Uganda was one of the earliest countries to experience a generalized AIDS epidemic [1–3]. Beginning in the late 1980s, the government, non-governmental organizations, and the international community launched an aggressive public health response to AIDS [4,5]. Perhaps as a result of these efforts, substantial declines in HIV prevalence have been observed since the mid-1990s . HIV remains common, however, across all age groups . Since the late 1980s, HIV/AIDS surveillance has provided data to monitor the epidemic and inform control programmes [2,8]. Surveillance has relied largely on the testing of antenatal clinic (ANC) attendees, supplemented by cohort studies [9,10] and cross-sectional surveys. These data, however, are usually confined to estimates of HIV prevalence, risk factors for infection, or the natural history of HIV infection. To inform HIV/AIDS control programmes that provide care and treatment to an increasing number of people living with HIV/AIDS (PLHA), estimates about HIV incidence, disease burden and the corresponding needs are essential. Whereas the efficacies of most biomedical interventions to prevent infection, morbidity, or mortality are well understood, their comparative effect and impact on the burden of disease over time on a population level is difficult to gauge. As in most resource-constrained settings, the health information system in Uganda records only a fraction of HIV-related events; in such circumstances software-based models and projections may be an acceptable substitute.
A recent HIV/AIDS serobehavioral survey provided accurate HIV prevalence estimates representative of Uganda's adult population . We used these and other data to estimate the burden of HIV, as well as the resulting need for and the potential effects of selected interventions on HIV/AIDS in Uganda.
HIV prevalence estimates
We used ANC-based HIV prevalence data (1989–2002) derived from unlinked anonymous testing of leftover blood collected routinely from pregnant women for syphilis screening (22 sites in 2002), following international guidelines . In 2004–2005, Uganda conducted a population-based and nationally representative HIV/AIDS Sero-Behavioral Survey (UHSBS) in which 18 525 participants were tested for HIV . ANC and UHSBS-based prevalence data were entered into the UNAIDS/WHO Estimation and Projection Package [EPP, 2005 edition, version 2.0e English (c)] software to obtain a modelled prevalence estimate over time. EPP is a software package recommended by the UNAIDS Reference Group for Modeling, Estimates, and Projections to model HIV prevalence over time . For the years 1981–2004, we accepted the official prevalence curve generated by the Ministry of Health's AIDS Control Programme. Briefly, ANC data were fitted using EPP for both rural and urban populations and then calibrated according to the UHSBS-based prevalence estimates for 2004 using the calibration feature in EPP. The adult (15–49 years) HIV prevalence curves for urban (HIV prevalence, 2004: 10.1%) and rural (5.7%) Uganda were then merged to obtain a weighted national HIV prevalence curve. This weighted HIV prevalence curve suggested a prevalence of 0.02% in 1981, rising to 12.3% in 1992, and declining thereafter to 6.4% in 2004. For the years 2005–2010 we kept the inputed HIV prevalence constant at 2004 levels. We thus refrained from speculating about any future change in prevalence, which allowed us to describe better the net effect of HIV/AIDS interventions. Of note is the fact that a stable prevalence does not necessarily imply a stable incidence, but merely a balance between AIDS-related mortality and new HIV infections. Only as an output did we allow HIV prevalence to be affected by interventions, recognizing, e.g., that antiretroviral therapy (ART) decreases AIDS mortality.
Using Spectrum (version 2.39 Beta2) we estimated the burden of HIV disease from the epidemic's onset, concentrating on the years 2005–2010. Spectrum is a modular program examining the consequences of HIV prevalence projections and selected interventions at the population level [14,15]. Table 1 lists several key input parameters used. We utilized country-specific data and estimates when available. These included UHSBS-based age and sex-specific HIV prevalences for 2005 (for 5-year age groups from 15 to 59 years) , the proportions of 15–19-year-old women never married (76.4%) and of married women in monogamous unions (67.8%) , as well as census data for total population size and fertility estimates . When country-specific estimates were unavailable, we accepted United Nations Population Division or default Spectrum parameters (Table 1). The Spectrum default determination for ART needs was used, i.e. HIV-infected individuals require ART 2 years before death and one year before AIDS onset. Children were defined as aged under 15 years. Orphans were defined as having lost one or both parents and aged 0–17 years.
