Report

Critical Review and Analysis of Voluntary Counselling and Testing Literature in Africa Prepared by: V. Solomon, H. Van Rooyen, R. Griesel, D. Gray, J. Stein and V. Nott

April 2004

Critical Review and Analysis of Voluntary Counselling and Testing Literature in Africa Written by: V. Solomon, H. Van Rooyen, R. Griesel, D. Gray, J. Stein and V. Nott HIV/AIDS Counselling Research and Evaluation Group School of Psychology, University of KwaZulu-Natal, Pietermaritzburg Campus

April 2004 Available from http://www.hst.org.za

Suggested citation: Solomon V, van Rooyen H, Griesel R, Gray D, Stein J, Nott V. Critical Review and Analysis of Voluntary Counselling and Testing -Literature in Africa. Durban: Health Systems Trust; 2004

Health Systems Trust 401 Maritime House, Salmon Grove, Victoria Embankment, Durban 4001 Tel: +27-(0)31-307 2954 Fax: +27-(0)31-304 0775 Email: [email protected]

Prepared for the Health Systems Trust, as part of Grant number 286/01: An Audit of Voluntary Counselling and Testing Services in South Africa. The information contained in this publication may be freely distributed and reproduced, as long as the source is acknowledged, and it is used for non-commercial purposes.

TABLE OF CONTENTS 1.

STUDY OVERVIEW ........................................................................................................1

2.

LITERATURE IDENTIFICATION, SOURCING AND DATA-BASING............................2 2.1 2.2 2.3

3.

Published Literature...............................................................................................2 Unpublished Literature ..........................................................................................2 Data-basing ...........................................................................................................3

STUDY 1: CONTENT LITERATURE REVIEW..............................................................3 3.1 3.2 3.3 3.3.1

Aims.......................................................................................................................3 Methods.................................................................................................................4 Findings .................................................................................................................4 The Efficacy of VCT for Achieving Prevention Goals ............................................4 The Methodological Limitations of the Efficacy Studies.........................................7 3.3.2 The Efficacy of VCT for Achieving Support Goals...............................................10 Service Delivery and Efficacy Goals....................................................................12 3.3.3 Factors Associated with the Demand for VCT......................................................14 Attitudes towards VCT .........................................................................................14 Perceived Vulnerability ........................................................................................16 Additional Factors in the Demand for VCT ..........................................................17 Critiques of VCT and HIV/AIDS Counselling .......................................................18 Critiques of the Counselling Paradigm from a Developmental Perspective ........18 Critiques of the Western Counselling Model........................................................19 Critiques of the Preventative Orientation of the Counselling Model ....................21 3.4 Conclusions .........................................................................................................23 4.

STUDY 2: A META-ANALYSIS OF VOLUNTARY COUNSELLING AND TESTING LITERATURE IN AFRICA............................................................................24 4.1 Aims.....................................................................................................................24 4.2 Methods.................................................................................................................25 4.3 Findings ..............................................................................................................26 4.3.1 Publication Profile ................................................................................................26 4.3.2 Author Characteristics .........................................................................................28 4.3.3 Research Literature .............................................................................................29 Research Focus ..................................................................................................29 Research Design .................................................................................................31 Populations and Sampling ...................................................................................32 Data Collection Methods......................................................................................36 Data Analysis.......................................................................................................37 Ethical Issues ......................................................................................................38 4.3.4 Research Quality .................................................................................................39 Quantitative Research Quality .............................................................................39 Qualitative Research Quality ...............................................................................41 Mixed Research Designs.....................................................................................42 4.3.5 Conclusions .........................................................................................................43

5.

STUDY 3: A QUALITATIVE ANALYSIS OF THE UNPUBLISHED (‘GREY’) VOLUNTARY COUNSELLING AND TESTING LITERATURE IN AFRICA.................44 5.1 Aims........................................................................................................................44 5.2 Methods................................................................................................................45 5.3 Findings ..................................................................................................................45 5.3.1 Constructions and definitions of VCT ...............................................................45 VCT as a prevention Strategy..............................................................................46 VCT as health promotive .....................................................................................48

VCT as psycho-emotional support.......................................................................49 VCT as a process of informed consent................................................................50 VCT and stigma reduction ...................................................................................52 VCT and human rights.........................................................................................54 5.3.2. Government Policies ..........................................................................................54 5.3.3 Implementing VCT..............................................................................................56 Access .................................................................................................................57 Continual upgrading of facilities...........................................................................58 Monitoring and evaluation of VCT sites ...............................................................58 Other factors ........................................................................................................59 5.3.4 Barriers to VCT Implementation ..........................................................................60 Internal Barriers: ..................................................................................................61 External Barriers: .................................................................................................64 5.4 Conclusions ...........................................................................................................67

LIST OF FIGURES Figure 1: Distribution of Articles across Year ....................................................................................... 27 Figure 2: Distribution across type of publication……………………………………………………………27 Figure 3: Institutional Affiliation of Primary Author………………………………………………………….28 Figure 4: Distribution of research literature across year………………………..………………………….29 Figure 5: Research Focus Areas.......................................................................................................... 29 Figure 6: Secondary Classification of Other Research ........................................................................ 30 Figure 7: Research Design……………………………………………………………………………………31 Figure 8: Quantitative Research Designs…………………………………………………………………...31 Figure 9: Qualitative research designs………………………………………………………………………32 Figure 10: Mixed research designs…………………………………………………………………………..32 Figure 11: Population Profile ................................................................................................................ 33 Figure 12: Population Gender .............................................................................................................. 34 Figure 13: Location of Research .......................................................................................................... 34 Figure 14: Sampling Strategies ............................................................................................................ 35 Figure 15: Data Collection Methods …………………………………………………………………………36 Figure 16: Quantitative Procedures Used ............................................................................................ 37 Figure 17: Type of Qualitative Procedures Used ................................................................................. 38 Figure 18: Ethical Issues Discussed .................................................................................................... 38 Figure 19: Quantitative Research Quality ............................................................................................ 40 Figure 20: Qualitative Research Quality............................................................................................... 41 Figure 21: Mixed Method Research Quality Rating ............................................................................. 42

1.

Study Overview

This investigation forms a component of a wider audit of the current status of Voluntary Counselling and Testing (VCT) services in South Africa, which aims to generate analysis and recommendations for the promotion, development and sustainability of these services at a national level. A primary aim of this audit is to provide an analysis of facilitative or obstructive factors that affect effective VCT service delivery in South Africa. As stated by Richter, Durrheim, Griesel, Solomon and van Rooyen (1999) “Despite the level of service provision and proportion of people reached by VCT, both in developed and developing countries, the focus of attention is definitely shifting from the urgency of ‘doing something’ to the questions “is it working?” and “how can it be improved?” (p.32). This stated aim is addressed in this component of the audit through a critical literature review and analysis of all VCT literature in Africa. This critical review was undertaken in order to:

1. Consolidate and integrate all literature on VCT in Africa - including published research, policy and strategic planning documents, unpublished literature and media reports – in order to extract lessons that inform and bolster the national VCT strategy. 2. Critically evaluate this literature in order to identify research trends, to critique these trends and identify important omissions in research. 3. To contribute to the development of a coherent national research agenda on VCT for the National Department of Health, researchers and communities alike. Three separate analytic studies were undertaken for the purposes of this review, i.e. 1. A qualitative content review of the published literature and unpublished research literature that focuses on providing an overview of what has been researched or discussed in the area of VCT and what recommendations or conclusions have been drawn as a result of this research or discussion. 2. A quantitative meta-analytic evaluation of both published literature and unpublished research literature that focuses on analysing trends and shifts in VCT research or discussion, evaluating the quality of research that exists on VCT and looking for key omissions or gaps in this research. 3. A qualitative thematic analysis of the unpublished literature that focuses on the way that the role and function of VCT is described across government reports, policy documents, unpublished research etc. In this chapter the results of each of these three literature analyses will be presented separately; ending in a discussion of this data in terms of the above-stated aims. It is anticipated that these results will provide both a wider background in which the audit can be situated, and be an important resource document for others working in the field of VCT. To this end, an annotated bibliography of all the literature collected for this component of the audit is also provided in a separate document to this report. Please note that due to the large amount of material reviewed for the studies presented below, a complete reference list has not been provided at the end of this document. Instead, please consult the annotated bibliography. In order to ensure easy accessing of particular articles of interest all referenced text will include the name of the authors, the date of publication and the reference number of the article in the bibliography. Where articles have been consulted, but are not included in the bibliography, the complete reference is provided as a footnote. References consulted for methodological help are examples of this type of referencing.

1

2.

Literature Identification, Sourcing and Data-basing

A number of different types of literature were collected for the purposes of this investigation, including published research reports and discussion documents, unpublished research reports, strategic documents and policies and media reports from both national and local newspapers. The method used for identifying, sourcing and databasing each type of literature used in the three studies is described below.

2.1

Published Literature

The published literature was identified and located using a number of local and international library search engines including PSYCH-LIT, EBSCO-HOST, ISI, Sabinet and MEDLINE. In all cases the search parameters entered were (HIV or AIDS) and Counseling as well as (HIV or AIDS) and Counselling. This search covered all years until end 2001. This initial search produced a total number of 13618 articles. However, numerous duplications were present due to the number of cross-searches, which were performed, and these were removed from the database, leaving around 3200 articles. The literature was then further filtered to identify those articles that would be most useful to the study. The criteria for inclusion employed were that firstly, the article had to relate directly to South Africa, Africa or developing country contexts and secondly, that the article had to be about counselling in the context of HIV-testing. In this second criterion it meant that articles to do with on-going counselling, bereavement counselling or other forms of HIV related counselling were not included and that articles relating to HIV testing but not dealing with the subject of counselling were not included (e.g. articles relating to medical aspects of HIV testing). Three hundred and eighty five articles passed these criteria for inclusion and it is these articles that were sourced and databased.

2.2

Unpublished Literature

The identification and location of the unpublished literature proved to be more difficult than the published literature, as this so-called “grey” literature is not generally listed in databases, nor is it easily accessible through libraries. Thus, in order to identify this literature, the following methods were employed:

1. A search of the NEXUS database was conducted in order to identify any relevant Masters or Doctoral research being conducted at a South African university. In this search, the same search terms and criteria for inclusion were employed as in the published literature. The search covered all years to end 2001. 19 theses or dissertations were identified using this method.

2. A search of all national and local newspapers in South Africa was conducted, using the services of SA Media at the University of the Free State. The search terms used were (HIV or AIDS) and Counselling and “Voluntary Counselling and Testing”. The search covered all years to end 2001. Again only those articles relating directly to counselling in the context of testing were included. 162 newspaper articles were located using this method.

3. Contact was made with all the Provincial Departments of Health to request any policy/strategy or guideline documents relating to VCT. In addition, we visited the National Department of Health in Pretoria so that we could search their library and archives and speak to key informants. Through both of these methods, a total of 74 articles were located.

2

4. A list of African research organisations with interests in HIV counselling or VCT research was drawn up from the National AIDS Database and an Internet search of organisational web pages was conducted looking for any relevant posted research. Each research organisation was also contacted via e-mail for the same purpose. Internet searches were also conducted on any international organisations, e.g. UNAIDS and WHO looking for appropriate literature. Finally, a request for information was also sent out on appropriate discussion lists, e.g. AF-AIDS, REPROHEALTH and CARE. A total of 83 articles were sourced using these three Internet methods.

5. Finally, a search was conducted through the abstract collections of the International AIDS conferences in 2000 (Durban) and 1998 (Geneva). Other relevant conference proceedings were also searched. 91 articles were located using this method. It is hard to determine the completeness of the unpublished literature that was collected. As we have no way to ascertain how much unpublished literature exists on VCT in Africa, we cannot determine how inclusive our database is. This can have obvious consequences for any discussion or conclusions drawn as a result of this data; as our analysis of this set of literature may not represent what would have been found with a more complete data set. Unfortunately, this is a primary problem with the use of “grey” literature - one that is hard to overcome. However, while we cannot claim to have collected all of the unpublished literature on VCT in Africa, what we can comment on is how representative our database may be. We would argue that the literature that was collected does comprehensively cover the different types of unpublished literature that do exist. Also, it would seem that the methods that were used were comprehensive in that we attempted to cover numerous potential sources. The general comprehensiveness of this method can be seen in the fact that many of the same articles that were sourced using one method were also sourced using another. Thus, while we cannot claim to have a complete data set, it would appear to be comprehensive across both the different types of unpublished literature and across the potential sources for unpublished literature that exist. It could therefore be said that the literature – although not necessarily complete – is generally representative of the unpublished data that is out there.

2.3

Data-basing

Once a hard copy of these published and unpublished articles had been sourced, all of these articles were databased using Reference Manager™. The articles were entered by type, assigned a reference ID and cross-referenced using key words such that the database would be searchable. A total of 754 references were entered into the database and it is these references that make up the annotated bibliography.

3.

Study 1: Content Literature Review

3.1

Aims

This literature review is aimed at providing an overview of the content of the body of published literature and unpublished research articles on VCT in Africa. More specifically, this review is aimed at providing a critical overview of: 1. What has been researched or discussed about VCT in Africa, i.e. the main focus areas. 2. How these focus areas have been researched.

3

3. The main conclusions and recommendations that have been drawn about VCT as a result of this research or discussion. This analysis will therefore answer the critical questions: 1. What do we know about VCT in Africa and what lessons are there for researchers, policy makers and service providers? 2. What critical gaps exist in our knowledge of VCT and what research will be needed to address this lack of understanding? In addressing these two questions, the results of this critical review will enable us to extract some lessons about VCT that can both inform the national VCT strategy and act to set a future research agenda for VCT in Africa.

3.2

Methods

Two types of literature were reviewed for this study, i.e. all published research, discussion documents and review articles and all unpublished research including theses, dissertations and research reports. Thus, the review is not limited to literature describing empirical research – either quantitative or qualitative – but also covers articles that are discursive and/or simply descriptive. Likewise, while the review focuses predominantly on articles from indexed international journals, other reports and unpublished material have been included. It should also be noted that for this study, while the review focuses on Africa, literature from elsewhere in the world is included where appropriate. As the method for collecting this literature has already been discussed, this issue will not be addressed here. The review structures the research into 4 main content areas, i.e. • • • •

The efficacy of VCT for achieving prevention The efficacy of VCT for achieving support Factors associated with the demand for VCT Critiques of the appropriateness of VCT counselling objectives and practices

Within each of these focus areas, the main findings and recommendations of the literature are discussed. In addition, a discussion of the main methodological limitations of the research reviewed for each section has been included where appropriate.

3.3 Findings 3.3.1 The Efficacy of VCT for Achieving Prevention Goals When discussing the efficacy of VCT for achieving prevention goals, two main preventative outcomes are usually addressed, i.e. 1. The efficacy of VCT in achieving primary prevention, i.e. facilitating behaviour change in those who test negative 2. The efficacy of VCT in achieving secondary prevention, i.e. facilitating behaviour change in those who test positive. Most investigations into these two areas, conducted in Western contexts, point to the efficacy of VCT for achieving secondary prevention goals, although it does not appear to be an effective means of primary prevention (Wilson, Levinson, Jaccard, Minkoff & Endias, 1996,

4

ID350; Landis, Earp & Koch, 1992, In ID 350; Ickovics, Morrill, Beren, Walsh & Rodin, 1994, In ID350). Similar studies conducted in the African context, have revealed varied outcomes with regards to the efficacy of VCT to achieve either primary or secondary prevention goals. For example, a study conducted in rural Uganda, showed that those who had undergone VCT showed no difference in sexual behaviour after one year when compared to those who had not been tested regardless of sero-status (Kipp, Kabagambe & Konde-lule, 2001, ID186). However, in direct contradiction, a study in Kampala, Uganda, showed that those who had been counselled and tested reported safer sexual behaviour as compared to those who had not been tested, regardless of their sero-status (Muller, Barugahare, Schwartlander, Byaruhanga, Kataaha, Kyeyune, Heckmann & Ankrah, 1992, ID95). One consistent finding across the literature is that VCT is more effective as a means of secondary than primary prevention in African settings, mirroring the international literature in this regard. This means that VCT is more effective at facilitating behaviour change in those who test positive than in those who test negative. For example, an efficacy trial conducted in Kenya, Tanzania and Trinidad, showed that infected individuals and infected couples were more likely than uninfected ones to reduce unprotected sex (Coates & The Voluntary Counselling and Testing Efficacy Study Group, 2001, ID34) and a study conducted in Rwanda, showed that women who were HIV-positive were more likely to adopt condom-use than HIV negative women (Van der Straten, King, Grinstead, Serufilira & Allen, 1995, ID143). Additionally, it has been shown that sero-negative women with untested partners are the least likely to discuss or attempt to negotiate condom-use (Allen, Serufilira, Bogaerts, Van de Perre, Nsengumuremyi, Lindan, Carael, Wolf, Coates & Hulley, 1992, ID351; Van der Perre et al, 1995, ID143). Findings that VCT does not appear to facilitate preventative behaviour change in seronegative individuals is confounded by the fact that increased condom-use, which is often used to measure behaviour change, may not be an appropriate outcome measure for seronegative individuals. This is because many individuals use VCT testing (or repeat HIV testing) to establish sero-status at the outset of a new monogamous relationship, with the implicit intention of ceasing condom-use. For example, a study conducted in Uganda, revealed that 27% of participants were taking an HIV test because of a planned marriage or a new relationship. This was even more common in the case of couples coming for VCT where 84% were planning a new marriage/relationship (Muller et al, 1992, ID95). When asked about behaviour change in response to a negative test result, the most common reported intentions were marriage (81%) followed by monogamy (17%). The use of condoms was only mentioned by 2% of respondents. This finding may well reflect the focus of the previous AIDS education campaign in Uganda, in which condom-use was de-emphasized. However, the fact remains that in these cases, VCT clearly functions as ‘a form of technological solution’ to the problem of safer sex (Watney, 1990; in Beardsell, 1994, ID 270). While this may be a rational strategy for monogamous couples, it has nonetheless been argued on the basis of such research that VCT may be counterproductive in so far as individuals then proceed to have unprotected sex (Wilson et al, 1992, ID350). Clearly, perceptions of low susceptibility may be reinforced by a negative test result. In terms of both primary and secondary prevention goals, it would appear that there are a number of factors that influence the effectiveness of VCT in African settings. One such factor would be whether the person being counselled or tested is a member of a couple that is either sero-concordant or sero-discordant. Another factor would be whether these couples are tested together or individually. Studies show that VCT is most effective in facilitating behaviour change in both positive and negative clients, when that client is one of an HIV sero-discordant couple who were counselled together, rather than individually (Kamenga, Ryder, Jingu, Mbuyi, Mbu, Behets, Brown & Heyward, 1991, ID68; Allen, Tice, Van de Perre,

5

Serufilira, Hudes, Nsengumuremyi, Bogaerts, Lindan & Hulley, 1992, ID342; Allen et al, 1992, ID351; Van der Perre et al, 1995, ID143; Mckenna, Muyinda, Roth, Mwali, Nga'ndu, Myrick, Luo, Priddy, Hall, von Lieven, Sabatino, Mark & Allen, 1997, ID176; Roth, Stewart, Clay, Van der Straten, Karita, Allen, 2001, ID177)1. In these cases, VCT has been shown to reduce unprotected intercourse and increase condom use among sero-discordant couples. This is especially so if the man is the HIV negative partner, i.e. at risk of infection. This may well be because condom-use largely depends on its acceptance by the male, rather than female partner (Allen et al, 1992, ID342; Van der Perre, et al, 1995, ID143). When looking at the efficacy of VCT for achieving secondary prevention goals in seropositive individuals who are tested individually, rather than with their partners, the results regarding behaviour modification are less optimistic. For example, some studies conducted in Gambia found little behaviour modification or disclosure on the part of asymptomatic seropositive participants after VCT and two additional follow-up sessions (Wilkins, Alonso, Baldeh, Chan, Corrah, Hughes, Jaiteh, Oelman & Oickering, 1989, ID159) and that VCT did not modify condom-use in Commercial Sex Workers regardless of whether they were HIV positive or HIV negative (Pickering, Quigley, Pépin, Todd, Wilkins, 1993, ID100). Similarly, a study conducted in Kenya, (Moses, Kiragu, Fusallah, Wamola & Piot, 1990, ID116) revealed that contraceptive use, condom use and pregnancy rates over the year following VCT were similar in HIV-positive and HIV-negative women; and that only 37% of HIV infected women had disclosed their HIV status to their partners. Finally, a study conducted in Zimbabwe found that HIV-positive people provided with ongoing counselling took some measures to prevent HIV spread but did not succeed in maintaining long-term, consistent condom use (Krabbendam, Kuijper, Wolffers & Drew, 1998, ID184) With regards to conclusions and recommendations that stem from these results, two main suggestions stand out, i.e. •

VCT appears to be more effective for prevention with sero-positive than with seronegative people, who are more likely to continue to engage in unsafe sexual practices. In order to be effective as a prevention tool with HIV sero-negative individuals and couples, post-test counselling for sero-negatives will have to be reconsidered. In particular, such interventions should aim to help individuals maintain a negative result and provide them with information to better assess their need for safer sexual practices in the future.

