The 2000 version of the DoH is completely restructured. There is now a section headed ‘Introduction’ comprising paragraphs 1–9 which sets out the scope of the document and some of the underlying principles. Although many of the statements in the ‘Introduction’ were present in previous versions of the Declaration, they have been re-ordered to present a more logical sequence of reasoning. Arguably one of the most important statements is the requirement in paragraph 5 that ‘In medical research on human subjects, considerations related to the well-being of the human subject should take preference over the interests of science and society’. By the end of the ‘Introduction’ the document has very clearly set up the dilemma that gives rise to the need for clear thinking about research ethics. On the one hand, it would be unethical not to challenge current methods in medical practice (paragraph 6) through research. On the other hand, it is wrong to simply use people as a means to an end (paragraph 5), particularly vulnerable people (paragraph 8). Having described this ethical tension in the ‘Introduction’, the DoH then seeks in the next two sections to articulate the guiding principles for deciding what research meets the ethical standards required and what does not.
There is no longer any specific section dealing with ‘Non-therapeutic’ research, which is often viewed as synonymous with ‘healthy volunteer’ research. There is specific reference to ‘healthy volunteers’ in three paragraphs of the Edinburgh (2000) revision. Paragraph 16 explicitly states that participation of healthy volunteers as research subjects is permissible. Were this not stated, then a certain way of interpreting paragraph 19 may lead to the conclusion that such research was now proscribed. In paragraph 18 healthy volunteers are identified as a group where the importance of prior weighing of the importance of research against its risks and burdens is especially important. Finally, Paragraph 8 in the ‘Introduction’ lists ‘those who will not benefit personally from the research’ among those groups that are vulnerable and in need of special protection.
Paragraph 29: 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 (See Appendix 3 for Note of Clarification) As already mentioned, the 1996 version of the DoH is the first version of the DoH to mention specifically the use of placebo in trials. Paragraph II.2 from the 1996 version stated ‘The potential benefits, hazards and discomfort of a new method should be weighed against the best current diagnostic and therapeutic methods’. This has been changed to the wording seen in the first sentence of paragraph 29 (above). The sentence which then followed in the 1996 version (and which formed the first sentence of paragraph II.3) stated ‘In any medical study, every patient – including those of a control group, if any – must be assured of the best proven diagnostic and therapeutic method’ has been eliminated. Finally, in the 2000 revision very little is changed in the actual sentence referring to placebo which is the second sentence in paragraph 29 (above); the words ‘inert placebo’ from the 1996 version are changed to ‘placebo, or no treatment’. In a careful reading of the two versions, however, it appears that very little has changed in the overall ethical guidance with respect to placebo use. Therefore, what is surprising is that the outcry following the 2000 revision far exceeded the response to the 1996 revision.
The overall effect of paragraph 29 would seem to rule out use of placebo wherever proven treatment exists. As mentioned, this raised such a cry of protest that the WMA took the unprecedented step of issuing, in 2001, a Note of Clarification to Paragraph 29. The Note of Clarification was formally adopted as part of the DoH in 2002, although the WMA has not described this as a ‘revision’ since the actual text has not been modified – only ‘clarified’!
However, the Note of Clarification certainly seems to modify the requirements and represents the first occasion where the WMA have issued explanatory text indicating the intent behind a specific paragraph. One of the best summaries with respect to placebo use in trials is that of Emanuel and Miller , who define three broad positions: placebo orthodoxy, active-control orthodoxy and the ‘middle ground’ (see Table 2 for definitions). It would appear that the Note of Clarification moves the stance of the DoH from what appears to be active-control orthodoxy towards the ‘middle ground’. The debate in the literature over the ethics of placebo controls has raged for at least the past decade between the proponents of ‘active-control orthodoxy’ such as Rothman, Michels and Weijer [19–21] and those supporting ‘placebo orthodoxy’ such as Levine  and Temple .