Using Spectrum we estimated key output parameters including HIV incidence, number of PLHA, and HIV-related deaths from the epidemic's onset in 1981 to 2010. We estimated the effect of selected interventions using existing coverage estimates (2000–2006) or planned (usually 2007–2010) programme coverage data according to the Ugandan Ministry of Health [21,22]. We also estimated the effect of these interventions for each percentage increase in coverage, and the maximum possible effect with a hypothetical 100% coverage. While estimating the net effect of each intervention, we kept all other interventions at their estimated (2000–2006) or planned (2007–2010) coverage. Table 2 provides selected coverage assumptions over time for pediatric and adult ART coverage, pediatric co-trimoxazole provision and the prevention of mother-to-child HIV transmission (PMTCT). We defined the need for pediatric co-trimoxazole provision as all HIV-exposed children until age 18 months and continuing for HIV-infected children thereafter. PMTCT was defined as the administration of single-dose nevirapine for both the pregnant mother and newborn, further assuming an average of 20+ months of breastfeeding [23,24] and no substantial or nationwide formula-feeding programme in place. We used the actual 2004 and 2005 uptake of nevirapine by pregnant mothers and assumed equal uptake by their newborn babies. To account for the difference between HIV-infected ANC clients who were offered PMTCT and those who actually received nevirapine, we applied the 2004 and 2005 ratios of nevirapine uptake over PMTCT coverage (availability of PMTCT services) to the estimated/planned PMTCT coverage for the years 2000–2003 and 2006–2010. In doing so, nevirapine uptake rose steadily from 0.6% (2000) to 12% (2005) and is planned to be 48% by 2010. We assumed that all interventions mentioned above had negligible coverage before 2000. Using Microsoft Excel, we also evaluated the potential effect of co-trimoxazole prophylaxis for HIV-infected adults, assuming a reduction in adult AIDS-related mortality of 46% [16,17] both in ART-naive and ART-exposed individuals. Adult co-trimoxazole prophylaxis was scaled up beginning in 2004, but no reliable coverage data for 2004 and 2005 and no planning figures for 2006–2010 were available. We rounded all estimates to the nearest 100 or 10 and refer to the year 2005 unless stated otherwise. For key estimates, we conducted a sensitivity analysis, exploring the effect of a rise in HIV prevalence from 6.4 to 8.4% by 2010 or conversely, a fall to 4.4%.
The estimated course of the HIV epidemic in Uganda from its onset in 1981 to 2010 is shown in Fig. 1. Population growth (3.2% per year) alone indicates that the number of PLHA will increase by 197 600 from 2005 to 2010, and by 2008 may reach the projected historic peak of 1.08 million PLHA in 1994.
Prevalent HIV infections
Estimates for Uganda's HIV/AIDS burden in 2005 are displayed in Table 3. An estimated 915 400 PLHA were living in 2005. The 95% confidence intervals (CI) for the national prevalence estimate correspond to 842 700 and 1 002 600 PLHA. The majority of PLHA were rural (79.8%, 775 400) or female (58%). In the UHSBS survey, 19.8% (women 22.8%; men 15.6%) of HIV-positive respondents aged 15–59 years reported having been tested and counselled before. Applying these sex-specific proportions to all adult PLHA, a minimum of 647 000 were unaware of their HIV infection. This estimate represents the lower limit, as some HIV-positive survey respondents may have tested HIV negative at their previous testing occasion. The number of HIV-infected children (0–14 years) nationwide was estimated at 109 000; this estimate was based on mother-to-child transmission and survival probabilities as well as the reported nevirapine uptake (Table 1).
