The 2011 National Antenatal Sentinel HIV & Syphilis Prevalence Survey in South Africa

Transcription

1 2011 The 2011 National Antenatal Sentinel HIV & Syphilis Prevalence Survey in South Africa health Health DIRECTORATE: Epidemiology & Surveillance National Department of Health Cluster: HIMME 1

2 2 The 2011 National Antenatal Sentinel HIV & Syphilis Prevalance Survey in South AFrica

3 Published by the National Department of Health, Civitas Building, Corner Struben and Andries Street, Pretoria 2012 Department of Health The information contained in this report may be freely quoted, distributed and reproduced, provided that the source is acknowledged and it is used for non-commercial purposes. Suggested citation: The National Antenatal Sentinel HIV and Syphilis Prevalence Survey, South Africa, 2011, National Department of Health. Search citation: South Africa antenatal sentinel HIV prevalence, 2008, 2009, 2010, 2011 HIV prevalence trends, antenatal sentinel HIV survey South Africa, HIV and AIDS Estimates SA, 2008, 2009, 2010, 2011, Syphilis trends. Prepared and obtainable free of charge from: Directorate: Epidemiology and Surveillance National Department of Health Private Bag X 828 Pretoria, 0001 Tel: i

4 EXECUTIVE SUMMARY The HIV sentinel surveillance data has helped to map the epidemic and monitor HIV infection trends in the country and has served as an advocacy tool, resulting in the mobilization of partners, resources and development of innovative approaches by the national response to HIV and AIDS. The 2011 HIV survey was the 22nd round to be conducted in the country by the Department of Health. The South African antenatal clinic survey is done annually in October to obtain an estimate of the point prevalence for that year. The data set generated from this survey is used by mathematical modelers such WHO/UNAIDS reference group and others to estimate the rate of new HIV infections(incidence) and HIV-associated deaths are derived through mathematical models applied to HIV prevalence estimates. The Department of Health will continue to sustain and enhance the use of this robust Unlinked anonymous HIV surveillance methodology to monitor the HIV epidemic trend in South Africa through conducting annual HIV antenatal sentinel surveillance at national, provincial and district levels. The HIV epidemic in South Africa has in the last 6 years shown stabilization, particularly among antenatal care first time bookers in their current pregnancy served in public health sector clinics. In 2011, a total of first time antenatal care attendees participated in the survey. This was a representative sample to make conclusive inferences on the HIV and syphilis occurrence at National and Provincial level and in all 52 Health Districts. The survey epidemiological design uses a cross-sectional unlinked anonymous testing method using blood samples collected for other purposes in selected sentinel primary health care facilitates. This sampling approach is convenient because, as part of the antenatal care services it is mandatory to routinely draw blood from the first bookers, and this minimises participation bias and reduces costs. In addition, pregnant women are universally the most common Sentinel Population for HIV and in South Africa the most common mode of transmission is the heterogeneous sexual route. It is not perfectly representative of all women and even less of men, children and non-pregnant women, but it is an important means of coverage for countries that have a generalized HIV epidemic (i.e. where HIV prevalence among pregnant women is >2%) and it also has a wide geographic coverage (urban; informal settlements and rural communities). Intravenous blood samples are collected from the pregnant first bookers in their current pregnancy, recruited from public sector antenatal clinics every month of October since Their full intravenous blood specimens were collected and sent to central pathology laboratories in the various provinces for HIV and syphilis analysis. The laboratory diagnostic test used were the highly specific and sensitive Enzyme Linked Immuno Sorbent Assay (ELISA) for HIV antigen testing and the Rapid Plasma Reagin (RPR) card test for active syphilis. Given that the sentinel sites were chosen on a probability proportional to size basis by district, the sampling period was fixed and the districts samples were self-weighting, the provincial prevalence estimates were calculated as the total of the results from the districts in the provinces. The national prevalence was weighted ii

5 according to the total number of women aged years in each province using the 2011 midyear population size estimates. The sentinel population (participants) who agreed to participate in the survey has increased from in 2010 to in 2011 of the targeted pregnant women attending antenatal care. The sample population realization rate in 2011 was 92.9% which exceeds 70% compliance as outlined in the survey protocol. THE NATIONAL HIV PREVALENCE The estimated 2011 national HIV prevalence was 29.5% (95% CI %) showing a slight drop of 0.7% from the 2010 national HIV prevalence. The 2011 confidence interval includes the 2010 point estimate of 30.2% and the 2011 estimate is also in line with estimates from This indicates a stable prevalence of HIV infections among pregnant women aged years and attending their first antenatal care during their current pregnancy in public health clinics in South Africa over the past 5 years as indicated below: 29.4% (95% CI: ) in % (95% CI: ) in % (95% CI: ) in % (95% CI: ) in % (95% CI: ) in According to the UNAIDS SPECTRUM model the estimated national HIV prevalence among the general adult population aged years old has remained stable at around 17.3% since In 2011, an estimated 5, [ ] people living with HIV resided in South Africa. The estimated number of new infections was 1.43% in 2011 compared to 1.63% new infections in HIV PREVALENCE BY PROVINCE KwaZula-Natal has recorded a notable decrease in HIV prevalence which is promising, whereas Mpumalanga has recorded an increase in the past four years which is worrisome. The HIV prevalence estimates across provinces is variable in year to year changes. There is however a notable drop in the 2011 HIV prevalence recorded in KwaZulu-Natal with an estimate of 37.4% (95% CI: %). The upper limit of the 2011 confidence interval is lower than the 2009 and 2010 estimates of 39.5% indicating a decline by 2.1% in HIV prevalence in this province. Mpumalanga province has shown an increase in estimated HIV prevalence of 2.0% from 34.7% in 2009 to 36.7% (95% CI %) in There was an increase in HIV prevalence in Free State from 30.6% in 2010 to 32.5% in 2011, and the North-West from 29.6% in 2010 to 30.2% in Limpopo is showing a steady increase from 21.4% in 2009 to 22.1% in iii

6 HIV PREVALANCE ESTIMATES PROJECTED IN THE GENERAL POPULATION The estimated provincial HIV prevalence in the general population (15-49 years) for 2010 and 2011 is shown below. Province Eastern Cape Free State Gauteng KwaZulu-Natal Limpopo Mpumalanga North West Northern Cape Western Cape South Africa HIV PREVALENCE BY DISTRICT The districts are clearly heterogeneous with respect to the epidemic, with prevalence rates ranging from a high of 46.1% in Gert Sibande in Mpumalanga, the highest district HIV prevalence ever recorded in this province, followed by Ugu and Mkhanyakude in KZN which recorded 41.7% and 41.1% respectively. The lowest HIV prevalence was 6.2% in Namaqua in the Northern Cape. When data are pooled over the five years this heterogeneity persists. The new Buffalo City district has recorded the highest HIV prevalence of 34.1% in the Eastern Cape. The district that recorded the highest HIV prevalence of 46.1% in 2011 was Gert Sibande in Mpumalanga. The HIV prevalence in this district increased by 7.9% from 38.2% recorded in In 2011 only 3 districts recorded HIV prevalence above 40% viz:ugu, UMkhanyakude and Gert Sibande in Mpumalanga compared to 5 districts in In 2011, twenty six (26) out of the 52 districts recorded HIV prevalence rates prevalence rates below the national average of 29.5% compared with 28 out of 52 districts that recorded below the national average of 30.2% in There were 24 out of 52 districts that recorded HIV prevalence rates above the national average in 2010; whereas there were 25 out of 52 districts that recorded HIV prevalence rates above the national average in There was considerable variation in HIV prevalence rates between the 52 health districts observed over the four-year period , particularly where the sample size in a district was small, where districts were merged (i.e. Tshwane and Metsweding) and where there are new district demarcations like Buffalo City, making it difficult to interpret any trends in the current report. HIV PREVALENCE ESTIMATE BY AGE There are distinctly different risk factors that lead to HIV infection among infants and adults, hence age is an important risk factor and is central to monitoring the epidemic among the highly sexually iv

7 active group, while at the other end of the age spectrum vertical transmission from the infected mother is the most important and significant risk factor. HIV prevalence estimate is the most important indicator used to provide empirical evidence when monitoring HIV incidence (new infections). The survey participants aged years accounted for 49.4% of the survey participants. HIV prevalence in this age group has been suggested as a proxy measure for average incidence in the youth because of sexual onset and hence prevalent infections are assumed to be recent while this age group is less likely to be affected by AIDS mortality. The HIV prevalence among the year old pregnant women was 21.8% (95% CI: ) in 2010 compared with 20.5 % (95% CI: ) in 2011, a decline of 1.3%. The specific AIDS MDG target is that by 2015 the expected HIV prevalence reduction should be 25% less than the baseline prevalence of 23.1% in The findings of monitoring trends in this age group in South Africa show that we should not relent of our collective efforts to achieve this AIDS MDG target. There was a slight increase in HIV prevalence among young women in the age group years from 13.7% in 2009 and14.0% in 2010, followed by a decline of 1.3% to 12.7% in 2011, however, these changes in prevalence were not statistically significant. There was an increase in HIV prevalence in pregnant women aged above 35 years. In year old women the 2011 estimated prevalence was 39.5% (95%CI: %) compared to the point estimate of 35.4% in SYPHILIS PREVALENCE The national prevalence of syphilis shows a 0.1% increase where the prevalence was 1.5% (95% CI: ) in 2010 to 1.6% (95% CI: ) in Mpumalanga syphilis prevalence increased from 2.1% in 2010 to 4.1% in 2011, which makes it the province with highest syphilis rate, whereas in the past four years the Northern Cape recorded the highest prevalence of syphilis. In Gauteng there was a slight drop of 2.9% in 2009 to 2.0% in 2011 From 2012 the monitoring of the trends in syphilis prevalence in this survey will be dropped. However, the routine monitoring of syphilis prevalence among pregnant women during the antenatal care will remain. In the 2012 survey we have started to pilot monitoring of Herpes Simplex HSV2 which usually causes genital herpes and is transmitted primarily by direct contact with sores, most often during sexual contact. The findings will be reported in the 2012 report. v

8 CONCLUSION The HIV prevalence of 29.5% in 2011 is in line with the prevalence observed in the past four previous years. To avoid a resurgence of the HIV and AIDS epidemic in South Africa, HIV prevention efforts need to be urgently strengthened and sustained. Furthermore, ecological correlations between the trends in HIV prevalence, and behavioral changes that will focus on reducing the incidence of infection exposure factors, especially in districts that record more than 30% HIV prevalence, is warranted. Further in-depth epidemiological investigations on what could be causing the variation between the districts and between provinces in the identified epicentres could assist in understanding the different patterns of the transmission potential of the virus. The 2011 report has shown beyond reasonable doubt that there is no significant correlation between HIV and Treponema palladium the aetiological agent for active syphilis as co-factor for HIV infection. In the 2012 antenatal sentinel HIV survey we will be piloting and testing for Herpes simplex and investigating whether infection is significantly correlated to increased risk of HIV infection. We will establish systems to track measures such as AIDS-related mortality by age, by sex, by district, by province and monitor loss to follow up, monitor number of patients on different regimes, monitor Pharmacovigilance (drug side effects), monitor drug resistant patterns and treatment failures and HIV incidence rate. vi

