This paper's findings are derived from the MATCH (Multi-country African Testing and Counseling for HIV) study, the first designed to systematically compare client experiences with HIV testing across countries and modes of testing. Numerous government and donor-supported initiatives have been implemented in Burkina Faso, Kenya, Malawi and Uganda, where HIV testing is increasingly conducted within health facilities. The four countries face similar challenges in expanding their testing regimes: insufficient resources and infrastructure, shortage of adequately trained staff, and difficulties with referral and follow-up. But there are differences in pace of increase (somewhat slower in Burkina Faso) and in the scale of resources invested (larger in Kenya and Uganda).
The MATCH study included a main survey conducted in 2008-09 among clients and providers at health facilities offering HTC services. The study was conducted in the capital region and one rural province or district in each country, where research teams purposefully selected the facilities to include a variety of types of health facilities. Research teams contacted those in charge at each facility, obtained clearance, and asked them to inform clients about the project. Interviews were planned on at least two different days of the week, with an aim of interviewing a total of 15 clients per facility. On the appointed days, interviewers approached clients, described the study, and invited them to participate. The fraction of respondents to be interviewed at each of the selected facilities was based on the expected numbers of respondents per day: at smaller facilities all clients were invited to participate, whereas at busier facilities, every nth client was. The teams had to adjust plans when attendance was low, or when health staff did not present at the clinic. While the full MATCH study includes both women and men, and respondents who have and have not tested for HIV, the current paper reports on the MATCH study's findings as they pertain to the subset of women who tested in an antenatal care facility or because of pregnancy.
The MATCH survey was administered in each country by a team of experienced researchers, who were responsible for training field interviewers and data entry staff, and have provided input into the survey design and interpretation of results. Prior to beginning the interview, respondents were told about the study and asked to provide consent. In most cases, written consent or a thumbprint was given by respondents; however, in a few situations where these methods of obtaining consent were deemed unacceptable by the country team, verbal consent was allowed. Field interviewers were project staff, and were not associated with the health facilities where the interviews were conducted. The interviews were conducted in French (Burkina Faso), English (Kenya, Malawi, Uganda) or local languages. The survey questionnaire included closed and open-ended questions and usually lasted from around 30 min to 1 h. Data were collected on respondents' socio-demographic characteristics and their experiences with HIV testing and counseling, using a checklist of interactions and services during pre- and post-test meetings and during follow-up care. Before entering the data into the database, all questionnaires were checked for consistency.
In this paper we describe the experiences of HIV-positive and HIV-negative pregnant women who were tested for HIV. The full MATCH client sample is comprised of 3,660 respondents. Of these, 2,117 respondents were tested for HIV in or after 2007. Out if this sub-set, 421 respondents were PMTCT testers. Women were categorized as PMTCT testers if they either reported testing in an antenatal care facility or cited pregnancy or PMTCT as their reason for testing in an integrated testing and medical facility, such as a hospital or general health clinic. We excluded 13 PMTCT testers with missing HIV status from the analysis. The final sample therefore included 408 women. These women were generally young (in their 20 s and 30 s) with no or only primary education. In Burkina Faso, educational level of the women was lowest: 40.4% of the respondents had no formal education (see Table ).
Age, educational attainment, household assets, marital status, and number of previous HIV tests by country, PMTCT testers
Two kinds of data were collected from this sub-set of 408 women: multiple choice responses to assess their views and experiences of HIV testing, and answers to open-ended questions which gave women an opportunity to expand on their views and experiences of counseling, consent and confidentiality (the 3Cs).
We used guidelines on HIV testing from major international, United States, and European organizations [25
] to select variables to measure women's views and experiences of the 3Cs. In light of the debate over the extent of necessary pre-test counseling, and the need to tailor post-test counseling to test results [25
], we examined pre- and post-test counseling separately (see Table ). In addition, the survey included measures for self-stigma, enacted and community stigma from the WHO's manual HIV Testing, Treatment and Prevention: Generic Tools for Operational Research
], see Table . We used Stata 10.1 SE to conduct a descriptive analysis of the 3Cs and measured significance of differences by country and by HIV status using Fisher's exact tests appropriate for the relatively small sample size with a threshold for significance at p
The 3C's variables used in our study
We closely examined experiences of stigma, disclosure, post-test counseling and access to follow-up psycho-social support in a sub-set of 63 HIV-positive women (6 in Burkina Faso, 8 in Kenya, 34 in Malawi and 15 in Uganda), using primarily their responses to the open-ended question. And, to gain further insight into these issues, we held 14 in-depth interviews with key informants recruited through support groups for HIV-positive individuals and 21 interviews with healthcare workers in PMTCT facilities. To further probe on the experiences of HIV-positive individuals, the teams conducted 20 focus group discussions with members of support groups (7 in Burkina Faso, 6 in Kenya, 5 in Malawi, and 2 in Uganda). The focus group discussions were facilitated by an experienced qualitative researcher. All interviews except the key informant interviews with HIV-positive women attending support groups and the focus group discussions were held in antenatal care facilities. The interviews and focus group discussions were transcribed and analyzed using NVivo 8 software.
We also reviewed PMTCT policies in the four countries through document study and key informant interviews to determine the guidance given to health workers on how to conduct counseling, and how to confront the challenges of consent, confidentiality and disclosure in pre- and post-test counseling.
The mixed-methods approach of this study allows us to present quantitative measurements as well as qualitative insights on the quality of counseling, consent and confidentiality, and the stigma and disclosure challenges faced by HIV-positive pregnant women. The study was not designed to provide a representative picture of PMTCT implementation in Burkina Faso, Kenya, Malawi and Uganda. It rather aimed to describe policies and practice as experienced by women in antenatal sites and to highlight the issues at stake in further scaling up and increasing access to PMTCT.
The study was reviewed by the ethics committees of the Amsterdam Medical Center of the University of Amsterdam, the World Health Organization, and relevant bodies in each of the four African countries.