Utility of antenatal HIV surveillance data to evaluate prevention of mother-to-child HIV transmission programs in resource-limited settings

Article Outline

• Abstract
• Materials and Methods
• Results
  • Kenya
  • Ethiopia
  • Zimbabwe
• Comment
• Recommendations
• Acknowledgment
• References
• Copyright

Prevention of mother-to-child human immunodeficiency virus (HIV) transmission (PMTCT) programs are expanding in resource-limited countries and are increasingly implemented in antenatal clinics (ANC) in which HIV sentinel surveillance is conducted. ANC sentinel surveillance data can be used to evaluate the first visit of a pregnant woman to PMTCT programs. We analyzed data from Kenya and Ethiopia, where information on PMTCT test acceptance was collected on the 2005 ANC sentinel surveillance forms. For Zimbabwe, we compared the 2005 ANC sentinel surveillance data to the PMTCT program data. ANC surveillance data allowed us to calculate the number of HIV-positive women not participating in the PMTCT program. The percentage of HIV-positive women missed by the PMTCT program was 17% in Kenya, 57% Ethiopia, and 59% Zimbabwe. The HIV prevalence among women participating in PMTCT differed from women who did not. ANC sentinel surveillance can be used to evaluate and improve the first encounter in PMTCT programs. Countries should collect PMTCT-related program data through ANC surveillance to strengthen the PMTCT program.

Key words: antenatal clinics, Ethiopia, human immunodeficiency virus, Kenya, prevention of mother-to-child human immunodeficiency virus transmission, surveillance data, Zimbabwe

In many resource-limited countries with generalized HIV epidemics, defined as HIV prevalence greater than 1% among pregnant women,¹ HIV sentinel surveillance is conducted in antenatal clinics (ANCs). About 70% (range 30-90%) of women in resource-limited countries make at least 1 ANC visit during pregnancy.² ANC sentinel surveillance has been used widely for more than 2 decades to estimate the HIV prevalence in the general population and to monitor the impact of HIV prevention programs. Unlinked anonymous testing (UAT) is the most common strategy used for ANC sentinel surveillance. With UAT, leftover blood routinely collected from women at the first antenatal care visit is stripped of all identifiers and tested for antibodies to HIV.¹ Testing is unlinked and anonymous; therefore, informed consent is not obtained and test results cannot be provided to participants.

Programs for the prevention of mother-to-child HIV transmission (PMTCT) are also conducted in antenatal clinics and target the same group, pregnant women, as ANC sentinel surveillance. With increased funding and attention,³, ⁴, ⁵ and in line with the goal of the United Nations General Assembly Special Session on AIDS in 2001 to reduce mother-to-child HIV transmission by 50% by 2010,⁶ many resource-limited countries have begun expanding PMTCT programs, although global PMTCT coverage is still low at 8%.⁷, ⁸ Expansion of PMTCT services increasingly includes clinics at which ANC sentinel surveillance is conducted, resulting in increased availability of HIV testing data from both PMTCT programs and ANC sentinel surveillance in the same clinic population.

Operational similarities and distinctions between PMTCT programs and ANC sentinel surveillance are outlined in Table 1. In summary, HIV testing in PMTCT programs is voluntary; with all ANC clients either routinely recommended HIV testing with a right to refuse testing (“opt-out”) or required to specifically request to be tested (“opt-in”).⁹, ¹⁰ Therefore, the acceptance or uptake for HIV testing varies widely by clinic and country.¹¹, ¹² Ideally, the HIV test is offered at the first ANC visit, because many women may make only 1 ANC visit during pregnancy, although this does not always happen on that first visit because of the lack of resources. The HIV test is usually performed within the PMTCT clinic or a related laboratory. Women who accept testing have an opportunity to receive their results and benefit from PMTCT interventions, including antiretrovirals for PMTCT and prevention, treatment, care, and support services for themselves and their families. Typically, PMTCT programs enter individual data into ANC registers and logbooks, which are aggregated as a monthly or quarterly report and sent to the Ministry of Health at the district and national levels, where they are compiled.

