UNAIDS and the World Health Organization (WHO) have called for universal access to HIV prevention, care, and treatment in sub-Saharan Africa . Positive prevention that aims to support HIV-infected persons to limit HIV transmission through behavioral and medical interventions has been recommended [2–4] but has not been widely implemented in Africa. Transmission risk from HIV-infected persons varies with the stage of infection and viral load , use of antiretroviral therapy (ART) , extent of risky sexual behavior , and the presence of other sexually transmitted infections (STIs) . Several positive-prevention interventions have proven effective, including HIV-testing for individuals  and couples , consistent condom use by HIV-infected individuals , and on-going prevention with positive interventions delivered in routine care and service settings . To develop effective positive-prevention approaches for Africa, however, an understanding of transmission risk behaviors within specific HIV-infected populations is essential.
Transmission risk behavior by HIV-infected persons has been described in a few African settings, [13–15] but these were not nationally representative. To inform the development of effective prevention interventions, we assessed factors associated with risky sexual behavior, including knowledge of one's own and one's partners' HIV sero-status, among HIV-infected adults who participated in the 2004–2005 Uganda HIV/AIDS Sero-Behavioral Survey (UHSBS).
The UHSBS was a nationally representative, population-based survey involving 15–59-year-old adults. Data were collected on behavioral, social, and demographic indicators. Blood samples were obtained for testing for HIV, herpes simplex virus type 2 (HSV-2) and hepatitis B. Individual test results were anonymously linked to individual and household questionnaires through bar-coded identification numbers. Details of the sampling design, survey implementation, data management, and laboratory procedures are presented elsewhere . Participants provided separate informed consent for interviews and blood sampling. Survey protocols were approved by the Uganda National Council of Science and Technology and the Centers for Disease Control and Prevention (CDC) and by the Institutional Review Boards of the Uganda Virus Research Institute and ORC Macro.
The survey collected information from 9529 households in 417 sample enumeration areas. A total of 11 454 women and 9905 men aged 15–59 years were eligible for interviews and blood sample collection. Of these, individual questionnaires were completed for 95% of women and 89% of men, and blood specimens were collected for 90% of women and 84% of men. From households where married or cohabitating spouses were both present and interviewed, we analyzed behavioral and demographic information at the couple level.
HIV testing was performed using two HIV enzyme immunosorbent assays (EIA), based on different antigens. Specimens with equivocal or discordant test results were retested with the same EIAs, and if warranted were resolved by western blot testing. For quality control, all positive specimens and 5% of negative specimens were retested in a different laboratory using the same testing algorithm. For HSV-2 testing, the Kalon HSV Type 2-specific IgG assay was used; this was a recombinant type 2 antigen (gG2) modified to eliminate reactivity arising from HSV type 1 infection, at the same time retaining the natural antigenic characteristics of HSV-2. The assay is based on the indirect enzyme-linked immunosorbent assay (ELISA) principle. A representative subsample was also tested for hepatitis B core antibodies (HBcAb) by EIA, indicating a lifetime infection; HBcAb-positive specimens were further tested by EIA for hepatitis B surface antigen, indicating current hepatitis B infection.
Participants with laboratory confirmed HIV-infection were categorized as knowing their HIV status if they reported that they had been tested and received their results. Participants were categorized as either knowing all their partners' HIV status, some of their partners' HIV status or not knowing any of their partners' HIV status. For multivariate analyses, we compared participants who knew the HIV status of all their partners with those who knew for only some or none of their partners. Spouses were defined as married or cohabitating partners, steady partners as noncohabitating and nonmarital regular partners, and casual partners as infrequent or nonregular partners.
HIV sexual transmission risk behavior was defined as having unprotected sex at last sex encounter with any partner in the previous 12 months. Participants reporting having a previous spouse who died were classified as having experienced previous widowhood. Those who reported having an STI or STI-related symptom in the previous 12 months were categorized as having any STI. Finally, those who knew there were drugs that could prolong the life of people living with HIV and who knew ART by name were categorized as having ART knowledge.
We assessed gender differences in demographic, knowledge, and behavioral characteristics. Five separate multivariate analyses were conducted to assess factors associated with knowing one's HIV status, with being sexually active, with using condoms, with having nonmarital partners, and with using contraception. All models were adjusted for age, sex, religion, education, occupation, and urban or rural residence. All analyses are weighted to account for the sampling and survey design. Statistical analysis was conducted using SAS version 9.1 (SAS Institute, Cary, North Carolina, USA). Full results of the condom use multivariate model are presented, but for all other models, only factors independently associated with each outcome are presented in the text (full results available).
