There is evidence that voluntary HIV counseling and testing (VCT) can affect sexual behaviors, [1–4], and a multi-country trial suggested that VCT is cost-effective and efficacious in promoting behavior change, particularly in high HIV prevalence settings [3,4]. There are examples of high-quality VCT services in low- and middle-income countries , but most are located in urban centers whereas populations living in rural areas generally have limited access to VCT .
In many VCT services, HIV testing requires at least two visits to a test site which reduces returns for test results . The AIDS Information Center in Uganda reported that, prior to the implementation of rapid HIV testing, up to 25% of the VCT clients did not learn their HIV results, either because they failed to return or because their test results were not available in time [8,9]. To improve compliance, alternative approaches have been developed to make VCT more accessible including result notification and counseling by phone , rapid HIV testing [8,9,11], and home delivery of HIV test results [12,13]. Studies in rural Rakai district, Uganda, show that home delivery of results increases utilization of VCT [12,13]. In this article we examine the correlates of VCT acceptance, and prospective risk behaviors and HIV incidence among VCT acceptors and non-acceptors in a rural population cohort. The provision of VCT in this cohort differs from conventional clinic-based programs in that pretest counseling and HIV testing is provided to all consenting adults, and participants can choose to receive or not receive their results and post-test counseling.
The Rakai Health Sciences Program (previously called the Rakai Project) has conducted annual community-cohort surveillance in Rakai district, southwest Uganda since 1994 [14–16]. For this analysis, we examined persons surveyed in 1999. A household census in 43 study communities identified eligible persons and consenting adults aged 15–49 years were visited in the home and enrolled into the cohort. Subjects completed an extensive socio-demographic, behavioral and health interview by same sex interviewers. Venous blood was collected from all consenting adults and serologic diagnosis of HIV was based on two enzyme immunosorbent assays (EIA) (Vironostika HIV; Organon Teknika, Charlotte, North Carolina and Cambridge Biotech, Worcester, Massachusetts, USA) with Western blot confirmation of discordant EIA results and HIV sero-converters (HIV WB; Bio-Merieux-Vitek, St Louis, Missouri, USA).
The Rakai Health Sciences VCT program has been described elsewhere . In brief, after each interview and blood draw, participants were provided with pre-test counseling information to encourage them to accept their HIV results and to help them make an informed decision regarding individual and couples’ counseling, provided free of charge by trained Rakai Program community counselors. The Uganda Ministry of Health policy on HIV testing allows testing without a priori agreement to receive HIV results.. Participants could request VCT, as individuals or couples, either at the time of interview, or during the inter-survey period. The resident counselors provided VCT in the home or another venue of the respondent's choosing. HIV counseling services have been integrated within the existing local government health-framework and most of the community-based counseling offices are located at health units.
At the time of results notification, counselors provided post-test counseling to explain the test results, to help respondents assess their own risks of HIV infection and/or transmission, and to plan risk reduction strategies. Information was also provided on health care, nutrition and social support. To insure privacy and confidentiality, the respondent's HIV status was indicated by a confidential code known only to the counselor, HIV results were communicated verbally and no documents disclosing HIV status were retained by the client. These measures to protect privacy increased trust in the confidentiality of the counseling services. Couples were encouraged to receive their HIV results together , and those who opted not to receive couples counseling were encouraged to share their test results with their partners. There was no involuntary disclosure of HIV results to partners without prior written consent, in keeping with the Ugandan Government's HIV testing policy .
Participants were free to refuse results, and this did not affect their access to Rakai Program services such as health education, condom promotion and mobile health clinics. All study participants had free and unlimited access to the community-based counseling, and community HIV/AIDS education sessions. Drama and video shows were conducted to promote VCT, especially before marriage. Participants were free to visit the community-based counseling offices for general counseling, irrespective of their VCT participation. The Rakai community cohort study, and the VCT services were reviewed and approved by institutional review boards at the Uganda Virus Research Institute (Science and Ethics Committee), Columbia University and the Johns Hopkins Bloomberg School of Public Health.
