There is a Corrigendum (June 2005) associated with this Commentary.
Integrating HIV testing into immunological studies of non-HIV-related diseases
Siske S Struik1, Caroline A Maxwell1, 2, Mwele Malecela-Lazaro2, Sue Eckstein3, John D H Porter1
& Eleanor M Riley1
1 Siske S. Struik, Caroline A. Maxwell, John D.H. Porter and Eleanor M. Riley are in the Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London WC1 E 7HT, UK.
2 Caroline A. Maxwell and Mwele Malecela-Lazaro are with the National Institute for Medical Research, Dar es Salaam, Tanzania.
3 Sue Eckstein is with the Centre of Medical Law and Ethics, King's College London, London WC2R 2LS, UK. eleanor.riley@lshtm.ac.uk
HIV testing is now required for non-HIV-AIDS-related immunological studies in areas of high HIV prevalence. Ethical guidelines for testing in these circumstances need clarification and sensitive protocols need to be developed.
As human immunodeficiency virus (HIV) prevalence rises, the possibility that apparently healthy volunteers enrolled in immunological studies have altered immune function due to HIV infection cannot be ignored, especially in sub-Saharan Africa, where a substantial portion of the population is infected (Fig. 1). HIV testing has become necessary for studies of host responses to pathogens that are not classically associated with HIV infection, but it brings with it myriad ethical and practical problems. Here we summarize the scientific, ethical and practical issues surrounding HIV testing in the context of non-HIV-related research and discuss options for the implementation of HIV testing. We illustrate this discussion with our experience of introducing HIV testing into studies of immunity to malaria in Africa.
Figure 1. Estimated HIV prevalence rate (%), adults, 2003.
Used with permission from the World Health Organization.
The need for HIV testing People with clinical symptoms of advanced HIV infection (AIDS) are unlikely to be inadvertently recruited into research programs or clinical trials However, in areas of high HIV prevalence, there is a considerable risk that HIV-infected but healthy, asymptomatic people may be recruited into such studies. Recruitment of HIV-infected people poses a substantial problem, as even in the asymptomatic phase they show defects in antigen presentation, intracellular signalling1, a shift toward T helper type 2 responses2 and reduced serological responses to vaccines3. Alterations in immune function precede CD4+ T cell depletion in peripheral blood1, and destruction of recently activated CD4+ T cells in lymphoid tissues leads to loss of memory responses to commonly encountered antigens. Thus, asymptomatic HIV infection is likely to alter ex vivo and in vitro immune responses to unrelated pathogens, but criteria such as clinical symptoms, CD4+ T cell count or CD4/CD8 ratio are insufficient for exclusion of infected people. Exclusion on the basis of risk factors such as profession, family circumstances and so on reduces the likelihood of recruiting seropositive people but risks introducing bias if the study sample is no longer representative of the population. If HIV affects only the outcome of the study (and not the exposure of interest), inclusion of HIV-infected people may lead to underestimation of size of the effect being studied. Increasing the sample size (without HIV testing) will, in theory, overcome this problem but can only be done if the prevalence of infection and its effect on the outcome of interest are known. In practice, if the chances of recruiting HIV-positive people are deemed substantial, it will be necessary to do HIV antibody testing and to restrict or stratify the study accordingly.
From a practical viewpoint, given that rapid diagnostic tests for HIV are cheap (by research standards) and sufficiently specific and sensitive for the purpose of excluding HIV-infected donors4, the introduction of HIV screening into a research protocol is neither expensive nor time-consuming. However, this ignores the myriad social, cultural, ethical, and personal and public health issues that surround a diagnosis of HIV infection. Counseling and treatment (with highly active antiretroviral therapy (HAART)) are well established in most high- and middle-income countries but, although their availability is increasing in many low-income countries, universal provision is still a long way off5,
6. Furthermore, the stigma and fear attached to an HIV diagnosis remains substantial, and HIV testing might easily compromise recruitment into non-HIV-related studies. Careful consideration of the process of informed consent, the provision of counseling and 'after-care', incentives, data protection and confidentiality, and logistical and budget implications is essential if a study is to achieve its objectives in an ethical and socially acceptable way.