9 FOREWORD Systems integration is critical to the AIDS response in South Africa. We have a generalized epidemic, but the highest prevalence of HIV in South Africa is among pregnant women. In our resolve to eliminate vertical HIV transmission we have to ensure that HIV and Maternal-Child health services are integrated. We have to treat patients holistically because although HIV is by far the biggest Public Health crisis in our country, it is not the only epidemic we are dealing with. We have a 63% Tuberculosis and HIV co-morbidity rate and these two diseases must be treated as two sides of the same coin. There is also a clear relationship between HIV and the pandemic of non-communicable diseases, such as diabetes mellitus and cancer, particularly cervical cancer. WHO/UNAIDS Reference Group estimated that million people were living with HIV and AIDS in South Africa in 2011, the highest number of people in any country. In the same year, it was estimated that South Africans died of AIDS-related causes, reflecting the huge number of lives that the country has lost to AIDS related deaths over the last three decades. Marking a change in South Africa s history of HIV the Government launched a major HIV counselling and testing campaign (HCT) in Since its implementation, the HCT campaign has had a notable impact on the availability and uptake of HIV testing and treatment. Confidently, the HIV epidemic in South Africa has stabilised albeit at a very high and an unacceptable level. According to the annual HIV sentinel surveys, stabilisation dates back to While this might be promising news, I am acutely aware of age, provincial and district disparities. I also take note that the HIV epidemic continues to rob the country of the socio-economically productive section of the population and disrupting the social fabric that holds our communities together. As you know, HIV prevalence is no longer a good measure of success for HIV prevention intervention programmes. Stabilisation of HIV prevalence means that HIV incidence (new infections) and mortality (AIDS related deaths) are balancing out. However, with one of the largest treatment programmes in the world we should see an increase in HIV prevalence as people living with HIV will be living longer. I, therefore, urge the public health sector to isolate out the contribution of the ART programme to the observed HIV prevalence rates on the findings of this report. True to the vision of Chris Hani, who once admonished: We cannot afford to allow the AIDS epidemic to ruin the realization of our dreams, on the 1st of December 2011, President Jacob Zuma launched the national HIV,STI and TB Strategic Plan that sought to galvanise South Africa s response to the dual epidemics up to This plan takes into account the United Nations High Level Meeting (HLM) targets and the investing for impact approach. The National Department of Health (NDoH) vii

10 strongly advocated for integration of the dual epidemics into a single strategic plan to efficiently respond. Most important to note is that the plan is founded on solid evidence of which HIV sentinel surveillance surveys has been a major contributor. Our response to AIDS goes beyond the Ministry of Health, it involves multiple sectors. We have made great progress towards reaching our treatment target goal of 80 percent coverage, classified by the WHO as being universal access. In mid-2011, following the launch of the HCT campaign in early 2010, the number of people on antiretroviral treatment had increased significantly from 923,000 in February 2010 to 1.4 million in May In October 2012, we reached the target of universal access to treatment as the total number of people receiving treatment reached 2 million (or 80 percent of all in need of treatment). The task before us is huge and tedious but we know that only by getting communities and community leaders involved can we successfully integrate supply with demand for HIV treatment and prevention services. And eventually vanquish the ravaging impact of HIV and AIDS to our society and economy. We will use the upcoming 29th Orange Africa Cup of Nations 2013 Soccer tournament towards the realisation of UNAIDS Three Zeros namely: Zero new infections, Zero discrimination, Zero AIDS-related deaths. DR. PAKISHE AARON MOTSOALEDI MINISTER OF HEALTH, (MP) DATE viii

11 AKNOWLEDGEMENTS I would like to extend my appreciation to all nurses in the public health clinics for their continued dedication and support over the past 22 years in the implementation of this survey and their professionalism in adherence to the survey protocol for the collection, handling and transportation of blood specimens to laboratories. A very special thanks to all the women who agreed to participate in the survey and made this report on HIV and syphilis trends possible. Special thanks goes to the National Department of Health s coordinating team, in particular Dr. Thabang Mosala, for her technical and managerial oversight of the survey and for providing strategic leadership in the compilation of this report. Many thanks also go to the Epidemiology and Surveillance Directorate staff, Ms. Manti Maifadi and Mr. Lusizo Ratya for taking the lead in coordinating the survey, compiling the tables and the figures and for conducting technical input in the review of the report; Mr. Anslen Koffman and Corrie Nagel for the administrative support. Our sincere gratitude also goes to the provincial survey co-ordinators: Mr. Z. Merile and Ms. Lindeka Mangesi (EC), Mr. R.A.Khajoane (FS), Dr. B. Ikalafeng (GP), Ms. N. Moodley and Ms. N. Mbana (KZN), Mr. E. Maimela (LP), Mr. M. Machaba (MP), Mr. M. Khumalo and Ms. T. Naicker (NC), Mr. L. Moaisi (NW), Dr. D. Pieneer and Dr. V. Appiah-Baden (WC), and their teams who coordinated of the survey in the respective provinces and districts. Sincerest gratitude, is also extended to the testing laboratories and coordinators: Ms. A. Burell (PE Complex Hospital), Mr. L. Hilderbrand and Mr. H. Van Der Merve (Pelonomi Hospital), Mr. B. Singh and Mr. L Reddy (Inkosi Albert Luthuli Hospital), Mr. T.J. Chephe and Ms. R. Diokana (MEDUNSA), Ms. R. Thompson (Middleburg Hospital), Ms. E. Weenink and Ms. B. Gool (Kimberly Hospital) and Mr. T. Stander and Ms R. Bester (Tygerberg Hospital) and all staff at these laboratories and Ms. B. Singh and Prof. A. Puren (NICD). Thanks to Mr. Calle Herdberg of HISP, who developed the new DHIS Antenatal Survey data capturing tool, collated all the national pooled dataset, validated the laboratory data and produced a clean 2011 dataset. The Department also acknowledges and expresses gratitude for the technical support from the HIV Surveillance Expert Task Team members and NDoH principals, who advised and assisted in the analysis and interpretation of the results, and provided scientific review of the report namely: Dr Yogan Pillay, Mr Thulani Masilela, Dr Thobile Mbengashe, Prof. Carl Lombard (MRC), Dr. Eleanor Gouws (UNAIDS), Dr Eva Kiwango (UNAIDS) and Prof. Rob Dorrington (UCT). MS. M.P. MATSOSO DIRECTOR GENERAL (HEALTH) DATE: ix

12 TABLE OF CONTENTS EXECUTIVE SUMMARY FOREWORD ACKNOWLEDGEMENTS ACRONYMS ii vii ix xii CHAPTER 1 1. INTRODUCTION 1 CHAPTER 2 2. METHODOLOGY Survey Design Sampling Sentinel population Selection of survey population Selection of sentinel surveillance sites Selection of Primary Sampling Units (PSU) Sample Collection Laboratory Methods Laboratory techniques Laboratory quality assurance Quality Control of Fieldwork Data Management and analysis Exclusions from analysis Calculation of confidence intervals Weighting Biases with antenatal data National Population Based Surveys vs. Sentinel Surveillance Sentinel series antenatal HIV surveys National population based surveys Extrapolation of HIV infection to the general population WHO/UNAIDS estimation process Ethical considerations Reliability of this report results 8 CHAPTER 3 3. RESULTS - HIV PREVALENCE Characteristics of Survey Population Participation The demographic characteristics of the sample Population National participation rate by age National participation rate by population group 8 x

13 3.2 National HIV Prevalence Trends ( ) HIV Prevalence by Province, HIV Prevalence by District, HIV Prevalence Trends by Age, HIV Prevalence Trends by Individual Province, The Eastern Cape Province The Free State Province The Gauteng Province The KwaZulu-Natal Province The Limpopo Province The Mpumalanga Province The North West Province The Northern Cape Province The Western Cape Province Extrapolation of HIV Infection to the General Population 56 CHAPTER 4 4. RESULTS - SYPHILIS PREVALENCE Syphilis Prevalence by Province 58 CHAPTER 5 5. CONCLUSION AND RECOMMENDATIONS 61 LIST OF REFERENCES 64 xi

14 ACRONYMS AIDS ANC ART BCC BoD BSS CCMT CI DHIS DoH EC ELISA EPP FS GA HCW HCT HIMME HIV HSRC HST KZN LP MDG MEDUNSA MP MRC NC NDoH NHC NHLS NICD NTP NSP NW PCR PAC PHC Acquired Immuno Deficiency Syndrome Antenatal Care Anti-retroviral Therapy Behavior Change Communication Burden of Disease Behavioural Surveillance Survey Comprehensive Care Management and Treatment 95% Confidence Interval District Health Information System Department of Health Eastern Cape Enzyme Linked Immuno Sorbet Assay Estimation and Projection Package Free State Province Gauteng Province Health Care Worker HIV Counseling and Testing Health Information Management Monitoring and Evaluation Human Immunodeficiency Virus Human Science Research Council Health Systems Trust KwaZulu-Natal Province Limpopo Province Millennium Development Goals Medical University of South Africa Mpumalanga Province Medical Research Council Northern Cape Province National Department of Health National Health Council National Health Laboratory Service National Institute for Communicable Diseases National Tuberculosis Programme National Strategic Plan for HIV, AIDS and STI North West Province Polymerase Chain Reaction Provincial AIDS Councils Primary Health Care xii

15 PMTCT PPS PSU QA RPR SA SANAC STI TB UCT UKZN UNAIDS UNGASS UNICEF UNISA USAID WC WHO Prevention of Mother-to-Child Transmission Probability Proportional to Size Primary Sampling Unit Quality Assurance Rapid Plasma Reagin (A screening test for syphilis) South Africa South Africa National AIDS Council Sexually Transmitted Infection Tuberculosis University of Cape Town University of KwaZulu-Natal United Nations Joint Program on HIV & AIDS United Nations General Assembly Special Session on HIV & AIDS United Nations Children s Fund University of South Africa United States Agency for International Development Western Cape Province World Health Organisation xiii

16 CHAPTER 1 1. INTRODUCTION The year 2012 marks 30 years since the first case of HIV case was reported, 15 years since treatment became a reality, 10 years since the United Nations General Assembly Special Session on HIV/AIDS and five years since our commitment to achieve universal access to HIV prevention, treatment, care and support. This report provides evidence of where we are and weighs that against our vision for the future to achieve zero new infections, zero discrimination and zero AIDS-related deaths. According to Kofi Annan (2012), the Millennium Development Goal of halting and starting to reverse the spread of AIDS by 2015 is still a formidable challenge. Mr Sidibe of UNAIDS (2012) further emphasized that while perspectives differ, one simple truth emerges, that we cannot break the arc of this epidemic, while we still have five people newly infected for every three we initiate on treatment in 2010, i.e. if we adopt a business as usual approach. South Africa has been more affected by the HIV and AIDS epidemic than any country in the world. AIDS has stolen the lives of thousands of our children, brothers and sisters and has left many orphaned and vulnerable children and even some child headed households. The epidemic has transformed our country in some ways for the better given the coordinated, integrated and cohesive response to mitigate the epidemic. Our losses have strengthened us under a committed and informed health Minister who embodies the uncompromising, stop agonizing and get it done approach. Our current Minister Motsoaledi further emphasizes that while we are learning, we are far from getting it right, but we know that only by getting communities and community leaders to get involved and participate in our planning processes will it be that we can successfully integrate supply with demand for HIV treatment and effective prevention services. Estimates of HIV prevalence in South Africa are mainly based on Department of Health surveillance among pregnant women attending sentinel antenatal clinics (ANC) and Human Sciences Research Council periodic national community survey every 3 years. The department of Health has collected this data on an annual basis since The antenatal survey is only one tool used to track the spread of HIV. A range of other community-based and incidence studies are employed in order to obtain a better understanding of the dynamics of the epidemic in the country. In addition, control of HIV in South Africa involves multi-sectoral approaches involving government, research and academic institutions, civil society, non-governmental organisations, community based organisations and the private sector. This report presents the findings of the 22nd National Antenatal Sentinel HIV and Syphilis Prevalence Survey conducted in 2011 in all nine provinces and 52 Health districts of South Africa and shows prevalence trends of the HIV epidemic from 1990 to The purpose of the survey The purpose of undertaking an annual sentinel antenatal point prevalence survey is to assess the HIV sero-prevalence amongst first time antenatal clinic attendees (seen as a particularly suitable sentinel group to represent most closely the HIV prevalence of the generally sexually active part of the population) and to assess trends in HIV prevalence over time. The antenatal service provides for pregnant women and serves as an important point of contact with the health care services with this female age group. So far this survey has the largest representative sample of the sexually active population being tested for HIV infection in this country. 1