Population and data
A total of 13 758 respondents consented to interviews and 10 694 (77.7%) provided a blood sample in the 1999 survey. Participants who gave blood and received pre-test counseling at the time of survey were eligible for free VCT services. The data presented in this paper are based on 10 618 (99.3%) participants for whom consistent and complete information was available. We excluded from the analysis 76 participants with indeterminate HIV results or with missing information on condom use or number of sexual partners 6 months prior to the survey.
To obtain information on prior VCT, study participants were asked whether they had ever previously received their HIV results. Participants were considered to have received prior VCT from the Rakai Program if study records documented their receipt of results. Individuals for whom no documentary evidence existed were categorized as ‘self-reported prior VCT recipients’. All other individuals who did not fulfill these conditions were considered not to have received prior VCT.
Participants were asked whether they had ever had sex, and if they answered in the affirmative, they were asked whether they had had sex in the 6 months prior to interview. Individuals who reported sex in the past 6 months were further asked about the number of sexual partners and whether they had used condoms during any or all of their sexual encounters. Consistent condom use was defined as use of condoms during all sexual encounters in the referent period. To assess HIV risk perception, participants were asked whether they thought they had been in any situation that could expose them to HIV infection, and responses were categorized as ‘likely’, ‘unlikely’, or ‘don’t know’.
To assess HIV acquisition and risk behaviors among VCT acceptors and non-acceptors, we assessed 6088 initially HIV-negative subjects surveyed in 1999, who provided follow up information on HIV status and risk behaviors at the subsequent annual study visit in 2000. This prospective sample constituted 67% of the HIV-negative persons observed in the 1999 survey. Losses were largely due to out-migration or absenteeism (15%), refusal to provide a subsequent blood sample for detection of HIV seroconversion (14%), or incomplete information on sexual behaviors (4%). Among the 6088 HIV-negative persons followed up in 2000, 3158 (51.9%) accepted VCT in 1999.
The proportions of persons requesting HIV results and those who ultimately received VCT were estimated using the number of those providing a blood sample as the denominator. We used cross tabulations and chi square tests (χ2) to assess the statistical significance of associations between VCT acceptance and respondents’ socio-demographic characteristics (gender, age, marital status, education), behavioral factors (HIV risk perception, number of sexual partners and condom use in the past 6 months), current HIV status in 1999, and previous receipt of HIV results. All the covariates were categorical. We used a log-binomial generalized linear model (GLM) to estimate crude and adjusted risk ratios (RR) of VCT acceptance, and 95% confidence intervals (95% CI). In the univariate analyses, we estimated crude risk ratios of VCT acceptance by all covariates. In the multivariate analysis we estimated adjusted risk ratios of VCT acceptance controlling for suspected confounders and checking for interactions. All variables with a P-value < 0.15, or with RRs greater than 2 or less than 0.5, or potential confounders identified in the univariate analyses, were included in the multivariate models.
HIV incidence among VCT acceptors and non-acceptors was estimated per 100 person-years (PY) from the number of initially HIV-negative respondents who seroconverted in 1999–2000, divided by the number of person-years of observation. Person-years were estimated as the time between two consecutive study visits, and HIV seroconvertors were ascribed half the interval of observation. All statistical analyses were done using STATA software Release 7.0 .
Requests for HIV results
Ninety three percent (9910 of 10 618) of respondents initially requested their HIV test results (Table 1). HIV result requests were higher among persons aged 15–24 years, those who had never married, persons with no formal education, HIV-negative individuals, and persons who reported no sexual partner in the past 6 months. There were no differences in result requests between females and males (93 and 94%, respectively). HIV result requests were highest among persons who had prior VCT from the Rakai Program (96.1%), and among those with self-reported VCT (95.7%), but lowest among individuals with no prior VCT (90.1%).
Receipt of HIV results and post-test counseling
Sixty-two percent (6602 of 10 618) of those who were eligible for VCT ultimately received their HIV results and post-test counseling. There were 4016 persons who requested their HIV results but did not subsequently receive VCT, of whom 48.8% (1959 of 4016) were unavailable at the time of the counselor's visit, 12.4% (499 of 4016) said they were no longer interested in receiving their results, 19.5% (783 of 4016) migrated out of the area before the counselor's visit, and 1.7% (67 of 4016) died before they received their results. In addition, 17.6% (708 of 4016) stated that they had not requested their results at the time of the survey.