TABLE 1. Comparison of PMTCT programs and ANC sentinel surveillance

In contrast, ANC sentinel surveillance consecutively samples leftover blood routinely collected for various pregnancy tests on all new ANC attendees until the target sample size is reached. Unique identifiers, such as name and hospital number, are removed before the HIV test is done. HIV testing is usually performed outside the clinic, usually in a central reference laboratory, with strong quality assurance measures. For each sampled ANC surveillance client, individual data, such as age, parity, and gravidity, are abstracted or transcribed into individual data forms without identifiers, entered electronically, and analyzed. The Figure shows the flow chart for conducting ANC surveillance within a PMTCT service site.

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE. 

  Flow chart for PMTCT program and ANC sentinel surveillance at dual sites*

There have been recent evaluations of the utility of PMTCT program data for the purpose of surveillance¹³, ¹⁴, ¹⁵, ¹⁶, ¹⁷ because of ethical concerns about the inability to provide UAT-based test results and services to ANC clients; the introduction of rapid syphilis testing with limited access to leftover blood; and the larger number of PMTCT sites and sample sizes, compared to ANC surveillance. However, the limited published results to date suggest that whereas PMTCT program data often include large numbers of women, their data quality, availability, and the uptake of PMTCT HIV testing must be considered when using them for surveillance.¹³ The variable HIV testing uptake in PMTCT programs may compromise the use of PMTCT data for ANC sentinel surveillance because not all ANC attendees accept to be tested for HIV.

On the other hand, given the uneven quality of PMTCT clinic data, the varying uptake of PMTCT services and the increasing availability of PMTCT services at ANC sentinel surveillance sites, it is useful to assess whether ANC sentinel surveillance data can be used to evaluate how PMTCT programs perform at the first ANC visit. Some countries have added PMTCT variables into their ANC sentinel surveillance data forms. These data can provide information on whether the HIV prevalence among women who participate in PMTCT or accept HIV testing (ie, acceptors) is lower or higher, compared with HIV prevalence in women who do not participate in PMTCT or refuse HIV testing as part of PMTCT services (ie, refusers). These data are not available from routine PMTCT records and can help countries estimate the number of HIV-positive women not identified by the PMTCT program and the related lost impact (ie, number of infant infections that could have been averted if all HIV-positive mothers were identified in the PMTCT program). This information can be useful for planning and setting PMTCT service targets. However, the experiences in using these data have not been documented. Presented here are case studies of efforts to use ANC sentinel surveillance data to monitor and evaluate PMTCT programs and a discussion of approaches to using ANC sentinel surveillance to enhance and plan PMTCT programs.

Back to Article Outline

Materials and Methods 

We analyzed data from Ethiopia, Kenya, and Zimbabwe. Ethiopia and Kenya added PMTCT variables to their ANC sentinel surveillance data form. Variables added were whether the client sampled for ANC sentinel surveillance accepted to participate in the PMTCT program (Ethiopia) or whether the client was offered an HIV test through the PMTCT program and the PMTCT-related HIV test result (Kenya). Zimbabwe does not collect PMTCT data as part of surveillance, but we were able to utilize the HIV prevalence estimates from ANC sentinel surveillance sites and the annual PMTCT program data for this analysis.

Using the ANC sentinel surveillance data from sites with PMTCT services, we estimated the overall burden of disease based on the HIV prevalence among all new ANC clients enrolled in surveillance. For Kenya and Ethiopia, we estimated the percentage of new ANC clients who participated or tested for HIV under the PMTCT program as well as the pooled and the median (interquartile range) HIV prevalence among clients who accepted PMTCT, compared with prevalence among those who were not offered PMTCT services or HIV testing or who refused the HIV test.