17 1.2 The general objective The general objective of the survey is to determine the distribution of HIV and syphilis infection among pregnant women attending public health antenatal clinics at National, Province and District level and disaggregated by age. 1.3 The primary objectives are: To assess HIV and syphilis sero-prevalence among women attending public sector antenatal clinics; To monitor HIV and syphilis trends over time among women attending public antenatal clinics and; To use this data for estimation and projection of HIV sero-prevalence trends and the burden of AIDS in the general population. 1.4 The secondary objectives are: To estimate the national, provincial and district level prevalence of HIV and syphilis infection among the adult population (15-49 years) using women attending antenatal clinics in selected public health sector clinics providing antenatal care services as a proxy; To identify trends in both HIV and syphilis prevalence in the country; To estimate the number of children, men and those HIV infected persons who need to be receiving ART 2

18 CHAPTER 2 2. METHODOLOGY During the month of October 2011 the 22nd National Antenatal Sentinel HIV and syphilis Prevalence Survey was conducted in South Africa, across the nine provinces and 52 health districts using the standard unlinked and anonymous methodology (WHO/UNAIDS). The survey is used as a proxy to estimate the trend in the prevalence of HIV and syphilis among pregnant first bookers presenting to a public sector ANC facility for the first time. A total of pregnant women were targeted to participate in The number of Primary Sampling Units (PSU) was 1 445, sentinel sites selected from all 52 public health districts in South Africa. Health Care Service-based HIV surveillance is recommended by WHO as an entry point, because PMTCT services are available in all sites to provide women an option of HIV Counselling and Testing and further care. The survey design is an unlinked anonymous testing method using blood samples collected for other purposes in selected sentinel primary health care facilitates. This sampling approach is convenient because, as part of the antenatal care services it is mandatory to routinely draw blood from the first bookers, and this minimises participation bias and reduces costs. In addition, pregnant women are universally the most common Sentinel Population for HIV and in South Africa the most common mode of transmission is the heterogeneous sexual route. It is not perfectly representative of all women and even less of men, children and non-pregnant women, but it is an important means of coverage for countries that have a generalized HIV epidemic (i.e. where HIV prevalence among pregnant women is >2%) and it also has a wide geographic coverage (urban; informal settlements and rural communities). 2.1 Survey design The national antenatal HIV and syphilis prevalence survey is an anonymous, unlinked, crosssectional survey targeting pregnant women attending antenatal clinics in the public health sector. Only first-time attendees are recruited, to minimize the chance of any woman being included more than once. Since 2006, this survey has expanded its sample population to target pregnant women recruited from ±1450 PSU compared with women recruited from 461 clinics in This has expanded geographic coverage considerably to include a representative sample from all 52 health districts in all the nine provinces and includes urban and rural comparison. 2.2 Sampling Sentinel population Pregnant women attending ANC services at public health facilities were used as the target population as they are sexually active; constitute an easily accessible and stable population, and are more likely than other groups to be representative of the general population. In addition, they obtain antenatal care at facilities that draw blood as part of routine medical services offered to this group Selection of survey population Inclusion criteria All pregnant women attending antenatal for the first time clinic during their current pregnancy were eligible for inclusion. Exclusion criteria Pregnant women who had previously visited the ANC clinic during their current pregnancy during 3

19 the survey period were excluded (to avoid duplicate sampling during the same month). No pregnant women were excluded from participation on the basis of their known HIV status Selection of sentinel surveillance sites The basic goal was to select sentinel surveillance sites representative of the population size estimate of the area to be surveyed. Sentinel sites were selected using the Probability Proportional to Size (PPS) method as this combines a random approach with a bias towards the larger clinics. By using this approach, it made the analysis easier as it introduced a natural weighting process Selection of Primary Sampling Units (PSU) The following are the criteria that were applied in selecting sentinel surveillance sites to be eligible for inclusion in the sample: Any randomly selected health establishment in the public health sector, providing antenatal care services and routinely drawing blood from attendees on the first visit of the current pregnancy with facilities to store sera at 4 C; The sentinel site should provide ANC services to sufficient first time antenatal clinic attendees to ensure that a minimum of 20 first time bookers be recruited over one month; Availability of transport arrangements in place that will allow for biological specimens to be taken to a reference laboratory within 24 hours or if the blood samples are centrifuged then transfreed to referral laboratory within 72 hours. The clinic staff must be willing to cooperate and have the capacity to conduct the survey. It should be noted that no other criteria were applied in selecting sites. In particular, sites were not selected specifically to monitor either high risk or low risk sub-populations, nor with the aim of monitoring interventions. These criteria are strictly adhered to in order to limit bias and promote comparability. 2.3 Sample collection Full blood analysis for pregnant first bookers at the ANC clinic was used as an entry point for HIV testing using anonymous unlinked procedures. One blood sample was taken by veni- puncture and labelled with the bar code number of the individual pregnant woman and stored at 4ºC. The demographic details of the participants, with the exclusion of any particulars from which it may be possible to ascertain the identity of a patient, were collected using a standardized collection form (Appendix A, 2010 HIV Report). The data collection form with the woman s demographic details was labelled with the same bar code number. At the close of each day the supervisors checked the forms against the blood samples for any mistakes and for completeness. The samples, together with the forms, were transported in a cooler box to the participating provincial laboratory where HIV and syphilis testing was done. 2.4 Laboratory methods Laboratory techniques In accordance with the recommendations of the WHO on HIV screening for surveillance purposes, blood samples were tested with one ELISA (Abbot Axysm System for HIV-1 HIV-2) assay. The samples were also screened for active syphilis using the RPR test. Participating laboratories included the National Health Laboratory Services (NHLS) laboratories in Bloemfontein, Johannesburg, Kimberley, Middleburg, Port Elizabeth and Stellenbosch, MEDUNSA, and the Virology laboratory of the University of KwaZulu-Natal. 4

20 2.4.2 Laboratory quality assurance Internal quality assurance was the responsibility of the individual laboratories. While most laboratories participate in external quality assurance programmes, for the purposes of this study the NICD was responsible for overall external quality assurance. The National Institute for Communicable Diseases (NICD) compiled a battery of 20 HIV positive and negative sera which was sent to each participating lab to test. Results were sent back to the NICD. In addition, each laboratory compiled a batch of 20 sera comprising HIV positive and negative sera, including some borderline cases. These were forwarded to the NICD for confirmatory testing. After completion of the survey the NICD produced a quality assurance report on the performance of the laboratories for HIV testing and University of Limpopo (Medunsa Campus) Microbiology Department for RPR testing. 2.5 Quality control of fieldwork District level monitoring of the sentinel sites was done weekly by a team from the district health office. Provincial coordinators also undertook provincial level monitoring and visited the sites in their province. The national team conducted supervisory visits to at least two districts per province. The main purpose of the visits, was to monitor that the protocol was being adhered to by observing practices and reconciling the number of submitted specimens to the calculated expected number which was derived from the routine data collection. A monitoring checklist was used to ensure that monitoring and supervision was standard for all sites. 2.6 Data management and analysis Raw data was captured at provincial level, using the Antenatal HIV and Syphilis Prevalence Survey DHIS 1.4 Patient Module. This database is designed with range restrictions to ensure that data captured are not out of range. Additionally extensive internal data consistency checks against the original data capture form were done by each provincial coordinator to ensure the data were accurate. After data were entered, frequency tables were produced for each data element to identify missing or inconsistent values that may have originated from incorrect entry of data into the computer. Further data cleaning and validation and quality assessment was done at the national office. Data analysis was carried out by independent Statisticians, Actuarial Scientists and Epidemiologists from the NDoH, MRC, UCT, WHO/UNAIDS. The analysis was mainly descriptive and focused on determining national, provincial, district and age group specific prevalence rates of HIV and syphilis. Data analysis was done as follows: 1. Generating basic frequency tables on the whole sample. 2. Determining sample weights at provincial level using the 2011 midyear populations estimates for women year old. No adjustments were made for refusals, inadequate blood samples or under/over sampling. 3. Estimating the national and provincial HIV and syphilis prevalence using women age years old in the sample. New district demarcation data where used as strata there was no difference from using old districts as strata observed. 4. Estimating age specific HIV prevalence. 5. Estimating HIV at the district level. 5

21 2.6.1 Exclusions from analysis The following entries were excluded from the analysis: Those which had no HIV status result was not indicated. Those with no age of the survey participant Calculation of confidence intervals For the 95% confidence intervals, the normal approximation to the binomial distribution was used. In a few cases where the sample size or prevalence was small, the exact binomial calculation was used and adjusted for the design effect of the domain Weighting The national estimate was weighted according to the total number of women aged years in the different provinces using the StatsSA mid-year population estimates current at the time of the survey. Given that the sentinel sites were chosen on a probability proportional to size basis by district, the sampling period is fixed and the districts are self-weighting, the provincial prevalence estimates were simply calculated as the total of the results from the districts in the provinces Biases with Antenatal data Only pregnant women and not all reproductive age women were tested Only pregnant women who attend public sector antenatal clinics are tested Clinics selected may not be representative In general terms, Antenatal survey data: underestimate prevalence in the general female population overestimate prevalence in the rural population overestimate prevalence in the (young and old adolescent) year old 2.7 The National Sentinel Antenatal HIV surveillance vs National Population Based (Community) HIV surveys Sentinel surveillance and population-based surveys each have strengths and weaknesses but taken together provide complementary information and can provide a clear picture of both overall trends and geographic distribution of HIV in South Africa Sentinel series antenatal HIV surveys Strengths Easy access to a cross-section of sexually active women from the general population using the public health sector facilities. Testing among pregnant women is a good proxy for prevalence in the general population. Provides data on trends in the HIV epidemic over time. Biases are recognized and can be corrected. Geographical coverage can be expanded. Weaknesses Women attending ANC may not be representative of all pregnant women ANC does not provide data on the prevalence among men. Estimates for men are based on assumptions about the ratio of male to female prevalence derived from community based surveys. 6

22 2.7.2 National population (community) based HIV surveys Strengths Can provide representative estimates of prevalence in the general population (for generalized epidemics) as well as for different subgroups. Results can be used to adjust estimates from antenatal sentinel surveillance Provides an opportunity to link HIV status with social, behavioural and other biomedical information. Weaknesses Sampling from households may not adequately represent high risk and mobile populations. Non-response can bias population-based estimates. Population based surveys are expensive and logistically difficult to carry out. 2.8 Extrapolation of HIV infection to the general population WHO/UNAIDS estimation process The Estimation and Projection Package (EPP) recommended by UNAIDS was used to estimate and project adult HIV prevalence from surveillance data and to estimate HIV incidence from prevalence and ART coverage data. After surveillance data from various sites and years showing HIV prevalence among pregnant women were included in the model, the package fitted the best epidemic curve, scaled to be consistent with estimates of the general population prevalence. Separate estimates and time trends were developed for each of the provinces, and combined within EPP to produce a national estimate for HIV prevalence and its trends over time. The resulting national estimated adult HIV prevalence was transferred to a demographic package (Spectrum: a computer modelling for demographic projections) to calculate the number of people infected and other variables, such as the number of adults and children who need to be receiving ART, and the estimated number of AIDS deaths and other information. Adjusting HIV prevalence curve using EPP: Adjusting for race-based relative attendance rates at ANC: Based on race-standardized prevalence. Adjusting for the use of HIV prevalence among pregnant women: Based on ratio of prevalence among adults in the general population, using data from the HSRC National population based HIV survey and prevalence among pregnant women. Required inputs in Spectrum are: Country data on Demographic data projected by age and sex over the time period of interest Adult prevalence / Incidence curve MTCT program description PMTCT coverage Adult ART coverage Child treatment coverage Epidemiologic assumptions Effect of HIV on fertility Progression from infection to need for treatment and AIDS death Age distribution of infections Sex ratio of incidence Mother-to-child transmission rates by regimen and feeding options Effect of child treatment 7