As shown in Table 1, receipt of results was lower among those aged 15–24 years compared with the older age groups (P < 0.0001), but did not differ by gender. Furthermore, VCT acceptance was lower among those who had never married (55.4%), individuals with post-primary education (60.1%), and persons who reported consistent use of condoms in the 6 months prior to interview (52.1%). Acceptance of VCT was highest among the currently married (65.5%), persons with no formal education (67%), and persons who reported one sexual partner in the past 6 months (63.7%).
Figure 1 shows prior receipt of VCT and acceptance of HIV results in 1999. A high proportion of the surveyed population (47.2%) had previously received VCT from the Rakai Program (5016 of 10 618), and of these individuals, 67% (3362 of 5016) requested and received their results again in 1999. The majority of these repeat testers were HIV-negative (89.6%; 3012 of 3362). There were 772 individuals with self-reported prior receipt of VCT (7.4%), of whom 488 (63.2%) received their results in 1999, and 86.1% (420 of 488) were HIV-negative. Among 4830 persons who had no prior VCT (45.5%), 2752 (57%) accepted VCT in 1999 and 89.5% (2463 of 2752) were HIV-negative.
Table 2 shows the proportions of persons accepting VCT by source of prior VCT and current HIV status. Overall, 64.7% (5895 of 9016) of HIV-negative and 46.8% (707 of 1512) of HIV-positive individuals received their HIV results in 1999 (P < 0.001). VCT acceptance was consistently higher among HIV-negative than HIV-positive individuals irrespective of prior receipt or non-receipt of results. VCT acceptance was lowest among persons with no prior VCT irrespective of HIV status.
Table 3 shows the crude and adjusted risk ratios of VCT acceptance by socio-demographic, behavioral characteristics, and prior knowledge of HIV status. In the adjusted analysis, marital status was significantly associated with increased VCT acceptance (adj. RR = 1.14; 95% CI, 1.08–1.20, for currently married; and adj. RR = 1.11; 95% CI, 1.04–1.18, for previously married, versus never married). The probability of VCT acceptance was significantly lower among those with primary education (adj. RR = 0.94; 95% CI, 0.90–0.99) and post-primary education (adj. RR = 0.91; 95% CI, 0.87–0.97) compared with those with no formal education. VCT acceptance was lower among persons who reported condom use in the past 6 months (inconsistent users, adj. RR = 0.95; 95% CI, 0.90–0.99; consistent users, adj. RR = 0.88; 95% CI, 0.82–0.95), in persons who had no prior VCT (adj. RR = 0.88; 95% CI, 0.85–0.90), and among HIV-positive persons irrespective of their previous receipt of results (adj. RR = 0.72; 95% CI, 0.68–0.76). VCT acceptance was not significantly associated with gender, age, number of sexual partners, or self-perception of HIV risk.
Prospective behavioral change and HIV acquisition
Table 4 shows the number of sexual partners and condom use reported by initially HIV-negative participants in 1999 and at the 2000 survey. HIV acquisition during follow-up is also shown. There were no significant differences in risk behaviors reported by persons accepting or declining VCT in 1999, or any significant change in risk behaviors during follow-up, among persons who accepted or refused VCT in 1999. Furthermore, there was no significant difference in HIV incidence among VCT acceptors (1.6/100 PY) and non-acceptors (1.4/100 PY, P = 0.6).
These data show that 62.2% of those who were eligible for VCT ultimately received their HIV results and post-test counseling despite free and accessible services. The main reason for non-receipt of results was absenteeism at the time of the counselor's visit (48.8%) and migration out of the area. The average interval between phlebotomy and receipt of results was 1 month, and all participants were offered their results within 3 months. The counselors made three or four call-back visits to absentees , and appointments were scheduled at venues convenient to the participant. Thus, it is unlikely that non-receipt of results was due to limitations of access.
Acceptance of VCT was significantly lower among those who had never married and the better-educated persons with no prior VCT, particularly if they were HIV-infected (Table 1). Acceptance was also lower in individuals who reported use of condoms in the 6 months prior to interview. There was no reduction in risk behaviors observed after receipt of VCT, and HIV acquisition was comparable among persons who received VCT and those who declined the service (Table 4).
