was again a low incidence of being neglected or disowned by their family, but that difference (3 versus 1%) was significant. Couple members who received HIV VCT were also more likely to experience emotional support from peers. Overall, the results suggest that couples receiving HIV VCT together were more vulnerable than individuals to both positive and negative effects of HIV VCT compared with receiving HI only.
Among individual participants, those who tested positive for HIV-1 were significantly less likely to report the strengthening of a sexual relationship and increased emotional support from peers, and were more likely to report the break-up of a marriage and being neglected or disowned by their family. The only positive life event that was more common among seropositive individuals was increased emotional support from health professionals, which would be expected as those who tested seropositive in the study were referred for any available medical care. Among couples, we had the opportunity to look at couple serostatus as opposed to the serostatus of couple members individually. Findings indicated that participants in seroconcordant negative couples were most likely to report the strengthening of a sexual relationship after VCT, followed by those in seroconcordant seropositive couples, serodiscordant female-negative couples and serodiscordant female-positive couples. We also found that participants in serodiscordant female-positive couples were most likely to report the break-up of a sexual relationship. This combination of findings points to a particular vulnerability among participants in serodiscordant female-positive couples, and suggests that these couples be given specific attention in post-test and subsequent counselling with the goal of mitigating potential negative consequences. Although it has not been empirically proved that the provision of additional couple counselling sessions, counselling for HIV-positive women in serodiscordant couples and referrals for ongoing social support services would mitigate negative consequences, our results suggest that it would be prudent to offer additional resources as available. Unfortunately, we were not able to estimate models including couple serostatus for the break-up of a marriage, physical abuse or being neglected or disowned by their family because of sparse data. We were also not able to stratify the couple serostatus analysis by sex.
Overall, participants reported a high level of serostatus disclosure. That 91% of couple members disclosed their serostatus to a sexual partner may be the result of a counselling protocol that promoted disclosure. Disclosure had a positive effect (strengthening of a sexual relationship) among both HIV-seropositive and HIV-seronegative men and among HIV-seronegative women, suggesting that disclosure is less likely to improve relationships among HIV-seropositive women. Both seronegative women who did not disclose and HIV-seropositive women who did disclose were more likely to report the break-up of a sexual relationship. Whereas the relatively higher rate of break-up of a sexual relationship among seropositive women who disclosed would be expected, the higher rate of break-up of a sexual relationship among HIV-seronegative women who did not disclose is surprising. It could be that women who test HIV seronegative choose to leave relationships for the purpose of risk reduction. Unfortunately, as a result of sparse data, we were not able to estimate the disclosure by serostatus interactions among couple members. Although some of the differences across subgroups in Table 5 appear to be relatively large, in almost all cases small sample sizes did not allow inferences to be drawn, and the reader should avoid over-interpretation of the data presented.
One limitation of our study was not collecting baseline rates of positive and negative life events. Overall, however, we found that positive life events were much more common than negative life events at follow-up. By far the most commonly reported negative life event was the break-up of a sexual relationship. Ending a sexual relationship may have served some participants as a risk reduction strategy. This hypothesis is supported by the fact that nearly a third of participants enrolled as individuals reported breaking up a sexual relationship. The break-up of a marriage, by comparison, was rare.
This study is unique in three respects. First, we examined both positive and negative life events after HIV VCT rather than negative life events only. Because this study was a clinical trial, we were also able to compare the occurrence of positive and negative life events in a comparable group who had not received HIV VCT. Finally, our sample included both men and women, and allowed comparison of participants who received HIV VCT as individuals and as couples.
Overall,our results suggest that attending HIV prevention services does not result in high rates of negative life events. Taken together with the results of the Voluntary HIV-1 Counselling and Testing Efficacy Study [1,2], the findings support the dissemination of HIV VCT services. Our findings also suggest that certain groups are vulnerable to negative life events, and encourage the provision of additional support and counselling services to couples, particularly serodiscordant couples with an HIV-seropositive female partner. Additional qualitative research with couples attending HIV VCT services would be useful to determine the most effective means of promoting such additional support services and their efficacy in preventing negative life events.
*Participants in the Voluntary HIV-1 Counselling and Testing Efficacy Study Group
Center for AIDS Prevention Studies, University of California San Francisco: Thomas J. Coates, Olga A. Grinstead, Steven E. Gregorich, David C. Heilbron, William P. Wolf, Kyung-Hee Choi, Julius Schachter, Peter Scheirer, Ariane van der Straten.