•

VCT appears to be more effective when it targets both sexual partners as opposed to only one. Since this finding occurs within a context of unequal power dynamics, it may also indicate that individual VCT services in the African context would be most effective if they actively targeted males rather than continuing to routinely access females at family planning and antenatal clinic facilities. Only one of the quantitative studies under review, apart from those assessing couple counselling, involved male participants (Machekano, Mcfarland, Mbizvo, Bassett, Katzenstein & Latif, 1998, ID168). This lack of research is important, as there is reason to believe that in African male-dominated societies, VCT with men may have more impact on a couple’s subsequent behaviour than VCT with women.

However, it is questionable whether any of these conclusions and recommendations about the preventative efficacy of VCT should be drawn from the body of literature reviewed here. This literature contains numerous methodological and conceptual limitations, making any claims of efficacy highly dubious, and making it all but impossible to make any clear-cut

1

Finding based on a total of 13 VCT quantitative efficacy trials and 1 qualitative study regarding HIV prevention in African settings.

6

policy recommendations. These methodological limitations, and their implications for research in this area, are discussed in some detail below.

The Methodological Limitations of the Efficacy Studies A main methodological limitation of those studies looking at the efficacy of VCT for achieving prevention goals is that most of these studies did not constitute randomised controlled clinical trials with the result that efficacy becomes extremely difficult to establish. Instead, as a result of the ethical requirements attached to the provision of VCT to anyone who requests it, the efficacy trials were forced to select a sample from those who had already requested HIV counselling and testing; using either the rate of sero-conversion in the general population (Muller et al, 1992, ID95), those who had not requested VCT, or those individuals who had received some other intervention (Coates et al, 2000, ID34) as a comparison group. There are a number of problems with this process: Firstly, in using those who volunteered for VCT, a strong self-selection bias is created – a bias that is especially problematic as there is good reason to suspect that those who volunteer for VCT are already fundamentally different to those who do not. It is highly probable that these volunteers are far more sexually responsible and at lower risk of HIV infection in the first place2. This self-selection bias is further compounded by the fact that participants must not only volunteer for VCT, but also for the research study and follow-up (cf. Higgings, 1991, In Beardsell, 1994, ID270). Thus, apart from participant bias, the research may itself constitute an intervention impacting upon behaviour change. Due to this self-selection bias, the two groups that are chosen for comparison (i.e. VCT and non-VCT) cannot be said to be similar along a number of important base-line characteristics, e.g. sexual behaviour, condom-usage etc. This is an important problem for any efficacy study, as it means that any behaviour changes that occur in the intervention group cannot be attributed to the VCT process; such changes could be as a result of the differences between the groups. In fact, well-controlled studies show that those who come forward for testing may have been motivated to make behavioural changes before the testing even took place, with the result that VCT did not, in fact, cause the behaviour change. For example, a study in South Africa (Abdool Karim, Hoosen & Susser, 1998, ID69) showed that when a baseline survey was conducted, VCT did not alter or increase participant’s reported intentions to practice safer sex. The authors found that this was because both before and after VCT, almost all women stated that they intended to use condoms and reduce the number of their partners. A second methodological limitation with the VCT efficacy studies is that efficacy is commonly measured in terms of only two behavioural outcomes, i.e. reported condom use and rates of sero-conversion. However, in an effort to establish the efficacy of VCT, researchers have tended to ignore the obvious: that outcome measures based on sexual behaviour are both unreliable and inappropriate. This is for three main reasons: 1. VCT, in and of itself, is unlikely to, and cannot be expected to, re-arrange human sexual behaviour. The relationship between VCT and sexual behaviour change is not a linear one and behavioural change is therefore an unrealistic outcome 2

The notion that it is those who volunteer for VCT are more likely to be HIV-negative than HIV positive is commensurate with study findings to the effect that the only variable significantly associated with failure to return for post-test counselling is a positive result. (Ladner, Lerol, Msellati, Nyiraziraje, De Clercq, Van de Perre & Dabis, 1996, ID71). The voluntary HIV-1 Counselling and Testing Efficacy study found – similarly, that men and women who reported unprotected sex were less likely to return for additional post-test counselling sessions.

7

measure. Any research that proves that such a direct relationship exists is inherently dubious. On the other hand, research that proves that such a relationship does not exist cannot conclude that VCT is a waste of time or resources since the prevention effect of VCT, in combination with other interventions and under certain conditions, may be substantial. 2. Self-reported behaviour change cannot be relied upon; behavioural measures, which are retrospectively self-reported, are biased by both recall and perceived social desirability. 3. To measure the success of a client-centred counselling approach against predefined objectives, as opposed to the client’s expectations and needs, is contradictory. The idea that specific behavioural outcomes are produced by VCT is in conflict with the goals and aims of counselling. A third methodological limitation of the studies under review is that in the studies reviewed the process of VCT has not been well defined, with the result that there are important differences across the studies in terms of key aspects of VCT (e.g. length and/or number of counselling sessions). In general, these studies provide little detail about the actual counselling procedures used, and even where standard guidelines for counselling are provided, there is little recognition of the fact that counsellors may disregard these guidelines and/or interpret them differently. The main problem with these differences is fairly obvious, i.e. if the stimulus of VCT is not uniform, efficacy studies evaluating “VCT” may all be evaluating different things. Efficacy outcomes may have more to do with the type and/or standard of counselling, or the personalities or objectives of counsellors themselves, than any other factor. The impact that different types of counselling or counsellors can have on the VCT process has been demonstrated in some detail in qualitative research by both Silverman (1989, ID393) and Stein (2001, ID279). In addition, quantitative studies have shown that there are considerable variations in the proportions of clients who decide to proceed with testing after counselling depending on the differences in location and/or counsellor characteristics (Beardsell and Coyle, 1996, ID413). In the studies under review, equally little attention is given to the context of VCT and how this may have determined study findings. Multiple social, economic and cultural factors, too numerous to list let alone measure, may interact with counselling to affect behaviour. For example, participant characteristics moderating the effectiveness of VCT have not, and indeed cannot be, adequately controlled. Studies under review have attempted to ascertain the impact of participant characteristics in terms of demographic and crude psychosocial categories such as gender, income and education. Unsurprisingly, such studies show inconsistent findings. Motivation to change on the part of the individual appears to be an uncontrolled factor for all relevant studies under review with the exception of Abdool Karim et al (1991, ID69). Clearly, change is dependent upon motivation and the pre-existence of motivation to change, or lack thereof, will influence behavioural outcomes. While there is increasing evidence that motivation to change is not necessarily a good predictor of any behaviour, especially sexual behaviour, the fact remains that many of the investigations under review conducted their studies in contexts such as ante-natal and STI clinics. In these settings, VCT is routinely promoted to all clients, making VCT less of a personal choice than a routine mandated testing procedure. It seems likely therefore, that a sample of clients from the general population actively seeking VCT, rather than those having the service recommended to them by authority figures such as nursing personnel, may yield very different findings.

8

On the basis of these methodological limitations, a number of recommendations for future research can be made. Firstly, it is clear that if we want to establish the efficacy of VCT for achieving preventative outcomes, more randomised trials are needed where extraneous factors can be controlled for. In addition, these studies need to better address the methodological limitations discussed above - especially the issue of establishing better outcome indicators for VCT. However, the variability that exists across study findings as to the efficacy of VCT for achieving either primary or secondary prevention goals raises the issue of whether exploring issues of efficacy is the right question for VCT research in African settings. Rather, it could be that the question that needs to be explored in this field is not so much whether VCT is effective for primary or secondary prevention in African contexts, but rather, under what specific conditions it is effective. Clearly, disclosure to partners is one of the determining variables in this regard. In the absence of disclosure to partners, participant’s sexual behaviour depends as much and in the case of women, arguably more, on the attitudes and intentions of their partners as on their own attitudes and intentions regarding safer sex. This could prove to be an important area for future research. Moreover, it is clear that even if available outcome evaluation data does provide valid, albeit inconsistent evidence as to the effects of VCT on prevention, rather than what has been referred to as ‘spurious statistical links’, they nonetheless provide little information as to how and why such outcomes, negative or positive, were achieved. For this reason, some authors advocate a move away from outcomes based research, which they claim, contributes nothing to the generation of theory relevant to VCT because such research neither identifies nor describes the factors leading to specific outcomes. These authors therefore support the idea that resources be dedicated to more qualitative process-based research which can identify the reasons why any given service is succeeding or failing. Only such research can provide an informed basis for amending and developing services (Beardsell and Coyle, 1996, ID413). One qualitative study under review (Meursing, 1991, ID161) is an example of the way in which process-based research can provide valuable information regarding the behavioural outcomes under review. In this study, counsellors transcribed all their sessions, thus creating a highly detailed record of client’s problems and coping strategies over time. Meursing’s research showed that HIV positive individuals did not succeed in maintaining long-term, consistent condom-use and her qualitative study design also meant that she was able to explain exactly why this was so.3 Given the wealth of experience which counsellors have, not to mention their access to clients, one must wonder why it is that only one study under review made use of counsellors own obvious capacity to elicit detailed contextualised data within counselling sessions which has obvious potential for challenging existing theories and generating new ones. It seems likely that this is because such ‘case-history’ methodologies do not approximate the scientific ideals of verification and generalisability. These values, despite the fact that they are arguably outdated and inappropriate, remain the guiding principals in defining research design and determining what research findings are published. As Beardsell & Coyle (1996, ID413) point out, this is despite the fact that quantitative studies seldom produce findings that can contribute to the framing of policy or the development of counselling practice. Additional qualitative research is urgently required regarding all aspects of the actual counselling and testing process, including the way in which informed consent is obtained, risk is assessed and risk reduction strategies developed, and the way that results are given 3

Meursing (1991, ID161) shows that HIV positive people encounter the same barriers to safer sex as the community in general; gender inequality, denial, and the desire for children. In addition, the fear of stigma and prejudice make negotiating safer sex particularly difficult.

9

to clients. The value of supplementary research to contextualise efficacy study findings cannot be over-emphasised in terms of their insights into what it means to do effective VCT, what factors influence the efficacy of VCT and how these can translate into the development of better VCT programmes.

3.3.2 The Efficacy of VCT for Achieving Support Goals In comparison to research into the efficacy of VCT for achieving primary and secondary prevention objectives, there is a noticeable paucity of research into the efficacy of VCT for achieving either treatment or psychosocial support objectives. Most available findings regarding the efficacy of VCT for achieving support objectives are qualitative, and this research is largely anecdotal. This lack of research into the extent to which the supportive objectives of HIV/AIDS counselling are achieved is cause for some concern as it suggests that those involved in HIV/AIDS research are much more concerned with public health goals, than with individual care and support. Most of the available studies that have focused on the impact of VCT on subsequent coping in HIV infected individuals have, by and large, concentrated on the question of disclosure. This is in all likelihood because disclosure is commonly promoted – not just as a means for achieving psychosocial support – but as a means of prevention as well. However, the evidence on the effect of VCT on disclosure rates is far from optimistic. The HIV-1 Counselling and Efficacy study conducted in Tanzania showed that only 27% of HIV positive women disclosed to their partners six months after testing (Grinstead, 2001, in Maman, Mbwambo, Hogan, Kilonzo & Sweat, 2001, ID172). Also, a study conducted in Dar es Salaam revealed a disclosure rate of only 22% in women two months after testing (Antelman, Daaya, Mbwambo, Fawzi, Msamanga, Hunter & Smith, 1999 in Maman et al, 2001, ID172). Similarly, of those in the PETRA – Perinatal Transmission - study, only 16.6% of women disclosed to their partners 18 months after testing (Kilewo, Massawe, Lyamuya, Kalokola, Semali, Karrisson, Mhalu, Biberfield, 1999 in Maman et al, 2001, ID172). An important aspect of the disclosure of HIV status to significant others, is the negative lifeevents that have been reported by a number of HIV positive individuals as resulting from sero-status disclosure. According to Temmerman, Ndinya-Achola, Ambani & Piot, (1995, ID295), 6% of HIV-infected pregnant women in Nairobi reported a violent event after learning their HIV status. Likewise, Grinstead, Gregorich, Choi & Coates (2001, ID329) reports that positive sero-status is associated with the break-up of a marriage and being neglected or disowned by family. The authors state: “the only positive life event that was more common among sero-positive individuals was increased emotional support from health professionals4.” In addition, discrimination and physical violence against African women, triggered by the disclosure of a positive sero-status, has been well documented by North and Rothenberg (1993, in Grinstead et al, 2001, ID329), Rothenberg and Paskey (1995, in Grinstead et al, 2001, ID329) and Temmerman, Ndinya-Achola; Ambini and Piot (1995, ID295). However, despite these reports of negative personal and social outcomes after VCT, these authors continue to advocate the use of VCT (Temmerman et al, 1995, ID295; Maman, Mbwambo, Hogan, Kilonzo, Sweat & Weiss, 2001, ID8798; Grinstead et al, 2001, ID329). For example, Temmerman et al promote VCT on the grounds that it “safeguards the health of 4

However, their results, which focus on the occurrence of positive and negative life events after VCT regardless of HIV status, partially obscure this finding. These authors suggest, somewhat disingenuously, that since those who underwent VCT overall were no more likely to experience negative life events than those who were assigned health information without testing, VCT should be promoted. However, it is clear that those who were positive experienced more negative life-events than those who underwent VCT and were negative.

10

the community” and Maman describes the lack of sero-disclosure after VCT as a “serious public health concern”. While these authors do not deny the need for post-test support for those infected, it is clear that the prevention outcomes of VCT are to be prioritised over support and care goals with the result that the positive prevention outcomes associated with disclosure are seen to outweigh the negative life consequences of disclosing. Further evidence of this kind of thinking can be seen in the work of Roux (2001, ID206). After citing ample research evidence to suggest that violence against women as a result of disclosure does indeed regularly occur, Roux (2001, ID206) somehow manages to conclude that what he considers to be ‘uniformly low rates of violence reported following disclosure of the diagnosis to the woman’s partner” should “encourage counsellors to support disclosure in all but relationships with a prior history of violence, drug abuse or homelessness”. He is mindful of the need to provide psychosocial support to those who are infected, but suggests that this is best achieved by facilitating disclosure in order that support can be sought from the families of affected persons. Roux acknowledges that there is a paradoxical tension between the need for emotional support from others and the stigmatisation attendant upon disclosure but argues that while fear of stigmatisation may be the motive for non-disclosure, non-disclosure also perpetuates stigma. He suggests that a modification to the consent process, involving the family in pretest counselling, might break the cycle of stigma and non-disclosure. Overall, however, it is clear that Roux (2001, ID206) is not overwhelmingly concerned about the support HIV positive women stand to gain, or lose, as a result of disclosure. His primary concern is that non-disclosure is a “serious obstruction to the flow of information regarding HIV and the risk of infection, with potentially lethal consequences to society.” In contrast to this view, Krabbendam et al (1998, ID184), point out that ‘the basic goal of counselling is to help people in dealing with the results of the HIV test’. These authors argue that the multiple stories of negative outcomes from women illustrate that the fear of stigmatisation is well founded and that the negative consequences of HIV infection are multiple and extremely difficult. This kind of context makes disclosure practically impossible, with the result that ongoing counselling for those who are infected is required if these difficulties are to be overcome. One study that did look at how well VCT met support and care objectives, was a detailed qualitative study conducted by Meursing (1999, ID161). This study showed that many people with HIV, as well as their family members, meet the news of HIV infection with a strong sense of helplessness, resignation or ‘deep fatalism’. Meursing argues that while VCT improves basic factual information regarding HIV/AIDS, it fails to address more subtle problems in patients’ conceptualisation and understanding of HIV that have far-reaching consequences for their preparedness to deal with their test results. She also shows that the expectation of social rejection works against disclosure goals, and that the psychological denial of infection, especially during symptom-free periods is common. Meursing (1999, ID161) argues that VCT programmes therefore need to go hand in hand with interventions that promote ‘living positively with HIV’ to the community as a whole and that counsellor training should include culturally appropriate interventions for dealing with denial and avoidance. Like Krabbendam (1998, ID184), she suggests that VCT in and of itself, is insufficient to prepare clients who are infected for effective coping. Despite showing that VCT, whether accompanied by disclosure or not, is often not adequately coped with by HIV positive individuals, authors such as those mentioned above do not go on to question the necessity for VCT in an African context or to suggest that the process of VCT be substantially modified. Rather, their emphasis is on the need for VCT to be complemented by ongoing counselling service provision.

11

On the basis of detailed interviews with HIV positive patients in Johannesburg, Stein (1996, ID389) argues that there are important differences between the realities of HIV infection in African countries as opposed to first world countries which determine the ways in which people cope, regardless of psychology or group culture. She maintains that if so-called ‘avoidance coping’ is prevalent in this context; then this is not necessarily because individuals are in denial but rather, because ‘active coping’ (i.e. the goal of supportive counselling) is less useful in an African context than in a Western one. Stein argues that avoidance coping in this context may be a rational strategy and a conscious act of will which, from the point of view of HIV positive people, is a pre-requisite for ongoing mental and physical health. Likewise, participants in her study view non-disclosure to be adaptive in the face of the stigma attendant upon the disease. In the absence of any meaningful medical treatment, social and/or welfare support, avoidance coping is an inevitable resort against helplessness and despair. That the counselling paradigm utilized in VCT and ongoing counselling alike are premised on a theory of coping that has a strongly individualistic and self-reflexive basis has been noted by numerous authors promoting the development of a more culturally appropriate counselling model for Africans. Stein (1996, ID389) argues, however, that it is not so much cultural differences per se, as a lack of adequate resources which makes the Western counselling paradigm, with its objective of ‘active coping’ so difficult to achieve. Clearly, avoidance coping and non-disclosure puts the HIV infected person in a tenuous relation to prevention, but Stein argues that if counsellors discredit and undermine such coping strategies, then this has more to do with the requirements for prevention than it has to do with the requirements for coping with HIV infection on an individual level. Clearly, inadequate coping with HIV and a lack of disclosure have negative implications for the efficacy of VCT for prevention. By implication, improving the efficacy of VCT for prevention entails improving the capacity of counselling and welfare services to provide ongoing support to those who are HIV infected. Overall then, there is a need for more research into the support and care outcomes of VCT, with the realisation that such outcomes are as important, if not more so, than the public health goals of prevention and behaviour change. However, there does appear to be some need for a wider definition of support and care – beyond the issue of promotion of disclosure. In fact the issue of disclosure promotion needs to be re-examined in terms of its effects on coping, rather than its effects on the prevention agenda that it supports. Thus, research needs to move beyond measuring the benefits of testing and instead focus on the benefit or harm to the individual being tested, keeping in mind that “medical practitioners should at least do no harm”. Research into how VCT helps an HIV-positive person accept their results, as well as how it can promote coping behaviour is desperately needed. Most authors seem to be advocating the view that this type of support and coping is only possible when VCT is properly located in a continuum of care where on-going support and care is possible. In fact, the literature above seems to endorse the view that while VCT should most certainly be available to those who choose it, it should not be actively promoted as a public health policy in the absence of an enabling package of both care and support facilities. As Keeling (1993, ID405) points out, the possibility of prolonging life by early intervention in HIV disease is abolished by the suicide of a person who was tested before he was ready to receive an HIV positive result.

Service Delivery and Efficacy Goals There is recognition in the literature on efficacy that certain aspects of the delivery process for VCT may affect how effective that VCT service is at achieving either prevention or support goals. For example, some South African research has shown that very little that takes place in the name of VCT is in fact voluntary – a fact which undermines both preventative and support outcomes. Fawcett (2001, ID335), in a study conducted at a district

12

hospital in South Africa, found that of the 368 patients counselled and tested between Jan 1999 and May 2000, none came voluntarily for testing. Rather all patients were referred to have VCT by a doctor and were often in an advanced stage of AIDS. The primary goal of VCT in such a context is the initiation of appropriate medical treatment, rather than either primary prevention or the provision of psychosocial support. Other organisational factors that can impede psychosocial and prevention outcomes are the time-constraints imposed by the multi-tasking expected of nurse-counsellors, which play a significant role in how patients are received and treated. In Fawcett’s (2001, ID335) study, the average number of patients receiving consultations with health care professionals was 6 per hour, averaging 10.4 minutes per patient. It is doubtful whether either coping or preventative messages can be conveyed in such contexts. Fawcett also notes that nurses’ perceived role and authority also hinder dialogically based activities founded on the principle of equality, thereby making most counselling prescriptive and advice-giving, rather than facilitative and coping directed. This could substantially affect psychosocial goals. Despite the perceived impact that service delivery factors can have on the efficacy of VCT, there is very little operations research that looks at the service delivery conditions that best promote the goals of VCT. However, numerous prescriptions abound as to what services should be introduced and what those services should look like. This gap in understanding is replicated at a policy level where numerous national plans exist, but little attention is given to the processes that are required to implement effective VCT services. Existing VCT policies tend to rely on universal blueprints developed out of context that are not rooted in the possibilities and constraints of local situations. Such policies tend to over-estimate implementation capacity on the ground (Schneider & Stein, 2001, ID390). Brown, Macintyre et al (2001, 352) provide an extremely useful overview of the VCT literature in Sub-Saharan Africa with an emphasis on findings guiding the development of operations research studies to improve VCT service delivery in the future. They list a large variety of questions that require ongoing research and note that the answers to these questions will differ depending on specific contexts. The multitude of questions they list relating to the demand and supply of VCT demonstrates the complexity and multiplicity of the factors under review. Some of the urgent questions requiring research that they list include: • • • •

What is the most appropriate model for provision of VCT services? Who is best placed to provide these services? How viable are community-level approaches to the provision of VCT? What level of resources will have to be devoted to VCT in order for it to have a significant impact?