Among participants who accepted VCT, 58.3% had previously received their HIV results, and 89.1% of these repeat acceptors were HIV-negative (Table 2 and Fig. 1). It is likely that this high rate of repeat VCT reflects a desire for reassurance among HIV-uninfected persons. However, focus group discussions suggest that HIV-negative repeat testers may assume they have been fortunate in their choice of partners, or believe themselves to be immune to HIV, and these misconceptions may lead to persistent risk behaviors as indicated in Table 4. There is, therefore, an urgent need to correct these misapprehensions through more intensive risk reduction counseling.
The finding that VCT acceptance was lower among persons with no prior VCT is disappointing, given the presumed benefits associated with knowledge of HIV serostatus [1–4]. An earlier evaluation of the Rakai VCT program in 1994–1995 found that high-risk groups were less likely to participate in VCT , and this is reflected in the lower acceptance rates among HIV-infected persons in 1999, irrespective of whether they had prior VCT (Table 2) Spielberg et al.  suggest that fear of receiving a positive HIV test result is one of the barriers to seeking VCT. Although persons who considered themselves likely to be at risk of HIV had relatively high rates of VCT acceptance (60.1%), the low rate of VCT acceptance among persons who actually were HIV-positive (46.8%) suggest that fear or denial might be important barriers to VCT acceptance, and there is need for interventions to assuage these concerns.
Receipt of VCT was significantly higher in ever-married persons, which is consistent with previous research in Rakai , and might reflect the impact of programs encouraging couples counseling . In addition, married persons are less mobile and easier to follow up for counseling and result receipt than the never married. For example, only 5.8% of those who were currently married had migrated out of the area by the time of the counselors’ visit, but among those who had never married, the proportion who had migrated out was 11.3%.
We did not find any significant difference in acceptance of VCT between males and females. Other studies have reported that fewer women accepted VCT, and fear of partner violence or marital dissolution have been cited as deterrents to uptake of VCT by women [19,21,22]. In Rakai, women may have overcome such barriers, and this suggests that access to free, home-based VCT could increase female participation, even in a male-dominated society in which the man is the main decision-maker. Surprisingly, we found that persons with education had a significantly lower VCT acceptance rates compared to those with no education (Table 3), which suggests a need to encourage VCT among the better-educated participants. The finding that condom users were less likely to accept VCT (Table 3) may reflect higher risk behaviors among condom users  who fear that they are HIV infected. There is, therefore, need for targeted approaches which enhance condom users’ personalized assessment of their HIV risk, and encourage their acceptance of VCT.
The high rates of VCT acceptance in this cohort have been attributed to the use of community-based resident counselors [12,13]. However, use of resident counselors and home delivery of results may not be possible or appropriate in all situations. For example, in Kabarole district, western Uganda, study participants preferred that HIV counselors should not be community residents due to concerns over confidentiality of HIV results notification . Furthermore, Fylkesnes et al. advocated the use of non-resident counselors in community settings, because it is assumed that there is less risk of breach of confidentiality. However, the Rakai Program has provided home delivery of results since 1990 [26,27], and although Program counselors live within the communities where they work, the Program has avoided breach of confidentiality by intensive counselor training and supervision, and by the strategy of providing results verbally, so as not to leave documentation of HIV status accessible to third parties. This may contribute to community trust in the Rakai home-based counseling service and explain why the majority of the study participants now request home counseling .
It is discouraging that we found no effect of VCT acceptance on subsequent risk behaviors or on HIV incidence in HIV-negative persons (Table 4). A separate analysis of the effects of repeat VCT on sexual risk behavior and HIV incidence among initially HIV-negative subjects in Rakai also found no significant impact of VCT on sexual risk-behavior and HIV incidence among persons who received VCT in comparison with those who had never received VCT . Rakai Program counseling messages are modeled on those of the other major HIV VCT programs in Uganda [the AIDS Information Center (AIC) and The AIDS Support Organization (TASO)]; Program counselors received training at AIC and TASO; and counseling staff receive regular supervision and retraining, so it is unlikely that the Rakai results reflect a lower quality of services in comparison with those in other Ugandan programs. It should be noted that the Rakai Program is unique in having follow-up data on behaviors and HIV incidence for both VCT acceptors and on non-acceptors, whereas most VCT programs do not have information on behaviors among non-VCT recipients. Other studies have also found a lack of impact of VCT on behaviors in HIV-negative individuals. In a US study of injection drug users (IDUs), knowledge of HIV serostatus motivated transmission-reducing behaviors among HIV-positive, but not among HIV-negative IDUs . A meta-analysis of 27 studies suggested that VCT reduced unprotected intercourse in HIV-discordant couples and in HIV-infected individuals, but had no effect on risk behaviors reported by HIV-negative VCT recipients . Randomized trials of VCT have observed reductions in self-reported risk behaviors, but not in HIV incidence [1,3]. Collectively, these findings suggest that VCT may have limited efficacy in promoting behavior change among uninfected recipients.