Characteristics of HIV-infected adults in Uganda
Of 18 525 Ugandans aged 15–59 years who had HIV serological testing, 1092 (5.9%) were HIV-infected, which after weighting gave a national estimate of 6.3%. Of HIV-infected adults, the median age was 33 years (36 for men and 31 for women, P < 0.001) and 64% were female. The majority of HIV-infected persons (57%) were currently married, 36% previously married (20% widowed and 16% divorced or separated), and only 7% had never been married. Although 77% of HIV-infected persons resided in rural areas and 63% had no or incomplete primary level education, 52% were in the two highest wealth quintiles (Table 1).
Of all HIV-infected people, 954 (82%) had knowledge of drugs that can prolong the life of an HIV-infected person, 132 (11%) could name ART specifically, 986 (85%) knew that a healthy looking person could be HIV-infected, 87% knew about vertical HIV transmission, and 69% knew about drugs to reduce vertical transmission. Only 21% understood that HIV discordance within a couple was possible. Men were more likely than women to have knowledge of HIV discordance (24 vs. 18%, P = 0.015).
Of HIV-infected respondents, 116 (10%) were ill for three or more of the previous 12 months, 25% reported having a genital sore or ulcer in the last year and 37% had genital discharge. A total of 48% reported having any STI or STI-related symptom in the last year; 69% of these sought treatment at a health facility. Overall, 84% were serologically positive for HSV-2 infection. Of those tested for Hepatitis B, 223(59%) had a positive Hepatitis B core antibody (HBcAb) and 31(14%) a positive Hepatitis B surface antigen (HBsAg). Of HIV-infected men, 18% were currently circumcised. Of all HIV-infected persons, 46 (6%) women and 5 (1%) men had had a blood transfusion in their lifetime.
Knowledge of HIV status
Of all HIV-infected persons, 267 (23%) had an HIV test done, with a median of two tests among those who had tested (range, 1–10), and overall 21% had received their test results. Of those tested, 97 (41%) had been tested within the previous year, 46 (17%) within the previous 12–23 months and 101(42%) within the previous 2 years or more. Reasons for never having tested included low perceived risk (27%), cost of testing or distance (23%), not knowing where to get a test (22%), not wanting to know if HIV-infected (16%), testing not being a priority (9%), and having no knowledge about HIV tests (8%) (Table 1). Overall, 100 (9%) of those who had ever had sex stated that they knew all their partners' HIV status and 2% knew some of their partners' HIV status. Of those who knew their HIV status and had a current partner, 86% reported having disclosed their status to their partner or spouse.
Knowledge of HIV status was related to illness; 33% of those with illness in the last year knew their status compared with 20% of those without serious illness (P = 0.002). Similarly, ART knowledge among HIV-infected people was overall 11%, 12% among those who had been ill in the previous year and 27% among those who knew their HIV status.
HIV-infected persons were significantly more likely to know their own HIV status if they were female [adjusted odds ratio (AOR), 1.9; 95% confidence limits (CI), 1.2–3.0], in the higher or highest wealth quintiles (AOR, 4.0; CI, 1.7–9.5 and AOR, 5.1; CI, 2.2–11.9), chronically ill in the last year (AOR, 2.7; CI, 1.6–4.8), had ART knowledge (AOR, 3.4; CI, 2.0–5.9), had ever been married (currently married AOR, 4.1; CI, 1.0–16.6, widowed AOR, 7.6; CI, 1.8–31.9, or divorced AOR, 4.4; CI, 1.0–19.2), or knew their partner's HIV status (AOR, 6.2; CI, 1.7–22.5 for those who knew the status of some partners and AOR, 4.9; CI, 2.8–8.7 for those who knew the HIV status of all their partners).
Overall, 99% of HIV-infected persons had had sex and 77% were sexually active in the last year, including 86% of men and 72% of women In multivariate analysis, being married was the strongest factor associated with current sexual activity (AOR, 13.8; CI, 6.3–29.9). Persons who were at least 35 years were less likely than 15–24-year-old persons to be sexually active (AOR, 0.2; CI, 0.1–0.4) as were women (AOR, 0.5; CI, 0.3–0.7) and persons with chronic illness (AOR, 0.5; CI, 0.3–0.9). These findings were similar even when the analysis was stratified by currently married or not-currently married.