The ethical framework Guidelines for ethical research on humans are continually evolving in response to changing health problems and new options for prevention and treatment7,
8,
9. The stigma attached to HIV, the chronic and fatal nature of the disease and the costs of providing lifelong care and treatment have raised unique ethical dilemmas. Although existing guidelines do address research in developing countries, they draw heavily on Western perceptions of autonomy and individuality. The HIV-AIDS pandemic, because of its rapid spread in the least developed countries of the world, has led to collaborative research between scientists trained to follow these guidelines and healthcare professionals and study populations who may have different socio-cultural expectations. Ethical review processes now exist in most countries10 and are key tools for the development of locally appropriate protocols, but limited expertise and resources, lack of clarity about what is expected and real or perceived imbalances of power can hamper dialogue between ethics committees of the sponsoring institution and the host country. The opportunity to gain access to HAART can be a major motivation for people in low-income countries to undergo HIV testing in the context of biomedical research projects, and this has led to a long, sometimes very contentious and unresolved debate about the long-term responsibilities of researchers in terms of HAART provision for their study participants11.
Informed consent is a cornerstone of ethical research. 'Anonymous' testing of samples for HIV without the knowledge of the participants is now considered justifiable only for the purpose of public health surveillance12,
13, as it otherwise undermines the trust of the participants and denies them access to important health information13. However, finding practical alternatives to anonymous testing poses new dilemmas. In the absence of access to effective treatment, the benefit of knowing one's HIV status is debatable and may simply cause distress and social exclusion; moreover, ethical guidelines stipulate that the autonomy of participants and their freedom to choose should be protected. In studies in which disclosure of HIV status to the person cannot be avoided, the consequences of participation require careful discussion at the consent stage and participants must agree to counseling and testing before enrollment. However, in studies in which HIV status needs to be taken into account only at the analysis stage and infected people do not need to be excluded at recruitment, disclosure of HIV status is not essential, and one option is to inform the participants that their samples will be tested and to offer them the choice of whether or not to receive their test results.
Case study The immunology of malaria infection is a good example of a research area that has developed its methodology against an essentially stigma-free background. Underlying HIV infection does not affect susceptibility to malaria in the same dramatic way as, for example, tuberculosis, and although evidence of interactions between HIV infection and malaria risk is emerging14, malaria is not closely associated with HIV in the public's perception. Integrating HIV testing into malaria research thus raises new ethical concerns and has logistical and budgetary implications. These issues are 'generic' rather than malaria specific and are likely to be encountered to some extent by any non−HIV research team that introduces HIV testing into their research protocol.
The purpose of our study was to test the hypothesis that acquired immunity to Plasmodium falciparum malaria is associated with the ability to regulate the proinflammatory cytokine responses that are linked to pathology15. We sought to compare in vitro cytokine responses of peripheral blood mononuclear cells from people living in African villages that are malaria-free versus those subject to endemic malaria transmission. Age-stratified but otherwise strictly randomized recruitment from a demographic census was essential. At the time of the study (2001−2002), the prevalence of HIV among adults in the study area was approximately 10% (ref. 16). HIV antibody testing to exclude seropositive samples was therefore required. As there was to be no long-term follow-up, it was not necessary to exclude HIV-positive subjects at recruitment. We therefore proposed to test blood samples on arrival in the laboratory (by rapid diagnostic test), discard the HIV-positive samples and do immunological assays on negative samples only. Recruitment targets were increased to allow for discarded samples.
The ethics committee of the London School of Hygiene and Tropical Medicine took the view that participants should be told about everything that was going to happen to their blood and that the subject's right to choose whether to know their test results should be respected. The Tanzanian ethics committee (Institutional Review Board of the Tanzanian National Institute for Medical Research, which is mandated to review all health research in the country) endorsed this approach and specifically commented that openness about HIV testing was the only, albeit difficult, way forward. Although we were not expected to provide HAART, both ethics committees explicitly pointed out our responsibilities in facilitating access to the best locally available voluntary counseling and testing (VCT) and after-care for those participants who wished to know their HIV results and in contributing to sustainable local development of VCT and after-care where possible.
Because we did not have the resources or expertise to provide VCT or patient care, we arranged to pay (via a voucher system) for VCT to be provided by local not-for-profit clinics, having established that the study community trusted these clinics to maintain confidentiality and valued the advice and care they provided. To ensure that test results could not be linked back to specific people, blood samples and questionnaires were labeled only with a study number. Participants who wanted to know their HIV status would thus need to be tested twice: once (anonymously) by the research team and once at the VCT clinic. The procedures were explained to individual participants during the process of informed consent.