Both the pooled and median HIV prevalence data are presented for several reasons. The pooled data are appropriate for use when limiting the interpretation only to the sentinel sites sampled, and is also useful in calculating infant infections averted. The median prevalence accounts for outlying larger clinics and has been used for years by sentinel surveillance to extrapolate to the general population, which we did not do here for Kenya and Ethiopia but show for comparative purposes. We calculated the number of additional infant HIV infections that could have been averted among HIV-positive women not identified by PMTCT (ie, women not offered PMTCT services or HIV testing or who refused the HIV test as part of PMTCT).

For the calculation of infant infections averted or that could have been averted, we multiplied the number of HIV-positive women by 35%, the standard estimated percentage of HIV-positive women who will transmit the virus to their infants without PMTCT interventions in resource-limited settings.¹⁸ Next, we multiplied the result by 41%, the efficacy of single-dose nevirapine (SDNVP)¹⁹ to prevent mother-to-child HIV transmission. SDNVP was the main antiretroviral (ARV) prophylaxis regimen for PMTCT in these countries during the reported surveillance rounds. The results from Kenya and Ethiopia were extrapolated from the 3-month surveillance period to a year to determine number of infections that could be averted in a year at these sites.

Zimbabwe does not collect PMTCT data on its surveillance form, but we used the 2004 PMTCT program summary data and ANC surveillance data from the 19 sentinel surveillance sites. We multiplied the HIV prevalence from the ANC surveillance data with the number of new ANC attendees in all PMTCT sites in 2004 to get an estimate of the number of HIV-positive pregnant women attending all PMTCT sites. We compared the estimated number of HIV-positive pregnant women based on surveillance data with the actual number of HIV-positive women identified through the PMTCT program to estimate the percentage of HIV-positive women who were not identified (not tested for HIV) by the PMTCT program. Based on our assumption of 35% transmission risk and 41% SDNVP efficacy, we estimated the number of infant infections that could have been averted among women not identified or missed by the PMTCT program in 2004.

Back to Article Outline

Results 

Kenya 

In Kenya, ANC sentinel surveillance is conducted annually.²⁰ Pilot programs for PMTCT began in 1999 and national scale-up began in 2002. In 2001, 3 of 35 sentinel surveillance sites offered PMTCT services. By 2005, there were more than 1100 PMTCT sites in Kenya, and all sentinel surveillance sites (n = 43) offered PMTCT services, which includes routinely recommending HIV testing to ANC clients with a right to opt out. Between 2004-2005, the following PMTCT variables were added to the sentinel surveillance data form: whether HIV testing was offered; whether the client accepted the test; and the PMTCT-related test result.¹⁸

During the 2005 survey period (approximately 3 months), a total of 13,026 women were consecutively sampled for ANC sentinel surveillance at the 43 sites. Of these, 9690 (76%) were offered and accepted HIV testing under the PMTCT program; 2988 (23.6%), either refused testing or were not offered HIV testing; and there were 348 women with missing data.

The overall pooled HIV prevalence from the ANC sentinel surveillance was 7.3%. The HIV prevalence was 5.4% among those who were not offered testing or refused the test, compared with 8.0% among women who accepted HIV testing from the PMTCT program (Table 2). Similarly, the median HIV prevalence among acceptors was higher than among refusers and women not offered testing, (Table 2). Of all HIV-infected women in the ANC sample, 17.3% (162/939) were not HIV tested as part of PMTCT during the surveillance period.

TABLE 2. Estimating HIV-positive women not identified through PMTCT in Kenya and Ethiopia in 2005 using ANC surveillance

