SciDev.Net

Summary

One of the most contentious issues in research ethics is the level of treatment or other medical intervention – usually referred to as the ‘standard of care’ – that should be given to participants in clinical trials. Much of this debate has focused on the treatment provided as part of a trial to members of control groups, particularly when such individuals have received a placebo (a 'treatment' known to be without effect), rather than the best current treatment.*

In the developed world, the accepted practice is that those in a control group should, in most circumstances, be given the best current treatment for the condition being studied, regardless of cost. The contentious issue is whether the same principle should be applied in developing countries, where local conditions may mean that the best current or ‘universal’ standard of care is not widely available.

The debate takes place in a world that is characterised by huge disparities in access to health care. Some conclude that one effective way of reducing these disparities is to identify medical interventions that are currently affordable to those with relatively few resources, even if doing so involves employing different standards of care when carrying out research in different settings. Others, however, warn that with such a strategy the risks of exploiting those without access to effective treatments are so great that one should instead work directly towards reducing health disparities.

Sharp disagreement over this issue has surfaced in recent debates over revisions to international guidelines on clinical trials. The guidelines strongly affirm that the ideal should be to provide the universal standard of care to patients in the control group of the trial. But almost all now allow for exceptions – while insisting that any exceptions require careful justification. And there is still considerable controversy what conditions need to be met for exceptions to be justified. There is also concern over whether research ethics committees have the expertise to judge whether an exception is justified, and whether effective mechanisms have been put in place to prevent abuses that could result from allowing such exceptions.

[* The term ‘standard of care’ is used in various ways in discussions of research ethics. Some use it to refer to all the care that a participant receives during a trial (for example in addition to the intervention being studied participants may receive diagnostic tests and treatment for other conditions). However, the term can also be used to refer merely to the intervention whose effects are being studied (which in the case of a control group may be a placebo). In this paper we use the term ‘standard of care’ in the second, narrower sense.]

Perinatal transmission of HIV/AIDS

Much of the recent controversy on standards of care was sparked off by trials carried out in 1994 of a treatment to prevent the transmission of HIV infection from mothers to their children, so-called ‘perinatal transmission’. These trials – which became known as the 076 trials after the number of the US National Institutes of Health (NIH) protocol – were conducted in the United States and France. They demonstrated that treating a mother with the drug zidovudine (AZT) during the last three months of pregnancy, then treating her intravenously during the delivery of her baby, and finally treating the newborn child, dramatically reduced perinatal HIV transmission.

However, such a treatment regime cost about US$800 per pregnancy, which made it unaffordable in sub-Saharan Africa, the region with the highest level of perinatal transmission in the world. In addition, many women in sub-Saharan Africa did not seek medical attention early enough in pregnancy, or live close enough to medical facilities, to enable them to receive ongoing treatment during the final three months of pregnancy and following the birth of the child.

There was therefore an urgent need to develop affordable and deliverable alternatives for those in this region, and a number of trials with shorter versions of the treatment regimen were carried out, conducted partly under the aegis of the Joint United Nations Programme on HIV/AIDS, the World Health Organisation, the US Centers for Disease Control and Prevention (CDC) and the NIH.

All of these new trials – apart from one in Thailand – used a placebo in the control group. The results showed that even the shorter period of treatment was signficiantly better than a placebo in reducing perinatal transmission of HIV/AIDS. However, the new trials were criticised in the New England Journal of Medicine and The Lancet. Editorials and an article in these journals argued that placebos should not have been used in the control group, as an effective treatment – namely the 076 regimen – already existed, and that depriving those in the control group of this treatment was ethically unacceptable.

Sacrificing standards for effective results?

The main argument of the critics was that the short course alternatives should have been compared with the 076 regimen, as acting otherwise meant withholding known effective interventions from trial participants. They pointed out that no researcher would have been allowed to do this in rich countries, and that permitting it to be done in poor countries therefore amounted to applying a double standard.

The critics also pointed out that the trials violated the Helsinki Declaration – the principal ethical guidelines for conducting clinical trials. The 1996 version of the Declaration stated that: “In any medical study, every patient – including those of a control group, if any – should be assured of the best proven diagnostic and therapeutic method”.

The main response of those who defended the use, under certain conditions, of placebos was that the standard of treatment used in clinical trials should reflect local conditions in order to ensure that the results are relevant to the country in which the trial takes place.

They pointed out, for example, that there was no reliable knowledge about the HIV transmission rate in African countries in untreated pregnant women, although it was known that the transmission rate had varied over time, even in the same location. In such circumstances, they argued, an ‘equivalence trial’ – comparing the long 076 treatment with the short course treatment – would not provide the data needed by developing-country medical authorities, namely how effective was the short course treatment compared to no preventative treatment at all.

The supporters pointed out that the trials using placebo controls would not leave anyone worse off, as participants would not otherwise receive effective treatment to reduce transmission. Also, such trials could potentially produce knowledge of substantial benefit to all those living in the countries in which the trials took place.

Answering the criticism that the interests of some of the trial participants were being sacrificed for the common good – a principle that is generally rejected in the context of research ethics – the defenders of the trials pointed out that this would also be true of an equivalence trial. For in such a trial, the short-term medication being received by the treatment group was already known to be inferior to the longer term treatment, in other words to an established standard of care. They argued that the equivalence trial would therefore violate the same ethical standards as the placebo trial – and without even producing results of relevance to the host country.