23 2.9 Ethical considerations Participation in the survey was voluntary, with informed consent for answering the questions on the forms and for collecting the blood samples. For reasons of confidentiality, testing was done on anonymous unlinked samples. A unique bar code was allocated to each of the participants and it is this number that was recorded on the form and also used for labelling the blood samples and linking laboratory results with demographic data. The bar code was used to link the demographic information with the laboratory results while maintaining anonymity of the survey participant. For future surveys, a revised proposal will be submitted to the MRC Ethics Committee to request that the results of all HIV positive women be communicated to the clinic to ensure that each woman is aware of her status and is provided with a choice to participate in the PMTCT programme, without the survey investigators knowledge of her identity Reliability of this report results To ensure that we publish a robust report which provides reliable scientific evidence the National DoH does the following: 1. Continuously liaises with the scientific HIV Surveillance Task Team 2. Uses a reliable DHIS data capturing and verification management tool to ensure data validity and plausibility. 3. Revises the protocol with the provincial survey coordinators and laboratory technicians annually before the implementation of the next survey 4. Receives critical technical inputs from experts in different fields of Public Health 5. The report goes through a thorough scientific peer-review process 6. Data is subjected to internal and external independent scientific analysis including internationally renowned Epidemiologists, Bio-Statisticians, DoH Epidemiology directorate scientific technical staff, WHO HIV Specialists, UNAIDS HIV M&E Specialists and Statisticians, MRC and UCT. 8

24 CHAPTER 3 RESULTS HIV PREVALENCE 3.1 Characteristics of survey population This section of the report firstly presents the characteristics of the survey population and will present the descriptive summary of HIV prevalence at national and provincial level, by district and by age distribution Participation Facility and individual level A total of out of the targeted pregnant women attending antenatal care sentinel clinics, at selected public health facilities in the nine provinces, representing the 52 health districts in South Africa participated in the survey during October 2011 (Table 1). The sample population realization rate was 92.9% with women recruited from sentinel clinics. Table 1: Sampled population distribution by province, 2009 to 2011 Province N % N % N % Eastern Cape Free State Gauteng Kwa-Zulu Natal Limpopo Mpumalanga North West Northern Cape Western Cape Total N = Realised sample size The demographic characteristics of the sample population National participation rate by age There are distinctly different risk factors that lead to HIV infection among infants and adults, hence age is an important risk factor and is central to monitoring the epidemic among the highly sexually active group, while at the other end of the age spectrum vertical transmission from the infected mother is the most important and significant risk factor. The HIV prevalence in the year age group is crucial when reporting the outcome of the AIDS Millennium Development Goal 6, Target 7, Indicator 18. The age pattern of the pregnant women recruited in the survey was similar to the previous five surveys (Table 2). The age distribution of pregnant women who participated ranged from girls aged 10 years to women aged over 50 years as shown below in Table 2 and Figure 1. The antenatal distribution in the past 5 surveys was concentrated in the year old age group, which was more than 30% of the survey population. Antenatal women older than 39 years and younger than 15 years were underrepresented in the survey compared to the 15 to 39 year old. 9

25 Table 2: Sampled population distribution by age group, 2009 to Age in years N % N % N % < > Not specified Total N = Realised sample size. Figure 1: National distribution of survey participants by age group, National participation by population group The distribution by race of the women recruited in the 2011 survey was similar to the previous three surveys as shown in Table 3. Eighty nine (89.9%) of the survey participants were African women, while 9.1% were Coloured. Asians and Whites together accounted for less than 1% of the total tested. The number of Asians and Whites who participated was too small to provide reliable estimates for these two population groups. More than 75.4% of the survey participants were single women, of which 30.1% were HIV infected, only of the pregnant women sampled were married and of these 24.6% were HIV infected. Furthermore, 102 divorced women who participated in the survey, 19.6% were also HIV infected (Table 4). 10

26 Table 3: Sampled population distribution by population group participation in 2009 to Population group N % N % N % African Asian Coloured White Not specified Total N = Realised sample size. Table 4: Association between the demographic and background characteristics and HIV outcome status of survey participants, 2010 and 2011, is presented below Variable Level N % HIV Prev. N % HIV Prev. Population African group Asian Coloured White Level of Education None Primary ,7 Secondary Tertiary Marital Status Single Married Widowed Divorced Parity None Number of live born children More than Age of Partner < > ** ** ** No participants in this age group 11

27 3.2 National HIV prevalence trends ( ) In 2011, the overall HIV prevalence amongst antenatal women was 29.5% (95%CI ) a decrease of 0.7% from 30.2% (95%CI: ) in The 2011 HIV prevalence estimate confidence interval is in line with estimates from The national HIV prevalence estimate amongst the women surveyed has remained stable around 29% over the past five years (Figure 2,3 and 4): 29.4% (95%CI: ) in % (95%CI: ) in % (95%CI: ) in % (95%CI: ) in % (95%CI: ) in 2011 Thus the 2010 overall HIV estimate gave a higher estimate than expected which was 30.2% (95%CI ). The HIV prevalence trends from 1990 to 2011, show that the Error Bars ( ) between 2007 to 2011 overlap, which indicate that there is no statistical difference in the HIV estimates in the past 5 years as shown in Figure 2 and HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 2: The HIV prevalence trends among antenatal women, South Africa 1990 to The estimates from 2006 are based on a different sample size to the previous years. 12

28 HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Year Figure 3: HIV prevalence epidemic curve among antenatal women, South Africa, 1990 to HIV prevalence (%) Year Figure 4: The five data set points that show the plateauing of the HIV curve from 2007 to

29 3.3 HIV prevalence by province The provincial HIV prevalence estimates vary between provinces (Figure 5), where some show a slight decrease, in Gauteng the HIV prevalence has decreased from 30.4% in 2010 to 28.7% in 2011 and most notably in KwaZulu-Natal from 39.5 % to 37.4%, a surprising decline of 2.1%. There was an increase in HIV prevalence in Free State from 30.6% in 2010 to 32.5% in 2011, and the North- West from 29.6% in 2010 to 30.2% in Limpopo is showing a steady increase from 21.4% in 2009 to 22.1% in 2011 and Mpumalanga HIV prevalence has increased from 35.1% in 2010 to 36.7% in Limpopo 22.1% North-West 30.2% Gauteng 28.7% Mpumalanga 36.7% Northern-Cape 17.0% Free State 32.5% KwaZulu- - Natal 37.4% Western Cape 18.2% Eastern Cape 29.3% Key HIV prevalence range <20% 20-30% >30% Figure 5: HIV prevalence distribution by province, South Africa, 2011 The results of the 2011 survey still show that the highest HIV prevalence rates are located in the Central and Eastern parts of the country, and the lowest prevalence rates in the Western Cape, Northern Cape and Limpopo. KwaZulu-Natal has the highest HIV prevalence followed by Mpumalanga, Free State and North-West with overall prevalence rates greater than 30.0%. Limpopo, Gauteng and the Eastern Cape recorded prevalence rates between 20.0% and 30.0% and Northern Cape and Western Cape are the only provinces that have HIV prevalence rates below 20.0% (Table 5 and Figure 6). 14

30 Table 5: HIV prevalence among antenatal women by province, 2009 to N % Prev. 95% CI N % Prev. 95% CI N % Prev. 95% CI SA EC FS GA KZN LP MP NW NC WC N = Realised sample size. Figure 6. HIV prevalence trends among antenatal women by province, SA, 2009 to

31 3.4 HIV prevalence by district In 2011 there was no single district that recorded prevalence below 5%. Eleven districts recorded HIV prevalence between 10% and 20%. The number of districts recording prevalence rates between 30% and 40% has increased from 14 out of 52 in 2010 to 19 out of the 52 districts in 2011 (Figure 7). It was observed that there was a decrease of the number of districts which recorded HIV prevalence greater than 40%. There were five districts in 2010 to 3 districts in 2011 which recorded HIV prevalence above 40%. In 2011, twenty six (26) out of the 52 districts recorded HIV prevalence rates prevalence rates below the national average of 29.5% compared with 28 out of 52 districts that recorded below the national average of 30.2% in There were 24 out of 52 districts that recorded HIV prevalence rates above the national average in 2010; whereas there were 25 out of 52 districts that recorded HIV prevalence rates above the national average in 2011 (Figure 7 and 8). There was considerable variation in HIV prevalence rates between the 52 health districts observed over the four-year period , particularly where the sample size in a district was small, where districts were merged (i.e. Tshwane and Metsweding) and where there are new district demarcations like Buffalo City, making it difficult to interpret any trends in the current report. The districts are clearly heterogeneous with respect to the epidemic, with prevalence rates ranging from a high of 46.1% in Gert Sibande in Mpumalanga, the highest district HIV prevalence ever recorded in this province, followed by Ugu and Mkhanyakude in KZN which recorded 41.7% and 41.1% respectively. The lowest HIV prevalence was 6.2% in Namaqua in the Northern Cape. When data are pooled over the five years this heterogeneity persists. The new Buffalo City district has recorded the highest HIV prevalence of 34.1% in the Eastern Cape. There was a notable decrease in HIV prevalence observed in J.T Gaetsewe from 27.5% in 2010 to 17.7% in 2011 (Table 12). 16

32 National HIV prevalence (29.5%) Eastern Cape Free State Gauteng KwaZulu-Natal Limpopo Mpumalanga Northern Cape North West Western Cape Figure 7 : HIV prevalence among antenatal women by district, South Africa,

33 Number of Districts per prevalence range , 2010 and <10% 10% - 20% 20% - 30% 30% - 40% >40% Below National Average HIV prevalence range Above National Average Figure 8 : Number of Districts per HIV prevalence range , 2010 and HIV prevalence by age HIV prevalence estimate is the most important indicator used to provide empirical evidence when monitoring HIV incidence (new infections). The survey participants aged years accounted for 49.4% of the survey participants. HIV prevalence in this age group has been suggested as a proxy measure for average incidence in the youth because of sexual onset and hence prevalent infections are assumed to be recent while this age group is less likely to be affected by AIDS mortality. The HIV prevalence among the year old pregnant women was 21.8% (95%CI: ) in 2010 compared with 20.5 % (95%CI: ) in 2011, a decline of 1.3%. The specific AIDS MDG target is that by 2015 the expected HIV prevalence reduction should be 25% less than the baseline prevalence of 23.1% in The findings of monitoring trends in this age group in South Africa show that we should not relent of our collective efforts to achieve this AIDS MDG target (Figure 9). The dot-plot (Figure 8A) of the HIV prevalence by age group for the nine provinces are shown. The national prevalence estimate of ( %) is plotted to facilitate comparison across provinces. The HIV prevalence is given as proportions. A notable feature across all provinces is the approximate linear increase in prevalences over the four youngest age groups from years of age. 18