AIDSCAP, Family Health International: Munkolenkole C. Kamenga, Michael D. Sweat, Isabelle De Zoysa, Gina Dallabetta.
Global Programme on AIDS, WHO and UNAIDS: Kevin R. O'Reilly, Eric van Praag, David Miller, Monica Ruiz, Samuel Kalibala, Ben Nkowane.
Kenya Association of Professional Counsellors and University of Calgary: Donald Balmer, Francis Kihuho, Stephen Moses, Frank Plummer.
Muhimbili Medical College, University of Dar es Salaam: Gloria Sangiwa, Margaret Hogan, Japhet Killewo, Davis Mwakigile.
Queens Park Counselling Centre, Trinidad, West Indies: Colin Furlonge.
1. Sangiwa G, Balmer D, Furlonge C, Grinstead O, Kamenga M, Coates T. Voluntary HIV counseling & testing (VCT) reduces risk behavior in developing countries: results from the voluntary counseling and testing study.12th World AIDS Conference
. Geneva, Switzerland, 1998.
2. The Voluntary HIV-1 Counselling and Testing Efficacy Study Group. Efficacy of Voluntary HIV-1 counseling and testing in individual and couples in Kenya, Tanzania, and Trinidad: a randomised trial. Lancet 2000, 356: 103 –112.
3. Sweat M, Gregorich S, Sangiwa G. et al.. Cost-effectiveness of voluntary HIV-1 counseling and testing in reducing sexual transmission of HIV-1 in Kenya and Tanzania: the Voluntary HIV-1 Counselling and Testing Efficacy Study. Lancet 2000, 356: 113 –121.
4. Furlonge C, Gregorich SE, Kalibala S, Grinstead OA, Coates TJ, O'Reilly KR. HIV-related risk factors in a population-based probability sample of north and central Trinidad. AIDS Behav 2000, 4: 49 –62.
5. Karim Q, Karim S, Soldan K, Zondi M. Reducing the risk of HIV infection among South African sex workers: socioeconomic and gender barriers. Am J Public Health 1995, 85: 1521 –1525.
6. Keogh P, Allen S, Almedal C, Temahagili B. The social impact of HIV infection on women in Kigali, Rwanda: a prospective study. Soc Sci Med 1994, 38: 1047 –1053.
7. van der Straten A, King R, Grinstead O, Serufilira A, Allen S. Couple communication, sexual coercion and HIV risk reduction in Kigali, Rwanda. AIDS 1995, 9: 935 –944.
8. Coker A, Richter D. Violence against women in Sierra Leone: frequency and correlates of intimate partner violence and forced sexual intercourse. Afr J Reprod Health 1998, II: 61 –72.
9. Maman S, Campbell J, Sweat M, Gielen A. The intersections of HIV and violence: directions for future research and interventions. Soc Sci Med 2000, 50: 459 –478.
10. Wolitski R, Rietmeijer C, Goldbaum G, Wilson R. HIV serostatus disclosure among gay and bisexual men in four American cities: general patterns and relation to sexual practices. AIDS Care 1998, 10: 599 –610.
11. Simoni J, Mason H, Marks G, Ruiz M, Reed D, Richardson J. Women's self-disclosure of HIV infection: rates, reasons, and reactions. J Consult Clin Psychol 1995, 63: 474 –478.
12. Mansergh G, Marks G, Simoni J. Self-disclosure of HIV infection among men who vary in time seropositive diagnosis and symptomatic status. AIDS 1995, 9: 639 –644.
13. Temmerman M, Ndinya-Achola J, Ambani J, Piot P. The right not to know HIV-test results. Lancet 1995, 345: 969 –970.
14. Baingana G, Choi K, Barrett D, Byansi R, Hearst N. Female partners of AIDS patients in Uganda: reported knowledge, perceptions and plans. AIDS 1995, 9 (Suppl. 1) : S15 –S19.
15. DHHS. HIV prevention counseling: a training program.
Centers for Disease Control and Prevention; 1993.
16. World Health Organization. Global Programme on AIDS, Division of Technical Cooperation and Health Care Support
. Guidelines for implementing HIV/AIDS counseling
. September, 1993.
17. The Voluntary HIV Counselling and Testing Study Group. The voluntary HIV-1 counselling and testing efficacy study: design and methods. AIDS Behav 2000, 4: 5 –14.
18. Diggle PJ, Liang K-Y, Zeger SL. Analysis of longitudinal data.
New York: Oxford University Press; 1994.