	Kenya				Ethiopia
	N	%	Median site %	Interquartile range	N	%	Median site %	Interquartile range
ANC surveillance sites	43				79
New ANC attendees enrolled in ANC surveillance sites	13026				28572
Dual sites⁎	43				36
New ANC attendees enrolled in surveillance at dual sites	13026	100%			28572	43%
Accepted PMTCT or HIV testing as part of PMTCT	12678	76.4%	84.2%	53.2%-99.3%	12316	46.9%	44.3%	24.8%-72.6%
Refused PMTCT participation, HIV testing, or service not offered	12678	23.6%	15.8%	0.7%-46.8%	12316	53.1%	55.7%	27.4%-75.2%
HIV prevalence among all clients enrolled at dual sites	13026	7.3%	5.3%	3.7%-8.4%	12316	7.4%	7.7%	2.5%-10.5%
HIV prevalence among clients accepting participation in PMTCT or tested for HIV in PMTCT	9690	8.0%	5.1%	0.0%-8.5%	5774	6.4%	7.0%	2.4%-11.5%
HIV prevalence among clients refusing participation in PMTCT, HIV testing, or service not offered	2988	5.4%	4.7%	1.4%-6.8%	6542	8.2%	7.9%	2.1%-11.2%
Proportion of HIV-positive women missed by PMTCT	939	17.3%	15.4%	0.0%-52.4%	906	59.2%	52.8%	25.6%-80.6%
Potential infant infections that can occur with 35% transmission risk in all HIV-positive women (number HIV-positive × 0.35):
3-mo surveillance period	329				317
Extrapolated to 1 y	1316				1268
Infant infections that could be averted with use of SDNVP with 41% efficacy (potential number of infant infections × 0.41):
3-mo surveillance period	135				130
Extrapolated to 1 y	540				520
Infant infections not averted among HIV-positive women missed by PMTCT (% missed × infant infections that could be averted with SDNVP)
3-mo surveillance period	23				77
Extrapolated to 1 y	93				308

Based on the data and assumptions depicted in Table 2, 329 HIV-positive women would be expected to transmit the virus to their infants and 135 infant infections would have been averted if all HIV-positive women had received SDNVP at dual sites (ie, sites with both PMTCT and sentinel surveillance). Given that 17% of the HIV-positive women did not test under PMTCT, we estimate that an additional 23 infant HIV infections could have been averted had the women’s serostatus been identified and they had received SDNVP. Extrapolating to a full year, about 93 infant HIV infections that should have been averted occurred at the 43 sites.

Ethiopia 

In Ethiopia, ANC sentinel surveillance is conducted biennially for a period of approximately 3 months. Ethiopia’s 2005 ANC sentinel surveillance data collection form²¹ included 1 PMTCT variable: whether the enrolled client participated in PMTCT (ie, agreed to be referred to receive PMTCT services). ANC sentinel surveillance was conducted at 43 rural and 36 urban sites. Of these, 36 (12 rural and 24 urban) had PMTCT services. These 36 sites with ANC sentinel surveillance and PMTCT services included 12,316 of the 28,572 women (43%) sampled by the national ANC surveillance system.

The percentage of clients participating in PMTCT at the 36 sites was 47% (58% among rural; 41% among urban sites). The remaining 53% of women were either not offered PMTCT services because of lack of availability of service (11%) or refused PMTCT participation (42%). As shown in Table 2, the overall pooled HIV prevalence from the ANC sentinel surveillance was 7.4%. The prevalence was higher, 8.2% (536/6542), among women who did not participate or for whom services were not available, compared with 6.4% (370/5774) among women who participated in PMTCT. Similarly, the median HIV prevalence among refusers and women not offered testing was higher than among acceptors (Table 2). Of all HIV-infected women in the ANC sample, 59% were missed by the PMTCT program at the time of surveillance.

Based on the data and assumptions depicted in Table 2, 317 HIV-positive women were at risk of transmitting the virus, and an estimated 130 infant infections could have been averted if all HIV-positive women had received SDNVP. Given that 59% of the HIV positive women did not participate in PMTCT, we estimate that about 77 infant HIV infections that could have been averted had the women’s serostatus been identified occurred; this extrapolates to 308 infections over 1 year for the 36 sites.