Adult HIV incidence
An estimated 135 300 individuals (56% women) became infected with HIV in Uganda in 2005 (Table 3). This corresponded to an overall adult (15–49 years) HIV incidence of approximately 0.96 infections/100 person-years. The Spectrum-provided plausibility bounds for adult HIV incidence were 0.55 and 1.64/100 person-years. The female-to-male incidence rate ratio among 15–49 year-olds was 1.5 and peaked at 38 among 15–19 year-olds (female HIV incidence 0.94/100 person-years; male 0.02/100 person-years). For women, new HIV infections probably peaked among 15–24 year-olds, for men, among 25–34 year-olds.
Pediatric HIV incidence and the effect of the prevention of mother-to-child HIV transmission
Of the 135 300 new HIV infections in 2005, approximately 14% were caused by mother-to-child transmission. Of the estimated 1.34 million pregnancies leading to live births, 102 800 pregnancies were HIV positive and led to 19 600 vertical infections, corresponding to 1.47% of all births. Without PMTCT, this proportion would have been 1.53%, whereas a 100% uptake of nevirapine would have lowered it further to 1.04% and prevented an additional 5700 vertical transmissions. The nevirapine uptake in 2005 (12%) probably prevented 780 vertical HIV infections (Table 2) and an additional 1030 infections during 2000–2004 when the estimated nevirapine uptake increased from 0.6 to 8.1%. The planned expansion to 80% PMTCT coverage (48% nevirapine uptake) may prevent 13 000 infections during 2005–2010, lowering the neonatal HIV prevalence to 1.19% by 2010. The existing ART programme, covering an estimated 27 500 women of reproductive age (2005), probably averted a similar number of vertical infections (840, 4%) and is expected to avert a total of 10 800 vertical infections during 2005–2010.
AIDS cases, deaths, and orphans
For the year 2005, we estimated 88 100 individuals living with AIDS and 76 400 AIDS deaths. The cumulative total of all HIV-related deaths from 1981 to 2005 was 1.37 million, or 16.1% of all deaths during that time period, and is projected to reach 1.77 million by 2010. The HIV-related mortality in 2005 for all ages combined was estimated at 3.01/1000 (adults 4.62/1000; children aged 0–4 years 1.91/1000). With an overall under-5 mortality rate of 128.7/1000 in Uganda, the bulk of under 5 mortality is due to non-HIV related causes. Considering the effect of HIV-related premature deaths in 2005, the overall life expectancy was estimated at 48.9 years, compared with a hypothetical 56.9 years in the absence of HIV. With AIDS deaths occurring at an average age of 32 years, HIV-related deaths in 2005 caused a loss of 1.90 million life-years (distributed over 2005 and future years). In 2005, HIV-related deaths accounted for 19.6% of all deaths, for 6.2% of deaths in children under 5 years, for 33.5% among adults 15 years and older, and for 45.1% among adults aged 15–59 years.
The number of HIV and non-HIV orphans combined reached 2.18 million in 2005, and 46.4% (1 009 345) of these were caused by HIV. The number of dual orphans, i.e. children having lost both parents, was estimated at 567 700 (26.1% of all orphans); among these, 80.9% were caused by HIV.