One of the authors (C.M.) had been working with Tanzanian colleagues in villages highly endemic for malaria for more than 10 years and had successfully conducted a study in one such village in which participants had been informed that their blood would be tested for HIV (anonymously and without VCT). The team had not previously worked in villages where malaria was entirely absent, but we were able to identify a highland village that was interested in participating.
District medical officers and regional HIV coordinators endorsed the study protocol. Notably, although they understood the principle of informed consent, they questioned the need to tell participants about the HIV tests. In their experience, people were not keen to know their HIV status (in part because there was poor local access to HAART) and telling them about the tests would, they felt, make people unnecessarily anxious about a study in which they would otherwise have liked to participate. The need for HIV testing also required considerable explanation to project staff. Staff who had not previously worked on HIV-associated projects found the ethical issues very challenging; extra training and a more proactive approach to staff management and supervision were required to safeguard satisfactory implementation of the research protocol.
As is usual for research in rural African communities, the study was discussed initially with community leaders. Given the sensitivity and complexity of the subject, both we and the village leaders felt that a public meeting was not an appropriate forum for the discussion of issues related to HIV testing. HIV testing was thus discussed in private during home visits as part of the informed consent procedure, and vouchers for VCT were offered to all participants or their guardians. In addition, participants were given a written information sheet and a consent form to sign, or the information was read to them and their consent was witnessed by a family or community member of their choice. A parent or guardian was asked for consent for children (those reported to be under the age of 14); as is usual in these communities, older family members were invariably involved in decisions taken by adolescents and younger adults. Those who consented subsequently visited the local clinic for completion of a questionnaire and donation of a blood sample (Fig. 2).
Figure 2. Study participants attend a Tanzanian health center for completion of a questionnaire and donation of a blood sample.
Outcome of the study In the highly endemic village, malaria was perceived as an important health problem and thus, although the immediate benefits of taking part in the study (free access to a mobile clinic for the duration of the study) were small, villagers were enthusiastic. Very high participation rates reflected the trust that had been built up between the research team leader and the village over many years. No-one took up the offer of a voucher for HIV testing. A feedback meeting with village elders confirmed that our study protocol met the needs of the participants.
In contrast, the malaria-free village was very remote and villagers were not familiar with medical research. As malaria is not a problem there, the study had little perceived relevance to potential participants and, apart from the clinics, brought no obvious benefits. Nevertheless, the elders were keen for their village to participate and suggested that everyone should be told their HIV test results. The reasons for this suggestion are not entirely clear and are probably very complex, but may have included a desire to seem broad-minded, an assumption that knowing one's HIV status would lead to changes in behavior and thus would benefit the community, and/or the wish to know the prevalence of HIV in their community. In contrast to the enthusiasm of the elders, villagers were uneasy, recruitment was slow, and lengthy consent procedures were required to ensure that participation was truly voluntary. Although HIV testing was not the single focus of concern, it did substantially increase anxiety; everyone declined a VCT voucher. In a follow-up meeting, village leaders expressed no concerns about the conduct of the study.
The lessons learned The introduction of HIV testing into non-HIV-related research necessitates a careful review of costs and benefits for individual participants and their community. Especially in studies in which HIV status needs to be disclosed to participants, the potential for individual harm is substantial and the consequences of leaving newly diagnosed HIV carriers without long-term support in the community need to be addressed at the ethical review stage.
HIV testing had a considerable effect on study design and substantially increased the resources required, particularly in terms of staff time and staff training. This was more notable in the malaria-free village, where it was compounded by unfamiliarity with research and the lack of a pre-existing relationship between researchers and villagers. We conclude that it is feasible to introduce HIV testing into ongoing non-HIV research if it is done sensitively and openly and in the context of an established and trusting relationship between researchers and the study community. Conversely, HIV testing in the absence of the established confidence and trust of the community can be fraught with problems, no matter how open, honest and sympathetic the approach.
Our study was typical of many in Africa in that the initial decision to participate was taken at the community level. When people are keen to participate this should not cause problems, but when they are reluctant to do so because of HIV testing, it can lead to a feeling of being under pressure to participate so that the community can benefit. The provision of basic health care for the duration of a study is common practice but can constitute a very powerful community incentive10. In our case, village leaders were paid locally appropriate 'per diem' allowances as recompense for their time in acting as liaisons between researchers and villagers; the opportunity to be paid, in a community in which cash is hard to come by, may also explain their enthusiasm for the study. This puts additional responsibilities onto the research team to ensure that individual consent is truly freely given and to provide an environment in which this can be achieved.