Zimbabwe 

In Zimbabwe, ANC sentinel surveillance is conducted biennially, with the most recent survey in 2004. In that year, the ANC sentinel surveillance included 19 antenatal clinics (7 urban, 5 periurban, and 7 rural), all of which provided PMTCT services. In total, 800 of 1383 antenatal clinics throughout the country offered PMTCT services in 2004, but only 265 offered comprehensive PMTCT services (HIV counseling, on-site testing, and SDNVP). The remaining clinics provided basic ANC services including information on PMTCT and referral to a clinic with services for an on-site HIV test and SDNVP or transport of blood for off-site HIV testing and SDNVP provided to HIV-positive clients.

We used ANC surveillance prevalence estimates to estimate prevalence among new ANC clients attending PMTCT sites because the 2004 ANC sentinel survey in Zimbabwe did not collect PMTCT information on the surveillance data forms. Table 3 shows the application of the 2004 HIV prevalence estimates from ANC sentinel surveillance for monitoring the PMTCT program in Zimbabwe. The HIV prevalence was multiplied by the number of new ANC attendees in all PMTCT sites (sites that provided any PMTCT service including clinics that could refer pregnant women to clinics with PMTCT) to provide an estimate of the number of HIV-positive pregnant women attending ANC in all PMTCT sites. By comparing this number to the number of HIV-positive women identified through the PMTCT program, we estimated that approximately half of the HIV-positive women attending PMTCT sites in Zimbabwe in 2004 were tested for HIV. By using a 35% transmission risk and 41% SDNVP efficacy, we estimated that 12,162 HIV-positive women attending PMTCT sites but not identified by the program would have averted an additional 1746 infant HIV infections in 2004 if they had received SDNVP.

TABLE 3. Estimating HIV-positive women not identified through PMTCT in Zimbabwe in 2004 using ANC sentinel surveillance and 2004 annual PMTCT program monitoring data

Back to Article Outline

Comment 

These case studies demonstrate the utility of ANC sentinel surveillance data for PMTCT programs in ANCs. The inclusion of simple PMTCT variables in the ANC sentinel surveillance data forms in Ethiopia and Kenya provided an effective tool to monitor and evaluate PMTCT programs in ANCs during the first antenatal visit of pregnant women. All sentinel surveillance sites in Kenya and Zimbabwe now have PMTCT services, providing an opportunity to directly compare the surveillance and PMTCT program data.

There are several advantages to collecting PMTCT-related variables in ANC surveillance. PMTCT is a high-impact program as measured by the number of infant HIV infections that can be averted.¹⁸ The estimates calculated from the number of HIV-positive women tested through UAT but not identified by PMTCT because they refused either PMTCT participation or HIV testing or were not offered PMTCT services permits a measure of the potential number of infections that could have been averted with a fully successful program at current sites. Adding these variables also allows for an estimate of HIV prevalence among refusers, women not offered HIV testing, or HIV-positive women not participating in PMTCT, and can be used by policy makers and clinic managers to determine which ANC sites are missing HIV-positive women who might benefit from PMTCT services.

Because the data can be stratified at the clinic level, it can help PMTCT program managers decide on where to intensify efforts. This kind of evaluation cannot be performed with routine PMTCT program data because HIV prevalence among those not offered or refusing PMTCT is not known. Assuming that ANC surveillance sites are representative of the larger number of clinics and general population, the estimates of the number of women missed and infections not averted can be generalized to all PMTCT sites if the HIV testing uptake is known for all PMTCT sites in country. Additionally, the ANC HIV surveillance estimates can be applied to the number of annual births to estimate the infant HIV infection burden nationwide and also the potential for preventing infant infections. This analysis can be done with each surveillance round to document trends in PMTCT uptake among HIV-positive women.