Effect of co-trimoxazole prophylaxis
The existing coverage of co-trimoxazole prophylaxis to HIV-seropositive infants in 2005 (15%) prevented an estimated 820 deaths (Table 2). The total number of infants requiring co-trimoxazole prophylaxis was estimated at 209 300. The full potential of this intervention was estimated as preventing 13 000 deaths in 2005 and a total of 54 800 during 2005–2010. This compares favorably with a total of 64 000 pediatric AIDS deaths avertable through a combined ART/co-trimoxazole programme, suggesting a 5-year co-trimoxazole programme efficacy of 86% compared with a full-scale ART programme alone. As the effect of co-trimoxazole in postponing mortality is not as long-lasting as ART, however, this efficacy declines over time (data not shown). Even in the presence of the existing ART programme, a (hypothetical) full-scale co-trimoxazole prophylaxis programme for all adult PLHA beginning in 2005 may avert an estimated 55 800 deaths to 2010. The number of averted deaths through adult co-trimoxazole prophylaxis would be highest in the first year (28 600) followed by smaller mortality reductions in subsequent years (2010: 1300 deaths), indicating that most co-trimoxazole-averted deaths are postponed for relatively few years, reflecting a 46% efficacy in reducing mortality that is well below 100%. We estimated the potential number of adult life-years saved through such a programme during 2005–2010 at 283 300. The provision of co-trimoxazole for all HIV-infected adults from 2005 onwards would probably also reduce the number of AIDS orphans, by as much as 26 600 in 2005, and 1000 by 2010, resulting in approximately 44 500 orphan life-years saved during 2005–2010.
Antiretroviral therapy programme needs and effects
In 2005, an estimated 194 900 PLHA required ART (adults: 160 400; children: 34 500). Therefore, the existing ART treatment programme in Uganda was covering approximately 34.6% (67 000) of the total need in 2005. The estimated ART coverage was lower for children (14.5%) than for adults (38.9%). As the ART programmes expand to cover approximately 160 000 PLHA in 2010, the number needing ART likewise will rise to 271 400 (Table 2), increasing coverage to 59%. Table 2 (Adult ART – effect on adult AIDS deaths) illustrates some of the likely effects of Uganda's adult ART programme until 2010. Whereas adult ART coverage is expected to more than double to approximately 150 000, the number of individuals requiring but not receiving ART will only modestly decline from 98 000 to 90 700. During the same time period, the number of adult AIDS deaths will essentially remain unchanged between 2005 (62 100) and 2010 (63 000). Across all age groups, the absolute number of PLHA requiring but not receiving ART will decline only modestly from 127 600 (2005) to 111 100 (2010) in the presence of the expanding ART programme.
The ART coverage prevented an estimated 14 400 AIDS deaths of all ages in 2005, 22 600 since the programme inception in 2000, and may prevent a total of 78 800 deaths during the entire decade until 2010. These averted deaths will raise adult HIV prevalence by 0.4% in 2010. Although a hypothetical full-scale ART programme from 2005 onwards would see AIDS deaths drop sharply to 11 600 deaths in 2005, it would nevertheless increase again to 42 700 by 2010. ART also has population-level effects on the orphan burden in Uganda. During 2000–2005, ART probably averted 32 400 AIDS orphans. A fully expanded ART programme (100% coverage) could reduce the number of future orphans by 356 600 by 2010.
Table 4 lists estimates for selected indicators (identical to those shown in Table 3) for the year 2010, exploring three different adult HIV prevalence scenarios: remaining stable at 6.4%, falling to 4.4% or rising to 8.4% (further small changes in HIV prevalence caused by the expanding ART programme are not shown here but were accounted for). A changing prevalence would involve substantial increases or decreases in new HIV infections, HIV incidence, HIV-positive pregnancies and live births. In contrast, little changes can be expected until 2010 with respect to the long-term consequences of HIV disease, including AIDS cases and deaths, ART needs, and orphans. The estimated total number of PLHA in 2010 varies substantially between 0.74 (corresponding to an HIV prevalence of 4.4%) and 1.34 million (HIV prevalence 8.4%).