It is clear that for achieving both prevention and support objectives further research is required in order to explore other models, as well as other aspects of the overall continuum of care. These objectives should not and cannot be delimited to the practice of counselling per se. Indeed, other models of intervention may well prove to be better placed to achieve prevention and support objectives in the African context, given the seeming lack of cultural fit, not to mention the lack of human and financial resources for counselling in Africa, in particular, ongoing counselling. At present, VCT and HIV/AIDS counselling tends to acquire all the roles and functions that rightfully belong to an entire social service delivery system and to a continuum of treatment and care. The result is that counselling is inevitably burdened with inappropriate kinds of prevention and support functions. Consequently, the specific strategic objectives of VCT in particular, and counselling in general, in a national HIV/AIDS campaign are compromised.

13

3.3.3 Factors Associated with the Demand for VCT Attitudes towards VCT Studies that have focused on people’s attitudes to VCT, in general, attempt to provide some measure of the acceptability of VCT in some population as an indicator of the demand for VCT services in that population. Such studies have shown high acceptance of VCT programmes, and HIV prevention services in general across a number of countries, including Abidjan, Kenya, Tanzania, Malawi, Zambia and Zimbabwe (Cartoux, Meda, Van de Perre, Newell, de Vincentzi, Dabis & The Ghent International Working Group on Mother-To-Child Transmission of HIV, 1998, ID40; Cartoux, Msellati, Meda, Welffens-Ekra, Mandelbrot, Leroy, Van de Perre,P & Dabis, 1998, ID42; Kiarie, Nduati, Koigi, Musia & John, 1999, ID118; Kipp et al, 2001, ID186; Pool, Nyanzi, Whitworth, 2001, ID102; Wilkinson and Wilkenson, 2001, ID189) However, the value of using acceptability as a measure of the demand for VCT services is somewhat questionable. A high level of acceptance at the level of attitudes, and a stated desirability for a service is not always equal to the intention to use it. Likewise, a stated intention to test is not equivalent to, but only a precondition for, undergoing the process of VCT. This is shown not only in the fact that there are low levels of HIV testing in areas where there is high acceptability and demand for VCT services, but also in the fact that even those who do come for testing and counselling do not always return for their results, showing that willingness to test does not always equal willingness to know. This point is well demonstrated by Fylkesnes, Haworth, Rosensvard and Kwapa (1999, ID185). These researchers explored the uptake of VCT in a general population sample of selected urban and rural areas of Zambia. Of those stating that they wanted testing only 9.4% actually came forward to be tested. Furthermore less than half of those who did get tested returned for their test results. These authors found that although self-perceived risk and high-risk behaviours were positively associated with initial (in principal) willingness to test, they were not associated with the actual use of VCT services, demonstrating that the use of VCT services may be predicated on far more complicated factors than attitudes towards these services. A more appropriate measure of the extent to which VCT is actually deemed desirable in any given population, may be the extent to which participants who undergo VCT return for their test results. Certainly, if counselling accompanies testing, then this is because it is generally acknowledged that only after a process of counselling are individuals in a position to make informed decisions regarding whether or not they want to undergo testing. Thus, in measuring the extent to which people make this informed decision, one should have a better indicator of the actual willingness to test in a population. Studies that have used return rates as an indicator, have on the whole found low rates of return for test results (Ladner et al, 1996, ID71; Sahly, Kassa, Agonafer, Tsegaye, De Wit, Gebremarium, Doorly, Spijkerman, Yeneneh, Coutinho & Fontanet, 1999, ID127; Kip et al, 2001, ID186). For example, in a study conducted in Ethiopia less than half (43.5%) of those who came for VCT returned for post-test counselling and their results (Sahlu et al 1999, 127). Similarly, a study conducted in Nairobi found that only 35% of women choose to return for their results (Temmerman et al, 1995, ID295). Other studies show less extreme, but still significant dropout rates (Ladner et al, 1996, ID71; Cartoux, Msellati, Meda, Welffens-Ekra, Mandelbrot, Leroy, Van de Perre & Dabis, 1998, ID42; Sahly et al, 1999, ID127; Kiarie, Nduati, Koigi, Musia & John, 2000. ID118; Kipp et al, 2001, ID186) This issue may, at first sight, appear to have lost much of its urgency with the advent of rapid testing. It has been demonstrated that uptake rates, i.e. the proportion of those returning for post-test counselling and receiving their results) can be improved significantly by simply shortening the period of time between pre and post-test counselling (Fylkesnes et al, 1999,

14

ID185). However, findings regarding return rates are still very important, in so far as they indicate that people who request their blood to be taken often change their mind if given the opportunity to do so. This should clearly be taken into account in the refinement of counselling procedures applicable to rapid testing protocols to ensure that those receiving testing really do want to know their HIV status. It seems likely that high-drop out rates can be attributed to counselling procedure as much as to individual or group variables. In other words, it is possible that the approach to, and method of, pre-test counselling used, is a determining factor of the extent to which individuals, especially high-risk individuals, return for their results. For example, in a London Health District, where high-risk patients were individually counselled for between 20 and 40 minutes, all of those who decided to undergo testing returned to receive their results (Bor, Miller & Salt, 1991, ID565). By comparison, in many of the African studies under review, pretest counselling was performed through group pre-test, followed by individual post-test counselling. It is clear that those who are part of this group pre-test session (e.g. women at ante-natal clinics) may feel pressured to “consent” to testing and it is therefore unsurprising that many of these people fail to return for their results. This pressure to “consent” for testing may be especially true of VCT efficacy trials and other research studies where participants may be tested not to know their status but for a myriad of other factors. For example, in the efficacy study in rural Uganda where there were extremely high drop out rates for post-test counselling, 74% of participants said that they underwent HIV testing not because they wanted to know their results, but because they wanted to be part of another research study (Kipp et al, 2001, ID186). Similarly, in a study of informed consent in VCT research, Abdool Karim et al. (1998; ID69) found that although women participating were assured that their participation was voluntary, and indeed, participants themselves stated that they ‘wanted to know’ their HIV status, 88% also maintained that they felt compelled to participate in the research. Of these, 28% stated that they agreed to the test in order not to compromise the care they had come to the hospital for in the first place. This subtle coercive element - which may operate in any research conducted at public health care setting or where health care professionals are held in high regard - has already been noted by other researchers. It is well documented that patients relinquish autonomy to professional medical authority in medical settings and it is therefore not surprising that informed consent sought under such circumstances may be less than voluntary. In such contexts, not returning for test results may be the only polite way for many women to decline testing. These findings reinforce the concern of Fylkesnes et al (1999, ID185) that the use of rapid testing in contexts such as STI clinics, antenatal clinics and TB clinics - where VCT is routinely offered to all clients - needs to be further examined. The promotion of VCT to all clients is very different to a situation where VCT service is simply made available to those who may request it. While the anxiety attendant upon waiting for results is significantly reduced with the advent of rapid testing, the client is given only limited time to reflect on the pre-test and to make a considered decision about whether they really wish to proceed with the testing procedure (Beardsell and Coyle, 1996, ID413). Clearly, rapid testing highlights the difficulties inherent in achieving genuine and well-considered informed consent. The implications for counsellors as well as clients must be considered if many of those who might well have chosen not to return for their results, are in fact not truly willing to entertain a positive result, yet will now be receiving it. Finally, some reflections on the research reviewed in this section. Firstly, it is worth noting that, although there are exceptions to the rule (e.g. Killewo, Kwesigabo, Comoro, Lugalla, Mhalu, Biberfeld, Wall & Sandstrom, 1998, ID331; Pronyk, Kim, Makhubele, Hargreaves, Mohlala & Hausler, 2001, ID334), the vast majority of studies under review regarding the acceptability of VCT services were conducted at antenatal clinics. This is not surprising given that VCT has been implemented most extensively in this context where the imperative of

15

preventing mother-to-child transmission (MTCT) is paramount. As a result, however, substantially less information is available regarding men’s attitudes to VCT or regarding attitudes to VCT in other community facilities or contexts such as PHC clinics, where the advantages of VCT for the prevention of MTCT do not predominate. The importance of researching men’s attitudes to VCT is reinforced by the fact that numerous studies also concur that women are less likely to present for VCT at general public testing facilities than men (Pronyk et al, 2001, ID334) Secondly, in reviewing the large amount of research that has focussed on the acceptability of VCT, the question must be asked of why it is that so much time and resource has been devoted to researching what is by definition not an invasive public health intervention but rather a voluntary and individual choice. Clearly, Abdool-Karim et al (1998, ID69) have identified a key issue when they look at whether VCT is genuinely ‘voluntary’, especially in medical settings in which it is routinely advised.

Perceived Vulnerability Many studies have shown that the one variable most significantly associated with failure to return for post-test counselling and results is the perceived probability of a positive HIV test result (Wilson et al, 1996, ID350). This is despite a widely held view that an awareness of risk is one of the primary motivators for individuals to get tested in the first place (Maman et al, 2001, ID172). For example, Cartoux et al. (1998, ID43) show that HIV–infected women in Burkina Faso, West Africa, were three times less likely to return for results than uninfected women. Similarly, Temmerman et al. (1995; ID295) show that only 35% of women with a positive test result at an ante-natal clinic in Ethiopia requested their results (Ladner et al, 1996, ID71 and Kiarie et al, 1999, ID118) report similar findings). This suggests that those who are willing to receive a test result are doing so because they have reason to believe that it will be negative. For this reason, Manson (1990, as cited in Beardsell, 1994, ID270) argues: “people who are unlikely to be infected are the ones who take the test, in droves”. Certainly, there has been a great deal of resistance to VCT in African settings on the grounds that confidentiality procedures may not be trusted, and that stigmatisation and discrimination results from disclosure. In addition, it has been acknowledged that knowledge of sero-status increases anxiety without major treatment benefits in low-income countries where antiretroviral therapy is not universally available (Ladner et al, 1996, ID71; Van Rooyen & Wood, 1998, ID142; Kipp et al, 2001, ID186). Such concerns may well outweigh the advantages of receiving a test result; especially for those who have reason to believe their result may be positive (cf. section 3.3.2 for the impact of VCT on coping). However, findings regarding the extent to which HIV-positive as opposed to HIV-negative people volunteer for VCT are not wholly conclusive. In Northern Uganda, differentials in attitudes towards HIV tests showed that ethnicity, marital status and gender, but not HIV status, were some of the determinants of willingness to be tested (Ayiga, Ntozi, Ahimbisibwe, Odwee & Okurut, 1999, ID15). Likewise, In Ethiopia, (Sahly et al, 1999, ID127) factors affecting post-test attendance did not include HIV status and results were requested equally often by sero-positive and sero-negative women; however these authors did suggest that this population may not have recognised themselves to be at risk. What is clear, however, is that the assumption that perceptions of personal risk or perceived vulnerability will motivate individuals to be tested is an erroneous one. Two studies specifically investigated this question and found that previous high-risk behaviour was not significantly associated with the decision to test (Sahly et al, 1999, ID127; Machekano, Mcfarland, Mbizvo, Bassett, Katzenstein & Latif, 1998, ID168). In fact, males who reported recent casual partners tended to be more reluctant to learn their test results. This could be attributed to the psychological factors of denial and avoidance coping (i.e. not wanting to

16

know) on the part of those who know they are at high risk of being infected (Wilson et al, 1996, ID350). Other research findings show that the relation between perceived risk, actual risk and HIV test results is an exceptionally complicated one. For example, Wilson et al. (1996, ID350) have shown that self-reported previous risk is not significantly associated with increased likelihood of infection. They found that perceptions of susceptibility might not be related to actual risk levels as measured by STI tests. Furthermore, Muller et al. (1992, ID95) in a comparison of HIV-positive and HIV-negative individuals found that previous risk, measured in terms of reported lifetime numbers of previous partners and reported number of casual sexual contacts in the previous 6 months, did not appear to substantially differentiate HIVpositive from HIV-negative individuals. This is an interesting finding, in so far as it suggests that the behavioural measures of risk commonly used by researchers may not be highly predictive of infection.

Additional Factors in the Demand for VCT Apart from a perception of risk, another salient factor that appears to influence the individual’s decision to be tested is the perceived benefits of testing. A number of studies have focused on this question by looking at who comes for VCT and, more importantly, why they do so (Muller et al, 1993, ID95; Gibney, Wade, Madzime & Mbizvo, 1999, ID54; Balmer, Grinstead, Kihuho, Gregorich, Sweat, Kamenga, Plummer, O'Reilly, Kalibala, Van Praag & Coates, 2000, ID183; Maman et al, 2001, ID172). An ability to plan for the future is perhaps the main perceived benefit of testing, either because it allows couples to proceed with unprotected sexual intercourse, marriage and pregnancy or because it allows individuals to prolong and modify their lives in the face of HIV infection (Gibney et al, 1999, ID54; Maman et al, 2001, ID172). Other researchers have focussed their attention on profiling the socio-demographic characteristics of who comes for VCT in an attempt to identify variables associated with the demand for VCT (e.g. education, age, gender, marital status, etc.). These studies have, by and large, provided a number of inconsistent findings (cf. Brimmer, Gillespie & Schomer, 1996, ID290; Ayiga et al, 1999, ID5; Marum, Barnaba, Rayfield, Campbell, Dillon, Fridlund & Msowoya, 2000, ID219) with the result that while these studies do provide a demographic and behavioural profile of VCT clients in a specific area at a specific time, they tells us nothing about the nature of the decision-making process involved in being tested for HIV. Beardsell and Coyle (1996, ID413) therefore argue that most quantitative studies regarding VCT users have merely succeeded in identifying a variety of socio-demographic factors that are statistically associated with their dependent variables, but have not shed real light on the question of who uses VCT and why. One of the studies under review did attempt to address this question (who comes for VCT?) from another perspective, i.e. that of the research participants themselves (Kipp et al, 2001, ID186). Respondents were asked to identify their own reasons for returning or not returning for a result. However, the full decision-making process was not explored and participants are generally simply asked to provide a reason for non-attendance. In fact, reasons offered by participants would have been better labelled ‘excuses’ than ‘reasons’. Such responses included being sick, visited relatives, or forgetting (see p.283 of Kipp et al, 2001, ID186). Maman et al. (2001, ID172) point out that gender inequality is a significant factor affecting the uptake of VCT in women. Fear of a partner’s reaction and partner’s negative attitudes towards testing will act to limit the number of women who choose to be tested. Thus, it would appear that unequal power relations that limit women’s control over their risk of infection in the first place also limit testing behaviour. There also appear to be important differences in the factors that motivate men and women to ask for a test. For example, it would appear that

17

while men tended to be tested to confirm that they are negative, women are more likely to have a perception of personal susceptibility and therefore will engage in testing to find out if they are positive (Maman et al, 2001, ID172).

Critiques of VCT and HIV/AIDS Counselling The value of VCT, and of HIV/AIDS counselling more broadly, has been widely debated in both the lay media and the academic literature. Much of the literature advocating the use of counselling for HIV prevention is concerned with showing why it is that counselling, as an individualized, non-didactic problem-solving exercise, is a more effective behavioural intervention than education alone. While these arguments are generally accepted, proponents and critics of VCT and counselling alike have nonetheless criticized what they perceive to be a Western, individualistic and bio-medical model of counselling. Questions have been raised in numerous discursive articles regarding counselling aims and objectives, as well as counselling methodologies and practices (Green, 1989, ID53; Lindegger and Wood, 1994, ID76; Campbell and Rader, 1995, ID39; Mapekula, 1996, ID243; Jones, 1999. ID62; Mkhize, 2000, ID307; Richter, Van Rooyen, Solomon, Griesel & Durrheim, 2001, ID338) Generally, there are three inter-related strands of debate that can be identified from this discussion, i.e. 1. Critiques of the individualistic orientation of the counselling model, whereby counselling addresses the problem of HIV from an individual rather than a developmental perspective 2. Critiques of the western orientation of the model, whereby counselling addresses the problem of HIV/AIDS within a Western treatment and care paradigm rather than African or indigenous one 3. Critiques of the preventative orientation of the counselling model, whereby counselling focuses on epidemiological goals - the prevention of spread of HIV -to the detriment of the counselling goals of care and support.

Critiques of the Counselling Paradigm from a Developmental Perspective Theories of behaviour change upon which VCT and counselling interventions are based have been criticized for conceptualising the individual as the ultimate determinant of behaviour; thereby relegating contextual, environmental and social factors as confounding factors or barriers to, rather than primary determinants of, a rational decision-making process (Lachenicht, 1993, ID75). Thus, while contextual factors impacting on the rational individual are acknowledged, they are nonetheless implicitly demoted to a lesser role in the decision making process. This individualistic, self-empowerment model of counselling that has been strongly criticised in so far as it is considered to be a misguided attempt to produce behaviour change through changing people, rather than changing the contexts they are in. These critics argue that socio-economic and cultural contextual factors may be the most important determinants of behaviour, rather than simply being factors which moderate that behaviour (Standing, 1992, ID119). This body of literature is not necessarily ‘against’ VCT or counselling in general, but it is concerned to show that counselling is not going to be the “magic bullet” for HIV prevention that everyone is looking for. It has often been noted that, in relation to HIV, counselling has become a dominant paradigm for both prevention and care in the absence of a medical cure (Albertyn, 1992, ID13). However, the limitations of an individual behaviour change approach

18

must be recognised in that such behaviour change cannot occur in isolation from a number of contextual and cultural factors, some of which are primary determinants of that behaviour (Padayachee & Evian, 1991, ID108). For this reason, many authors call for widespread structural change as an alternative to behavioural and educational solutions to HIV/AIDS. As Lachenicht (1993, ID75) puts it, “a solution may only be found in the restructuring of society – or at least in solutions based on a deeper understanding of why the sexual behaviour one wishes to alter takes the form it does.” Similarly, Wood (1993, ID148; 1994, ID149) argues for the adoption of communityoriented models of prevention and support that will lead to the challenging of pervasive power inequalities in society. According to Wood, this leads to a social transformation model that seeks to enhance health and well being by bringing about far-reaching social change. Stein (1996, ID389) tackles the problem of the appropriateness of the HIV/AIDS counselling paradigm by analysing the history and origins of HIV/AIDS counselling from a Foucauldian perspective. Stein points out that, ‘there is hardly any aspect of HIV infection for which the individual is not held accountable by himself and others”, a viewpoint which is unsustainable in poorer communities where contexts and situations provide little self-determination. Like many other authors, she argues that self-empowerment and self-determination are only possible in a context of increased gains on the level of the individual’s ‘freedom of choice’. If HIV/AIDS functions as a social barometer, highlighting social imbalances within societies, and also across countries and continents, then it should not be surprising that the ultimate solution revolves around reducing developmental disparities (Barnett & Blaikie, 1992 in Lindegger &Wood, 1995, ID76). It seems certain that influencing the underlying causes of the epidemic will do much more to control the spread of HIV infection than the best education and counselling programmes (Moses & Plummer, 1994, ID 401). More conceptual or theoretical work is required to locate counselling convincingly as an effective and necessary behavioural solution to augment, rather than replace, the more fundamental structural social transformations required. Clearly, empowerment or ‘helping the person to help themselves cannot be achieved unless the contextual constraints placed on individual decision-making are removed. However, the question remains whether counselling as it is presently understood is an appropriate form of support and prevention in African settings. In this regard, other issues, such as the social acceptability of VCT, and of its explanatory paradigm and objectives, also remain to be addressed (Van Dyk, 1992, ID138).

Critiques of the Western Counselling Model The appropriacy of a western model of counselling in the African context hinges on the critique of what is essentially an individualistic rather than community paradigm, and a biomedical rather than indigenous paradigm for counselling. Such issues are invariably raised in the published literature – both on the level of counselling aims and objectives, and on the level of counselling practices (see Green, 1989, ID53; Lindegger and Wood, 1994, ID76; Mapekula, 1996, ID243; Jones, 1999. ID62; Mkhize, 2000, ID307; Richter et al, 2001, ID338) Some of the debates in this area are: •

The psychological paradigm within which VCT operates is devoid of the religious and spiritual conceptualisations underlying African interpretations of sickness and disease.

19

•

The individualistic orientation of Western psychology contradicts the community orientation of Africans, such that the goals of increased insight and self-actualisation could be culturally inappropriate.

•

The Western discourse of sexuality, its understanding of sexuality, sexual risk and sexual health do not concur with African cultural reality.

•

The western bio-medical model of and explanation for HIV/AIDS cannot be assimilated to the African model of sickness and disease.

•

The individualistic orientations of Western culture upon which the human rights principals of confidentiality and informed consent are based are wrongly assumed to be globally appropriate.