Table 2. Emanuel and Miller's three ethical positions with respect to placebo-controls 
|Active-control orthodoxy||Placebo orthodoxy||Middle ground|
|‘ Whenever an effective intervention . . . exists, it must be used in the control group . . . placebo controls are inappropriate because the clinically relevant question is . . . whether [a new drug] is better than standard treatment’||‘ When effective treatments exist, there must be compelling methodological reasons to conduct a placebo-controlled trial’||‘ Without a placebo group to ensure validity, the finding that there is no difference between the investigational and standard treatments can be misleading or uninterpretable’|
The Note of Clarification lists two situations where placebo is acceptable: where there is a scientifically compelling reason, or where the condition under study is minor and the subject at no increased risk of serious or irreversible harm. These two situations are linked by the word ‘or’ which has been questioned by Macklin . She asserts that the connector should be ‘and’ (i.e. both conditions must be fulfilled). The risk otherwise is that scientifically compelling reasons could be used to justify an increased risk of serious harm through use of placebo and this is argued to be inappropriate. This would be in line with the introductory principle of paragraph 5 that ‘considerations related to the well-being of the subject should take preference over the interests of science and society’. The counterarguments are both that valuable research may be prevented  and that placebo-controlled trials often require a much smaller sample size and follow-up time and therefore expose fewer people to any risks inherent in the research .
A further issue with respect to paragraph 29 has been the interpretation of the words ‘best current’ as the standard of comparator arm. Does this mean best in existence or best available in a local context? The Note of Clarification does not address the issue. The UK Nuffield Council on Bioethics argues the issue extensively, recognizing that ‘The Declaration of Helsinki (2000) is the primary source of guidance on which the majority of other guidance draws’. Their conclusion regarding the interpretation of ‘best proven’ is that ‘the minimum standard of care that should be offered [in the control arm] is the best intervention available as part of the national public health system’.
There is still considerable discussion around the circumstances in which placebo control is ethically acceptable. It seems clear that for some serious conditions where there is often ‘one chance’ at cure – such as many forms of cancer – placebo-controls should be ruled out. At the other end of the scale, except for the most extreme adherents to ‘active-control orthodoxy’, minor and self-limiting conditions seem to present little problem regarding placebo use. It must be remembered that paragraph 29 refers to ‘proven’ treatment, not ‘active’ treatment. Just because a pharmaceutical agent is shown to have pharmacological ‘activity’ does not mean it has been properly ‘proven’ to be superior to placebo. Indeed, such proof may never be forthcoming in some conditions where placebo response is either high or greatly variable. Symptoms of chronic stable angina, for example, can show a highly variable placebo response  and this condition is selected by Emanuel and Miller  as an example where a well-designed placebo-controlled trial should be satisfactory on ethical grounds provided patients are well monitored for worsening symptoms, that appropriate ‘rescue’ or ‘escape’ medication is available, and participants are fully aware of their right to withdraw from the trial at any time.
In the middle of these extremes are many clinical scenarios where the issue of whether placebo-controlled research is acceptable or whether serious or irreversible harm is risked needs to be undertaken on a ‘disease-by-disease’ basis. Among the conditions which have given rise to recent debate in this regard are hypertension , depression , schizophrenia  and postmenopausal osteoporosis . Taking osteoporosis as one example, Brody and colleagues  have pointed out that there are groups of patients in whom placebo-controlled trials clearly do not violate paragraph 29. They specifically identify as suitable for placebo-controlled trials: ‘competent, well-informed patients [who] refuse approved therapies for sound reasons’, situations where ‘there is a reasonable basis for substantial disagreement or lack of consensus among professionals about whether approved treatments are better than placebos’, or ‘subjects are refractory to known effective agents’. It should be noted, however, that this approach may introduce biases.
A person consenting to participate in any blinded randomized controlled trial is effectively agreeing not to be given information that most individuals would want to receive; that is, to know what treatment they are receiving at any one time. This agreement not to know such information is not unique to trials using placebo-controls. Placebo-controls are not deemed unethical in and of themselves by paragraph 29. What is called into question is the potential harm to research participants who may not receive otherwise available proven treatments during the course of a placebo-controlled study.