By triangulating HIV prevalence estimates with census and survey data, transmission probabilities, and mortality assumptions, we estimated the HIV-related burden of disease in Uganda and projected it forward to 2010. Although UNAIDS regularly publishes key HIV-related estimates for individual countries including Uganda , our work attempted to do so in greater detail, and examined the likely effects of selected biomedical interventions on a population level. A quarter century after the HIV epidemic took hold in Uganda, AIDS remains a leading cause of adult disease and death in the country. Our data suggest that every day in 2005, an estimated 370 individuals became HIV infected (54 through vertical transmission), approximately 290 became ill enough to require ART initiation, and 210 died of AIDS. Although current prevention and treatment programmes have a measurable effect on the burden of disease, they may not substantially alter the face of the epidemic in the near future. Despite the massive scale-up of ART programmes, intensified prevention efforts to slow adult HIV transmission are warranted to halt or reverse the expansion of Uganda's HIV epidemic, as the prevention of adult HIV transmission remains key in minimizing the burden of HIV in Uganda. Persisting HIV incidence combined with high population growth make a goal of universal ART access very difficult to achieve, suggesting that family planning services should be seen as an integral part of HIV control efforts in Uganda.
In the absence of health information systems that fully capture HIV-related events, modelling such events with tailored software packages may fill some of these information gaps. We used standard methodologies and official estimates whenever available, thus these estimates may be considered reasonable approximations of the true burden of HIV disease in Uganda. Such analyses and projections may also appear feasible for other countries with similar data availability and planning needs and can greatly help in understanding the population-level dynamics of the epidemic and the effects of selected interventions. The limitations of ANC data for use in surveillance are well known  and were partly overcome through calibration with the more precise general population-based 2004/2005 prevalence estimate while leaving the suggested trends intact. We kept HIV prevalence constant from 2004 onwards and thus did not speculate on the future course of the epidemic, which in turn facilitated showing the net effect of biomedical interventions over time. Our sensitivity analysis indicates that although a change in prevalence by 2% over 5 years implies substantial changes in the number of acute events (e.g. incident HIV infections), long-term outcomes (e.g. adult morbidity and mortality) would not change dramatically during this time period. Only continued surveillance will, however, show the true trend in HIV prevalence. Other limitations in our work include our inability to examine the effect of behavior-related interventions such as counselling and testing, information and education campaigns, abstinence, or condom use. Also, some estimates shown here did not consider interactions with other programmes, such as non-HIV-related excess deaths as a result of formula feeding.
The goal of this exercise was to provide a comprehensive range of estimates useful for policy guidance and programme planning. Our data indicate that even if incidence remains below 1% annually, approximately 135 000 Ugandans contract HIV annually. This estimate is double the number of PLHA who took ART in 2005. These annual cohorts of newly infected individuals will require care and treatment in years to come, adding strain to an already overstretched healthcare system. Furthermore, our incidence estimates are based on the assumption of a stable HIV prevalence. Any change in incidence could substantially alter the future burden of HIV disease in Uganda.
Clinical trial data that provided relatively precise estimates of the efficacy of nevirapine allowed us to project the impact of the existing PMTCT programme and its planned expansion. Nevertheless, our estimates are likely to be volatile as the number of vertical infections is a function of many factors, including actual transmission rates, breastfeeding parameters and PMTCT programme characteristics. Given Uganda's high fertility rate, the relative simplicity of the current PMTCT intervention and the number of life-years and subsequent ART costs saved per averted vertical infection, this programme holds huge potential for both the HIV-infected mother and child. Our data indicate that currently only a fraction of preventable infections are realized and that mother-to-child transmission-related infections accounted for 14% of all new HIV infections in 2005. PMTCT programmes not only prevent pediatric HIV infections but also deliver testing and counselling services, and direct HIV-infected pregnant women (and their partners) to prevention-with-positives, care and treatment programmes, warranting their full expansion. Of note is the fact that the existing ART programme in 2005 probably prevented more vertical infections than the nevirapine-based PMTCT programme. The recent policy recommendation by the Ugandan Ministry of Health to use short-course ART  may further increase the number of vertical infections averted. Family planning should also be regarded as an extended tool for PMTCT and, in the long run, to curb the overall burden of HIV disease in Uganda. Antenatal and postnatal healthcare services in particular should be better utilized to help avert any unwanted future pregnancies.