There is a marked discrepancy within this body of literature between the extent to which these issues are discussed and the extent to which they are directly addressed through empirical research studies. Several authors have remarked on the noticeable lack of local research to examine and identify culturally acceptable counselling models (cf. Richter et al, 2001, ID338). Likewise, very little formative research into the pro-active assimilation of traditional beliefs and attitudes into HIV/AIDS counselling, or the impact of these beliefs on counselling efficacy, was identified by this review. Overall, there is little empirical research that has systematically compared and contrasted the efficacy of western, as opposed to ‘indigenous’ models for prevention and support for HIV/AIDS. This is unfortunate in so far as such research may allow us to tease out the universalities of the work of HIV/AIDS counselling from those aspects that reflect particular cultures and client groups (Green, 1989, ID53). Despite the paucity of research, the call for the integration of traditional healers and traditional healing paradigms into HIV prevention and care activities, and counselling in particular, has been loud (Abdool Karim, 1993, ID163; Lindegger and Wood, 1995, ID76; Berger, Porter, Mekisini & Courtright, 1994, ID27). Some attempts have been made to elaborate these ‘alternative beliefs and practices’ (van Dyk, 1994, ID379; Mapekula, 1996, ID243; Burnett, Baggaley, McMillan, Sulwe, Hang'omba & Bennett, 1999, ID24). However, in practice there is little tolerance for the possible implications of such alternative paradigms on prevention and treatment goals. This is unsurprising given that such alternative beliefs and practices include, for example, the belief that babies a few weeks old can be protected from lust, promiscuity and rape by undergoing a ritual called ‘ukuguda’, whereby a string around the abdomen is pushed into the babies anus and pulled out by twisting of the two strands of the string (Abdool Karim, 1993, ID163). Mapekula, (1996, ID243) shows that counsellors remain aware of, but do not accept, the traditional approaches used by their clients to explain and treat their illness and that they judge these beliefs to be irresponsible and irrational and in need of correction rather than acknowledgement or integration. In particular, the belief many healers hold in their ability to cure HIV/AIDS is of great concern, especially to those who see it as a potential cause for the spread of the disease (Abdool-Karim, 1993, ID163). As a result, calls for the authentic integration of traditional healing paradigms remain, for the most part, empty rhetoric. It is argued that even the NGO and government programmes which aspire to be communitybased and culturally relevant, fall into the trap of adopting a medical model and individualistic paradigm and that counsellors become acculturated or ‘culturally encapsulated in the ideology of their training’ (Mkhize, 2000, ID307). Certainly, guidelines on ‘cross-cultural counselling’ often offer little in the way of advice on alternative approaches to VCT. Such guidelines advise measures such as the use of skilled interpreters, learning to greet appropriately, learning basic vocabulary and learning about 20

and respecting alternative beliefs and practices (South African Medical Journal, 1994, ID319). Similarly, attempts to utilize traditional healers as HIV/AIDS counsellors have tended to try to modify traditional beliefs through training in the bio-medical model (e.g. Berger et al, 1994, ID27). However, attempts to genuinely develop local models of counselling which challenge Western ones do exist, and more research into their efficacy is required (Rankin & Gilbert, 2000, ID262). The failure of Western biomedicine to integrate spirituality and faith into the healing and coping paradigm for terminal illnesses is not specific to HIV/AIDS. Medical anthropologists have shown that there are two inter-related, but separate, healing functions, i.e. control of sickness and the provision of meaning for individual’s experience of sickness. These authors contend that biomedicine’s emphasis on disease, as opposed to the sick person’s illness experience, often results in problems of clinical management. It is this deficit which is felt most keenly by those with HIV/AIDS (and other illnesses) who complement a bio-medical approach with more indigenous healing systems, which purport to treat both the physical and psycho-social causes and consequences of illness within a unitary conceptual paradigm. This is clear from the qualitative research conducted by Mapekula (199, ID243) with HIV positive patients. Mapekula shows that patients often use two or more forms of explanation, as well as treatment, for HIV/AIDS. While counselling attempts to address this, by helping the patient to develop a constructive approach to his or her illness experience, counselling and medical treatment remain discreet in the western bio-medical model - their parallel paths entrenched within the mind-body dualism of western consciousness.

Critiques of the Preventative Orientation of the Counselling Model The theoretical inconsistency and paucity of the HIV/AIDS counselling model developed by major international funders such as UNAIDS or WHO, has been criticised in a great deal of the international counselling literature (Balmer, 1992, ID267). More specifically, it is argued that counselling as defined by international health agencies for implementation in developing countries, and as defined by those specializing in the discipline of counselling, refers to two vastly different sets of activities. This disjunction stems from the fact that the WHO/GPA model of counselling is essentially a medical response to the epidemic, and is directed by a disease-centred approach and epidemiological conceptions of infection control which do not ‘readily activate or accommodate the practice and theory of counselling”(Balmer. 1992, ID267). Thus, there is a mismatch in the anticipated goals and outcomes of a counselling intervention in the context of AIDS. For example, some of the counselling objectives listed by Bor, Miller, Scher and Salt (1991, ID215) for HIV positive clients are: • • • •

To place the responsibility for problem-solving with those who define the problem (i.e. the client); To help the client find meaning, or a new understanding of what it is to have AIDS; To normalise the views, feelings, experiences of the client; To help clients to feel that they have choices.

However, in comparison, WHO/GPA have formulated the following aims for counselling: • •

The prevention of HIV infection; Psychosocial support for those people affected by the virus.

Thus, a critical difference between the GPA counselling model for HIV/AIDS and conventional facilitative counselling is that facilitative counselling tries to help people cope

21

with their problems as they see fit rather than trying to get people to modify their behaviour in accordance with pre-defined health promotion goals such as disclosure, condom-use and treatment compliance. A major outcome of the GPA model is to promote preventative behaviour, whereas a major outcome of a facilitative counselling model is to promote coping. This emphasis on prevention in HIV counselling has been criticised by a number of authors. For example, Balmer (1992, ID267) argues that counsellors have not prevented the spread of the virus, nor will they and that it is not good management to give counsellors an aim they cannot achieve. According to Balmer, the available evidence indicates that counselling changes self-concept and improves self-esteem, making such self-perception changes appropriate and realistic counselling aims. This emphasis on the preventative outcomes of VCT, can be seen both in the stated objectives for VCT and in practice. For example, in the training HIV/AIDS counsellors receive and how this translates into actual counselling practice. Qualitative research conducted by Stein, Allwood, Karstead and Brouard (1997, ID388) with nurse-counsellors at Baragwaneth Hospital in Soweto, South Africa, shows that in the absence of adequate counsellor training in the theory and methodology of facilitative counselling, the HIV/AIDS counselling goals of prevention and support are inevitably set at variance and counselling becomes a prescriptive advice-giving activity, albeit an individualized one tailored to the specific circumstances of each client. These authors point out that this is ironic given that counselling was introduced into HIV/AIDS care in the light of the now well-established failure of didactic or prescriptive health promotion techniques to achieve health promotion objectives. For this reason, some authors have argued that psychosocial counselling is best situated within the community rather than a hospital or medical setting. However, the medicalisation of VCT is not only determined by its relative proximity to health services. Thus, for example, community health nurses - and indeed, lay-counsellors - may also lack the skills or the personality required for effective facilitative counselling (Grinstead & van der Straten, 2000, ID330). Furthermore, where counselling is located, i.e. a community centre as opposed to a hospital or clinic does not make VCT into a “community” activity. VCT is still provided to individuals within a community, as opposed to the ‘community’ per se. Perhaps the most important distinction to be drawn between “Western” counselling and “African” counselling, is that specialist counsellors largely carry out the counselling of HIV positive individuals in the West. Clearly, developing countries with soaring infection rates cannot build counselling systems for HIV positive people based on specialist workers and primary health care staff must be trained to provide such counselling (Green, 1996, ID395). Most health workers presently providing VCT have minimal training or experience with a client-centred counselling approach (Moses & Plummer, 1994, ID401). Arguably, the integration of counselling is, not in and of itself, a bad thing. It allows for the large-scale and widespread integration of mental health services at a primary health care level. In addition, the integration of counselling training into the curriculum and skills base of medical practitioners can only be advantageous, given the complex and inter-connected relationship between mental and physical health, not to mention social well-being. It is this inter-connectivity, in fact, which has been neglected by the western medical model, to its detriment, and which gives traditional healing systems their potency. Paradoxically, if VCT and HIV counselling exists because the Western model of care is disaggregated into mental and physical health care, then VCT ends up picking up the flack for failures in the medical response to the disease, not only on the level of prevention, but also on the level of treatment and cure.

22

Bor, Miller & Perry (1988, ID214) argue that research into the area of VCT, and HIV/AIDS counselling more generally, has neglected a wealth of experience and resources from other areas of counselling. Psychologists in Africa should be rising to the theoretical and practical challenge of integrating HIV/AIDS counselling into primary health care, rather than taking a back-seat in policy formation, programme planning and AIDS research in South Africa with the result that the medical approach prevails. This is especially given that the task of meeting the bio-psychosocial needs of HIV positive people falls upon non-specialist health care workers.

3.4

Conclusions

As public awareness of HIV increases, so will public demand for VCT (Brown, Macintyre, Rutenberg & Hassig, 2001, ID352). According to Brown et al. (2001, ID352) Uganda is an example of an African country with relatively high demand for VCT in contrast to other countries. The imperative to provide testing, which is the entry-point for the provision of medical care and psychosocial support, cannot seriously be questioned. Apart from providing the opportunity to prolong life, VCT also potentially facilitates prevention and the destigmatisation of the disease. While there have been heated debates about almost every aspect of VCT, two questions dominate the research field: Firstly, there is the question of whether VCT is useful as a prevention strategy and secondly, the elucidation of factors preventing people from seeking VCT. While the implications of VCT for coping with HIV/AIDS have been addressed, it is considerably less well researched. In addition, many other issues, especially critical ethical issues, such as the implications of home test kits and rapid tests, the role of public policy and the medical establishment in promoting testing, the routine testing of specific groups such as pregnant women, the abuse of testing by insurance companies, etc, have not received the attention they deserve. Thus, consideration of the questions being asked and the methodologies used to generate answers in the literature shows that research agendas are seldom driven by theoretical or ethical questions, no matter how urgent, and that standard methodologies are duplicated to produce what are often less than useful results. Quantitative research has established, somewhat unsurprisingly, that VCT works to achieve prevention to varying extents across different contexts and populations. However, this research also suggests that, in the absence of adequate training and on-going support services, VCT may achieve neither its prevention nor its support goals. Additional quantitative research is needed to better understand and monitor the prevalence of HIV/AIDS in populations and to monitor the impact of interventions. However, qualitative research - still regarded as ‘soft’ research - is perhaps even more urgently required to analyse the motives underlying sexual behaviour, document the practice of VCT, and establish the preconditions for its efficacy as a preventative and supportive tool. Only in this way can VCT be improved and other necessary interventions developed. If VCT is to be a useful primary prevention strategy, considerably more information regarding how to make the service as effective as possible is required. At present, quantitative data dominates in the published literature. Whether this is because of its superior quality and usefulness needs to be both seriously questioned and addressed. The next decade of AIDS prevention and support programmes will have to locate VCT more specifically and more exactly, as part of a much broader strategy for dealing with HIV/AIDS. This broader strategy must involve securing a better and more equitable quality of life for Africans in general. More specifically, it must include improvements in housing, education, health, and gender equity. However, it must also include linking VCT to adequate care, support and welfare services.

23

The imperative to address gender inequality in the African context cannot be overemphasised. In the meantime, VCT promotion, objectives and practices in the African context should take cognisance of the potentially negative consequences of disclosure for women. There is ample evidence to suggest that, despite the fact that the socially sanctioned unsafe behaviour of men is the most likely cause of infections, women tend to be blamed for bringing HIV into the family – especially if they are tested first. The promotion of couple counselling in this regard requires specific investigation especially given that, at present women are more likely to get tested than their partners. Similarly, as the demand for VCT increases, the particular needs of children and youth in relation to VCT and HIV/AIDS counselling will have to be explored and addressed. There is practically no research at present that addresses this issue. Moreover, despite the plethora of educational initiatives targeting youth, VCT services for the youth remain under-developed (Brown et al, 2001, ID352). It seems possible that, like HIV/AIDS education, VCT may have a greater impact as a secondary prevention strategy with youth initiating their sexual careers, than with adults whose sexual behaviours are already entrenched. Finally, policy analysis research such as that described by Walt and Gibson (1994, ID147) regarding national and provincial strategies for VCT and counselling is also sorely absent. While prescriptions abound about what services should be introduced, little attention is given to the process of policy reforms required to do so. Much more research is required into the processes required to implement effective VCT strategies.

4.

Study 2: A Meta-analysis of Voluntary Counselling and Testing Literature in Africa

4.1

Aims

This meta-analytic study is aimed at providing some “analysis of analyses” (Glass, 19765). By this we mean that this study represents an attempt to bring some quantitative and statistical understanding to the large collection of literature on VCT collected for this study, such that some overall – or larger -understanding can be sought from these results. In general then, there are three main aims to this study, i.e. 1. To provide a quantitative analysis of the key types, areas, as well as any trends and shifts in research conducted on VCT in Africa. 2. To provide a critical understanding of key omissions and/or gaps in this research. 3. To provide a quantitative analysis of the quality of research on VCT in Africa through the development of a quality rating scale. In this way, this study does not follow the format of a more traditional meta-analysis, most of which are aimed at providing some integration to research findings such that some conclusions about the efficacy or effectiveness of specific interventions can be made. This study is not aimed at providing an idea of the efficacy of VCT as shown through the integration of the results of various randomised controlled trials. Instead, we have adopted the view of Hedges (1986, p. 3536) who calls for a broader definition of meta-analysis as any “literature review that makes explicit use of quantitative methods to express the results of 5

Glass, V. (1976). Primary, Secondary and Meta-Analysis of Research. Educational Researcher, 5(10), pp. 3-8. 6 Hedges, L.V. (1986). Issues in Meta-analysis. In E. Z. Rothkopf (Ed.). Review of Research in Education 13. Washington DC: American Educational Research Association.

24

studies or to combine those results across studies”. This broad conception of meta-analysis serves to clearly identify such activities as a variety of research review that involves quantitative methods. This would entirely fit with the primary goals of this research that aims at providing a quantitative profile of the research, both in terms of type and in terms of trend. Thus, it is anticipated that this analysis will both complement and extend the qualitative literature review presented above, as well providing some critical contribution to the development of a VCT research agenda in Africa.

4.2 Methods A total of 341 articles were submitted for analysis for this component of the literature study. This database of literature was comprised of all published research and discussion articles, as well as all theses and dissertations, unpublished research, organisational research and reports and conference papers. Three types of literature were excluded from analysis for a number of reasons: 1. All newspaper and other media articles were excluded on the basis that they could not be profitably analysed using this method. These articles did not usually report on findings or recommendations for VCT, rather they dealt with general understandings of VCT or reports on the current status of VCT. As they dealt with a different subject matter to the published and unpublished research and discussion documents, it was felt that the inclusion of this material would simply confuse findings. 2. All abstracts were also excluded from analysis. In some cases a full-length article could not be obtained - either because of where the journal was located or because the article was a conference presentation. In these cases where an abstract was available but not a full length article the abstract was not submitted for analysis on the grounds of lack of information. 3. Finally, some papers were not analysed because they were not located and/or databased in time for the analysis. This is particularly true of articles published after 2001. The literature was classified and coded along a number of dimensions including type of research, characteristics of the primary authors and characteristics of the research (sampling, data collection etc.). The complete coding schedule is provided as Appendix A of this document. For those articles that were classified as research articles (119 in total), a rating scale was also developed in order to provide some measure of the quality of research that exists on VCT. Quality was defined as “consistency with generally accepted standards of what it means to do good quality research”. In order to ascertain what these generally accepted standards were, a number of standard social science research methods texts were consulted and where issues of quality were discussed the proposed indexes of good quality research were recorded. Across these texts, four particular factors were said to constitute good quality research in that they might allow a researcher to judge the quality of research that he/she is consulting. These four areas were: 1. Overall coherence and appropriateness of the chosen research procedure in terms of the study aims 2. Validity and reliability of design and results. 3. Careful consideration of ethical issues 4. Clear reporting of all aspects of the research that would allow an independent observer to clearly decipher the entire research procedure. 25

Fourteen Likert-type questions were then designed to cover areas 1 to 3 above. The Likerttype scaling ranged from 1 to 4, where for each quality question the rater was asked to classify the research as being either poor (1), inadequate, adequate or good (4). The fourth quality area, i.e. the reporting of results was tested implicitly, i.e. for each question the rater had to judge whether the authors had provided enough information to allow for a quality rating. Where not enough information was provided a rating of 0 was given. Of the 14 questions designed, three questions tapped directly into the adequacy of research design (sampling, data collection procedures etc) and 4 questions looked at the larger research process in terms of coherency or the way that the elements of the research hang together in relation to the research aims. In addition, 2 questions focused directly on quantitative analysis, 2 on qualitative analysis, 1 question on the credibility of the results and finally 1 question on ethical issues7. The scores on each of these questions were then summed to provide an index of quality. However, the way the summated score was obtained differed depending on the type of research that had been conducted, i.e. a qualitative design, quantitative design or a mixed design. As stated above, the scale contained 2 questions that related directly to quantitative research and 2 that related to qualitative research alone. For each of these designs then, only those scores relating directly to the respective research design were summed. Therefore, if the research was either qualitative or quantitative, only 12 of the 14 variables were summed to provide a score. In the case of qualitative research designs, the questions relating to quantitative analysis were dropped and vice versa. In the case of the mixed designs, on the other hand, as answers were possible for all 14 questions, all 14 variables were summed to provide a quality score. In order to complete both the classification of the literature and the quality rating of the research literature, two independent researchers were employed. These two researchers had both training and experience in social science research methods. Once the researchers completed the classification and rating of the literature, the data was entered into SPSS™ and quantitatively analysed. For the classification variables, this analysis mainly entailed the production of percentages and proportions on a variable-byvariable basis. These results are discussed below. The quality rating scale, however, was also subjected to some scale analysis in order to establish its reliability and validity. This analysis is discussed below in the section on research quality.

4.3

Findings

4.3.1 Publication Profile The literature collected for the study spanned the period 1980 to 2001. As can be seen in Figure 1 below, there has been a steady increase in articles on the subject of HIV/AIDS Voluntary Counselling and Testing across this time period - from the one article published in 1980 to around 33 articles published in 2001. The most popular year for research and/or discussion articles in this area was 1994 where 40 articles were written about VCT in Africa.

7

The complete rating scale is provided as Appendix B

26

Figure 1: Distribution of Articles across Year

50

40

30

20

Frequency 10

0 1980

1985

1983

1988

1986

1990

1989

1992

1991

1994

1993

1996

1995

1998

1997

2000

1999

2001

Year of publication/reporting

The literature was classified into different types of articles, i.e. research articles, discussion or position articles, policy documents, best practice guidelines, information documents or review articles. In this case, a research article was defined as any article that reported on an original research study. Therefore, articles that reviewed other research only without any research contribution of their own were not included as research articles. Rather, they were grouped along with meta-analyses and historical analyses as review articles. Discussion documents or position papers refers to any article that provides critical discussion about VCT in Africa in terms of either definitions, types or models etc. that are appropriate to this context. The category of policy documents includes both policy articles and articles which provide an analysis of or discussion of this policy. Finally, information documents are those documents which simply provide information about VCT e.g. pamphlets etc and the best practice guidelines refers to any article which provides a “how to” for VCT or the provision of VCT services. The distribution across these categories is shown in Figure 2 below. What this distribution shows is that most of the articles about VCT take the form of a discussion or position paper. In fact, 43% of all articles written from 1980 to 2001 were articles of this form. Research papers formed the second largest category with 32%, followed by review articles with 11% and information documents with 8%. Finally, best practice guidelines made up 5% and policy articles 1% of the total number of articles on VCT. Figure 2: Distribution across type of publication

Review article Policy Best practice guidelines Research Information document

Discussion/Position

27

This distribution highlights a number of gaps that exist in the body of literature about VCT in Africa. For one thing, there is substantially less research about VCT in Africa than discussion about VCT. Also, although there has been some discussion in this literature about the best way to go about doing VCT (best practice guidelines), there has been very little analyses or discussion of larger policy issues. Analysis showed that the majority of this literature is available in peer-reviewed journals (nearly 85%), meaning that a lot of the literature about VCT is fairly available to those interested in VCT. However, the difficulty of locating and accessing unpublished data should be kept in mind as the distribution of published to unpublished could be severely skewed by our sampling methods, or just the unavailability of the unpublished material.

4.3.2 Author Characteristics In the classification scheme, we tried to provide some kind of profiling of the primary authors of the papers in terms of various characteristics, e.g. gender, institutional affiliation etc. Unfortunately this classification did not prove to be particularly easy as this kind of information is not routinely available across all types of literature. Thus, a category of “not enough information to tell” was introduced into the analysis. This category makes any conclusions based on this data tentative at best, as it could be that the information we are missing could alter the relationships in the data entirely. This is particularly true of gender as gender could not be determined in 154 cases. For this reason, no data is presented on this variable. The results that are available about the primary authors show that most authors were university affiliated (34%) as opposed to being affiliated to non-governmental organisations (14%), research organisations (7%), government (8%), international bodies (2%) or the private sector (2%). Figure 3: Institutional Primary Author

Affiliation

of

Private sector Not enough information International Body NGO

Government

Research organisation

University institute

or

other

tertiary

When we analyse the country of affiliation of the primary author, there is a wide distribution across a number of countries, including both developing and developed nations. The majority of authors on VCT in Africa are, however from South Africa (42%), followed by the USA (12%), the United Kingdom (8%), Kenya (3%), Zimbabwe (2%) and Zambia (2%). 22% of the time the country affiliation of the author could not be classified. The final 9% is comprised of authors from Australia, Burkino Faso, Burundi, Canada, France, Germany, India, Malawi, Namibia, The Netherlands, Switzerland, Tanzania, Trinidad, Uganda and Zaire.