In conclusion, our findings in this community-based VCT program in rural Uganda suggest selective uptake of services by persons who are less likely to be HIV infected, high rates of retesting among HIV-negative persons which may be associated with a false sense of security, and a lack of impact of VCT on risk reduction and HIV incidence among initially HIV-uninfected individuals. There is, therefore, a need for improved strategies to promote risk reduction in HIV-negative persons to promote primary HIV prevention.
We would like to thank Dr S.K. Ssempala, the Director of Uganda Virus Research Institute, for his support.
Sponsorship: This study was supported by grants RO1 A134826 and RO1 A13426S from the National Institute of Allergy and Infectious Diseases; grant 5P30HD06826 from the National Institute of Child Health and Development; Grant HMJF 5686/CFDA#12.420 from the Henry M. Jackson Foundation for the Advancement of Military Medicine and the United States Department of Defense, and The Johns Hopkins University, Bill and Melinda Gates Institute of Population and Reproductive Health.
1. Kamb ML, Fishbein M, Douglas JM, Rhodes F, Rogers J, Bolan G, et al
. Efficacy of risk-reduction counseling to prevent Human Immunodeficiency Virus and sexually transmitted diseases. JAMA 1998; 280:1161–1167.
2. UNAIDS. UNAIDS Policy on HIV Testing and Counseling
. New York Joint United Nations Program on HIV
/AIDS; August 1997.
3. The Voluntary HIV
-1 Counseling and Testing Efficacy Study Group. Efficacy of voluntary HIV-1 counseling and testing in individuals and couples in Kenya, Tanzania, and Trinidad: a randomized trial
4. Sweat M, Gregorich S, Sangiwa G, Furlonge C, Balmer D, Kamenga C, et al
. Cost-effectiveness of voluntary HIV
-1 counseling and testing in reducing sexual transmission of HIV
-1 in Kenya and Tanzania. Lancet 2000; 356:113–121.
5. UNAIDS. Report on the Global HIV/AIDS Epidemic
. New York Joint United Nations Program on HIV
/AIDS; July 2002.
6. Killewo JZ, Kwesigabo G, Comoro C, Lugalla J, Mhalu FS, Biberfeld G, et al
. Acceptability of voluntary HIV
testing with counseling in a rural village in Kagera, Tanzania. AIDS Care 1998; 10:431–439.
7. UNAIDS. Report on the Global HIV/AIDS Epidemic
. New York Joint United Nations Program on HIV
/AIDS; December 1997.
8. Downing RG, Otten RA, Marum E, Biryahwaho B, Alwano-Edyegu MG, Sempala SD, et al
. Optimizing the delivery of HIV
counseling and testing services: The Uganda
experience using rapid HIV
antibody test algorithms. J Acquir Immune Defic Syndr 1998; 18:348–388.
9. UNAIDS. Knowledge is Power: Voluntary HIV Ccounseling and Testing in Uganda
. New York Joint United Nations Program on HIV
/AIDS; June 1999.
10. Tsu RC, Burm ML, Gilhooly JA, Sells CW. Telephone versus face-to-face notification of HIV
results in high-risk youth. J Adolesc Health 2002; 30:154–160.
11. Keenan PA, Keenan JM. Rapid HIV
testing in Urban Outreach: a strategy for improving posttest counseling rates. AIDS Educ Prev 2001; 13:541–550.