Pulling apart the scientific justification

In the discussion about the perinatal transmission trials, one argument used by those defending the trials was that the variable transmission rate meant that the use of a placebo in the control group was necessary to provide relevant information. A central disagreement in the debate about standards of care is over whether – in order to answer questions that are relevant to the social and economic environment of a host country – it is ever necessary to carry out a trial using less than a ‘universal’ standard of care for patients in a control group.

Admittedly such an argument cannot by itself justify the use of a local standard of care in any particular proposed trial. It can only show that, in certain circumstances, a trial has to be carried out in a certain way to answer a particular question.

There may, or course, be other reasons for not doing the trial, such as a wish to avoid exploitation of the patients involved, or the fact that the questions being addressed are not sufficiently relevant to the host country to justify abandoning the use of a universal standard of care for control groups.

These counterarguments are considerably weakened, however, when a trial represents the only realistic way of identifying and implementing interventions that will provide considerable health gains for the host country. That is why the committees and commissions that have examined this question have all concluded that there are indeed circumstances in which it is acceptable to allow exceptions to the general rule, namely that a universal standard of care should be provided to members of control groups.

Revisions of guidelines

The controversy that blew up in 1997 over the perinatal transmission trials prompted a lengthy debate over revisions to established guidelines for the conduct of clinical trials. In particular these include the World Medical Association's (WMA) Helsinki Declaration and guidelines published by the Council for International Organisations of Medical Sciences (CIOMS), which are designed to be used by countries when defining national policies on the ethics of biomedical research involving human subjects, and in applying general ethical standards to their local circumstances.

The outcomes of these debates have differed. In 2000, for example, the WMA adopted a revised version of the Declaration of Helsinki that essentially affirms the commitment to a universal standard of care that was set out in the 1996 revision. Article 29 of the declaration now states that: "The benefits, risks, burdens and effectiveness of a new method should be tested against those of the best current prophylactic, diagnostic, and therapeutic methods. This does not exclude the use of placebo, or no treatment, in studies where no proven prophylactic, diagnostic or therapeutic method exists".

Those responsible for other international guidelines, however, have taken a different position. In each case these guidelines state that, in general, researchers have an obligation to provide participants in the control group with a universal standard of care. But they also accept that, under certain circumstances, exceptions may be allowed. (Click here for a summary of international ethical guidance on standards of care.)

Even within the groups responsible for these guidelines, however, there still appear to be significant differences of opinion. In 2003, for example, the European Group on Ethics – an advisory group to the president of the European Commission, whose views are used for assessing clinical trials funded by the commission – issued guidelines that reflected this split. Although a majority of the members of the group accepted that exceptions to the universal standard of care should be permitted, a minority disagreed.

The evidence suggests a similar split among those responsible for the 2002 version of the CIOMS guidelines. The specific guideline that deals with the choice of control in clinical trials does not mention the possibility of the type of exception considered above. However, the commentary that accompanies the revised guidelines indicates clearly that ethics committees can, in exceptional situations and after careful consideration, approve studies with less than universal standards of care.

In particular, the CIOMS guidelines now indicate that an ethics committee "can approve a clinical trial in which the comparator is other than an established effective intervention, such as placebo or no treatment or a local remedy" if it is satisfied:
(i) that the trial addresses the health needs of the local population;
(ii) that there is assurance that the medical intervention being assessed will be made reasonably available to that population if it proves to be safe and effective; and
(iii) that the established effective intervention cannot be used as a comparator because its use would not yield scientifically reliable results relevant to the health needs of the study population.

Avoiding abuse

Although it seems generally accepted that, under certain stringent conditions, exceptions to the general rule of providing a universal standard of care are allowed, substantial disagreements remain over what these conditions should be. For there remain some differences between the guidelines in the types of reasons they consider acceptable for allowing an exception to a universal standard of care, and some variation in the minimum standard of care that they require to be provided during a trial.

Furthermore there remains the need to address the main concern of those who are critical of allowing standards of care in clinical trials to be adapted to local conditions, namely that such a strategy will lead to an exploitation of those participating in clinical trials in poor countries. The worry of such critics is that allowing exceptions to a universal standard of care would mean that studies that are forbidden in rich countries – either because they are too risky, or because they deny participants established interventions – may be carried out in countries with less stringent procedures.

Linked to this is concern that clinical trials will be approved and carried out in poor countries whose results will primarily benefit those in rich countries. This is not just a theoretical possibility; there are a number of trials in which critics argue that this has already happened. For example, a recent US study revealed that a third of US researchers and nearly half of researchers abroad believe that the interventions being tested in their research are unlikely to be available to most host country residents in the foreseeable future.

In principle, however, the problem should not arise; all the guidelines referred to above that allow exceptions do so on the understanding that the approved trial is needed to obtain knowledge of direct use to the community in which the trial takes place. The crucial need is to ensure that those asked to evaluate the ethical basis of a proposed trial can discriminate between trials that meet this criterion, and those that do not.

Another need is to ensure that an ethics review committee is able to determine whether an intervention being tested is likely in practice to be made available if it is found to be safe and effective. Again this requires ensuring that the criteria for assessing a proposed trial are sufficiently specific to allow such a reliable judgement to be made. Detailed procedures for assessing the benefits of research would help here. For they might reduce the disagreements between those sceptical of allowing different standards to be applied in different circumstances, and others who consider a variation in standards necessary to tackle urgent health problems across the world.