Of all sexually active HIV-infected adults, 80% reported only one sexual partner in the previous year. Within the subset of married HIV-infected persons, 86% reported having had sex only with their spouses in the last year, including 75% of men and 96% of women (P < 0.001). Of all married HIV-infected persons, 13% reported only one sexual partner in their life (1% of men and 23% of women, P < 0.001). Of the 81% of HIV-infected married persons who did not understand that HIV discordance was possible within couples, 92% did not know the HIV status of their spouse's. Factors independently associated with having partners outside of marriage in the previous year included being male (AOR, 8.2; CI, 3.5–19.3) and having used a condom (AOR, 1.9; CI, 1.0–3.7).
Sexual activity of widowed, divorced, and separated HIV-infected persons
History of widowhood was overall 43% (N = 344) and 49% for women. Among those who were previously widowed, 63% of men and 19% of women had remarried. Of those remarried, 92% of men and 95% of women had a spouse of unknown HIV status. Of the 67 (20 women, 47 men) remarried individuals with confirmed laboratory diagnoses for their spouses, 48% of men and 57% of women had an HIV-negative spouse. Widowed, separated, and divorced HIV-infected persons, although representing only 36% of the total HIV-infected population, accounted for 51% of those who had unprotected sex with a nonmarital partner.
Of last sex acts of HIV-infected persons, 83% were unprotected and 84% of these took place with married and cohabitating partners, 13% with steady partners and 3% with casual partners. Among those who had had sex, 67% reported never having used a condom. Of nonusers, 49% said their reason for not using a condom at last sex was that they trusted that their partners were not infected; of these, 87% were unaware of their own HIV infection and only 9% knew all their partners' HIV status. Other reasons for not using a condom included: not liking condoms (18%), condoms not being accessible (5%), partner insisted on not using condoms (13%), and lack of knowledge about condoms (7%).
Currently married HIV-infected persons were much less likely to report using condoms than those who had never married (AOR, 0.1; CI, 0.0–0.2), and urban dwellers were twice as likely to report using condoms than their rural counterparts (AOR, 2.0; CI, 1.3–3.0). The odds of condom use among those who had had an HIV test were three times as high as among untested survey participants (AOR, 3.0; CI, 1.9–4.7). Likewise, the odds of condom use among those who knew their partners' HIV status were 2.3 times as high as among those who did not know their partners' HIV status (AOR, 2.3; CI, 1.2–4.3). Those with knowledge of ART reported higher rates of condom use than those who did not know about ART (30 vs. 14%) but this difference was not statistically significant in multivariate analysis (Table 2).
Pregnancy risk behavior
Of HIV-infected women, 24% were currently using contraception. Of these, 78% used hormonal injections, 23% oral contraceptives, and 8% had a tubal ligation. HIV-infected women not using contraception had a median of three children and 67% were sexually active. In multivariate analysis, HIV-infected women using contraception were more likely to live in urban areas (AOR, 2.7; CI, 1.6–4.5), more likely to have had three or more children (AOR, 47.8; CI, 3.8–608.7) and more likely to be less than 35 years of age (AOR, 5.0; CI, 1.67–10.0). At the time of their last pregnancy, 237 (89%) of HIV-infected women attended an antenatal clinic (ANC), 56% received education about AIDS transmission from mother to child, 50% received information about HIV prevention, 73% about family planning and 16% were tested for HIV during an ANC visit. When asked about their last pregnancy, 49% of HIV-infected women said their last pregnancy was unplanned.
In a nationally representative sample of HIV-infected Ugandan adults, 79% did not know their HIV status and 91% did not know their partner's status. Of the 77% who were sexually active, 84% had unprotected sex at their last encounter; nearly all were with spouses or steady partners. The odds of reporting condom use among those who knew their HIV status were three times as high as among those who did know their status, and among those who knew their partners' status, the odds were two times as high. Among chronically ill HIV-infected individuals, two-thirds had never been tested. Finally, despite fears about ART-associated risk compensation , we found no evidence that knowledge of ART was associated with risk behavior, and ART programmes in Uganda have been associated with decreased transmission risk . These findings suggest that interventions to support HIV-infected persons to learn their own and their partners' status could increase access to care [18,19] and treatment  and reduce HIV transmission in Uganda.