34 Eastern Cape Free State Gauteng KwaZulu-Natal Limpopo Mpumalanga North West Northern Cape Western Cape HIV Prevalence Figure 8A : Comparison of age specific HIV prevalence profiles between provinces Vertical dotted line is the overall HIV prevalence 19

35 Figure 9: The HIV prevalence among the year old pregnant women in South Africa, 2001 to Source: Antenatal Sentinel HIV Survey, National Department of Health Table 6: HIV prevalence among antenatal women by age groups (years), South Africa, 2009 to Age group (Years) N % Prev. 95% CI N % Prev 95% CI N % Prev. 95%CI * N = Realised sample size; CI= Confidence Interval. 20

36 Figure 10: HIV prevalence trends among antenatal women by age group, South Africa, 2009 to Nationally, the HIV prevalence among women in the age group years remains the highest increasing from 41.5% in 2009 to 42.2% in This age group constituted 14.9% of the sampled survey population. The age groups years, years and years show a small decrease in HIV prevalence whereas the prevalence in the women 35+ years continues to increase significantly (Figure 10 and Table 6). The HIV prevalence has increased in the 3 age categories as summarized below: Among the years the HIV prevalence has increased from 35.4% in 2009 to 39.4% in 2011, by 4.0%. Among the years the HIV prevalence has increased from 25.6% in 2009 to 31.7% in 2011, by 6.1%. Among the years the HIV prevalence has increased from 23.9% in 2009 to 30.4% in 2011, by 6.5%. 3.6 HIV prevalence trends by individual province For each province, comparisons of the provincial, district level HIV prevalence rates are reported from 2009 to Due to the smaller sample size in some districts, the sampling error is much larger than at the provincial level. Therefore changes of 4% in either direction between the years within a district can be expected simply due to chance, if the sample size was less than 500, and even greater for smaller sample sizes. NB: It is difficult to make any inferences about trends in prevalence in provinces where there are significant changes in municipal boundaries, like Gauteng and Eastern Cape,hence the maps showing changes in distribution pattern in those provinces are not presented in this report. 21

37 3.6.1 THE EASTERN CAPE PROVINCE In 2011, the Eastern Cape provincial HIV prevalence amongst antenatal women was 29.3% (95% CI: ). The estimated overall HIV provincial prevalence in this province has decreased by 0.6%, when compared to 29.9% in 2011 (Figures 11 and 12). The trends in district prevalence rates from 2009 to 2011 are presented in Table 7. The new Buffalo City District recorded an HIV prevalence of 34.1% of the 552 pregnant women recruited and is the only district in this province which recorded HIV prevalence above 30% HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 11: HIV prevalence trends among antenatal women, Eastern Cape, 1994 to HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 12: The HIV prevalence epidemic curve among antenatal women, Eastern Cape, 1994 to

38 Table 7: HIV prevalence among antenatal women by district in the Eastern Cape, 2009 to N % Prev. 95% CI N % 95% CI N % 95% CI Prev. Prev. Provincial Alfred Nzo Amatole , Buffalo City Cacadu Chris Hani N.M.M O.R. Tambo Joe Gqabi N = Realised sample size; CI= Confidence Interval. As might be expected, significant year-on-year changes were observed in the districts with smaller sample sizes). The HIV prevalence range from lowest in Cacadu district from 25.8% in 2011 to 34.1% in Buffalo City (Figure 13 and Table 7). Figure 13: HIV prevalence trends among antenatal women by district, Eastern Cape, 2009 to

39 Table 8: HIV prevalence among antenatal women by age group, Eastern Cape, 2009 to N % Prev. N % Prev. N % Prev. * > ** ** ** * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size. ** No participants in this age group Figure 14: Sampled population distribution by age group, Eastern Cape,

40 3.6.2 FREE STATE PROVINCE In 2011, the Free State provincial HIV prevalence amongst antenatal women was 32.5% (95%CI: ) an increase of 2.4% when compared with 30.1%recorded in 2009 (Figures 15, 16 and Table 9). Table 9: HIV prevalence among antenatal women by district, in the Free State, 2009 to N % Prev. 95% CI N % Prev. 95% CI N % Prev. 95% CI Provincial Fezile Dabi Lejweleputswa Mangaung Metro T.Mofutsanyana Xhariep N = Realised sample size; CI= Confidence Interval HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 15: HIV prevalence trends among antenatal women, Free State, 1994 to

41 40.0 HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 16: The HIV prevalence epidemic curve among antenatal women, Free State, 1990 to Figure 17: HIV prevalence trends among survey participants in the by district in the Free State, 2009 to

42 The Xhariep district s HIV prevalence decreased significantly by 8.7%, from 25.7% in 2009 to 17.0% in The prevalence in this district has jumped up again in 2011 to 26.1%. Three districts of the Free State in 2011, compared with four districts in 2010, recorded prevalence above 30%. The interpretation of HIV prevalence in districts with sample size less than 400, need to be interpreted with caution. Motheo is the only district that has shown a decrease in prevalence from 32.1% in 2010 to 29.9% in 2011(Figure 17 and Table 9). Lejweleputswa recorded an increase of 4.2% from 30.0% recorded in 2010 to 34.2% in 2011, followed by Fezile Dabi with an increase of 2.7% from 32.9% in 2010 to 35.6% in The changes in HIV infection and distribution in the Free State from 2008 to 2011 is shown in Figure 19. Table 10: HIV prevalence among antenatal women by age group, Free State, 2009 to N % Prev. N % Prev. N % Prev. * >49 ** ** ** ** * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size. ** No participants in this age group Figure 18: Sampled population distribution by age group, Free State,

43 FS 2008 Lejweleputswa 33.8% Fezile Dabi 34.5% Thabo Mofutsanyana 33.1% 2009 Lejweleputswa 33.4% Fezile Dabi 27.9% Thabo Mofutsanyana 31.3% Xhariep 26.9% Mangaung 31.6% HIV prevalence 20-30% 30 40% Xhariep 25.7% Mangaung 27.8% HIV Prevalence 20-30% 30 40% 2010 Lejweleputswa 30.0% Fezile Dabi 32.9% Thabo Mofutsanyana 30.7% 2011 Lejweleputswa 34.2% Fezile Dabi 35.6% Thabo Mofutsanyana 31.9% Xhariep 17.0% Mangaung 31.1% HIV prevalence 20-30% 30 40% Xhariep 26.1% Mangaung 29.9% HIV Prevalence 20-30% 30 40% Figure 19: HIV prevalence distribution among survey participants in the by district in the Free State, 2008 to

44 3.6.3 GAUTENG PROVINCE In 2011, the Gauteng provincial HIV prevalence amongst antenatal women was 28.7% (95%CI ). The overall prevalence in Gauteng has decreased from 29.8% in 2009 to 28.7% in 2011(Figures 20, 21 and Table 11). Table 11: HIV prevalence among antenatal women by district in the Gauteng, 2009 to N % Prev. 95% CI N % Prev. 95% CI N % Prev. 95% CI Provincial City of JHB Ekurhuleni Metsweding Sedibeng Tshwane West Rand HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 20: HIV prevalence trends among antenatal women, Gauteng, 1994 to

45 40.0 HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 21: The HIV prevalence epidemic curve among antenatal women, Gauteng, 1994 to The results of this survey suggest that the overall HIV prevalence in Ekurhuleni and Tshwane has decreased from 2009 to The highest HIV prevalence of 32.3% was recorded in West Rand in 2011, followed by Ekurhuleni with a prevalence of 30.1%(Figure 22 and Table 11). Figure 22: HIV prevalence trends among survey participants by district in the Gauteng, 2009 to

46 Table 12: HIV prevalence among antenatal women by age group, Gauteng, 2009 to N % Prev. N % Prev. N % Prev. * >49 ** ** ** ** * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size. ** No participants in this age group Figure 23: Sampled population distribution by age group, Gauteng,

47 3.6.4 KWAZULU-NATAL PROVINCE In 2011, KwaZulu-Natal provincial HIV prevalence amongst antenatal women has decreased from 39.5% to 37.4%. However, KwaZulu-Natal remains the province with highest HIV prevalence since 1990 (Figures 24 and 25). Two districts have shown a large HIV decrease in prevalence,i.e., UMzinyathi from 31.1% in 2010 to 24.6% in 2011 and ILembe from 42.3% in 2010 to 35.4 % in UMzinyathi still has the lowest HIV prevalence, (Table 13 and Figure 26). In 2011, six out of 11 districts in KwaZulu-Natal viz: Amajuba, Sisonke, Zululand, uthugela, Ethekwini and umgungundlovu have shown a slight decrease in HIV prevalence estimates below 40%. The 2011 survey has again recorded the same two districts with prevalence rates over 40%: Ugu and Umkhanyakude recording 41.7% and 41.1% respectively Table 13: HIV prevalence among antenatal women by district, in KwaZulu-Natal, 2009 to N % Prev % CI N % Prev. 95% CI N % Prev. 95% CI Prov Amj Snk Ugu Umk Umz Utng Uthk Zul ethk ILe Umg N = Realised sample size; CI= Confidence Interval. Amj = Amajuba; Snk = Sisonke; Umk = UMkhanyakude; Umz = UMzinyathi; Utng = UThungulu; Utkh = UThukela; Zul = Zululand; ethk = ethekwini; ile = ilembe; Umg = umgungundlovu 32

48 HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 24: HIV prevalence trends among antenatal women, KZN, 1990 to HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 25: The HIV prevalence epidemic curve among antenatal women, KZN, 1990 to HIV Prevalence (%) Ama Snk Ugu Umk Umz Utk Uth Zul eth ile umg KZN Districts Amaj = Amajuba; Snk =Sisonke; Umkh =Umkhanyakude; Umzny = Umzinyathi; Utgl = Uthungulu; Utkl = UThukela; Zulu = Zululand; etkn: = ethekwini; ilem = ilembe; umg = umgungundlovu Figure 26: HIV prevalence trends among survey participants by district in the KZN, 2009 to

49 The changes in HIV prevalence distribution by district in KwaZulu-Natal from 2008 to 2011 are shown in Figure 28. Table 14: HIV prevalence among antenatal women by age group, KwaZulu-Natal, 2009 to N % Prev. N % Prev. N % Prev. * > * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size Figure 27: Sampled population distribution by age group, KwaZulu-Natal,

50 KZN 2008 Sisonke 35.8% Amajuba 34.7% UThukela 38.6% Umzinyathi 29.2% UMgungundlovu 45.7% Zululand 36.1% Ilembe 35.8% Ethekwini 40.3% Uthungulu 36.1% Umkhanyakude 39.9%. KZN 2009 Amajuba 37.3% UThukela 46.4% Umzinyathi Ugu 28.2% Zululand 36.7% Ilembe UMgungundlovu 40.6% Sisonke 40.9% EThekwini 35.2% 41.5% Uthungulu 37.7% Umkhanyakude 39.7%. Ugu 40.6% Kilometers HIV Prevalence 20-30% 30 40% >40% >40% 40.2% 150 Kilometers HIV Prevalence 20-30% 30 40% >40% >40% KZN 2010 Sisonke 37.2% Amajuba 35.9% UThukela 36.7% UMzinyathi 31.1% UMgungundlovu 42.3% EThekwini 41.1% Zululand 39.8% ILembe 42.3% UThungulu 36.9% Umkhanyakude 41.9%. KZN 2011 UThukela 39.0% 39.9% Sisonke 35.9% Amajuba 35.3% UMzinyathi 24.6% UMgungundlovu 39.8% EThekwini 38.0% Zululand 39.3% ILembe 35.4% UThungulu 33.4% UMkhanyakude 41.1%. Ugu 41.1% Key HIV Prevalence 20-30% Ugu 41.7% HIV Prevalence 20-30% 30 40% 30 40% 200 >40% Kilometers >40% 0 Kilometers >40% >40% Figure 26: HIV prevalence distribution among survey participants by district in KwaZulu-Natal, 2008 to