28

4.3.3 Research Literature As discussed above, only 33% of the total literature on VCT in Africa is comprised of original research articles. This makes a total of 109 research studies spanning the period of 19982001. Although there is a similar steady increase in research output across these years, the year distribution is slightly different when we consider research articles alone. For example, as seen in figure 4 below, there has been a substantial rise in research publications in the area of VCT across the years 2000 and 2001. In fact publications in this period account for 32% of the total research output on VCT in Africa. Figure 4: Distribution of research literature across year 20

10

Frequency

0 1988

1990 1989

1992 1991

1994 1993

1996 1995

1998 1997

2000 1999

2001

Year of publication

Research Focus In order to provides some understanding of the main areas that VCT researchers have focused on, the studies under review were classified as being either a VCT impact or efficacy study, a PMTCT study, a Meta-analytic review, a study which concentrated on research methods or tools testing, Ethics studies and finally an other category was also provided for any study which did not fit this classification scheme. In addition, a secondary classification scheme was developed for these studies on the basis of a content analysis of the raters’ qualitative descriptions of the main focus area of these studies. This classification scheme was also used to provide a classification for those studies that fell into the other category. Figure 5: Research Focus Areas 70

60

50

40

30

Frequency

20

10 0 VCT efficacy/impact

Meta-analysis

PMTCT study

Ethics

Research methods and

Other

Research Focus

29

As can be seen in figure 5, most of the research on VCT fell into the category of VCT efficacy or impact studies and/or PMTCT efficacy/impact studies. Twenty-four studies were classified as concentrating on the efficacy or impact of VCT. In the secondary classification of these studies, it was found that within this category, research looked at efficacy either in terms of sexual behavior change indicators (sero-conversion rates or condom use), or in terms of other outcomes such as disclosure of results in couple counselling. Of these a significant majority focuses on secondary prevention outcomes rather than primary prevention outcomes associated with the first outcome behavior change. 13 of the 24 studies efficacy studies fell into this category. 4 of the 24 studies looked at disclosure of HIV status to partners. Finally, 2 studies were focused primary on the efficacy of VCT in terms of service delivery factors. Only, 1 study was focused on VCT in terms of support goals, 1 study on the profiling of those clients who either came or did not come for VCT, 1 study on profiling attitudes to VCT and lastly 1 study on the cost-effectiveness of VCT. The 5 meta-analytic studies can also be classified here as efficacy studies, as they were concerned with establishing the efficacy of VCT in producing preventative outcomes across a number of studies. In the 7 studies classified as being PMTCT studies, a primary focus was on profiling the attitudes of pregnant women towards VCT – most in terms of gauging the acceptability of such services. Other focal areas in this category were: the efficacy of PMTCT for producing behaviour change in pregnant women (1), profiling those clients who either did or did not use VCT services (1) and finally 1 study in this category focused specifically on barriers to behaviour change in pregnant women. The 7 studies that focused on ethical issues in VCT, were primarily concerned with issues of informed consent and confidentiality, both in terms of VCT service delivery (4) and in terms of VCT research (1). However, another primary area of ethical interest for researchers was the ethics of promoting disclosure in the face of risk factors to women (2). Finally, a few of the studies under review were primary concerned with methods or tools testing for research about VCT. In figure 6 in can be seen that the majority of studies that were classified did not fit the primary classification scheme developed for this study. In fact, 58 of these studies were classified by the raters as falling into the “other” category. For this reason, further analysis of this category was needed. When this category was further analysed, the following distribution was found: Figure 6: Secondary Classification of Other Research 30

20

Frequency

10

0 Counselling Methods

Client Characteristi

Barriers

Service Delivery

Attitudes

Media Analysis

Secondary Focus

30

The majority of studies that were classified as “other” were focused on investigating either counselling methods or counselling practices. In fact 27 studies fell into this category making it the primary focus for research on VCT across all studies reviewed in this literature study. A secondary focus was the quality of service delivery (13). Other studies were mainly concerned with either profiling clients (6), profiling attitudes towards VCT (7), looking at barriers to sexual behaviour change (2) and finally two studies were concerned with analysing media responses to VCT.

Research Design Of the research that has been done on VCT in Africa, the majority (62%) is quantitative in nature. In contrast, qualitative research articles account for only 18% of the total research output, while the final 19% is made up of articles which employed both quantitative and qualitative research methods, i.e. so-called mixed methods research. In line with this preference for quantitative research, the most popular design for research was survey (62%), followed by cohort studies (13%) and then by case studies (8%) and observational studies (6%). Figure 5 below shows this distribution. Figure 7: Research Design 70 60 50 40

Frequency

30 20 10 0 Case study

Observational Survey

Case Control

Randomised control

Other

Cohort

Design

In some cases, the research could not be rated as one of the categories provided, in which case a rating of “other” was given. This did not occur often (in only 5% of cases). However, when a rating of other was given the following categories of research were found: metaanalytic studies, secondary sources studies, media analysis, and economic modelling and evaluation research. When we look at the quantitative research alone, we see approximately the same distribution as above. Once again this is because of the strong preference for quantitative research designs. Most of the research is survey in nature, followed by cohort designs. This was then followed by 2 studies that fell into the category of randomised control trials. Figure 8: Quantitative Research Designs Frequency Valid

Percent

Valid Percent

44

67.7

67.7

Cumulative Percent 67.7

Case study

1

1.5

1.5

69.2

Randomised control trial

2

3.1

3.1

72.3 73.8

Survey

Case Control

1

1.5

1.5

Cohort

12

18.5

18.5

92.3

Other

5

7.7

7.7

100.0

Total

65

100.0

100.0

31

Looking at figure 9 below, this distribution is not reflected across qualitative research studies. While the majority of the research is also survey in design, case studies and observational types of designs are also substantially represented. In terms of the large representation of qualitative studies in the “Other” category, a number of qualitative studies used secondary source material and were therefore categorised as being secondary source analyses. Figure 9: Qualitative research designs Frequency Valid

Percent

Valid Percent

Cumulative

Observational

5

20.8

20.8

20.8

Survey

9

37.5

37.5

58.3

Case study

5

20.8

20.8

79.2 100.0

Other

5

20.8

20.8

Total

24

100.0

100.0

In terms of those studies which were categorised as “mixed designs, the following distribution was seen. Figure 10: Mixed research designs

Valid

Observational

Frequency 1

Percent 4.5

Valid Percent 4.5

Cumulative Percent 4.5

Survey

9

40.9

40.9

45.5

Case study

2

9.1

9.1

54.5

Randomised control trial

2

9.1

9.1

63.6

Case Control

2

9.1

9.1

72.7

Cohort

1

4.5

4.5

77.3

Other

5

22.7

22.7

100.0

Total

22

100.0

100.0

Again, surveys make up a substantial proportion of the total research output in this area, as do case studies, randomised control trials and case control studies. In this instance, the large number of research studies categorised as “other” were made up of studies classified as evaluation research. An example of a study which would fall into this category would be research which is concerned with evaluating a local VCT service provider in order to illuminate a point about VCT, or to provide best practice guidelines for VCT.

Populations and Sampling In terms of the populations that were chosen for these studies, the following profile was found:

32

Figure 11: Population Profile 40

30

Frequency

20

10

0 ts en ud st ity rs ve ni i U at rm fo er in th O gh ou en ot N es ut tit n os tio Pr la pu po ay G rs se U en g om ru w D cl nt s/ na nt ie eg at Pr n lP ta tio pi la os pu H po al er en G

Population Group

Most of the studies under review defined either the general population, the population of hospital/clinic patients, or pregnant women as their population of interest. A small number of studies focused on other populations of interest such as prostitutes, the gay population, university students or drug users. Again, in a substantial number of cases not enough information was provided to classify the main population of interest. In a number of cases, the population was classified as other. In these cases, where other was selected, some of the populations that were included in research were health care workers, People living with AIDS (PWLA’s), employees (e.g. transport workers), or youth. Health care workers made up the majority of this category, and youth formed the smallest proportion. In fact, it should be noted that of all the research rated for this study, only 2 studies were found which focused directly on youth. In terms of gender, most of the research included both male and female respondents. This category formed 57% of the total number of studies rated. A substantial number of studies also used female only samples (20%). In contrast, however, there were very few studies that focused directly on the male population. In fact only 5 studies of this type were found, which accounts for 4% of the total sample. When the gender distribution was compared to the population groups used, it was found that those studies were both men and women were used were studies involving the general population. Similarly, those studies that involved women alone were either studies of hospital patients or clinic attendees, prostitutes or studies of pregnant women. Those studies that involved only men were either studies that involved hospital patients, or studies where the focus was on employees. No study on the general male population or the general female population alone was found.

33

Figure 12: Population Gender 70

60

50

40

30

Frequency

20

10 0 Male

Both male and female Female

Not enough informati

Population gender

It should be noted that in a substantial number of cases the gender of the population could not be classified due to a lack of information. It could be surmised that these could be cases where both men and women were sampled and therefore neither was specified, however there is no way to substantiate this claim. In most cases where the gender could not be specified, neither could the population of interest meaning that we could not determine if these studies were drawing people from the general population or some other population of interest. In terms of what populations have been selected for study, i.e. the locations of the studies, it was found that the majority of the studies on VCT in Africa have taken place in Southern African countries. As can be seen in figure 13 below, South Africa, Zambia, Zimbabwe, Botswana and Malawi account for almost half of all studies done in Africa with South Africa being the most popular location for research of this type. Figure 13: Location of Research Place of Study South Africa Zambia Zimbabwe Malawi Botswana Uganda Kenya Rwanda Tanzania Democratic Republic of Congo Ethiopia Ghana The Gambia Nigeria Burkino Faso Trinidad Multi-site Not enough information Total

Frequency 29 4 10 1 1 8 6 5 4 2 1 1 1 1 1 1 15 15 106

Percentage 27.4 3.8 9.4 .9 .9 7.5 5.7 4.7 3.7 1.9 .9 .9 .9 .9 .9 .9 14.2 14.2 100.0

34

Central East African Countries – Uganda, Kenya, Rwanda, Tanzania and Ethiopia account for 24 studies overall and the Democratic Republic of Congo a Central African country for 2. Very few of the studies took place in Northern African countries – only four overall. Moreover, where studies were conducted they concentrated mainly on West African countries such as Ghana, The Gambia, Burkino Faso and Nigeria. There were 15 studies across the literature that used a number of different countries in their research and these were then classified as being multi-site. This category also accounts for all meta-analysis as these studies do compare across a number of studies conducted in a number of different locations. These meta-analytic studies account for 5 of the multi-site ratings. The rest of the studies classified as multi-site appear to be comparison studies of some kind where either the efficacy and/or the practice of VCT are compared across different countries. For example, the Voluntary Counselling and Testing Efficacy Studies fall into this category as they conducted a number of studies looking at the efficacy of VCT in Kenya, Tanzania and Trinidad. It should also be noted that in 15 cases not enough information was provided about the location of the studies to classify them. Again it is hard to know what to say about these studies, except that they are part of a general lack of description about the populations chosen for the research. When the sampling strategies across the studies were analysed, the following results were found: Figure 14: Sampling Strategies 40

30

Frequency

20

10

0

r fo er in th O gh ou en ot ce N en ni ve on C y ar pl

en

d

ve si e po tiv ta

s re

er

em Ex

r Pu

st

m

ie

do

tif

ep R

lu C

an

ra St

R

m i at

Sampling Strategy

In this distribution, it would appear that most of the studies analysed for this review did not provide enough information for their sampling strategy to be rated. However, for those studies where the sampling strategy was described, most of the studies were rated as using convenience type sampling methods. In fact, convenience sampling accounted for 26% of all sampling strategies chosen across the studies. This convenience sampling tended to take the form of self-selecting volunteerism. In terms of other sampling strategies used, random sampling accounted for 11% of the total, purposive sampling for 9.5%, representative for 7.5% and exemplary for 4%. There are obvious differences across different types of research designs as well. For example the quantitative research accounts for the random, stratified and representative

35

types of sampling, while the qualitative research accounts for the purposive and exemplary sampling. The majority of the studies that used convenience sampling were quantitative, however since there were more quantitative studies overall the significance of this cannot be commented on. A small proportion of the studies were also rated as falling into the other category. Studies that fell into this category were studies that used more than one type of sampling (e.g. random and stratified), or studies that used snowball sampling. The sample sizes chosen tended to have a fairly wide range with a minimum of 3 people to a maximum of 21 900 for a large scale, multi-site survey. The overall mean was 1239 people, however due to the wide variability (a standard deviation of 2808) it is hard to determine the meaningfulness of this figure. Overall, nothing particularly startling emerged from the analysis of the sample size. It would appear that qualitative studies used less people than quantitative studies and mixed studies chose a sample size somewhere in between.

Data Collection Methods In line with the preference for quantitative survey type research, the majority of studies reviewed for this research used questionnaires as their primary method of data collection. Interviews were the second most popular method, followed by observations and then focus groups. In one study psychological scales or measures were used. Interestingly enough, in a high proportion of cases for this variable, raters were not able to classify the primary data collection method used in the research due to a lack of information within the publication as to the particular method followed. Figure 15: Data Collection Methods

Valid

Questionnaires Observation Sheets

Frequency 36

Percent 34.0

Valid Percent 34.0

Cumulative Percent 34.0

4

3.8

3.8

37.7

27

25.5

25.5

63.2

Focus Groups

4

3.8

3.8

67.0

Psychological Scales/measures

1

.9

.9

67.9

21

19.8

19.8

87.7 100.0

Interviews

Not enough Information Other

13

12.3

12.3

Total

106

100.0

100.0

As would seem evident, questionnaires were more popular in quantitative survey research and the qualitative studies tended to use interview methods accounting for the high proportions in each of these two categories. However, interestingly there were a number of quantitative studies that did use interviewing as a primary method of data collection. It could be hypothesized that such interviewing was either used as a survey method the results of which were then quantified, or that interviews were used in the formulation of the questionnaire. Unfortunately the analysis presented here does not allow for confirmation of either hypothesis. Mixed method studies tended to use both of these methods however, questionnaires still formed a primary method of data collection, followed by interviews. Mixed method studies also account for the majority of the observational methods used while qualitative studies accounted for the majority of the focus group methods. Another interesting finding in this section was that on the whole those studies that did not provide enough information to classify the methods used were quantitative. In fact, 16 of the 21 cases were quantitative studies.

36

For those studies that fell into the “other” category, the primary data collection method used was the collection of medical samples. The collection of reports or case material also account for a substantial proportion of this category. Other data collection methods used, were the collection of self-reported usage rates (e.g. condom usage rates) and participant observational methods. The collection of medical samples was found only in quantitative or mixed methods types of research, while the collection of data through participant observation or the use of case reports was found exclusively in qualitative studies.

Data Analysis In order to provide an understanding of the types of analyses used in the studies under review, three questions were asked of the quantitative data. Firstly, raters were asked to record measures of reliability and validity, secondly they were asked to record instances of assumptions testing, and then they were asked to classify the primary procedures used in the research. Some of these classification schemes worked better than others. For example, the issue of validity and reliability was seldom dealt with in the published articles reviewed. In fact, the reliability and/or validity of the instruments and measures used in research were reported on only twice. Thus, while we can say something about the way that research is reported we can say nothing about the general reliability or validity of research measures used. A similar issue was found in the assumptions testing. Across all the studies rated for this literature review there was not one instance where any researcher reported on the assumptions testing for his or her analytic procedures. Thus, again we cannot say anything about whether researchers did or did not test their assumptions before performing analytic procedures. In some respects this lack of assumptions testing information is not surprising given that the majority of studies analysed their research using descriptives such as percentages and proportions for which assumptions testing need not be done. Figure 16: Quantitative Procedures Used 40

30

Frequency

20

10

0 Hypothesis testing

Categorical analysis

Descriptives

Correlations/regress

Multivariate statist

Not enough informati

Quantitative procedures used

However, the same cannot be said for the correlational analysis, or the hypothesis testing – both of which were used in a substantial number of studies. Few studies reviewed went beyond simple hypothesis testing or categorical analysis to the use of more advanced statistical tests or analytic procedures.

37

In terms of the qualitative analysis done, the raters were merely asked to classify the research in terms of the type of qualitative analysis used. The distribution across these categories is seen in figure 17 below: Figure 17: Type of Qualitative Procedures Used 20

Frequency

10

0 un ro

d de

y ph ra og hn s si Et y al an i nt at te rm on fo C in gh ou a en qu ot ric N ne ge al rm s fo si ly In a an ic i at ys em al an Th n io at rs ve on ry eo th

C

G

Type of qualitative analysis used

The majority of studies that employed some kind of qualitative analysis (i.e. either qualitative studies or mixed method studies) used an informal generic qualitative analytic method that has no explicit theoretical and/or methodological backing. In fact these studies accounted for 53% of all qualitative studies. Other types of qualitative analysis used are: thematic analysis (12%), content analysis (9%), conversation analysis (6%), grounded theory (3%) and ethnography (3%). In 14% of the cases, the type of qualitative analysis used was either not specified or not enough description of the analysis was provided to allow the raters to classify it.

Ethical Issues The discussion of ethical issues was generally poor across the studies. Only in 43 research articles could any discussion of these issues be rated at all and only in 17 studies could more than one ethical issues discussed be seen. Similarly, only in 10 cases were more than two ethical issues discussed. Within those articles where ethical issues were reviewed however, the following issues were discussed: Figure 18: Ethical Issues Discussed 20

Frequency

10

0 Informed consent

Debriefing

Confidentiality/anon

Ethical Issues

Reporting of results

Other

38

Informed consent and the issue of confidentiality and anonymity represent the most discussed ethical issues across all the research articles. Other issues discussed to a lesser degree were confidentiality in the reporting of results and/or issues around the debriefing of subjects. In terms of those studies rated as falling into the other category, those tended to be research articles that dealt directly with ethical issues that are particular to HIV/AIDS. Examples include issues of stigmatisation or discrimination, mandatory HIV testing, the reporting of HIV/AIDS test results, participant/researcher relationships (especially when health care workers are the researchers) and finally the risk of participating in the research.

4.3.4 Research Quality The research quality scale was used in this study, to give some index of the quality of research that has been conducted on VCT in Africa. The original scale developed for this study was a 5-point Likert-type scale where the raters were asked to give an evaluation of a research article along four dimensions. However, when inter-rater reliability indices were computed on this scale, they showed a very low level of inter-rater agreement. Further analysis revealed that this disagreement between raters was not in terms of direction, but rather in terms of degree. That is, when one rater gave a negative review for an article along some dimension then so did the other, however, there was some disagreement whether the rating should be poor or inadequate. After consultation with the raters, this was seen to be a limitation of the scale design as there was not enough differentiation between the terms “poor” and ‘inadequate” and similarly between the terms “good” and “adequate”. For this reason, the original 5-point scale was recoded into a 3-point scale with the points good, poor and not enough information to tell. After this recode, the inter-rater reliability for across the three scales was 0.416 for the rating of mixed method research types, 0.56 for quantitative method research and 0.661 for qualitative research. This indicates a moderate to high level of agreement between researchers on their rating of the scale. However, this agreement is not very high, a fact which does impact on the proposed reliability of the scale. It could be surmised that this moderate level of agreement across the raters may have something to do with how the raters were trained to use the scale or with the fact that quality ratings is a somewhat qualitative and subjective thing and therefore a high level of agreement could not be expected. Both of these possibilities should be explored in any other research that may use this scale, and indeed should be kept in mind when reading the results. High levels of internal consistency were found across the scales of both raters (and alpha of 0.926 for rater 1 and 0.931 for rater 2), indicating a high degree of consistency in the way that each individual rater rated the quality of the research.

Quantitative Research Quality The quantitative research quality scale is a 12-item scale, which was coded in such a way that a positive score indicates better quality and a lower or negative score indicates a rating of poor quality research. The maximum possible score on this scale is 12 and the minimum is –12, where 12 is the highest possible quality rating and –12 the lowest. It should be mentioned at the outset, that there were a number of papers where too little information was provided in order to rate them along the dimensions of research specified by the scale. In these cases, a score of zero was given for that dimension of quality for this paper. This issue will be discussed in more detail later.