Another benefit is the capacity to compare the prevalence among those who accept HIV testing under PMTCT and those who either refuse or who are not offered the test. HIV prevalence was higher in those who were not offered or refused HIV testing in Ethiopia. Conversely, in Kenya, prevalence among acceptors or women participating in PMTCT is higher, which is similar to findings in the literature,²², ²³ and which may reflect an increasing awareness and acceptance of the benefits of PMTCT services.²³

Additionally, HIV testing performed under the PMTCT program can be compared with HIV testing conducted under UAT sentinel surveillance, which is often done in a central reference laboratory. If the testing algorithms are similar in PMTCT program and ANC surveillance and all test results are available, this type of comparative analysis can be done. Kenya collects the related PMTCT test results on the surveillance form, but not all results were available in 2005 and thus not presented in this paper. The comparison of the 2 HIV test results (done under PMTCT and UAT) for each client can provide some information on the quality of test done by the PMTCT program and used for improvement.

The difference between HIV test acceptance in Ethiopia and Kenya (48% vs 81%) may be attributed to Ethiopia’s opt-in testing strategy in contrast to Kenya’s opt-out strategy. The opt-out approach to HIV testing results in greater acceptance and uptake for PMTCT.²⁴ Ethiopia is currently transitioning to the opt-out policy and should consider similar analyses of the data a year after this new policy takes effect to evaluate its impact on HIV testing uptake; prevalence between participants and nonparticipants of PMTCT services; and the impact on aversion of infant HIV infection.

Although Zimbabwe did not collect PMTCT data during ANC surveillance, PMTCT and ANC sentinel surveillance data sets were compared, permitting the estimation of the number of HIV-positive pregnant women attending ANC and the percentage of HIV-positive women not identified by the PMTCT program. As ANC surveillance becomes more robust, population-based estimates and extrapolations to program data will become more accurate. Ideally, countries should include PMTCT variables on their ANC surveillance forms; however, at sites at which this is not available, the type of analysis presented here for Zimbabwe can provide useful estimates.

There are a number of limitations to these analyses. Although ANC surveillance sites have been used for decades to provide information on trends and serve as a reasonably robust and representative sample for estimating HIV prevalence among pregnant women, sites are conveniently selected and are not a true random sample. For Zimbabwe, we extrapolated the prevalence to all PMTCT sites because we had the HIV testing uptake data; however, we realize that heterogeneity of HIV prevalence at all sites and knowledge of HIV prevalence at each site are preferred. In addition, some women who were not HIV tested by the PMTCT program during the surveillance period might already have a documented HIV-positive status. Furthermore, others may have been tested at subsequent ANC visits, which could bias the analyses toward an underestimate of PMTCT uptake and an overestimate of infant infections that could have been averted. We relied on multiple assumptions and were unable to extrapolate to all PMTCT sites because data on HIV testing uptake for each site were unavailable. The variable from Ethiopia, “participation in PMTCT,” is vague and does not provide detailed information on the number of women who actually agreed to the test. More accurate results can be obtained if countries utilize more specific variables like those from Kenya.

Despite these limitations, these estimates provide important and useful guidance for monitoring and evaluating PMTCT programs.

Back to Article Outline

Recommendations 

Comparing ANC sentinel surveillance and PMTCT program is relatively simple and can provide important information. Countries should be encouraged to incorporate PMTCT variables in ANC sentinel surveillance. We suggest several PMTCT variables and their related indicators (Table 4). These recommendations are in line with recent recommendations from the World Health Organization Regional Office for Africa’s Technical Network Group on sexually transmitted infections and HIV surveillance, which will be released in mid-2007.

TABLE 4. Suggested PMTCT variables to include in ANC sentinel surveillance forms and PMTCT indicators measured by these variables

We also recommend that countries should prioritize PMTCT implementation at ANC sentinel surveillance sites. This will ensure that ANC clients sampled for UAT are given the opportunity to learn their HIV status and receive appropriate services. However, efforts should be taken to ensure that the protocols for implementing ANC sentinel surveillance are retained when PMTCT programs are introduced. Anecdotal evidence suggests that introduction of PMTCT programs at ANC sentinel sites creates additional work for staff and can cause confusion in sampling, which can bias the HIV prevalence results. Additional resources with improved supervision, training, and support during the duration of ANC sentinel surveillance will aid in preventing sampling confusion and the risk of related bias.