The UHSBS did not collect information on the frequency of unprotected sex by partner type, and we had to assess transmission risk based on condom use at last sex encounter rather than by proportions within total sex acts by HIV-infected persons. Therefore, we probably underestimated the proportion of risk taking place within cohabiting partnerships, in which sex is more frequent. Also, our measure of knowledge of HIV status did not ascertain whether people knew they were HIV-infected and some may have reported on testing prior to seroconversion. Thus, we may have underestimated the effect of knowing one's HIV-positive status. Our measure of knowledge of partner status did not distinguish between knowledge based on assumed concordance or on partner's disclosure of test results. Therefore, we may have underestimated the effect of knowledge of partner's status in our couple's analysis.
Despite these limitations, our findings suggest several intervention priorities. Most risk behavior occurred within cohabiting couples. HIV testing by HIV-infected persons  and partner testing within discordant couples has been shown to decrease transmission risk [10,21,22]. Widespread implementation of the Ugandan Ministry of Health and WHO HCT policy  recommending routine and provider-initiated HIV counseling and testing should occur. The HIV-infected population should be encouraged to disclose their HIV status to partners, to reduce unprotected sex with nonconcordant partners, and, for widowed persons, to test partners prior to remarriage. Although condom use reduces transmission risk , 67% of HIV-infected persons in Uganda had never used a condom, highlighting the importance of condom promotion. Routine STI screening and treatment for HIV-infected persons and their partners might reduce HIV transmission risk in Uganda. Acyclovir for the 84% population with HSV-2 coinfection could reduce transmission within HIV-discordant couples ; trials are currently evaluating this. In addition, unmet need for contraception was high and family planning services should be integrated into care and treatment programs. Finally, HIV testing, prevention of mother-to-child transmission, and contraceptive service utilization was especially low for the rural and low-income, HIV-infected population, highlighting the need for expanded service access for these disadvantaged groups.
Nearly 25 years after HIV was first recognized in Uganda , a country well known for its HIV interventions, only 21% of HIV-infected people know their status, and 9% know their partners' status. Many other countries in sub-Saharan Africa also have low HIV testing coverage and are missing critical opportunities for positive prevention . Expansion of HIV counseling and testing is urgently needed in sub-Saharan Africa, not only for increasing access to care and treatment, but also to support HIV-infected persons to reduce transmission.
The authors thank all participants in the UHSBS and recognize the contributions of the Ugandan Ministry of Health and collaborator staff who conducted data collection, laboratory testing, and data management for the UHSBS. Support for this anaylsis was provided by the Emergency Plan for AIDS Relief in Uganda.
The findings and conclusions in this paper are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
1. UNAIDS. Intensifying HIV prevention (UNAIDS policy position paper). UNAIDS/05.18E. Geneva, Switzerland; 2005, pp. 1–54.
2. Janssen RS, Holtgrave DR, Valdiserri RO, Shepherd M, Gayle HD, De Cock KM. The serostatus approach to fighting the HIV epidemic: prevention strategies for infected individuals. Am J Public Health 2001; 91:1019–1024.
3. Centers for Disease Control and Prevention. Incorporating HIV prevention into the medical care of persons living with HIV. Recommendations of CDC, the Health Resources and Services Administration, the National Institutes of Health, and the HIV Medicine Association of the Infectious Diseases Society of America
. MMWR Recomm Rep
4. International HIV/AIDS Alliance. Positive prevention: prevention strategies for people with HIV/AIDS. Available at: International HIV/AIDS Alliance. Positive prevention: prevention strategies for people with HIV/AIDS; July 2003, pp. 1–93.
5. Quinn TC, Wawer MJ, Sewankambo N, Serwadds D, Li C, Wabwire-Mangen F, et al
. Viral load and heterosexual transmission of human immunodeficiency virus type 1. N Engl J Med 2000; 342:921–929.
6. Castilla J, del RJ, Hernando V, Marincovich B, Garcia S, Rodriguez C. Effectiveness of highly active antiretroviral therapy in reducing heterosexual transmission of HIV. J Acquir Immune Defic Syndr 2005; 40:96–101.
7. Gray RH, Li X, Wawer MJ, Gange SJ, Serwadda D, Sewankambo NK, et al
. Stochastic simulation of the impact of antiretroviral therapy and HIV vaccines on HIV transmission; Rakai, Uganda. AIDS 2003; 17:1941–1951.