39

The mean quality rating given to the quantitative research was 6.4 and the distribution was as follows: Figure 19: Quantitative Research Quality 10

8

6

Frequency

4

2

0 .00

2.00 1.00

4.00 3.00

6.00 5.00

8.00 7.00

10.00 9.00

12.00 11.00

Quality rating

What this figure shows, is an overall positive rating across the quality dimensions, for those research papers that used quantitative research strategies. No papers were evaluated negatively and in most cases a moderate to high level of quality was attributed to this particular area of research. In addition to computing an overall score for quality, each individual dimension of quality measured was also analysed. For these dimensions, the four questions that looked at the overall coherency and appropriateness of the research were summed (question 1,2, 12 and 14), as well as the four questions that looked directly at research design issues (questions 3,4,5 and 6). The three questions that focussed on the validity and reliability of the analysis procedures (questions 7,8 and 9) were also summed, and finally, the 1 question about ethics was simply entered into analysis on its own. The results from this analysis show high quality ratings across the dimensions of coherency and appropriateness and in terms of the design issues. Slightly lower (moderate) quality ratings were found for the validity and reliability of data analysis procedures. However, very poor levels of quality were found for the discussion of ethical issues in this body of research. For this question, only 24% of the papers rated received a rating of good. For all the other papers, they were either rated as poor or that there was not enough information to tell in the paper itself. The final dimension of quality that was measured implicitly throughout the scale was the adequate reporting of results. This did prove to be a problem across the research papers rated, as there was quite often not enough information to judge the appropriateness of the design, the coherency of the research process or the validity of the research findings. For the quantitative research papers, three problem areas of reporting were found – the reporting of design issues (e.g. the choice of subjects or the sampling procedures used), the reporting of reliability and validity statistics for scales or measures used and the discussion of the quantitative procedures used in analysis and finally the discussion of ethical issues in their research. For example, across the studies, the sampling procedure could not be rated in nearly 43% of the time, as this issue was not discussed in any detail in the paper (if at all). Similarly, the appropriateness of the statistical procedures used could not be rated around

40

40% of the time. This is a problem as it is this information that allows others to judge the appropriateness of the research procedures used in any study and therefore the credibility of the research results.

Qualitative Research Quality The qualitative research quality scale is the same scale as the quantitative research quality scale (a 12-item scale) except that the two questions on quantitative analysis were substituted for 2 questions on qualitative analysis. Figure 20: Qualitative Research Quality 3.5

3.0

2.5

2.0

1.5

Frequency

1.0

.5 0.0 -6.00

.00 -1.00

2.00 1.00

5.00 4.00

8.00 6.00

10.00 9.00

12.00 11.00

Qualitative quality ratings

Overall, the qualitative research has a slightly lower mean quality rating than the quantitative research. The mean in this case being 5.78. Again there most of the research was given a positive rating. However there were two papers that were negatively rating indicating a poor level of quality for these papers. Again rating scores were obtained for the four dimensions of quality. These four dimensions were computed in the same way as those for the quantitative research, except that questions 9 and 10 were substituted for questions 7 and 8, as it is these questions that pertain directly to analysis. When we examine these four dimensions, the results are much the same as those for the quantitative research. High quality ratings were obtained along the dimensions of coherency and appropriateness and in terms of the design issues. Also, slightly lower quality ratings were found for the validity and reliability of data analysis procedures. Again, poor levels of quality were found for the discussion of ethical issues as only 32% of papers received a rating of good along this dimension. The problems associated with good reporting, were also found for this body of research. However, unlike the quantitative research, there did not seem to be any problems with the reporting of design issues in this body of literature. The reporting of sampling and data collection procedures was detailed enough to be rated. However, problems were found with the discussion of the data analysis procedures. For example, the type of qualitative analysis used in the study could not be rated nearly 22% of the time and the analysis itself could not

41

be rated nearly 18% of the time. However, overall the reporting of the qualitative studies was better than that of the quantitative studies.

Mixed Research Designs Unlike the quantitative and qualitative quality scales, the quality scale for those studies that employed a mixed design is a 14-item scale. This is because it includes the questions on analysis from both the quantitative scale and the qualitative scale (meaning 2 extra questions are included). The maximum possible score on this scale is 14 (indicating a high quality rating) and the minimum is –14 (indicating a low quality rating). The following distribution was found for this body of research: Figure 21: Mixed Method Research Quality Rating 5

4

3

Frequency

2

1

0 -3.00

.00 -1.00

3.00 2.00

7.00 6.00

10.00 8.00

14.00 12.00

Mixed Research Quality Ratings

Overall, the mean quality rating for this body of literature is 5.3, indicating a lower rating of quality than either the quantitative or qualitative research designs. This is reflected in the distribution in Figure 21, which shows a number of papers with negative ratings or low quality ratings. Looking at the four dimensions of quality, we see low quality ratings for the overall coherence of the research and for both the quantitative and qualitative analysis dimensions. However, slightly higher ratings were found for the design dimension. Again very poor ratings of quality were found for the question on ethical issues, as papers were rated as poor on their discussion of ethical issues nearly 48% of the time. Again the reporting of these issues in research was found to be a problem, especially with regards to the discussion of analysis issues. For example, the quantitative procedures used in these studies were not discussed in 95% of the papers. Similarly, the qualitative procedures were not discussed nearly 43% of the time. Ethics was another problem area, as these discussions could not be rated nearly 34% of the time.

42

4.3.5 Conclusions It can be seen from the analysis presented above, that there has been a steadily increasing interest in voluntary counselling and testing in Africa over the last 30 years or so. It is also clear that this interest has taken on a particular profile and there are a number of immediately identifiable gaps in the body of literature on VCT. •

Much of the literature on VCT in Africa takes the form of discussion or position papers about the nature of counselling, the nature of testing and/or the role of voluntary counselling and testing in South Africa or Africa more generally. Therefore, an important finding of this research is that the majority of the available published literature on VCT in Africa does not comprise of research articles but rather of a combination of critical discussion, information, policy and review articles. Overall, there is a lack of research in Africa that focuses specifically on VCT. This lack of research constrains what can be said about the development and implementation of VCT services nationally

•

Within the research that does exist, although VCT efficacy is defined in terms of both primary and secondary prevention goals a greater emphasis is placed on the secondary prevention efficacy outcomes. More African research is needed to demonstrate the primary prevention efficacy of VCT – although this is not to say that secondary prevention goals are not important.

•

The identification of the role of VCT as preventative has largely led to a lack of research that looks at other goals/roles of VCT – e.g. VCT as support or the impact of VCT on longer-term coping. Where such research does exist, it is primarily concerned with the effects of disclosure on the lives of HIV positive people and does not directly address issues of support or coping.

•

Similarly, this outcome focused research means there is a lack of research that is process orientated – i.e. more qualitative research on the process of counselling or the decision to test. This process-orientated research is what can be used to influence policy and to inform counselling practice.

•

There are also number of biases and methodological limitations in VCT research in Africa: • Most research in Africa is quantitative survey type research that is reliant on selfreport questionnaires or some kind of medical data (e.g. sero-conversion rates). This reliance on self-report data is problematic in efficacy studies as they can be biased by both recall and perceived social desirability –especially when questioning about sexual behaviour. • Most of these studies also use sampling methods that are either convenience or selfselection. This self-selection bias is problematic as those who volunteer for VCT (or for the study) may be more sexually responsible than those who have not. More rigorous sampling methods need to be employed in these studies. • There is a distinct gender bias in VCT research where females are targeted as pregnant women or mothers and males are not targeted at all. • There is a geographical bias towards research being located in either South Africa or Southern African regions. Indeed where other African countries are focused on, it is usually in a multi-site study. Finally, something must be said about the quality of research done on VCT in Africa. Overall, the research was rated as being at a moderate to high level of quality. However, several problem areas were identified in this literature. The two main ones being: gaps in the reporting of analysis and results and the poor ratings given to the discussion of ethical issues across all three types of research papers. These two issues are important, as the discussion of analytic procedures allows researchers to gauge the credibility of the research and the

43

discussion of ethical issues, some of the limitations of the findings. It would appear that ethics are even problematic when they are discussed, as most researchers seem to concentrate only on informed consent processes and confidentiality or anonymity of results. I would argue that ethics play an important part in the voluntary counselling and testing process and that better reporting and further research in this area should be encouraged.

5.

Study 3: A Qualitative Analysis Of The Unpublished (‘Grey’) Voluntary Counselling And Testing Literature In Africa

5.1 Aims This qualitative analysis was undertaken in recognition of the fact that while there is an extensive body of published literature about VCT in Africa, a substantial proportion of the writings about VCT fall into the category of unpublished or “grey” literature. This body of literature includes: • • • • • • • • •

Newspapers Conference workshops and proceedings Pamphlets (produced by NGO’s and Government) National and Provincial policy documents and strategic plans Government Reports Organisational research reports Best practice guidelines Research proposals Theses and dissertation

While some of this literature is in fact published, in that it is available in the public domain (e.g. newspapers), for the purposes of this review all of these types of literature are regarded as unpublished. This is because, this literature has not gone through any kind of peer-review process and is generally not widely available to all in the manner that a published journal article would be. Also, these types of literature have not been included in more traditional reviews of literature, leaving a gap in our understanding of VCT in Africa. The analysis of this literature is seen to be important to this study for a number of reasons, i.e. 1. Firstly, the analysis of the unpublished literature is needed to provide a more complete picture of the current status of VCT and VCT research in Africa. To ignore these works would simply bias our understanding towards those conclusions and recommendations found in peer-reviewed journals. Thus, this analysis is aimed at providing an overview of the facilitative factors for VCT and the barriers that prevent effective function of this system.

2. Secondly, the analysis of this unpublished literature can provide an understanding of the way VCT is defined and understood in more practical or “everyday” types of writing. These types of writings reflect both the understanding of those who create and implement VCT systems and those who utilise it. Policy documents, pamphlets, newspaper articles, adverts etc all provide an indication of how VCT stakeholders define the role and function of VCT, while at the same time this is also the medium through which the general public are reached and made aware of the existence of VCT services in the country. Therefore, this literature plays a vital role in the

44

formation of the attitudes and approaches of the general public to the VCT system. Analysis of this literature can thus provide some indication of the current understanding of VCT that is placed out there by VCT stakeholders for the formation of current public attitudes and opinions about VCT.

5.2 Methods 338 articles were included in the analysis for this investigation. This body of literature included 19 theses and dissertations, 162 newspaper articles and/or advertisements, 74 government policies, pamphlets or reports and 83 unpublished organisational research, pamphlets or reports. While the conference proceedings of the international AIDS conference in Geneva and Durban were searched and 91 abstracts extracted, these were not included in the analysis as they were thought to provide too little information for a qualitative analysis. This body of unpublished literature was analysed using NVIVO™. The selected analytic method was a comparative thematic analysis. In order to perform this analysis, all the articles were classified and categorised according to type, i.e. newspaper, government report etc. Then within each category, the following questions were asked: 1. 2. 3. 4. 5.

How is VCT constructed and defined? What is the stated role of VCT? What is understood to be the facilitative factors for VCT? What is the literature’s view on the VCT implementation process? What critique does the literature offer with reference to the barriers affecting the VCT system?

Once this analysis was complete, then the different types of literature were compared in order to identify the main differences or similarities between the different types of literature. This comparison especially focused on government reports, newspaper reports and research reports with regards to the construction of the definition and role of VCT as a part of the HIV/AIDS strategic plan. It was hoped that in doing this comparative analysis we might be able to make some comment on the way in which the public understanding of VCT differs from that of the government and/or from recommendations made in the research literature. The findings from this analysis are presented below. Please note that where appropriate direct extracts from the literature have been included. These extracts are numbered and appear in the text in italics. Also, an indication has been provided of what type of literature these extracts were taken from.

5.3 Findings 5.3.1 Constructions and definitions of VCT In the thematic analysis of the unpublished literature, a primary focus area was to look at how different types of unpublished literature defined, constructed or used the term “VCT”. In this analysis, it became clear that the concept voluntary counselling and testing is both multidimensional and complex, making it difficult to define or describe in any clear-cut way. VCT is made up of a number of different process elements, the main one’s being voluntariness, counselling and testing. However, each of these process elements carries its own set of definitional issues, and when these concepts are combined they allow for a number of different perceptions and standpoints on what it means to do VCT.

45

Some of the constructions of VCT that emerged from the literature were: • • • • •

VCT as a prevention strategy for HIV/AIDS VCT as health promotive VCT as a psycho-emotional support VCT as a process of informed consent VCT and stigma reduction

Each of these definitions is discussed in more detail below. Please remember that although each of these constructions of VCT is discussed separately, they do in fact all form components of a general understanding of what VCT is. Each of these components is integrally linked, and have been artificially separated here for the purposes of clarity.

VCT as a prevention Strategy The first definition of VCT, and perhaps the most important in terms of the plan to reduce the impact of HIV/AIDS, is a definition that describes the role and function of VCT as primarily preventative. This preventative function is described in two ways. Firstly, VCT is seen to preventative in terms of inhibiting and/or stopping the spread of HIV/AIDS between people. In this construction of VCT, the primary and direct benefit of promoting HIV counselling and testing is its ability to prevent the spread of HIV. The second preventative function of VCT is the prevention of further illness in people who are diagnosed as HIV positive. This prevention is directly related to early detection and health promotion, both of which are considered in a later section. For now, let us consider only the first preventative function of VCT, i.e. as a preventative strategy for the spread of HIV. Consider the following extracts: Extract 1 (Government) VCT plays a role in the prevention of spreading the HIV virus by means of tests knowing sero-status, encouraging disclosure of status, information, resources, promoting precautions against spreading the disease Extract 2 (Government) For those who turn out negative, knowledge of this will reinforce efforts to remain uninfected Extract 3 (Newspaper) The department of health has emphasised the importance of HIV testing to promote health and prevent the transmission. Extract 4 (Newspaper) The article looks at the importance of counselling as an effective method of promoting prevention through assisting the mother to understand the implications of HIV. Extract 5 (Newspaper) Dit is uiters noodsaaklik dat mense wat HIV/vigs het van hul status bewus is sodat hulle nie in onkunde voortgaan on die virus te versprei nie. Extract 6 (Research) VCT has an important role in HIV prevention to help people make changes in their sexual behaviour so as to avoid transmitting HIV to sexual partners if sero-positive, and to remain sero-negative if negative

46

Extract 7 (Government) The Province regards counselling as a medium for prevention. Counselling provides the clients with the opportunity to make informed decisions, in terms of behavioural changes. It also offers care and support to the infected and affected. Counselling is both educational and informative; it also helps people to know their HIV status and means to prevent re-infection and spread of the epidemic. In these extracts, the preventative goals of VCT are clear. In all four instances, we see VCT being described as a strategy that is important for reducing, and that has indeed helped to reduce, the spread of HIV/AIDS. Across these extracts however, different mechanisms for prevention are identified. Some of these extracts hinge preventative outcomes on the process of HIV testing and knowing one’s status and others on the process of HIV counselling that is seen to provide preventative messages and transmit important knowledge about HIV and its implications. Research focuses on counselling as the basis that enables VCT to function as a preventive approach. Counselling is the means by which prevention education is communicated to both sero-positive and sero-negative individuals. Therefore, the training of counsellors to successfully deliver quality counselling is a vital component to ensure VCT can play a role in prevention. In both views, however, there is an assumption that once people are made aware of their own risks and how these risks can be changed (either through counselling or through testing) then the spread of HIV should be substantially reduced. Thus, in this view, both the process of counselling and the process of testing become primarily “prophylactic”, or medical tools for preventing HIV transmission between people. However, while this view is common, it is never explicitly stated how this prevention mechanism might work. In the literature, prevention is seen to be a primary outcome of VCT as opposed to being a variable indirect benefit of going through the VCT process, e.g. an indirect side-effect or function of increasing knowledge or encouraging people to abstain from immoral sexual behaviours. Thus it is simply taken as a given that VCT is preventative. This confidence in the preventative outcomes of VCT is also shown in discussions about the lack of available medical treatment for HIV/AIDS at the moment. There is a general view across the literature that VCT is the best strategy that can be followed at this time, because there is no way to medically prevent HIV transmission and no generally available antiretrovirals to treat those who are HIV positive. Consider the following extract: Extract 8 (Newspaper) We have to face that in the next decade we will not have a vaccine deliverable to people on a large scale. Prevention is the most important thing to focus on. This importance placed on prevention has landed largely on the shoulders of VCT. As VCT is seen to be the best current solution for containing the spread of HIV it has become an alternative to medical treatment, a move that places VCT at the centre of any HIV strategy. However, despite this emphasis that is placed on VCT as a means of prevention, there is some discussion in the literature of the need for a more “comprehensive” treatment package than is currently offered. Drug therapy is considered to be an essential future component of the VCT process Of all the literature consulted in this analysis, newspapers placed the most emphasis on this prevention definition of HIV. This is important, as the media is the most generally accessible form of writing about VCT, making it the primary means through which people come to understand the role and function of VCT. However, the question must be asked whether those people who make use of VCT services are displaying this understanding. Do people

47

who come for VCT do so before they suspect they might be HIV positive, or do they only do so once they suspect they have already got HIV/AIDS (i.e. when they are symptomatic)? Indeed, it would seem that this preventative message is contrary to an understanding of medical services in general where you only go to a clinic once you are sick – not before. The consequences of such general understandings need to be understood before the impact of prevention messages can be determined.

VCT as health promotive The second definition or construction of VCT is that it is health promotive. This construction of VCT as health promotion is linked to the earlier discussion of VCT as preventative and falls into the second category of prevention described above, i.e. that VCT can prevent the spread of illness in people with HIV. Within this construction, VCT is described variously as health promotive, an entry-point into care and as an early detection service. Consider the following extracts: Extract 9 (Newspaper) Early detection of HIV infection means that preventive and other forms of treatment can begin immediately. This lengthens and improves quality of life Extract 10 (Research) HIV voluntary counselling and testing (VCT) has been shown to have a role in both HIV prevention and, for people with HIV infection, as an entry point to care. VCT provides people with an opportunity to learn and accept their HIV sero-status in a confidential environment with counselling and referral for ongoing emotional support and medical care. Extract 11 (Newspaper) People are aware of AIDS, but many are not educated on issues that could prolong their lives or save the lives of others. He says:" We cannot give people who have tested positive a false hope, but we can arm them with information that could prolong their lives for years. This early detection message ties in with both prolonged life and improved quality of life for those living with HIV/AIDS. Early detection allows for the early prevention or treatment of opportunistic infections, monitoring of health status and also a means of arming people with information about their own health and health improvements. Detection of HIV status at an early stage can be crucial for prevention strategies to work, reducing the cost of opportunistic infections and the cost of medication, boosting immune systems and promoting good quality of life. VCT is thus seen to promote “wellness” in people. Obviously this early detection and health promotion message does feed back into the prevention message discussed above. If people can be persuaded to know their positive status this should not only improve their health, quality of life and prolong their life for years to come, it should also prevent the transmission of HIV and thus reduce spread. It could be surmised that this is why these two messages are often provided together – creating a construction that VCT is not only for the benefit of others, but also for your own benefit as well. The advantages of early detection are highlighted throughout all the different types of literature. This appears to be means of encouraging people to make use of VCT services before they “think” they have AIDS. Thus, the entry-point into medical care, prolonged quality of life and improved health are benefits of VCT that clients can expect if they come for VCT early enough.

48

Consider the following extracts: Extract 12 (Research) Also, if we find out at an earlier stage (that is, before we are sick) that we are infected with HIV, we will be able to gain information about what the disease entails, find necessary resources and learn about preventing the spread of HIV. Extract 13 (Government) Whether you are infected with HIV or not, because finding out how to live positively with HIV is one of the main benefits of VCT, we also give you some information on how to stay as healthy as possible for as long as possible. Extract 14 (Newspaper) The most important aspect is to arm your patient with factual, accurate and current knowledge. Doctors HIV and AIDS say doctors need to have a more normal relationship with patients. Counselling needs to be a life-long process This health promotive message envisages a much more comprehensive role for VCT, seeing it not only as a process that promotes prevention, but also as a total package of care that promotes health in those who are positive. It is a life-long process of providing clients with the necessary and correct information that enables them to manage their disease. Thus, the process of what it means to do VCT becomes much more holistic – a continuum of care geared towards the on-going support and care of individuals.