Using ANC sentinel surveillance to monitor PMTCT should not undermine routine PMTCT program monitoring systems. PMTCT program managers should improve the quality of routine PMTCT data collection systems through training and proper staff supervision. As the PMTCT data quality improves and HIV testing uptake continues to increase with more countries implementing an opt-out testing strategy, evaluations should be conducted to determine how well the PMTCT program data can complement or replace the ANC sentinel surveillance data.

At this time, ANC sentinel surveillance continues to play an important role in monitoring the HIV epidemic in many countries. ANC sentinel surveillance can be used to monitor, evaluate, and enhance PMTCT programs by including PMTCT variables in ANC surveillance data collection forms or by comparing both data sets and performing simple analyses as shown in the Zimbabwe example in which we were able to compare data from surveillance and PMTCT program We propose that these types of analyses be used to evaluate PMTCT uptake; estimate the number of HIV-positive women missed by the PMTCT program at first ANC visit; assess HIV prevalence among acceptors and refusers; and estimate PMTCT program impact and missed program impact (ie, the potential number of infections that could have been averted). This would also require consideration of the efficacy of various ARV regimens for PMTCT that are increasingly available in many of these countries. The findings from the country case studies can be followed up over time to determine improvement in PMTCT programs and used by policy makers to guide PMTCT scale-up and program planning.

Back to Article Outline

Acknowledgment 

The authors acknowledge the input and support from the Ministry of Health officials from Kenya, Ethiopia, and Zimbabwe who provided the data for the analysis and Jelaludin Ahmed of the Centers for Disease Control and Prevention Ethiopia, who provided technical input, support, and guidance on this paper. In addition, the authors thank Dr Theresa Diaz, Dr Nathan Shaffer, Ray Shiraishi, and Meade Morgan (Centers for Disease Control and Prevention/Global AIDS Program Atlanta) and Dr Garcia Calleja, Jesus Maria, and Dr Peter Ghys (World Health Organization/Geneva) for their technical reviews, suggestions, and comments.