51 3.6.5 LIMPOPO PROVINCE In 2011, the Limpopo provincial HIV prevalence amongst antenatal women was 22.1% (95%CI: ). The overall provincial HIV prevalence in Limpopo increased slightly, by 0.2%, between 2010 and 2011 (Figures 29 and 30). Table 15: HIV prevalence among antenatal women by district, Limpopo, 2009 to N % Prev. 95% CI N % 95% CI N % 95% CI Prev. Prev. Provincial Capricorn Mopani Sekhukhune Vhembe Waterberg N = Realised sample size; CI= Confidence Interval HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 29: HIV prevalence trends among antenatal women, Limpopo, 1994 to

52 30 HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 30: The HIV prevalence epidemic curve among antenatal women, Limpopo, 1994 to Figure 31: HIV prevalence trends among survey participants by district in Limpopo 2009 to 2011 Sekhukhuni and Vembe districts have shown a slight decrease in HIV prevalence between 2010 and The Waterberg district s HIV prevalence increased significantly by 4.2%, from 26.1% in 2010 to 30.3% in Sekhukhune has shown erratic changes in prevalence from 16.6% in 2009 to 20.2% in 2010 to 18.9% in 2011 (Figure 31 and Table 15). 37

53 Table 16: HIV prevalence among antenatal women by age group, Limpopo, 2009 to N % Prev. N % Prev. N % Prev. * >49 ** ** ** ** * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size. ** No participants in this age group Figure 32: Sampled population distribution by age group, Limpopo, 2011 The changes in HIV prevalence distribution by district in Limpopo from 2008 to 2011 are shown in Figure

54 LP 2008 LP 2009 Vhembe 14.7% Vhembe 14.3% Waterberg 23.6% Capricorn Mopani 25.2% 21.0% Waterberg Capricorn 28.8% 23.8% Mopani 26.2% Sekhukhune 21.8% Sekhukhune 16.6% HIV prevalence 10-20% 20-30% 30-40% LP 2010 LP 2011 Vhembe 17.0% Vhembe 14.6% Waterberg 26.1% Capricorn Mopani 24.9% 23.7% Capricorn Mopani Waterberg 25.2% 30.3% 25.3% Sekhukhune 20.2% Key HIV prevalence 10-20% 20-30% Sekhukhune 18.9% Figure 33: HIV prevalence distribution among survey participants by district in Limpopo, 2008 to 2011 HIV prevalence 10-20% 20-30% Key HIV prevalence 10 20% 20-30% 39

55 3.6.6 MPUMALANGA PROVINCE In 2011, the Mpumalanga provincial HIV prevalence amongst antenatal women was 36.7% (95%CI: ). This is a slight increase from 35.1% in 2010 as shown in Figures 34 and 35. HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 34: HIV prevalence trends among antenatal women, Mpumalanga, 1994 to HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 35: The HIV prevalence epidemic curve among antenatal women, Mpumalanga, 1994 to

56 Two districts in Mpumalanga, viz., Ehlanzeni and GertSibande, recorded the 6th and 7th highest HIV prevalence among the 52 health districts in the country in However, in 2011 Gert Sibande recorded a significant increase of 7.9% from 38.2% in 2010 to 46.1%, the highest HIV district prevalence recorded in the country in Nkangala recorded an increase in HIV prevalence from 27.2% in 2010 to 29.6% in Enhlanzeni declined from 37.7% in 2010 to 35.8% in 2011 as presented in (Figures 36 and Table 17). The HIV prevalence estimates in all three districts in Mpumalanga are above 29%. Table 17: HIV prevalence among antenatal women by district, Mpumalanga, 2009 to N % Prev. 95% CI N % Prev. 95% CI N % Prev. 95% CI Provincial Ehlanzeni Gert Sibande Nkangala N = Realised sample size; CI= Confidence Interval. Figure 36: HIV prevalence trends among survey participants by district in Mpumalanga, 2009 to

57 Table 18: HIV prevalence among antenatal women by age group, Mpumalanga, 2009 to N % Prev. N % Prev. N % Prev. * > ** ** ** ** * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size. ; ** No participants in this age group Figure 37: Sampled population distribution by age group, Mpumalanga, 2011 The changes in HIV prevalence distribution by district in Mpumalanga from 2008 to 2011 are shown in Figure

58 MP 2008 MP 2009 Ehlanzeni 34.9% Nkangala 31.8% Ehlanzeni 33.8% Nkangala 32.6% Gert Sibande 40.5% Key HIV Prevalence 30 40% Gert Sibande 38.2% >40% MP 2010 MP 2011 Ehlanzeni 37.7% Nkangala 27.2% Nkangala 29.6% Ehlanzeni 35.8% Gert Sibande 38.8% Key HIV Prevalence Gert Sibande 46.1% 30-40% 20-30% Figure 38: HIV prevalence distribution among survey participants by district in Mpumalanga, 2008 to Key HIV Prevalence 30 40% Key HIV Prevalence 20-30% 30-40% >40% 43

59 3.6.7 NORTH-WEST PROVINCE In 2011, the North-West provincial HIV prevalence amongst antenatal women was 30.2% (95% CI: ) in 2011 (Figures 39 and 40) HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 39: HIV prevalence trends among antenatal women, North-West Province, 1994 to HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 40: The HIV prevalence epidemic curve among antenatal women, North-West Province, 1994 to

60 Findings of the 2011 survey showed that in the Dr Kenneth Kaunda district, there was a slight decrease of 1.0% in the antenatal HIV prevalence, from 37.0% in 2010 to 36.0% in Bojanala district recorded a large increase from 29.3% in 2010 to 33.9% in 2011, an increase of 4.6% (Figure 41 and Table 19). Table 19: HIV prevalence among antenatal women by district, North West, 2009 to N % 95% CI N % 95% CI N % 95% CI Prev. Prev. Prev. Provincial Bojanala Dr. R.S. Mompati Ngaka M. Molema Dr. K. Kaunda N = Realised sample size; CI= Confidence Interval. Figure 41: HIV prevalence trends among survey participants by district in the North West, 2009 to

61 Table 20: HIV prevalence among antenatal women by age group, North West, 2009 to N % Prev. N % Prev. N % Prev. * >49 ** ** ** ** ** ** * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size. ** No participants in this age group Figure 42: Sampled population distribution by age group, North West, 2011 The changes in HIV prevalence distribution by district in the North West province from 2008 to 2011 is shown in Figure

62 NW 2008 NW 2009 Bojanala 31.8% Ngaka ModiriMolema Dr. R.S. Mompati 28.2% 28.1% Dr. K. Kaunda 35.2% Bojanala 34.9% Ngaka Modiri Molema Dr. R.S. Mompati 25.1% 25.7% Dr. K. Kaunda 29.2% HIV Prevalence 20-30% 30 40% HIV Prev 20-30% 30 40% NW 2010 NW 2011 Ngaka Modiri Molema Dr. R.S. Mompati 25.9% Bojanala 29.3% Ngaka Modiri Molema Dr. R.S. Mompati 24.9% Bojanala 33.9% 24.3% Dr. K. Kaunda 37.0% 20.5% Dr. K. Kaunda 36.0% Key HIV Prev 20-30% 30 40% HIV Prevalence 20-30% 30 40% Figure 43: HIV prevalence distribution among survey participants by district in the North West, 2008 to

63 3.6.8 THE NORTHERN CAPE PROVINCE The Northern Cape provincial HIV prevalence amongst antenatal women has decreased by 1.4% from 18.4% (95% CI: ) in 2010 to 17.0% (95% CI: ) in 2011 (Figures 44 and 45) HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 44: HIV prevalence trends among antenatal women, Northern Cape, 1994 to HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 45: The HIV prevalence epidemic curve among antenatal women, Northern Cape, 1994 to

64 In 2011, the antenatal HIV prevalence in Namaqua was 6.2%, and because the sample size in this district was small, it becomes difficult to discern any trends. Hence a great variation in prevalence over time is observed. Namaqua district has recorded the lowest antenatal HIV prevalence among the 52 health districts in South Africa. In 2011 the Francis Baard district recorded an HIV prevalence of 18.4% compared to 20.1% in The J.T. Gaetsewe district recorded a significant decrease in HIV prevalence from 27.5% in 2010 to 17.7% in 2011, a decline by 9.8%. Siyanda recorded the highest HIV prevalence estimate of 19.1% in the Northern Cape for This district also showed a 2.7% increase from Table 21: HIV prevalence among antenatal women by district, Northern Cape, 2009 to N % Prev. 95% CI N % Prev. 95% CI N % Prev. 95% CI Province F. Baard J. T. Gaetsewe Namakwa Pixley ka Seme Siyanda N = Realised sample size; CI= Confidence Interval. Figure 46: HIV prevalence trends among survey participants by district in the Northern Cape, 2008 to

65 Table 22: HIV prevalence among antenatal women by age group, Northern Cape, 2009 to N % Prev. N % Prev. N % Prev. * >49 ** ** ** ** * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size. ** No participants in this age group Figure 47: Sampled population distribution by age group, Northern Cape, 2011 The variation in HIV prevalence distribution in the Northern Cape province from 2008 to 2010 is shown in Figure

66 Figure 48: HIV prevalence distribution among survey participants by district in the Northern Cape, 2008 to

67 3.6.9 WESTERN CAPE PROVINCE In 2011, the Western Cape provincial HIV prevalence amongst antenatal women was 18.2% (95% CI: ). The overall HIV prevalence remained stable around 18% in the past two years (Figures 49 and 50) HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 HIV prevalance Year Figure 49: HIV prevalence trends among antenatal women, Western Cape, 1994 to HIV prevalence (%) '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 `10 `11 Year Figure 50: The HIV prevalence epidemic curve among antenatal women, Western Cape, 1994 to The Cape Metropole carries the heaviest burden of HIV in the Western Cape, with more than 70% of the HIV infected pregnant women in the province. The Cape Metro recorded a slight decrease, by 0.4% from 20.2% in 2010 to 19.8% in Overberg and Central Karoo recorded increases in antenatal HIV prevalence rates from 2010 to 2011, of 4.1% and 3.3% respectively (Figure 51 and Table 23). 52

68 Table 23: HIV prevalence among antenatal women by district, Western Cape, 2009 to N % % % 95% CI N 95% CI N Prev Prev Prev. 95% CI Provincial C. Winelands Central Karoo Eden Metropole Overberg West Coast The Cape Metropole carries the heaviest burden of HIV in the Western Cape, with more than 70% of the HIV infected pregnant women in the province. The Cape Metro recorded a slight decrease, by 0.4% from 20.2% in 2010 to 19.8% in Overberg and Central Karoo recorded increases in antenatal HIV prevalence rates from 2010 to 2011, of 4.1% and 3.3% respectively (Figure 51). Prevalence in the West Coast has remained at or below 10% for the past four years. The Overberg recorded a 4.1% increase from 17.3% in 2010 to 21.4% in Figure 51: HIV prevalence trends among survey participants by district in the Western Cape, 2009 to

69 Table 24: HIV prevalence among antenatal women by age group, Western Cape, 2009 to N % Prev. N % Prev. N % Prev. * >49 ** ** ** ** ** ** * The age group years is an indicator for Goal 6 of the Millennium Development Goals (MDG) N = Realised sample size. ** No participants in this age group Figure 52: Sampled population distribution by age group, Western Cape, 2011 The variation in HIV prevalence distribution in the Western Cape province from 2008 to 2010 is shown in Figure