8. Cohen MS. Sexually transmitted diseases enhance HIV transmission: no longer a hypothesis. Lancet 1998; 351(Suppl 3):5–7.
9. Marks G, Crepaz N, Senterfitt JW, Janssen RS. Meta-analysis of high-risk sexual behavior in persons aware and unaware they are infected with HIV in the United States: implications for HIV prevention programs. J Acquir Immune Defic Syndr 2005; 39:446–453.
10. Allen S, Tice J, van de Perre P, Serufilira A, Hudes E, Nsengumurenyi F, et al
. Effect of serotesting with counselling on condom use and seroconversion among HIV discordant couples in Africa. BMJ 1992; 304:1605–1609.
11. Weller S, Davis K. Condom effectiveness in reducing heterosexual HIV transmission. Cochrane Database System Review
12. Crepaz N, Lyles CM, Wolitski RJ, Passin WF, Rama SM, Herbst JH, et al
. Do prevention interventions reduce HIV risk behaviours among people living with HIV? A meta-analytic review of controlled trials. AIDS 2006; 20:143–157.
13. Olley BO, Seedat S, Gxamza F, Reuter H, Stein DJ. Determinants of unprotected sex among HIV-positive patients in South Africa. AIDS Care 2005; 17:1–9.
14. Nakayiwa S, Abang B, Packel L, Lifshay J, Purcell DW, King R, et al
. Desire for children and pregnancy risk behavior among HIV-infected men and women in Uganda. AIDS Behav 2006; 2:437–448.
15. Kiene SM, Christie S, Cornman DH, Fisher WA, Shuper PA, Pillay S, et al
. Sexual risk behaviour among HIV-positive individuals in clinical care in urban KwaZulu-Natal, South Africa. AIDS 2006; 20:1781–1784.
16. Ministry of Health (MOH) [Uganda] and ORC Macro. Uganda HIV/AIDS sero-behavioral survey 2004–2005
. Calverton, Maryland, USA: Ministry of Health and ORC Macro; 2006.
17. Bunnell R, Ekwaru JP, Solberg P, Wamai N, Bikaako-Kajura W, Were W, et al
. Changes in sexual behavior and risk of HIV transmission after antiretroviral therapy and prevention interventions in rural Uganda. AIDS 2006; 20:85–92.
18. Mermin J, Lule J, Ekwaru JP, Downing R, Hughes P, Bunnell R, et al
. Cotrimoxazole prophylaxis by HIV-infected persons in Uganda reduces morbidity and mortality among HIV-uninfected family members. AIDS 2005; 19:1035–1042.
19. Mermin J, Bunnell R, Lule J, Opio A, Gibbons A, Dybul M, et al
. Developing an evidence-based, preventive care package for persons with HIV in Africa. Trop Med Int Health 2005; 10:961–970.
20. Weidle PJ, Malamba S, Mwebaze R, Sozi C, Rukundo G, Downing R, et al
. Assessment of a pilot antiretroviral drug therapy programme in Uganda: patients' response, survival, and drug resistance. Lancet 2002; 360:34–40.
21. Allen S, Meinzen-Derr J, Kautzman M, Zulu I, Trask S, Fideli U, et al
. Sexual behavior of HIV discordant couples after HIV counseling and testing. AIDS 2003; 17:733–740.
22. Kamenga M, Ryder RW, Jingu M, Mbuyi N, Mbu L, Behets F, et al
. Evidence of marked sexual behavior change associated with low HIV-1 seroconversion in 149 married couples with discordant HIV-1 serostatus: experience at an HIV couselling center in Zaire. AIDS 1991; 5:61–67.
24. Celum CL, Robinson NJ, Cohen MS. Potential effect of HIV type 1 antiretroviral and herpes simplex virus type 2 antiviral therapy on transmission and acquisition of HIV type 1 infection. J Infect Dis 2005; 191(Suppl 1):S107–S114.
25. Serwadda D, Mugerwa RD, Sewankambo NK, Lwegaba A, Carswell JW, Kirya GB, et al
. Slim disease: a new disease in Uganda and its association with HTLV-III infection. Lancet 1985; 2:849–852.
26. Bunnell R, Mermin J, De Cock KM. HIV prevention for a threatened continent: implementing positive prevention in Africa. JAMA 2006; 296:855–858.