At least 647 000 (∼80%) of HIV-infected adults were unaware of their infection in 2005, indicating that many HIV-infected Ugandans remain unexposed to counselling and testing, prevention-with-positive education, as well as care and treatment programmes. This gap in HIV case detection calls for a rapid expansion of counselling and testing in all feasible settings. Routine HIV counselling and testing in clinical settings together with large-scale home-based counselling and testing programmes for more remote areas, initiated in several districts already, will hopefully increase the proportion of HIV-infected Ugandans who are aware of their infection.
Our estimate that in 2005 HIV caused every third adult death and a fifth of all deaths confirms that HIV continues to be a leading cause of mortality. Our HIV-attributable mortality estimate (45% for 15–59-year-old adults) appears consistent with cohort-based data from the Rakai District in southwestern Uganda in 2000 (15–59 years: 73.5%)  after considering the higher adult HIV prevalence in this cohort (2000: 16%, compared with a modelled national year 2000 estimate of 9.2%). Co-trimoxazole prophylaxis and ART prevented or delayed many deaths in the same year. Despite the planned expansion of these programmes we anticipate that the number of HIV-related deaths will not decline in the near future because of a continued increase in the number of PLHA requiring ART and because ART does not prevent all HIV mortality. Our projections over a 6-year period suggest that co-trimoxazole prophylaxis for HIV-exposed and infected children has a large potential to postpone morbidity and mortality  even in the absence of ART. Its low cost, low toxicity and broad beneficial effects should make co-trimoxazole-based programmes a priority. Our data also indicate that both co-trimoxazole prophylaxis and ART have the potential substantially to lower the number of orphans, half of which are caused by AIDS.
Whereas Uganda's HIV epidemic is one of the oldest in the world and HIV prevalence has declined significantly, the burden of HIV disease remains very high. The current HIV/AIDS control activities clearly mitigate the epidemic's effects; however, increased efforts in prevention, care, and ART will be necessary to change the course of the epidemic substantially.
The authors are grateful for the comments and editorial assistance provided by Rand Stoneburner, Willi McFarland, and George Rutherford. They also thank Justine Nankinga and Elizabeth Namagala for providing programme coverage estimates and Mary Glenn Fowler for advice on transmission probabilities.
W.H. reviewed available data and literature, analysed data, and wrote the manuscript. J.S. provided technical advice on analysis. W.K., A.O., and J.M. provided data and editorial comments on the manuscript. R.B. and F.K. provided editorial comments. J.M. provided editorial comments and guidance for analysis. The corresponding author had full access to all data in the study and takes final responsibility for the decision to submit the paper for publication.
Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Conflicts of interest: None.
1. Bond C. AIDS lays waste to Uganda.New Afr
2. Carswell JW, Lloyd G. Rise in prevalence of HIV antibodies recorded at an antenatal booking clinic in Kampala, Uganda
. AIDS 1987; 1:192–193.
3. Sewankambo NK, Wawer MJ, Gray RH, Serwadda D, Li C, Stallings RY, et al
. Demographic impact of HIV infection in rural Rakai district, Uganda
: results of a population-based cohort study. AIDS 1994; 8:1707–1713.
4. Elangot F. Uganda
: an AIDS control programme. AIDS Action 1987: p 6. Available at: http://www.aidsaction.info/aa01.html
. Accessed: November 2007.
5. Okware SI. Towards a national AIDS-control program in Uganda
. West J Med 1987; 147:726–729.
6. Stoneburner RL, Low-Beer D. Population-level HIV declines and behavioral risk avoidance in Uganda
. Science 2004; 304:714–718.
7. ORC Macro. Uganda
HIV/AIDS Sero-Behavioural Survey 2004/5. Kampala, Uganda
. Ministry of Health; 2006.
8. Berkley S, Okware S, Naamara W. Surveillance for AIDS in Uganda
. AIDS 1989; 3:79–85.
9. Mbulaiteye SM, Mahe C, Whitworth JA, Ruberantwari A, Nakiyingi JS, Ojwiya A, Kamali A. Declining HIV-1 incidence and associated prevalence over 10 years in a rural population in south-west Uganda
: a cohort study. Lancet 2002; 360:41–46.