70 WC 2008 WC 2009 West Coast 9.3 % CMC 17.9% C. Winelands 12.0% Overberg 15.9% Central Karoo 14.8% Eden 13.0% HIV Prevalence < 10% 10 20% West Coast 9.5 % City of CPT 18.0% C. Winelands 13.2% Overberg 20.8% Central Karoo 11.8% Eden 18.2% WC 2010 WC 2011 West Coast 10.0% City of CPT 20.2% C. Winelands 14.9% Overberg 17.3% Central Karoo 8.5% Eden 18.0% Key HIV Prevalence < 10% 10 20% West Coast 9.9% CMC 19.8% C. Winelands 15.7% Overberg 21.4% Central Karoo 11.3% Eden 16.1% Figure 53: HIV prevalence distribution among survey participants by district in the Western Cape, 2008 to HIV Prevalence < 10% 10 20% HIV Prevalence < 10% 10 20% 20-30% 55

71 3.7 EXTRAPOLATION OF HIV INFECTION TO THE GENERAL POPULATION According to the UNAIDS SPECTRUM model the estimated national HIV prevalence among the general adult population aged years old has remained stable at around 17.3% since 2008 (Table 25). In 2011, an estimated 5, [ ] people living with HIV resided in South Africa. The estimated number of new infections in South Africa was 1.43% in 2011 compared to 1.63% new infections in The number of newly infected children aged 0-14 years fell by 56.2%, from in 2008 to an estimated in More than 95% of women in need of PMTCT services were estimated to be covered in According to the UNAIDS estimates, between 2008 and 2011, the number of people dying from AIDS-related causes in South Africa declined by 28.7%, from [ ] to [ ]. However, an estimated [ ] children were AIDS orphans in 2011 compared to approximately [ ] in 2008 representing a 9.5% increase. Table 25: Selected HIV estimates for South Africa, UNAIDS SPECTRUM, Indicator Total HIV population (adults and children) UNAIDS [ ] HIV+ adults (15+) [ ] Adult (15-49) HIV prevalence (%) Adult prevalence (15-24) female population (%) Adult prevalence (15-24) Male population (%) Adult HIV+ female population (15+) HIV population (children <15) 17.2 [ ] 13.1 [ ] 5.7 [ ] [ ] [ ] Total annual AIDS deaths [ ] AIDS orphans [ ] Total number of new HIV infections New infections (Children birth -14yrs) Total need for ART among adults (15+)* [ ] [ ] [ ] Children needing ART [ ] Mothers receiving PMTCT (excluding snvp) * (CD4 350 starting 2009) [ ] [ ] 17.2 [ ] 12.6 [ ] 5.5 [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] ] [ ] [ ] [ ] [ ] [ ] 17.3 [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] 56

72 Tables 26 and 27 indicate considerable variation in provincial prevalence, incidence and new infections according to UNAIDS estimates. In 2011, HIV incidence ranged from 0.34% in the Western Cape to 2.32% in KwaZulu-Natal. In Northern Cape an estimated new infections occurred in the adult population aged years old in 2011 compared to approximately in KwaZulu- Natal. In 2011, more than 30.8% of adults years who acquired new infections live in KwaZulu- Natal followed by Gauteng (20.3%) and Eastern Cape (12.1%). Table 26: Estimates of new infections, , UNAIDS SPECTRUM Year Estimated number of new infections (adults years) Estimated rate of new infections (adults years) % per annum Estimated number of new infections (Children birth -14 years) Eastern Cape Free State Gauteng KwaZulu-Natal Limpopo Mpumalanga North West Northern Cape Western Cape South Africa Table 27: Provincial HIV prevalence of adults years (%), , UNAIDS SPECTRUM Eastern Cape Free State Gauteng KwaZulu-Natal Limpopo Mpumalanga North West Northern Cape Western Cape South Africa

73 CHAPTER 4 4. RESULTS - SYPHILIS PREVALENCE The 2011 estimated national syphilis prevalence was 1.6% (95%CI: ) nearly unchanged from the 1.5% (95%CI: ) in The trend in syphilis prevalence among attendees of antenatal clinics from 1997 to 2011 is shown in Figure 54. Figure 54: National syphilis prevalence trends among antenatal women, South Africa, 1997 to Syphilis prevalence by province The estimated highest prevalence of syphilis (4.1%) in 2011 was recorded in the Mpumalanga. This is the highest syphilis prevalence Mpumalanga has ever recorded. The syphilis prevalence in this province has more than doubled from 2.1% in 2010 (Figure 55 and Table 28). Figure 55: Syphilis prevalence trends among antenatal women by province, South Africa 2009 to

74 KwaZulu-Natal, which has the highest HIV prevalence in the country, continues to record the lowest syphilis prevalence in the country (Figure 55). Decreases in syphilis prevalence were noted in the Eastern Cape and Gauteng provinces. Table 28: Syphilis prevalence among antenatal women by province, South Africa, 2009 to N % Prev. 95% CI N % Prev. 95% CI N % Prev. 95% CI SA EC FS GA KZN LP MP NW NC WC N = Realised sample size. ; CI= Confidence Interval 59

75 Table 29: Syphilis prevalence among antenatal women by age group, South Africa, 2009 to Age group N % Prev. 95% CI N % Prev. 95% CI N % Prev. 95% CI N = Realised sample size. ; CI= Confidence Interval The 2011 report has shown beyond reasonable doubt that there is no significant correlation between HIV and Trepanema palladium the aetiological agent for active syphilis as co-factor for HIV infection. In the 2012 survey we have started to pilot monitoring of Herpes Simplex HSV2 which usually causes genital herpes and is transmitted primarily by direct contact with sores, most often during sexual contact. 60

76 CHAPTER 5 CONCLUSION AND RECOMMENDATIONS South Africa has gone through different phases of national HIV and AIDS responses from widespread denial in the late 1980 s to mid 1990 s, adoption of a multi-sectoral policy in 1997, and strong government leadership and political commitment since 2005, when cabinet approved the rollout of anti-retroviral treatment (ART) to those who need it. Linking various interventions such as behaviour change communication (BCC) and HIV counselling and testing (HCT) with prevention of mother-tochild transmission (PMTCT) and antiretroviral therapy (ART), has created a continuum of prevention and care services. The government, knowing that the war against the HIV and AIDS pandemic has yet to be won, will continue to lead the fight. One goal is to reduce new infections by 50% by 2011 and increase the PMTCT coverage using dual therapy to 100%. Other key programmes include blood safety, early diagnosis through the know your status campaign and prompt syndromic treatment of sexually transmitted infections (STIs), home-based care and support for the infected and affected, developing a strong link between TB and HIV programmes, and effective management of opportunistic diseases. One of the important observations of the 2011 findings is that the higher prevalence in older age groups could be partly explained by increased access to ART; however in the younger women (15-24 years old), the MDG target group, there is no decline in the number of HIV positives and this age group should not be much affected by access to ART. The ANC survey does not collect information on ART coverage and hence the contribution of ART to the increase in HIV prevalence cannot be directly determined. However, it is recommended that data on ART be collected and triangulated with in all future ANC surveys. It is becoming very important to start collecting age specific information on patients receiving ART for those women who attend ANC clinics. We observed that the HIV prevalence is higher in older women. If HIV prevalence were to increase in the context of a large ART programme, then it would mean that life expectancy has increased for those who got infected in earlier years. The negative association between HIV and syphilis prevalence re-affirms the empirical evidence that syphilis prevalence is not a useful factor to positively correlate with HIV. Hence, an appropriate STI needs to be identified for the co-infection. Conclusions that can be drawn from the 2011 findings are as follows: 1. South Africa has an established generalized HIV epidemic with an estimated 17.3% prevalence in the general population and an estimated prevalence of 29. 5% in the antenatal population. 2. KwaZulu-Natal is promising by showing a notable decrease in HIV prevalence from 2010 to 2011, however, the gradual increase in HIV prevalence in Mpumalanga is worrisome. 61

77 3. There is a important decrease in HIV prevalence among the year old from 14% in 2010 to 12.7% in The HIV prevalence among the year old pregnant women was 20.5 % (95% CI: ) in 2011 compared to 21.8% (95% CI: ) in 2010, a decline of 1.3%. The specific AIDS MDG target is that by 2015 the expected HIV prevalence reduction should be 25% less than the baseline prevalence of 23.1% in The findings of monitoring trends in this age group in South Africa show that we should not relent of our collective efforts to achieve this AIDS MDG target. 5. It is crucial that the department conduct pilot analytical (in-depth), epidemiological surveys in high prevalence (>40%) and low prevalence districts (below 10%), in order to investigate potential risk factors that drive the epidemic and determine the type of HIV strains (sub types) that could be circulating in these districts. 6. The 2011 report has shown beyond reasonable doubt that there is no significant correlation between HIV and Trepanema palladium the aetiological agent for active syphilis as co-factor for HIV infection. In the 2012 survey we have started to pilot monitoring of Herpes Simplex HSV2 which usually causes genital herpes and is transmitted primarily by direct contact with sores, most often during sexual contact. 7. From 2012 the syphilis surveillance will be dropped from the survey, while routine monitoring of syphilis among pregnant women attending antenatal care remain. The following recommendations can be made from the implications of the findings: To conduct further research and triangulation of other HIV surveillance data within the public health sector to further understand the potential risk factors for high risk groups, i.e. to reduce high teenage pregnancy and to improve the department s focus to strengthen reproductive health through the Integrated School Health Programme. To inform the nurses in the clinics about the outcome of the HIV+ results from the laboratories in order for a provider initiative HIV testing in case the woman did not participate in the prevention of mother to child transmission (PMTCT) programme, while the identity of the participants remains anonymous to researchers. To report on HIV prevalence distribution by geotype (rural vs. semi-rural vs. urban), because in generalized heterosexual epidemics the standard practice is to categorise populations by geographic subdivisions. To publish a separate scientific paper on the risk factors associated with HIV status of pregnant women, and to do multiple regression analysis of the relationship between risk factors and HIV outcome. To monitor the association between the Human Papilloma Virus (HPV) prevalence and HIV infection in order to assess the risk of cervical cancer in pregnant women and monitor trends in HPV and HIV co-infection. To monitor the association between and Hepatitis B prevalence and HIV infection in order to assess and monitor trends in Hepatitis B and HIV prevalence co-infection. 62

78 To determine the presence of Anti-retroviral drugs in all blood samples of the sentinel survey participants. To conduct this survey on a monthly basis from April 2013 in order to report on trends at national, provincial and district level disaggregated by age in the 2014 report. To find out if the pregnant women ever participated in the PMTCT. To establish systems to track measures such as AIDS-related mortality by age, sex, district, province, monitor loss to follow-up, number of patients on different regimes, pharmocovigilance, drug resistance patterns, treatment failures and HIV incidence rate. For the past 21 years this study has been conducted annually in October, however, we believe that during this period we miss out teenage contact with pregnant teenagers because they are writing examinations. For this reason we will start testing all pregnant women every month starting from April 2013 in order to report on the HIV prevalence rate while maintaining the current antenatal sentinel report in October. Our target is to test pregnant women of all ages annually. The first results will be reported in