Discussing HIV testing in private during home visits rather than during community meetings seemed to work well, and our openness about HIV testing was clearly appreciated. Home visits allow greater privacy, greater openness on part of both researchers and participants and more time for detailed discussion; the drawback is that it adds to the air of secrecy about HIV testing that can propagate fears and rumors. As the extent to which HIV is openly discussed varies among different communities, it is important to explore what is likely to work best in any particular setting. In this community, as in many African societies, in addition to individual consent, the consent of a spouse or older family member was often required, even for mentally competent adults. This may account in part for the poor demand for vouchers for VCT, as having to 'express an interest' could be seen as an admission of promiscuity. With hindsight, a voucher should simply have been given to every participant; demand for vouchers might also have been increased if we had explained that the voucher could be used either by the person to whom it was given or by a friend or relative. However, demand for VCT may still have remained low if people fear being seen attending the VCT clinic.
Collaboration with a local VCT service, rather than setting up one's own, avoids duplication of services and undermining of existing provision. It minimizes conflict of interest between recruitment and provision of impartial information about HIV testing and separates concerns about participation in the study from anxieties surrounding attendance for HIV testing. It also allows for a simple and transparent data protection protocol. This suggests that research involving HIV testing should, as far as possible, be done in areas where some form of VCT and HAART is already available.
Conclusion There is an increasing scientific need for HIV testing in the context of non-HIV-related research. However, in the absence of freely available treatment, testing is likely to be unpopular with both participants and research staff. Those who fund non-HIV-related research are unlikely to provide resources for HAART or other long-term care. It would follow from the debate on HIV-related research that on pragmatic grounds, lack of provision of HAART is acceptable in the context of non-HIV-related research, but there is no literature specifically addressing this issue, and guidance is sorely needed.
Our study indicates that where a trusting partnership between researchers and study populations already exists, HIV testing can be introduced into non-HIV-associated projects with relative ease; conversely, where such a partnership does not exist, it can be hard to forge one in the present climate of anxiety and mistrust surrounding HIV testing and research. This has substantial implications for funding agencies, because short-term, stand-alone studies are unlikely to have the time or resources to build a sufficiently trusting partnership. At present, 'best practice' requires that participants be informed that their blood is being tested for HIV and be offered optional access to their results. How this is best implemented will depend on the local context, and there is clearly a need to strengthen the dialogue between researchers, ethics committees and, where they exist, national HIV care and treatment programs to develop locally appropriate and ethically sound protocols. We hope that our experiences may inform future discussions and the development of a 'generic' framework for this type of research.
Obaro, S.K., Pugatch, D. & Luzuriaga, K. Lancet Infect. Dis.4, 510−518 (2004). | Article | PubMed | ChemPort |
World Health Organization. Rapid HIV Tests, Guidelines for Use in HIV Testing and Counselling Services in Resource-Constrained Settings 17−20 (World Health Organization, Geneva, 2004).
Moatti, J.P. N'Doye, I., Hammer, S.M., Hale, P. & Kazatchkine, M. Nat. Med.9, 1449−1452 (2003). | Article | PubMed | ChemPort |
"3 by 5" Progress Report. December 2004/WHO and UNAIDS 11−16 (World Health Organization, Geneva, 2005).
National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research (National Institutes of Health, Bethesda, Maryland, 1979).
World Medical Association. Declaration of Helsinki: recommendations guiding physicians in biomedical research involving human subjects, 1964. Bull. Med. Ethics162, 8−11 (2000).
The Council for International Organizations of Medical Sciences (CIOMS). International Ethical Guidelines for Biomedical Research Involving Human Subjects (CIOMS, Geneva, 2002).
The Ethics of Research Related to Healthcare in Developing Countries 1−190 (Nuffield Council on Bioethics, London, 2002).
Wendler, D., Emanuel, E.J. & Lie, R.K. Am. J. Public Health94, 923−928 (2004). | PubMed |
Nicoll, A., Gill, N., Goldberg, D. & Peckham, C. Br. Med. J.320, 1275 (2000). | ChemPort |
Doyal, L. Br. Med. J.314, 1107−1111 (1997). | ChemPort |
Acknowledgments Conducted as part of the Joint Malaria Programme, a collaboration between the Tanzanian National Institute for Medical Research, Kilimanjaro Christian Medical College (Moshi, Tanzania), the London School of Hygiene and Tropical Medicine and the Centre for Medical Parasitology, University of Copenhagen (Copenhagen, Denmark). Funded by The Wellcome Trust (060115).