12. Matovu JK, Kigozi G, Nalugoda F, Wabwire-Mangen F, Gray RH. The Rakai
Project counseling programme experience. Trop Med Intl Health 2002; 7:1064–1067.
13. Matovu J. Coverage of voluntary counseling and testing in a rural population-based cohort, Rakai, Uganda
. XIV International Conference on AIDS
, Barcelona, 2002 [abstract MoPeF4021].
14. Wawer MJ, Sewankambo NK, Serwadda D, Quinn T, Paxton LA, Kiwanuka N, et al
. Control of sexually transmitted diseases for AIDS prevention in Uganda
: a randomised community trial. Lancet 1999; 353:525–535.
15. Wawer MJ, Gray RH, Sewankambo NK, Serwadda D, Paxton LA, Berkley S, et al
. A randomized community trial of intensive sexually transmitted disease control for AIDS prevention, Rakai
. AIDS 1998; 12:1211–1225.
16. Gray RH, Wabwire-Mangen F, Kigozi G, Sewankambo NK, Serwadda D, Moulton LH, et al
. Randomized trial of presumptive sexually transmitted disease therapy during pregnancy in Rakai
. Am J Obstet Gynecol 2001; 185:1209–1217.
17. AIDS Control Program (ACP). HIV Testing Policy
. Entebbe: Ministry of Health; October 1992.
18. Anon. STATA Statistical Software
, Release 7.0. College Station, Texas: STATA Corporation; 2001.
19. Nyblade LC, Menken J, Wawer MJ, Sewankambo NK, Serwaadda D, Makumbi F, et al
. Population-based HIV
counseling and testing in rural Uganda
: Participation and risk characteristics. J Acquir Immune Defic Syndr 2001; 28:463–470.
20. Spielberg F, Kurth A, Gorbach PM, Goldbaum G. Moving from apprehension to action: HIV
counseling and testing preferences in three at-risk populations. AIDS Educ & Prev 2001; 13:524–540.
21. Maman S, Mbwambo J, Hogan NM, Kilonzo GP, Sweat M. Women's barriers to HIV
-1 testing and disclosure: Challenges for HIV
-1 voluntary counseling and testing. AIDS Care 2001; 13:595–603.
22. Temmerman M, Ndinya-Achola J, Ambani J, Piot P. The right not to know HIV
-test results. Lancet 1995; 345:969–970.
23. Ahmed S, Lutalo T, Wawer MJ, Serwadda D, Sewankambo NK, Nalugoda F, et al
incidence and sexually transmitted diseases prevalence associated with condom use: a population study in Rakai
. AIDS 2001; 15:2171–2179.
24. Kipp W, Kabagambe G, Konde-Lule J. HIV
counseling and testing in rural Uganda
: Communities’ attitudes and perceptions towards an HIV
counseling and testing programme. AIDS Care 2002; 14:699–706.
25. Fylkesnes K, Haworth A, Rosenvard C, Kwapa PM. HIV
counseling and testing: over-emphasizing high acceptance rates a threat to confidentiality and the right not to know. AIDS 1999; 13:2469–2474.
26. Lianjo B, Wawer MJ, Lutalo T, Sewankambo N, Kelly R. Use of HIV counseling in rural Uganda
. Tenth International Conference on AIDS
, Yokohama, 1994 [abstract Pd0734].
27. Konde-Lule JK, Musagara M, Musgrave S. Focus group interviews about AIDS in Rakai
district of Uganda
. Soc Sci Med 1993; 37:679–684.
28. Matovu JK, Kigozi G, Nalugoda F. Repetitive VCT, sexual risk behavior and HIV incidence, Rakai, Uganda
. Paper presented at the Uganda
Virus Research Institute (UVRI), Entebbe, Uganda
. November 28, 2003.
29. Ouellet LJ, Bailey SL, Thorpe LE. Effect of knowing HIV serostatus on risk behaviors among injection drug users in Chicago
. Paper presented at the 128th Annual Meeting of APHA, 14 November 2000 [abstract 9466]. Retrieved from: http://apha.confex.com/apha/128am/techprogram/paper_9499.htm
on 9 March 2004.
30. Weinhardt LS, Carey MP, Johnson BT, Bickham NL. Effects of HIV
counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985–97. Am J Public Health 2000; 90:1152–1153.