79 LIST OF REFERENCES 1. Anderson and May, Infectious diseases of Humans Dynamics and Control. Oxford University Press. New York. 2. Chin J, Mann J, Global Surveillance and Forecasting of AIDS (1989). Bull World Health Organ, Death Notification Report. StatsSA, Department of Basic Education, Teenage pregnancy among school going learners, 2009 report. 5. Department of Health South Africa (2007). National HIV and Syphilis Antenatal Prevalence Survey, South Africa, Pretoria. 6. Department of Health, HIV & AIDS and STI Strategic Plan for South Africa Pretoria. 7. Department of Health, Report: National HIV and Syphilis Antenatal Prevalence Survey, South Africa, Pretoria. 8. Department of Health, Report: National HIV and Syphilis Antenatal Prevalence Survey, South Africa, Pretoria. 9. Department of Health, Report: National HIV and Syphilis Antenatal Prevalence Survey, South Africa, Pretoria. 10. Department of Health South Africa (2008). National HIV and Syphilis Antenatal Prevalence Survey, South Africa, Pretoria. 11. Department of Health South Africa (2009). National HIV and Syphilis Antenatal Prevalence Survey, South Africa, Pretoria. 12. Department of Health. Comprehensive HIV and AIDS Care, Management and Treatment Plan. South Africa, Pretoria. 13. Jackson, H. AIDS in Africa Continent in crisis. SAfAIDS. ISBN UNAIDS (2007). Comparing adult antenatal-clinic based HIV prevalence with prevalence from national population based surveys in sub-saharan Africa. UNAIDS presentation. 15. UNAIDS (2009). AIDS Epidemic Update 2008: Special Report on HIV/AIDS: December UNAIDS (2001). Declaration of Commitment on HIV/AIDS: UN General Assembly Special Session on HIV/AIDS, June UNAIDS Reference Group on Estimates, Modeling, and Projections (2006). Improving parameter estimation, projection methods, uncertainty estimation, and epidemic classification. Report of a meeting of the UNAIDS Reference Group on Estimates, 18. Modeling, and Projections, Prague, Czech Republic, 29 Nov 1 Dec. Report/2007/2006prague_ report_en.pdf. 19. UNAIDS Reference Group on Estimates, Modelling and Projections (2002). Improved methods and assumptions for the estimation of the HIV/AIDS epidemic and its impact: recommendations of the UNAIDS Reference Group on Estimates, Modelling and Projections. AIDS, 16: W1 W WHO, UNAIDS, UNICEF (2007). Towards universal access: scaling up priority HIV/AIDS interventions in the health sector: progress report. April. Geneva. ISBN UNAIDS & WHO (2007). AIDS epidemic update: December UNAIDS, Geneva UNAIDS/07.27E/ JC1322E. ISBN UNAIDS & WHO (2006). Guidelines for measuring national HIV prevalence in population based surveys. UNAIDS, Geneva. ISBN UNAIDS & WHO (2006). AIDS epidemic update: December UNAIDS, Geneva UNAIDS/06.29E. ISBN UNAIDS & WHO (2005). AIDS Epidemic Update UNAIDS & WHO (2003). Working group on HIV/AIDS & STI surveillance. Guidelines for 2nd Generation HIV surveillance. 26. UNAIDS & WHO (2000). Working group on HIV/AIDS & STI surveillance. Guidelines for 2nd Generation HIV surveillance. 27. UNAIDS & WHO Global Programme on AIDS (1989). Unlinked anonymous screening for the public health surveillance of HIV infections. International Guidelines. 28. WHO (2003). World health report: 2003: shaping the future. Geneva. ISBN

80 29. Shisana O, Rehle T, Simbayi L, Zuma K, Jooste S, et al. (2009) South African National Prevalence, Incidence, Behaviour and Communication Survey, A Turning Tide Among Teenagers? Cape Town, South Africa: HSRC Press 30. Crainiceanu, CM, Diggle PJ, Rowlingson B. (2006). Bivariate Binomial Spatial Modelling of LOA loa Prevalence in Tropical Africa. Johns Hopkins University; Dept. of Biostatistics Working Papers: Paper Kleinschmidt I, Pettifor A, Morris N, MacPhail C, Rees H. (2007). Geographic Distribution of Human Immunodeficiency Virus in South Africa. American Journal of Tropical Medicine and Hygiene. 77(6), Kleinschmidt I, Ramkissoon A, Morris N, Mabude Z, Curtis B, Beksinska M. (2006). Mapping indicators of sexually transmitted infection services in the South Africa public health sector. Tropical Medicine and International Health. 11(7), Corbett EL, Watt CJ, Walker N, et al. (2003). The growing burden of tuberculosis: global trends and interactions with the HIV epidemic. Arch Intern Med, 163: Browne, W. J., Goldstein, H., and Rasbash, J. (2001). Multiple membership multiple classification (MMMC). Statistical Modelling, 1, Cantwell MF and Binkin NJ (1996) Tuberculosis in sub-saharan Africa: a regional assessment of the impact of the human immunodeficiency virus and National Tuberculosis Control quality. Tuber Lung Dis, 77: Carey V, Zeger SL, Diggle P. (1993). Modelling multivariate binary data with alternating logistic regressions. Biometrika; 80: Corbett EL, Watt CJ, Walker N, et al. (2003). The growing burden of tuberculosis: global trends and interactions with the HIV epidemic. Arch Intern Med, 163: Feltbower RG, Manda SOM, Gilthorpe MS, Greaves MF, Parslow RC, Kinsey SE, Bodansky HJ, Patricia A McKinney PA. (2005). Detecting small area similarities in the epidemiology of childhood acute lymphoblastic leukaemia and type 1 diabetes: a Bayesian approach. American Journal of Epidemiolo, 161; John-arne R, Cameron W, Garnett GP (2001). A systematic review of epidemiologic interactions between classic sexually transmitted diseases and HIV: how much really is known? Sexually Transmitted Diseases; 28: UNAIDS/WHO (20003). Guidelines for conducting HIV sentinel serosurveys among pregnant women and other groups / UNAIDS/WHO Working Group on Global HIV/AIDS and STI Surveillance. WHO Library Cataloguing-in-Publication Data. Geneva. 41. Johnson LF, Coetzee DJ, Dorrington RE (2005). Sentinel surveillance of sexually transmitted infections in South Africa: a review. Sexually Transmitted Diseases; 81: Leyland AH, Langford IH, Rasbash J, Goldstein H. (2000). Multivariate spatial models for event data. Statistics in Medicine; 19: Manda SOM and Leyland A. (2005) Maximum likelihood and Bayesian estimation methods for multivariate disease outcomes in spatial epidemiology. (Submitted to Statistics in Medicine) 44. Monteiro EF, Lacey CJN and Merrick D. (2005). The interrelation of demographic and geospatial risk factors between four common sexually transmitted diseases. Sexually Transmitted Infections; 81: Over M and Piot P. (1993). HIV infection and sexually transmitted diseases. In Disease Control Priorities in Developing Countries (eds. D.T. Jameson, W.H. Mosely, A.R. Measham and J.L. Babadilla), pp New York: University Press. 46. Department of Health (2009). National Antenatal Sentinel HIV and Syphilis Prevalence Survey in South Africa 2008 Report. South Africa: Department of Health. 47. Shisana O, Rehle T, Simabyi LC, Parker W, Zuma K, Bhana A, Conolly C, Jooste S, Pillay-van-Wyk V, et al. (2005). South African National HIV Prevalence, HIV incidence, Behaviour and Communication Survey Cape Town: HSRC Press. 48. Shisana O, Rehle T, Simabyi LC, Parker W, Zuma K, Bhana A, Conolly C, Jooste S, Pillay-van-Wyk V, Mbelle N, Van Zyl J, Parker W, Zungu NP, Pezi S & the SABSSM III Implementation Team (2009). South African National HIV Prevalence, HIV incidence, Behaviour and Communication Survey 2008: A turning tide among teenagers? Cape Town: HSRC Press. 65

81 49. Sandøy IF, Gunnar Kvale G, Michelo C, Fylkesnes (2006). Antenatal clinic-based HIV prevalence in Zambia: Declining trends but sharp local contrasts in young women. Tropical Medicine and International Health, 11 (6) pp UNAIDS (2009) Report on the global AIDS epidemic. Available at < KnowledgeCentre/HIVData/GlobalReport/2008/>. 51. UNAIDS (2008). Status of the global HIV epidemic jc1510_2008_global_report_pp29_62_en.pdf 52. Day C, Monticelli F, Barron P, Haynes R, Smith J, Sello E, editors. The District Health Barometer 2008/09. Durban: Health Systems Trust; May Meidany F, Horikoshi Y, Rohde J (2000). HIV prevalence rate and population density: Eastern Cape experience, South Africa International Conference on AIDS. 54. Spiegelhalter D, Thomas A, Best N. and Lunn D (2004). BUGS: Bayesian Inference Using Gibbs Sampling, Version 1.4}. MRC Biostatistics Unit: Cambridge. 55. Pettifor AE, Kleinschmidt I, Levin J, Rees HV, MacPhail C, Madikizela-Hlongwa L, Vermaak K, Napier G, Stevens W and Padian NS. (2005). A community-based study to examine the effect of a youth HIV prevention intervention on young people aged in South Africa: results of the baseline survey. Tropical Medicine and International Health, 10 (10) pp William B, and Dye C (2003). Antiretroviral Drugs for Tuberculosis Control in the Era of HIV/AIDS. Communicable Diseases, World Health Organization, 1211 Geneva 27, Switzerland 57. Cohen, D (1998). Socio-economic causes and consequences of the HIV epidemic in Southern Africa: The case of Namibia, UNDP Issues Paper No. 31, Zierler S, Krieger N, Tang Y, Coady W, Siegfried E, DeMaria A, Auerbach J. (2000). Economic Deprivation and AIDS Incidence in Massachusetts. American Journal of Public Health. 90; Montana L, Neuman M, Mishra V and Hong R. (2005). Spatial Modeling of HIV Prevalence in Cameroon, Kenya, and Tanzania. Population Association of America Annual Conference, Montana L, Neuman M, and Mishra V. (2007). Spatial Modelling of HIV Prevalence in Kenya 61. DHS Working Papers. No 27, Calverton. 62. Browne WJ, Goldstein H, Rasbash J. Multiple membership multiple classification (MMMC). Statistical Modelling 2001; 1, Hargrove J. (2008). Migration, mines and mores: the HIV epidemic in Southern Africa. South African Journal of Science 104, January/February Yahya-Malima KI, Evjen-Olsen B, Matee MI, Fylkesnes K, Haarr L. (2008). HIV-I, HSV-2 and syphilis among pregnant women in a rural area of Tanzania: Prevalence and risk factors. BMC Infectious Diseases 2008, 8: Mullick S, Beksinksa M, Msomi S. (2005). Treatment for syphilis in antenatal care: compliance with the three dose standard treatment regimen. Sexually Transmitted Infections. 81: Ramkisson et al (2004). National Baseline Assessment of Sexual Transmitted Infection and HIV services in South African public sector health facilities. Summary report Lurie, M.N., Williams, B.G., Zuma, K., Mkaya-Mwamburi, D., Garnett, G. P., Sturm, A. W., Sweat, M. D., Gi ttelsohn, J. & Abdool Karim, S.S. (2003) The impact of migration on HIV-1 transmission in South Africa: a study of migrant and nonmigrant men and their partners. Sexually transmitted diseases. 30 (2): Montana, LS, Mishra V, Hong R (2008). Comparison of HIV prevalence estimates from antenatal 69. Sex Transm Infect : i78-i84 66

82 Notes 67

83 The 2011 National Antenatal Sentinel HIV & Syphilis Prevalance Survey in South Africa 3

84 unaids 4 The Photo UNAIDS / S. Montanari National Antenatal Sentinel HIV & Syphilis Prevalance Survey in South AFrica