10. Whitworth JA, Mahe C, Mbulaiteye SM, Nakiyingi J, Ruberantwari A, Ojwiya A, Kamali A. HIV-1 epidemic trends in rural south-west Uganda
over a 10-year period. Trop Med Int Health 2002; 7:1047–1052.
11. Mishra V, Opio A, Cross A, et al.The 2004–2005 Uganda HIV/AIDS Sero-Behavioural Survey: methods and impact of non-response bias.AIDS
: In press.
12. UNAIDS/WHO Working Group on Global HIV/AIDS/STI Surveillance. Guidelines for conducting HIV sentinel serosurveys among pregnant women and other groups. Geneva, Switzerland: WHO; 2003.
13. Ghys PD, Brown T, Grassly NC, Garnett G, Stanecki KA, Stover J, Walker N. The UNAIDS Estimation and Projection Package: a software package to estimate and project national HIV epidemics. Sex Transm Infect 2004; 80(Suppl. 1):i5–i9.
14. Stover J. Projecting the demographic consequences of adult HIV prevalence trends: the Spectrum Projection Package. Sex Transm Infect 2004; 80(Suppl. 1):i14–i18.
15. Stover J. AIM Version 4, A computer program for making HIV/AIDS projections and examining the social and economic impacts of AIDS. Washington, DC, USA: The Futures Group International; 2005.
16. Mermin J, Lule J, Ekwaru JP, Malamba S, Downing R, Ransom R, et al
. Effect of co-trimoxazole prophylaxis on morbidity, mortality, CD4-cell count, and viral load in HIV infection in rural Uganda
. Lancet 2004; 364:1428–1434.
17. McNaghten AD, Hanson DL, Jones JL, Dworkin MS, Ward JW. Effects of antiretroviral therapy and opportunistic illness primary chemoprophylaxis on survival after AIDS diagnosis. Adult/Adolescent Spectrum of Disease Group. AIDS 1999; 13:1687–1695.
18. Guay LA, Hom DL, Mmiro F, Piwowar EM, Kabengera S, Parsons J, et al
. Detection of human immunodeficiency virus type 1 (HIV-1) DNA and p24 antigen in breast milk of HIV-1-infected Ugandan women and vertical transmission. Pediatrics 1996; 98:438–444.
19. Jackson JB, Musoke P, Fleming T, Guay LA, Bagenda D, Allen M, et al
. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda
: 18-month follow-up of the HIVNET 012 randomised trial. Lancet 2003; 362:859–868.
Population and Housing Census. Kampala, Uganda
Bureau of Statistics; 2002.
21. Namagala ED. AIDS Control Programme. Kampala, Uganda
: Ministry of Health; 2006.
22. Nankinga J. AIDS Control Programme. Kampala, Uganda
: Ministry of Health; 2006.
23. ORC Macro. Demographic and Health Survey 2000–2001. Entebbe, Uganda
Bureau of Statistics; 2001.
Bureau of Statistics (UBOS) and MACRO International Inc. Uganda
Demographic and Health Survey 2006. Calverton, Maryland, USA; 2006.
25. UNAIDS. Uniting the world against AIDS, Uganda
. Geneva, Switzerland: UNAIDS; 2007.
26. Zaba B, Boerma T, White R. Monitoring the AIDS epidemic using HIV prevalence data among young women attending antenatal clinics: prospects and problems. AIDS 2000; 14:1633–1645.
27. STD/AIDS Control Programme. Policy guidelines for prevention of mother-to-child transmission of HIV. Kampala, Uganda
: Ministry of Health; June 2006.
28. Sewankambo NK, Gray RH, Ahmad S, Serwadda D, Wabwire-Mangen F, Nalugoda F, et al
. Mortality associated with HIV infection in rural Rakai District, Uganda
. AIDS 2000; 14:2391–2400.