VCT as psycho-emotional support Closely linked to this idea of VCT as health promotive in terms of the provision of medical services is the idea of VCT as health promotive in terms of the provision of emotional and psychological support for those who are HIV positive as well as their friends, families and peers. VCT is not only seen to be an entry point into medical care, but also an entry point into mental health care and a whole host of support services. Indeed across all the literature this is seen to be an essential component of any VCT service. This understanding of VCT comes through strongly in the research literature where VCT is seen to be a comprehensive and on-going process, which includes both the medical care and the psychological support of AIDS sufferers. In order for VCT to work, there has to be both treatment and patient support available. For example, see the following extracts: Extract 15 (Research) In addition disclosure may not be what people wish to do if there is no treatment available. VCT can only work where there is some support and treatment following diagnosis. Extract 16 (Research) Various studies have proved that good HIV counselling has assisted people to make informed decisions such as whether to have an HIV test; helped many other people living with HIV or AIDS to cope better with their condition and lead more positive lives; and helped prevent HIV transmission

49

Extract 17 (Research) HIV voluntary counselling and testing (VCT) has been shown to have a role in both HIV prevention and, for people with HIV infection, as an entry point to care. VCT provides people with an opportunity to learn and accept their HIV sero-status in a confidential environment with counselling and referral for ongoing emotional support and medical care. One of the primary outcomes of the provision of support to HIV positive people and their families is the promotion of quality of life in patients as well as prolonging life through the generation of hope and the will to live. Support and care help overcome the psychological burden of coming to terms with one’s status and of living with AIDS and in addition promotes healthy living, and “wellness” which leads to longer and better life. Extract 18 (Newspaper) If they are positive, we let them know that HIV is a disease that can be lived with. Although there is no cure, your quality of life can be as good or as bad as your attitude. However, despite the importance placed on care and support in defining what VCT is, there is some indication in the literature that the lack of such services in some areas is proving to be a discouragement to testing. Extract 19 (Research) In many high - prevalence countries VCT is not widely available and people are often afraid of knowing their sero-status because there is little care and support available following testing This lack of back-up services is also held to be a primary reason for repeat testing. People are not supported through the first positive result and are therefore going to be tested repeatedly with the hope of receiving a negative result. This form of denial is precisely the function of VCT described above, suggesting that while such services are seen to be important they have not been implemented well. Finally, this definition of VCT as an entry-point into “support” or care can also be extended beyond ideas of illness or wellness into other aspects of coping with illness. For example, patients who are HIV positive often have to cope with much more than being ill, often they also have to cope with various socio-economic and survival issues as they become sicker. There is recognition in the literature that those people who are most vulnerable for HIV are also the most vulnerable in terms of poverty, access to healthcare and the availability of support structures. These social and economic conditions are thought to exacerbate the progression of the disease and interfere with prevention efforts. There is therefore some feeling in the literature that the government needs to provide facilities and support structures to those in these areas, and that VCT will then play a role in providing people with support or referrals to other support structures so that they can receive assistance with their wider social and economic problems.

VCT as a process of informed consent Across all of the literature, there is a common construction of VCT as being a process of informed consent, in that VCT is defined as a service that provides a complete “package” of important information that equips the client to make an informed decision about whether or not they wish to go ahead with the HIV test. This view of VCT emphasises the need for

50

people to know and understand the process of testing fully, in order to make a voluntary decision about being tested. See the following extracts: Extract 20 (Newspaper) VCT stands for voluntary counselling and testing. VCT is when a person chooses to undergo HIV/AIDS counselling so that they can make an informed decision about whether to be tested for HIV. Extract 21 (Government) VCT is a process of informed decision - support and confidentiality as part of a decision process of whether to take the test - it is more than just a test Within this view, VCT is seen to be a means of facilitating decision making in clients themselves through the provision of knowledge and support about HIV and the testing process. Thus, the counselling aspect of VCT is primarily geared towards promoting decision making about testing. This conception of VCT as an informed consent process redirects attention away from the anticipated public health goals of VCT (prevention and support) and provides a much more practical and grounded view of the primary function of the VCT process. However, questions must again be raised about how this function of VCT might work in actuality. Although clients are provided with valuable information as to the testing process, one wonders if the client has not already made up his or her mind before approaching the clinic for a test. How much does the actual counselling process contribute to the testing decision? In addition, while the counselling focuses on the outcomes of VCT, does it actually explicitly explore this issue of whether or not people are ready to be tested? Also, the issue of voluntariness in HIV counselling is a two-part concept. Firstly, voluntary refers to the counselling process, which should enable an individual to make an informed choice about being tested for HIV. Secondly, voluntary refers to the manner by which people come to make use of VCT services. In many instances, people approach VCT services freely and without force. In other circumstances, people are referred for HIV testing by health care workers for treatment and/or diagnostic purposes. Under these conditions, it is unlikely that persons coming for VCT have done so freely, and without coercion. Questions must be raised about this function of VCT in settings where people are referred for testing and are therefore not necessarily given an option as to whether they do or do not want to be tested at all. This definition of VCT places voluntariness at the centre of the VCT process – people must be informed so as to make their decisions on their own. See the following extract: Extract 22 (Research) The aim of VCT is not persuading people to be tested. VCT stands for voluntary, which means a decision taken by the person them self. However, both research and newspaper articles mention that testing takes place at times, by doctors or the nurses in South Africa without any counselling at all, let alone counselling for permission to test. Although the concept of VCT should be understood as a voluntary process it appears that some health workers construct it as a “routine test”.

51

Extract 23 (Newspaper) When Jabu Molefe was diagnosed HIV positive in 1993, she received no counselling before her blood was taken. Neither the doctor nor nurses at the hospital asked for the permission. Extract 24 (Government) How voluntary is the VCT in the medical sites? Are people accessing the sites because they are already in the clinics? How much self-referral is there? Perhaps the question is asked at times, why should consent be sought for HIV/AIDS tests if all other test for diseases are done routinely without consent? However, as research has found, there does seem to be a concern for situations where no pre-test counselling takes place. It was found that people receiving their HIV status without counselling and consent often have to deal with the trauma of been tested HIV positive by themselves and revert to avoidance tactics and associated behaviours as a coping mechanism Extract 25 (Research) Wood says since he joined ATICC, he has had to deal with numerous requests from medical practitioners and insurance brokers to pick up the pieces of insensitive and failed attempts to treat HIV as "as just another illness". Doctors who have "routinely tested for HIV " without any pre-test counselling have assured me that their distressed patient will be coming to see me for post-test support. Not one of these cases has turned up for their appointment at ATICC. They are out there somewhere, disbelieving, shocked, unsupported and denying their situation. Often such denial means deliberately not introducing changes in behaviour, for this would be admitting vulnerability Issues relating to mandatory or non-voluntary testing are intensely complex. The paradoxical feature attached to the HIV/AIDS epidemic of needing to know the sero-status of as many people as possible on the one side, and protecting the human rights and choices of people on the other hand contributes to immense frustration for health practitioners as well as clients of the system. For example, in PMTCT sites, mothers are often routinely tested as part of a package of pre-natal tests and medical services. While it is acknowledged that it is better if mothers volunteer for HIV, this is not always followed because of the risk to the child and the availability of treatment to prevent mother to child transmission. Similarly, rape victims would feel strongly that their aggressors should be tested. Overall, it would appear that volunteerism is regarded as one of the “corner stones” of any definition of VCT. There is still a feeling that if VCT is voluntary then more people will come forward to be tested and these people will be better able to deal with their results. However, this is clearly not always the case (or it is not perceived to be the case). Thus, in this literature the role that V should play in VCT is contentious with voluntariness acknowledged as better in theory, but not always carried out in practice.

VCT and stigma reduction A final construction of VCT has to do with the idea of “normalizing” HIV/AIDS as disease and thereby reducing the stigma associated with being HIV positive. In the literature, it is generally held that if more people are tested for HIV, then this should have the desired effect of lifting the associated fear and stigma that surrounds the disease – making it more like other chronic and terminal illnesses.

52

Extract 26 (Government) The government is encouraging all of us to come forward to be tested for HIV. They believe that if many of us get tested, even though we may not be sick, this will help to lessen the amount of stigma associated with the HIV test Extract 27 (Newspaper) They believe that if many of us get tested, even though we may not be sick, this will help to lessen the amount of stigma associated with the HIV test. Extract 28 (Newspaper) Widespread VCT contributes towards the destigmatisation of the disease Government literature advocates VCT as a place in a community that has a normalizing and realistic type impact about the disease within that community. The physical occurrence of the VCT institution may bring about the realization that the disease actually does exist in the community and that the community needs to support each other, just as they would in the case of any other disease. Extract 29 (Government) When the service is known to exist within the community, members get used to it and HIV/AIDS becomes a "normal" issue and stigma and discrimination are reduced. It becomes an easy subject to discuss, thereby causing that all people know and protect themselves A major advantage of reducing stigma is seen to be the promotion of prevention messages. As more people know about HIV and it becomes more “normal” in society, so more people will speak out against risk behaviours and provide more means of educating the general society about HIV and its implications. Extract 30 (Newspaper) Unlike in the past where people died without disclosing their HIV status, victims now want to educate others to prevent the pandemic from reaching uncontrollable proportions. This idea of stigma reduction is linked to concepts of confidentiality and secrecy in the process of VCT. Confidentiality is held to be a significant aspect in achieving the aim of drawing people in for testing with the aim of normalizing HIV/AIDS as a disease. VCT is defined in terms of a free and confidential HIV/AIDS service. However, some suggest that the confidentiality associated with the testing process is problematic, implying that if the secrecy attached to the disease were lifted then this should help to lift the stigma associated with HIV. See the following extracts: Extract 31 (Newspaper) One of the problems with HIV/AIDS is that of the confidentiality associated with testing for the disease. The patient must give permission for the blood test to be done. If the blood test is done without permission then it is illegal. This does not apply to any other diseases. Perhaps by making people aware of their infection and not being too concerned with confidential issue the aspects of secrecy can be lifted. Extract 32 (Newspapers) Perhaps by making people aware of their infection and not being to concerned with confidential issue the aspects of secrecy can be lifted. Thus, there is a tension in the literature with regards to the role that confidentiality should play in the definition of VCT. This tension is further compounded by the fact that there are some indications in the literature that the results of HIV counselling and testing process are

53

not always kept confidential. For example, many of the research articles consulted highlight the fact that the quality of confidentiality in the VCT context varies greatly. Extract 33 (Research) Furthermore, the quality and benefits of VCT, in particular with regard to confidentiality, counselling and access to clinical and social support, vary enormously. It is difficult to determine exactly what role confidentiality and secrecy play in the definition of VCT. Some authors place it firmly at the centre of any VCT strategy with the firm understanding that confidentiality is important for the VCT process to function effectively. Others, however, see confidentiality as secrecy – secrecy that simply compounds stigma and therefore works against a primary defining focus of the strategy as a whole. How can we normalize any disease that is not treated in a normal manner?

VCT and human rights Across several of the other definitions of VCT discussed above, lies another integral definition of VCT – that it is not only a health service but also a human right. People have the right to know their status or not know their status. Similarly, they have the right to be counselled and supported through both the decision to test and their results and people also have a right to be educated and informed about HIV and its implications. In this way, the provision of VCT is see not only to be an essential medical service or means of prevention, but also an ethical and moral obligation in terms of the rights of the general population to have access to such services. Consider the following extracts: Extract 34 (Newspapers) The medical world has reacted with surprise to a report that the Department of Health intends to make it a criminal offence for a medical practitioner not to offer counselling to a patient with HIV/AIDS. Extract 35 (Newspapers) Everyone has the right to be tested and to receive counselling, and access to National VCT resources and facilities. VCT provides a necessary human right. Are VCT facilities available to all? In addition to VCT being an essential human right in terms of the provision of services, it also crosscuts a number of other human rights issues, e.g. the protection of those who test positive from discrimination, stigma or violence. This brings us back to issues of confidentiality and secrecy mentioned above as well as the reduction of stigma and the issue of mandatory testing. Each of these is a human rights issue and each is complicated in its own right. For example, how does one protect the confidentiality of patients while still trying to reduce stigma in communities? Moreover, how can we protect the rights of those who are positive, thereby placing negative patients at risk? How VCT feeds into these human rights issues needs to be carefully examined and explored. Policies and laws have not adequately considered and dealt with these issues and have therefore not impact significantly at a ground level in terms of redressing some of the institutional human rights violations.

5.3.2. Government Policies In light of the above definitions and constructions of VCT, it was felt that another useful way to analyse the VCT literature would be to focus specifically on how governments’ view VCT in terms of both public policy and implementation strategies. The goals of VCT, be they prevention, emotional or social support, health promotion, stigma reduction or human rights

54

protection will fall squarely on the shoulders of national and provincial governments who will have to develop policies that address these goals and plan implementations that meet these goals. In this analysis we will look at how these governments have met the challenge of policy development and how they have addressed the issue of service implementation. The former issue will also be discussed in the next two sections when we look at factors that the literature has identified as either a facilitating factor for, or barrier to, the implementation of VCT services. VCT represents only one aspect of the overall strategic plan that the government has developed in relation to HIV/AIDS. Instead, the government has assumed a comprehensive and coherent policy approach where a national strategic plan exists to guide the entire country’s response to the epidemic. As can be seen in the extract below, this national plan is also not confined specifically to the health sector, but rather adopts a multi-sectoral approach where the needs across governments and departments are addressed. Extract 36 (Government) This document is a broad national strategic plan designed to guide the country's response to the epidemic. It is not a plan for the health sector specifically, but a statement of intent for the whole country, both within and outside for the whole country. It is recognised that no single sector, ministry, department or organisation is by itself responsible for addressing the HIV epidemic. It is envisaged that all government departments, organisations and stakeholders will use this document as a basis to develop their own strategic and operational plans so that all our initiatives as a country can be harmonised to maximise efficiency and effectiveness The guidelines developed and determined by HIV/AIDS policy are based on research of previous experiences. There is a clear aim across the literature to try and improve and build on those experiences to develop more efficient practices. Strategic concerns, addressed by Policy guidelines, are therefore, formulated through constant revision of previous and existing strategies in order to improve and encourage growth in the HIV/AIDS prevention effort. Extract 37 (Government) The guidelines are intended to reflect experiences and insights that have been gained in the field of HIV/AIDS counselling since the first version of WHO AIDS Series 8 was published in 1990. As countries develop their counselling services, new ideas and experience will in turn guide the further refinement of services and revision of Within this holistic and comprehensive approach to HIV/AIDS policy making, there is a clear understanding that testing people for HIV plays an important role in any national strategy aimed at combating HIV/AIDS. As reviewed in the constructions of VCT above, knowing one’s status is a cornerstone for prevention, support and care, health promotion, stigma reduction and the protection of human rights abuses. Extract 38 (Research) Policies strategies should address, education and awareness of HIV transmission and safer sex practices, strategies to ensure access to condoms, access to HIV testing and counselling services. In the government policy literature, there is a clear understanding that HIV is a complex disease that has legal, economic, psychological, ethical and medical implications for both those infected and affected as well as the health care sector. Within this holistic understanding of the nature of the HIV pandemic, is the strategic understanding that HIV testing needs to be conducted under a certain set of conditions that accounts for the wider implications of being an HIV positive person in a health care setting. These conditions are

55

set out in national policy as being what is generally understood to be VCT – i.e. HIV testing must take place only when the rights of persons are upheld through processes of informed consent and the regulation of pre-and-post-test counselling. See the following extracts: Extract 39 (Government) Testing for HIV infection presents serious medical, legal ethical economic and psychological implications in the health care setting. Because HIV infection is a life threatening condition, reasonable persons or health care workers will attach significance to the outcome of an HIV test, especially a positive diagnosis. For these reasons, and in accordance with the constitutional guarantees of freedom and security of the person, and the right to privacy and dignity HIV testing policy shall constitute national policy. This policy applies to persons able to give consent, as well as to those legally entitled to give proxy consent to HIV testing in terms of the law. The circumstances under which HIV testing may be conducted are given as well as informed consent, pre-test counselling and post-test counselling regulations. Extract 40 (Newspaper) The department's draft National Policy on Testing for HIV recommended that all HIV tests should involve voluntary consent with pre and post-test counselling. Thus, while HIV testing is the strategic goal, VCT is the implementation of this goal in such a way that the protection of the rights of persons is upheld. Due to the nature of the disease, one cannot simply be tested; rather one must be tested under conditions of voluntariness and where pre and post-test counselling has been provided. There is strong consensus across the government literature that this is the implementation goal across the country. However, as we have reviewed in the previous section questions about the meanings of these terms can be raised. What it means to be voluntary and what the outcomes of counselling should be are not always uniform, making policy implementation decisions that much harder. It is essential, in the effort to ensure effective functioning that agreement exists in the construction and definition of concepts such as voluntary, testing and counselling in the HIV/AIDS context. Policy can fulfil this role by providing standardised explanations for these concepts. In doing so, it can also provide a means of structuring the implementation approach by providing pre-determined parameters against which implementation can be measured and evaluated.

5.3.3 Implementing VCT National HIV/AIDS policies emphasise the need for counselling linked to voluntary testing as a means for people to find out their sero-status. In accordance with this strategic aim, the implementation of VCT is recognised as an important strategic aim across all the literature analysed for this study. In this analysis, it became clear that there were certain aims and outcomes that are regarded as priorities in the VCT implementation process. For example, precedence is placed on the promotion of prevention methods and messages, on the encouragement of safe-sex practise and on the provision of knowledge and information to people such that they can make a voluntary informed choice about whether or not to be tested. In terms of service provision outcomes, a number of factors were identified across the literature that seem to be regarded as essential elements of the implementation process. We have termed these factors facilitating factors for the purposes of this analysis, as they are elements of the VCT implementation process that either contribute to better service delivery

56

or are factors that are seen to be essential to the success of the programme as a whole. Each of these factors is discussed in some detail below:

Access The ability of the Department of Health to promote the expansion of VCT services seems to be an important aspect to consider when implementing a VCT programme. Expansion allows for more people to be exposed to the services of VCT, thereby promoting prevention messages and “normalising” the disease to decrease stigma and discrimination. Indeed, some seem to regard this expansion to be a primary factor in the reduction of HIV incidence rates across the country. The two extracts that follow emphasize the importance of “access for all” to any VCT strategy. Extract 41 (Newspaper) The rate of HIV infection in the North West province had dropped by 0,2% from 23,1% to 22,9%. MEC for Health Molefi Sefularo says his department has embarked on several programmes in an attempt to stem the HIV/AIDS tide. The department has increased its voluntary HIV/AIDS counselling and testing sites from 11 to 55 and plans to open more in the near future. Extract 42 (Government) A list has been compiled and will be updated regularly to accommodate expansion of VCT sties. VCT sites have successfully expanded. Out of 495 VCT sites identified, 359 are operational. Government seems committed to implement and provide VCT facilities to a large section of the population through consistently increasing the number of VCT sites over time. Moreover, in terms of implementing services, it is clear that this means a targeted intervention whereby areas that have been previously underserved will be the first to receive services. There is an understanding that improving access in rural and youth communities is a critical aspect of realising intended VCT outcomes: Extract 43 (Government) To provide HIV counselling and testing to 12.5% of the adult population ages 15-49 years within three years targeting the youth and rural communities through improving linkages between health, welfare and education sectors at district level. Extract 44 (Newspaper) The decision on the number of VCT sites should be based: • On the population of that district • HIV prevalence rate • Linkages with welfare services • Capacity in term of infrastructure For better access to VCT services the VCT sites should be spread proportionally in terms of the geographic location and the selection of sites in rural, urban, peri-urban should be proportional In these two extracts, we can see the importance of access within the VCT strategy has been linked to redressing imbalances in the health services delivery system at the moment. There needs to be a better spread of services across all populations and geographic locations thus ensuring better access for all. However, while this notion of access is seen as an important success factor for any VCT programme there is also some sense in the literature that VCT sites should start off small

57

and not expand too quickly. In beginning with so-called “model” or “demonstration” sites, the continued success of the programme could be better guaranteed as it allows for on-site learning before VCT is expanded to the entire country. The following extract highlights this viewpoint: Extract 45 (Research) In dealing with the issue of the provision of quality of VCT services in the operational sites, it is proposed that model sites be established. These sites will be used as demonstration sites. This will provide Provinces with an opportunity of learning about the best practices. It is envisaged that the learning process will be about the following: • The provision of and maintenance of quality counselling services. • A best approach to the supervision and support of HIV/AIDS counsellors. • The use of a well-developed system of data collection • Conducting local marketing interventions for promoting local VT services • The use of monitoring tools to obtain the necessary data Thus, across the literature, there is a subtle tension about this facilitating factor for VCT. While access to all is regarded as a cornerstone of the VCT strategy, there is also some feeling that expanding too quickly could be a barrier to success.

Continual upgrading of facilities Within the literature, there is recognition of the fact that there needs to be consistent and regular improvement of existing VCT services as well as expansion of these services to new sites. Activities that can strengthen already existing sites are workshops, continual counsellor training workshops and a continual upgrading of the knowledge base about HIV and VCT. This continual improvement is seen to be a feedback learning process whereby stakeholders in the process of setting up VCT sites are continually able to contribute their experiences about what works or doesn’t work in the VCT environment. These experiences should then be implemented in a real work setting as VCT is expanded to more sites and to more settings. See the following extract Extract 46 (Government) A one day VCT workshop which aims at exploring the implications of voluntary HIV testing and counselling and sharing experiences with organisations / programmes that had already started the programme has been planned for the 29. 06.00. It is hoped that the province will learn a lot from this discussion before it embarks on a full scale programme.

Monitoring and evaluation of VCT sites An important facilitating factor for VCT appears to be regular monitoring and evaluation of VCT sites and their operations. Regular monitoring and evaluation is a valuable procedure that ensures that certain standards of quality are maintained at VCT sites across the country, thereby feeding back into the need to continually upgrade these sites. Monitoring and evaluation allows for emerging problems to be dealt with immediately to avoid cumulative consequences and various VCT sites can be compared around the country as a means of learning from each sites experience. Some of the factors that are described in the literature as feeding into this monitoring and evaluation function are:

58

• •

Appropriate and on-going supervision of counsellors – this refers to both technical and emotional supervision The development of research tools for measuring VCT outcomes

In essence, this monitoring and evaluation function is about quality control – both of the counsellors and the counselling and of the entire VCT service delivery process. Thus, the emphasis is placed on determining progress and in monitoring change. The key to effective monitoring and evaluation is seen to be time – whereby results must be quickly available such that the programme becomes reactive to changes in the environment. See the following extracts: Extract 47 (Government) Monitoring and Evaluation aims at describing the data, tools and analysis of the key information obtained on progress and effectiveness of the VCT services. This should be done timeously and be able to disseminate the required information about the performance of the programme. The purpose of Monitoring and Evaluation is: • To help the programme focus on key areas of implementation. • To assist in tracking changes over time. • To be able to compare progress with other countries. • To be able to measure how the programme is progressing. • To help in measuring long-term outcome indicators i.e. behaviour change, attitude, knowledge etc.