The issue of placebo-control, probably more than any other, highlights the need for delicate considerations to balance ethical tensions which often exist between research which seeks to obtain answers as efficiently as possible (and there is nothing inherently wrong with that) and the well-being of participants in research. The DoH, particularly in paragraph 11 but also in other places throughout the document, affirms that unless research constitutes ‘good science’ it is unethical. However, as already mentioned, paragraph 5 places an ethical onus on the doctor never to sacrifice the interests of the individual in the interests of science and society. At the same time paragraph 6 (and others) place an ethical duty on doctors to undertake research. Taking any of the paragraphs to an extreme while ignoring the other paragraphs risks either endangering the well-being of participants or placing catastrophic barriers in the way of medical advance, which has the potential also to rebound to harm the individuals. The process of independent ethical review (paragraph 13) and adequate informed consent (paragraphs 22–26) must serve to protect the participants. Ethics committees are charged with deciding what kind of control group is ethically justified in individual protocols and ought to do so in full appreciation of the ethical tensions described above.
So, despite the adoption of the note of clarification, there is considerable work to be done in clarifying in what circumstances placebo-controlled studies are ethically acceptable. It would be useful to see evidence-based guidelines like those developed for mood disorders  undertaken for a wide variety of conditions. This would greatly assist those designing research protocols and ethics committees in their required assessment of the risks and benefits (paragraphs 16–19). Of course, such guidelines, to be useful, would need to be frequently updated to take into account medical advances.
Even after carefully thought out debate it is likely that there will still be those who would wish to see the Declaration interpreted in a way that would place greater restriction on use of placebo . As Macklin cautions, ‘Two paragraphs (29 and 30) . . . remain controversial and would still be controversial if changed to meet criticisms’.
Paragraph 30: At the conclusion of the study, every patient entered into the study should be assured of access to the best proven prophylactic, diagnostic and therapeutic methods identified by the study In the most recent edition of their highly successful textbook, The Principles of Biomedical Ethics, Beauchamp and Childress make the following observation: ‘Until the 1990s, the paradigm for ethical analysis focused on the risks and burdens of research (emphasis theirs) . . . and on the need to protect potential and actual research subjects from harm, abuse, and exploitation. . . . However, a paradigm shift recently occurred . . . As a result, justice as fair access to research (both participation in research and access to the results of research) became as important as protection from exploitation’. The most recent revision to the DoH, in particular paragraph 30 but also reflected in paragraph 19 (see below), would seem to bear this out. Nicholson asserts regarding paragraph 30 that ‘this is potentially the most far-reaching of all the changes to the Declaration’. Concerns about the implications of paragraph 30 have led to the WMA assembling a Workgroup to consider either an amendment to the paragraph or the addition of a note of clarification. The report of the Workgroup was presented to the Council meetings which preceded the most recent WMA General Assembly (10–14 September 2003 in Helsinki) and it was decided that no amendment or clarification would be undertaken but that the Workgroup's deliberations would be continued and consultations widened [35, 36]. Although this decision has drawn criticism , we argue that it represents a ‘sensible and measured’ approach to the situation .
The debate centres around the issue of what happens to patients in a trial once the trial is over. Capron has characterized this as an example of the larger question ‘who owes what to whom and why?’ In contrast to paragraph 29, where the critical question has been characterized as ‘are participants worse off in the trial than they were before the trial?’, the question here is ‘are participants worse off after the trial than they were during the trial?’. Those who see paragraph 30 as imposing too great a burden on researchers emphasize the benefits which accrue to patients during a trial where there was no access to treatment beforehand and assert that nothing is lost (compared with the pretrial situation) if, at the end of the trial, the status quo resumes and access is lost. In contrast, those supporting paragraph 30 as it is emphasize the additional trauma and distress caused to patients who, after treatment for a duration of the trial, learn what is possible for them, only to be deprived of access when the status quo resumes post trial. They argue that these patients are, indeed, worse off after the trial than they were before. There is no easy way towards consensus on this and the WMA press release regarding the DoH following the 2003 General Assembly noted ‘sharp differences of opinion over how to protect human participants in medical research’.