Back to Article Outline

References 

1. Guidelines for conducting HIV sentinel sero-surveys among pregnant women and other groups. Geneva: Joint United Nations Programme on HIV/AIDS/World Health Organization Working Group on Global HIV/AIDS/STI surveillance; 2003;Anonymous
   • View In Article
2. World Health Organization. Antenatal care in developing countries (Promises, achievements and missed opportunities). 2003;Available at: http://www.who.int/reproductive-health/docs/antenatal_care.pdf. Accessed March 19, 2007
   • View In Article
3. Presidents emergency plan for AIDS relief. Available at: http://www.usaid.gov/our_work/globalhealth/aids/pepfar.html. Accessed July 1, 2006.
   • View In Article
4. Global Fund. Available at: http://www.theglobalfund.org/en/about/aids/default.asp. Accessed July 1, 2006.
   • View In Article
5. World Health Organization. (WHO) 3 by 5. Available at: http://www.who.int/3by5/en/. Accessed July 1, 2006.
   • View In Article
6. United Nations General Assembly Session on HIV/AIDS. 2001;June 20-21, Available at: http://www.ungass.org/index.php/en/ungass/declaration_of_commitment/ungass_declaration_of_commitment. Accessed July 1, 2006.
   • View In Article
7. Grulich AE, Kaldor JM. Evidence of success in HIV prevention in Africa. Lancet. 2002;360:3–4
   • View In Article
   • Full Text
   • Full-Text PDF (55 KB)
   • CrossRef
8. De Cock KM, Mbori-Ngacha D, Marum E. Shadow on the continent: public health and HIV/AIDS in Africa in the 21st century. Lancet. 2002;360:67–72
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (93 KB)
   • CrossRef
9. Joint United Nations Programme on HIV/AIDS Global Reference Group on HIV/AIDS and Human Rights. Joint United Nations Programme on HIV/AIDS/World Health Organization policy statement on HIV testing. Available at: http://www.who.int/hiv/pub/vct/en/hivtestingpolicy04.pdf. Accessed July 1, 2006.
   • View In Article
10. World Health Organization. The right to know. New approaches to HIV testing and counseling. World Health Organization/HIV2003.08. Accessed July 1, 2006.
    • View In Article
11. Buyse D, Nuwaha FN, Karlin T, Wilfert C. Prevention of mother to child transmission of HIV: from research to action. In: Presented at the XIVth International Conference on AIDS, Barcelona, Spain, July 7-12. 2002;(Abstract TuPeF5413)
    • View In Article
12. Temmerman M, Quaghebeur A, Mwanyumba F, Mandaliya K. Mother-to-child HIV transmission in resource poor settings: how to improve coverage?. AIDS. 2003;7:1239–1242
    • View In Article
13. Hladik W, Masupu K, Roels T, et al. Prevention of mother-to-child transmission and voluntary counseling and testing programme data: what is their utility for HIV surveillance?. AIDS. 2005;19:S19–S24
    • View In Article
14. Baryarama F, Bunnell RE, Ransom RL, et al. Using HIV voluntary counseling and testing data for monitoring the Ugandan HIV epidemic, 1990-2000. J Acquir Immune Defic Syndr. 2004;37:1180–1186
    • View In Article
    • MEDLINE
    • CrossRef
15. Seguy N, Hladik W, Munyisia E, Bolu O, Marum L, Diaz T. Can data from programs for the prevention of mother-to-child transmission of HIV be used for HIV surveillance in Kenya?. Public Health Rep. 2006;121:695–702
    • View In Article
    • MEDLINE
16. Mpairwe H, Muhangi L, Namujju P, et al. HIV risk perception and prevalence in a program for prevention of mother-to-child HIV transmission: comparison of women who accept voluntary counseling and testing and those tested anonymously. J Acquir Immune Defic Syndr. 2005;39:354–358
    • View In Article
    • MEDLINE
    • CrossRef
17. Fabiani M, Nattabi B, Ayella E, Ogwang M, Declich S. Using prevalence data from the programme for the prevention of mother-to-child-transmission for HIV-1 surveillance in North Uganda. AIDS. 2005;19:823–827
    • View In Article
    • MEDLINE
18. De Cock KM, Fowler MG, Mercier E, et al. Prevention of mother-to-child HIV transmission in resource-poor countries: translating research into policy and practice. JAMA. 2000;283:1175–1182
    • View In Article
    • MEDLINE
    • CrossRef
19. Jackson JB, Musoke P, Fleming T, 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 randomized trial. Lancet. 2003;362:859–868
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (130 KB)
    • CrossRef
20. Republic of Kenya, Ministry of Health. National HIV/STD surveillance form X. Nairobi, Kenya: Ministry of Health; 2005;
    • View In Article
21. National HIV/AIDS sentinel surveillance protocol for Ethiopia; 2004. Addis Ababa, Ethiopia: Federal Ministry of Health, Ethiopia; 2004;
    • View In Article
22. Viani R, Araneta M, Ruiz-Calderon J, et al. Perinatal HIV counseling and rapid testing in Tijuana, Baja California, Mexico: seroprevalence and correlates of HIV infection. J Acquir Immune Defic Syndr. 2006;41:87–92
    • View In Article
    • MEDLINE
    • CrossRef
23. Nuwaha F, Kabatesi D, Muganwa M, et al. Factors influencing acceptability of voluntary counseling and testing for HIV in Bushenyi district of Uganda. East Afr Med J. 2002;79:626–632
    • View In Article
24. Centers for Disease Control and Prevention. Introduction of routine HIV testing in prenatal care—Botswana, 2004. MMWR Morb Mortal Wkly Rep. 2004;53:1083–1086
    • View In Article