Other factors Some of the other factors that were mentioned as facilitating factors for the implementation of good quality VCT services were: • •

The promotion/marketing of VCT facilities and benefits to the general public Sound financial management

The first of these factors has an obvious facilitating impact on VCT services. If people are not aware of what VCT is, where they can go for VCT or what the process will entail then the programme will have failed, as no one will use the services. Thus, a major facilitating factor for VCT would be to promote these services to the general public in such a way that they will want to come and be counselled and tested. In general, it is felt that a benefits based approach could accomplish this goal best. With regards to the second factor, it is generally felt across the literature that to run a VCT programme effectively, it is essential to have staff with management and particularly financial management skills. Accurate record keeping, accurate statistics and accurate monetary accounting are all seen to be factors that will enable the successful running of VCT services. Extract 48 (Research) Given the lack of coherent strategic plans with provincial HIV/AIDS units, these skills seem to be lacking. As many of the interventions being introduced at provincial level are new, well functioning and activity sensitive financial control and monitoring systems need to be put in place to evaluate and improve the interventions and the spending patterns behind them. The structure of the financial information management systems is such that detailed activity reporting is difficult.

59

As can be seen in the extract above, there is also some concern in the literature that these skills are not present in most sites. Also, there is some concern that the financial reporting structure as a whole may not be facilitating sound financial management or reporting either.

5.3.4 Barriers to VCT Implementation When one attempts to look at the realities of the implementation of VCT nationally, it is immediately apparent that there is no clear picture across any of the literature as to where this implementation process stands at the moment. There are numerous contradictions in the literature as to the current status of VCT services in South Africa. For example some government reports give the impression that VCT services are running smoothly and proficiently, so much so that they are looking at the possibility expanding VCT facilities to non-medical sites. However, other reports acknowledge the inadequacies that exist within the VCT system. Consider the following extract: Extract 49 (Government) Whereas enormous resources have been allocated from time to time to build the counselling service, HIV/AIDS counselling in South Africa remains underdeveloped, constrained by issues of competence, policy obstacles, lack of coordination, and lack of resources. Research seems to hold a critical stance towards the implementation strategies proposed by government. Within this body of literature there is a clear feeling that in order for the successful expansion of VCT services to occur, existing strategies have to be reviewed and adapted. According to statistics, HIV/AIDS seems to be on the increase, which indicates that VCT plans are possibly not producing the envisaged results. Moreover, research literature indicates that the implementation process has been misinterpreted and incorrectly applied by various VCT management structures. Health management seems to be communicating a distorted idea of VCT objectives to their employees. Some of these views are reflected in the extracts below: Extract 50 (Research) HIV prevalence seems to increase despite a comprehensive national HIV/AIDS strategic plan for SA, which indicates inadequate implementation of plan Extract 51 (Research) Like many other policies, the concept of VCT has not yet been well interpreted by many stakeholders and a modified version is cascaded downwards throughout the system as most top-down HIV interventions. The aim of VCT is not persuading people to be tested. VCT stands for voluntary, which means a decision taken by the people themselves. Across the literature, a number of factors have been identified as barriers to the effective implementation of VCT. These barriers can be divided into two categories, i.e. those that are internal and those which are external. Internal factors are those that are directly related to the VCT implementation plan and to the service delivery of HIV testing and counselling. External factors, on the other hand, refer to those factors that stem from the potential service users and their communities. As can be seen in the discussion below, those barriers that are internal can be addressed and rectified by reviewing strategies and policies. However, those that are external are attached to perceptions and fears relating to the disease making changes in these areas slow and difficult. Some of these internal and external barriers are discussed below:

60

Internal Barriers: 1. Access A key barrier to successful VCT implementation is inadequate access to VCT service centres. As discussed above, there are certain communities, which have been traditionally neglected in terms of health care and support service delivery. Thus, new government policies have focussed particularly on this issue in an attempt to redress the imbalances in the provision of services. However, even though the policy decisions are in place, until such redresses have been dealt with they still constitute a significant barrier to the use of VCT services. Many people who want to be tested either cannot or will not travel to VCT facilities either due to lack of time, lack of motivation or lack of funds. The need to make VCT generally available to all is reflected across the literature. Consider the following: Extract 52 (Newspaper) The mother must be identified as HIV positive first. This means the system to encourage voluntary testing with pre- and post counselling should be made more accessible. Extract 53 (Government) The Northern Province with a population of 5,5 million has a prevalence of 11,4 % as depicted in 1999 HIV prevalence survey. Voluntary counselling and testing is neither readily available nor accessible to the communities. Extract 54 (Government) Besides the lack of facilities, the financial burden of finding and financing transport to health facilities and payment for health services acted as barriers to access to care. Many rural hospitals had very limited access to medical doctors and medicines were not always available at public health facilities and expensive. There is also recognition in the literature that questions around access to VCT services, need to be extended to access for all services. Underserved areas require more than health services and many of these other support services will be needed for VCT to be effective. For example, in terms of on going care and support of clients and their families a number of support-type services will be needed. Housing, social services, family services, home-based care to name but a few all contribute to the efficacy of VCT in any given area. At the very least, there is a strong call for on going counselling for patients in the same area in which they live. 2. Demand vs. Quality Another barrier to effective VCT service delivery, has to do with service quality where often the need to keep up with the demand for HIV/AIDS services results in a compromised quality of service. For example, there is some concern that more attention and effort is spent establishing new VCT sites, than in improving the quality of services or upgrading resources at existing sites. In addition, the monitoring and evaluation processes seen to be vital to effective VCT functioning are generally not implemented or the information not used. Thus, there is no feedback of information about quality that can be used to improve the service delivery. This view is shown in the extracts below: Extract 55 (Research) This drive towards quantity is a threat to the provision of quality services as more effort is put to open more sites rather than to ensure the quality of counselling service provision.

61

Extract 56 (Research) Setting up VCT and ensuring a quality that will create demand is a considerable challenge. Building in self-assessments, monitoring and regular evaluation is an important tool to enhance the quality of VCT.

3. Human Factors HIV/AIDS counselling is a difficult and stressful job. It is often not easy for counsellors to maintain professional boundaries in the relationships that develop between them and their clients. Counsellors have to deal with intense emotional issues of dealing with people living with a life threatening disease, severe illness and death. The lack of a medical treatment for HIV/AIDS patients is an added frustration that counsellors are faced with. Extract 57 (Newspapers) The stressful part of this job is to see them die after you get to know them. We're supposed to be friendly, but not friends. But as human beings you develop relationships. Extract 58 (Newspapers) In the Haven, nurses find it very hard. They're dealing with people who are hurting. It's not easy. Extract 59 (Newspapers) Counsellors claim they have also been banned from even advising rape survivors where to find anti AIDS medicines or how to use them. The lack of clarity as to professional roles and boundaries leads to stress and burnout placing further burdens on the system. Moreover, the confusion as to what kind of relationship one should or should not have with a client impacts directly on the nature of the counselling that is performed and the quality of care the patient receives. Counsellors may generally not be suited to handling the stress of working with HIV positive patients and this kind of personality trait is often not detected in the short training period. The confusion that most counsellors’ experience as to their relationship with their client is further exacerbated by a tension between counselling roles that the counsellor has to negotiate through. Counsellors have to perform dual roles of therapist and educator. Counsellors are expected to not only provide basic information and advice, but also to provide support and care to a client. This kind of dual role is hard to negotiate and counsellors are not always trained to handle both kinds of counselling simultaneously. Extract 60 (Research) Tension between the role of counselling being one of psychosocial support and one of educating on prevention transmission of VCT and the third being counsellor is characteristic of VCT. At most VCT sites nurses are required to do counselling work in addition to their designated medical duties. The added workload and lack of financial compensation leaves nurses feeling frustrated and despondent. These negative factors have an impact on the quality of counselling outputs. A supervisory and mentorship system is urgently required to monitor service quality and support counsellors with emotional and technical elements of their work.

62

Extract 61 (Government) There is an urgent need to provide support and mentorship / supervision to counsellor’s in-order to prevent burn out and increase skills thereby increase quality of service provision. Extract 62 (Research) The impact of the workload of nurse counsellors and the consequent quality of counselling that patients receive. The alleged reluctance of nurse to undertake counselling without financial remuneration. There is recognition in the literature that often the VCT system does not take into account the fact that counsellors bring their own issues into any counselling situation with them. For example, the personality or values of the client may clash with those of the counsellor and counsellors are often not taught to deal with those situations either.

Extract 63 (Research) Staff bring to work their own needs arising from being personally affected by the epidemic an/or find their jobs increasingly stressful as they feel themselves to be helpless in the face of the growing incidence of HIV. A lack of resources is hampering the fulfilment of potential improvement of quality care. There does need to be more recognition of the problems that counsellors face on a daily basis as there is a perception in the literature that there is no enough support for those who are on the “front-line” so to speak and that their issues are often lost in the general problems that face VCT programmes.

4. Service Delivery Factors There are a number of service delivery factors that can be seen as barriers to effective VCT service delivery, most because they are factors which impact on exactly how the process of VCT gets to take place. For the most part, these factors result from the tension between the medical and the non-medical components of VCT. For example, because the HIV test is a medical test, only a trained medical person can do it. As a result, this means that most VCT is done in medical rather than non-medical settings as the problems of finding trained medical people to work in non-medical sites is prohibitive. In addition, the fact that only a medically trained person can do the test means that either the counselling and testing components of VCT must be split between two people or a nurse must be made to do both of these components. Neither of these solutions is ideal as typically nurses cannot do the counselling due to an overload of clients and yet having two people do one test is illogical and creates problems with clients as well as between medical and non-medical staff – especially in terms of roles, functions and responsibilities. Extract 64 (Government) Using existing staff to do HIV counselling as an add on job and has implications for sustaining counselling services. In non-medical sites the issue of non-health care workers not being allowed to perform the test and yet it is difficult to get a nurse on loan for a non-medical site. A major problem is the fact that HIV counselling is a non-medical service often performed by lay and non-professional people that takes place in a medical setting, which is occupied by professional doctors and nurses.

63

Extract 65 (Government) Counselling is also caught in a paradox of attempting to render a psycho- social service in the bio-medical context of health care settings with preferred outcomes for such counselling sessions. This particular problem impacts on the delivery of the complete VCT services. Ideally, VCT sites should have sufficient people trained in counselling and in testing that do not have to undertake any other medical duties as well. However, such a system does not seem probable. 5. Resources Another major apparent problem related to effective VCT service delivery is an apparent inconsistency between strategic objectives and service delivery realities. There is some discussion across the literature that VCT objectives cannot be met in reality due to inadequate resources such as testing kits, finances, infrastructure or even basic medication. This is a quality of care issue and there have been reports of extremely poor VCT service delivery for a variety of reasons: Extract 66 (Research) Common negative experiences of care in general clinics where victimisation of HIV positive users, callous treatment from Health providers, poor confidentiality, overcrowding, long waiting times, limited hours of service and shortage of drugs In some senses, this is also a staffing issue as discussed earlier. Staff cannot perform their duties effectively and yet they are also not supervised properly through the process. Strategies need to be reviewed and effectively implemented with regards to duty allocations and in particular performance outputs to ensure VCT services are delivered effectively to the public as originally intended. However, better resource allocation and management is also needed to ensure the resources needed to run a good quality VCT service are made available to those working in VCT sites. It cannot be forgotten that a major part of quality improvement is improved resources. Extract 67 (Research) With so much attention now given to improving health service performance and reducing costs, there is a danger of neglecting the care and prevention demands generated by HIV infection.

External Barriers: 1. Survival issues This category includes aspects of poverty that govern the poor person's perception of knowing their HIV status or even caring about it. Many people in South Africa, and particularly those that are vulnerable to HIV/AIDS have to deal with numerous socioeconomic burdens on a daily basis. Whether they are HIV-positive or negative is of little consequences for them. It might just mean the realization that they will not be able to support themselves or their families, when they become sick. Often they would rather not know. HIV/AIDS preventative mechanisms have little impact. Many women with low socioeconomic status have no choice regarding their sexual behaviour. They are governed by their partner’s demands. VCT has a greater chance of success in more developed and sophisticated communities. Socio-economic development strategies need to be put into place before VCT can reach its full potential.

64

Extract 68 (Research) The programmatic response fails to take account of the social context that influences behaviour. Many South Africans have little control over their lives and circumstances. Thus, messages to promote safe behaviour are often inappropriate and ineffective Extract 69 (Research) The issue has to be addressed beyond its biomedical features (and solutions). Social factors that drive behaviour must equally be recognised and addressed. Extract 70 (Government) The information in this supplement provides most of what we know already about HIV/AIDS. However, many more people have died of AIDS in Africa than in countries like America or Britain or other so-called "first world" countries. This has caused our leaders to begin asking questions about why so many Africans die as a result of AIDS. Questions that seem easier to answer in America or Europe about the HIV virus are harder to answer when we think about Africa. 2. Partner Violence A barrier in the prevention role that VCT tries to perform is that many women are scared to reveal their HIV-positive status to their partners in fear of a violent reaction. The extract below taken from a research study suggests that women have no control whether to be tested or not; these decisions are made by their partners. Extract 71 (Research) An important component of HIV voluntary counselling and testing VCT programs is encouraging clients to inform partners of their serostatus. Findings of this study revealed that many women lack autonomy to make decisions about HIV testing. Disclosure to partners by HIV-positive women has increased over time but is still significantly less than that for HIV negative women. Partner violence is a serious problem among female VCT clients. A small proportion of women who disclosed their serostatus to partners reported a negative reaction. Women's HIV status is strongly associated with partner violence. 3. Stigmatisation The fear of being rejected by one’s own family, community, workplace, schools and even church is a major factor preventing people from coming forward and finding out their HIV/AIDS status. One of the objectives of VCT is to make people aware that people living with HIV/AIDS are as normal as any other person living with a life threatening disease. The stigmatisation barrier affects HIV victims and many of the HIV/AIDS workers, who are at times ostracized for the work they do. The question arises: do people want to know their status? Knowing one’s sero-status, means not only living with a terrible secret and facing rejection, but also knowing that there is no available HIV/AIDS treatment. An element of distrust about counselling confidentiality might explain why some counselling services have been slow to get off the ground. Extract 72 (Newspapers) People infected with HIV find themselves increasingly marginalized due to the public's ignorance about the disease, despite the fact that the virus was first diagnosed as long ago as the early 1980s. "People who are infected and affected by HIV live in isolated world without any hope, a world ruled by fear of being rejected once they disclose they are living with HIV-AIDS or are related to someone who is HIV positive, " said Pam Kisten, managing director of Open Door, a drop-in HIV-AIDS care centre.

65

Extract 73 (Newspapers) We dare not do it openly. There is great prejudice in the region against people who have AIDS. Employers will not keep them, schools will not admit their children, their homes are often burned down and they are either killed or chased into the wilds” She said what was perceived to be a "simple act of kindness" could result in gruesome death. Extract 74 (Research) Evidence from Zimbabwe indicates stigma attached with use of VCT services. In addition disclosure may not be what people wish to do if there is no treatment available. Extract 75 (Research) Counselling services have been slow to gain acceptance in many countries, especially where HIV is heavily stigmatised and access to services and support for the HIV-infected is limited. 4. Fear of results The VCT goal is often defeated when people do not come back for testing after receiving initial pre-test counselling. The fear of having to face being HIV-positive contributes to people not coming forward voluntarily for testing. Testing often only takes place when the victim starts showing signs of being ill. This factor inhibits the VCT objectives of promoting prevention and healthy lifestyles. Extract 76 (Newspapers) Since the MTCT pilot project began in June, 379 pregnant women have agreed to come for counselling but 255 were unwilling to undergo the testing, saying they needed time to think about it but they have never come back" she said. Mvemve said that some women are afraid to know their status. Extract 77 (Newspapers) People are scared to come to the counselling centre voluntarily, they come when they have to, when they start showing symptoms. Some when they discover their partners have been promiscuous or when people close to them die. Pre-test counselling is an important part of her work. Clients are unable to cope with the results of positive tests. 5. Cultural Factors From the extract taken from a newspaper one can realize the impact that traditional medicine has on the health system in our country. Traditional health care is an integrated and influences the lives of people. Many people would rather make use of the “Sangoma” or “Nyanga” than attending the local VCT clinic. An option for VCT planning could be to possibly examine the traditional health system to establish its popularity and to incorporate some ideas and even traditional health workers into the HIV/AIDS prevention system.

Extract 78 (Newspapers) Traditional medicine represents the largest indigenous resource base for caring for the people of Africa. It is the oldest art of healing on this continent and is therefore integrated in the socio-cultural and religious worldviews of many of our people," said Msisika.

66

Extract 79 (Newspapers) Tsotetsi said it was time for traditional healers to start attending HIV/AIDS workshops, where they would get all the necessary knowledge about the epidemic. "Many traditional healers still don't understand the importance of holding workshops and seminars with HIV/AIDS organisations and even local clinics. Extract 80 (Research) 50 % of people in Africa that are known to be HIV positive resort to traditional healers to find solutions to their conditions. What is the logic used by the traditional leaders? How do they require their knowledge?

5.4

Conclusions

This analysis strongly demonstrates a construction of VCT as fundamentally primary prevention and secondarily, health promotion or secondary prevention. The construction of VCT as preventative hinges on two related themes. Firstly, VCT is constructed as a preventative intervention based on the assumption that a linear relationship exists between knowledge of HIV status and behaviour change. Secondly, the prevention construction hinges on the educational outcomes of the counselling process that are assumed to yield behavioural changes. So, an essential element in controlling the spread of HIV/AIDS is through the mechanism of having everyone know their serostatus and being educated. The discourse in this regard is not coercive, i.e. the intention is not to force people to undergo a mandatory test. Rather, people are encouraged, whether sick or not, to come forward to receive adequate information to make a voluntary choice about being tested. In this way the spread of the disease can be curbed and people can be educated in prevention and behaviour change methods. The secondary prevention or health promotion construction of VCT is rooted in the early detection and treatment functions of testing and counselling and the access to treatment for opportunistic infections. A sub-theme in the secondary prevention construction includes the notion of psychosocial support. The secondary/health promotive construction conveys a twopart health message of benefit to individuals that simultaneously has the capacity to yield primary prevention outcomes for others. The construction of VCT as primary prevention prioritises the population whereas the secondary construction prioritises the well-being and health status of HIV positive individuals. The prominence of the primary prevention construction may be driven by several factors including the urgency of the pandemic, the absence of a vaccine and the unavailability of ARV treatments for the population. The secondary prevention construction is necessarily limited to treatment for opportunistic infections and access to support and a pro-social prevention outcome goal of limiting the spread of the infection. It is important to note here the disjuncture between these constructions of VCT and the VCT research literature reviewed above. For example, the fact that many studies have not supported the primary prevention outcomes for VCT. Secondly, the health promotion construction hinges on assumptions about the integrity and competence of the health service that in many instances may not match policy aims. The unpublished literature in general reflects a conditional optimism in respect of these two constructions of VCT. The primary conditions being, service and infrastructure capability, enabling policies and the efficacy of communication and media campaigns geared towards community preparedness and uptake of services. Disclosure and openness to achieve prevention, support and de-stigmatisation outcomes through VCT, also represents a central theme in the unpublished literature. The more people 67

are tested and the more people talk openly about having the disease, the more chance there is of destigmatising and normalising HIV. The unpublished literature again presents a conditional optimism that this will encourage communities to give each other support and care. This optimism rests on the assumption that the process of VCT yields both interpersonal and social communication and educational outcomes that diffuse the essential prevention messages through society, together with a reduction in negative social processes such as stigma and discrimination. The unpublished literature recognises that Government has formed policies and strategies to deal with the HIV/AIDS virus. VCT is one of the components in this plan. Although VCT has been implemented and some success is evident, there appears to be quite a number of barriers inhibiting the smooth and effective functioning of the initial VCT strategy intentions. These barriers should be addressed and appropriate remediation implemented to improve VCT’s capacity to function as an efficient service in the drive against HIV/AIDS. A recommendation in this body of literature is that Government should continue with a process of targeted, phased implementation of VCT sites, with more attention and effort spent on improving the quality of services or upgrading resources at existing sites, before establishing new sites. The unpublished literature highlights the naiveté of a “magic bullet” construction of VCT disconnected from social, economic and development contextual factors. VCT has a greater chance of success if the broader conditions that increase communities’ vulnerabilities to sickness and ill-health are addressed and socio-economic development strategies need to be put in place before VCT can reach its potential as one key dimension in an integrated strategy. In the unpublished literature and the study as a whole, the lack of agreement and the tensions between varying constructions of VCT was identified as an important barrier to implementation. An implementation strategy rooted in a confused conceptual base generates difficulties down the implementation line. Agreement needs to exist between the construction and definition of voluntary, counselling and testing on all levels within the HIV/AIDS context.

68