2. The classical systematic review

A systematic review answers a specific empirical research question. The question has a set form: it must refer to the population, the intervention or exposure, the comparison and the outcome. One question might be: ‘[POPULATION] In humans aged 18 or over, [INTERVENTION] does smoking more than 12 cigarettes a day, [COMPARISON] versus not smoking, [OUTCOME] increase the lifetime incidence of lung cancer?’ The systematic review answers the question based on the entire literature and tells us how confidently we should accept the answer; alternatively, it concludes that the question is not yet settled: further research is needed.

To write a systematic review, one first conducts a search designed to identify all the relevant publications. This is called an ‘exhaustive search’. The term is misleading, as a search designed to be exhaustive may fail to be so. Publications are considered relevant if, and only if, they meet pre-decided and explicit conditions for inclusion. The search should be reproducible and the written review should describe it in sufficient detail to enable its reproduction. The result of this search should be the entire literature that addresses the research question.

Next, one extracts from each selected publication the conclusion that it draws (its answer to the research question) and other data. One assesses the degree to which we should believe the answer/conclusion; sometimes, and as recommended when writing a systematic review that will inform clinical guidelines, one grades the answer/conclusion to reflect this degree.⁵ A particular study might give us strong reason to believe that smoking slightly increases cancer incidence; another might give us weak reason to believe that smoking greatly increases cancer incidence. The systematic review then concludes whether smoking increases cancer incidence, based on all the relevant publications, taking into account the extent to which we should accept each individual publication's answers/conclusions.

The primary purpose of the systematic review is to improve decisions: to enable decisions that are maximally informed and minimally biased. The need for systematic reviews arises because, for most empirical questions, the relevant literature is extensive. Decision-makers, for example clinicians and policymakers, lack the time and skills to retrieve all the relevant literature, appraise it and synthesize it in order to identify the relevant literature's all-things-considered answer to the research question. Without a systematic review, decisions are likely to be based on a subset of publications, which may not be representative of the whole literature, and the risk arises that the reviewers will consciously or unconsciously ‘cherry-pick,’ that is, select publications best supporting their views. Systematic reviews undertake the substantial project of identifying, assessing and synthesizing the literature using techniques that minimize bias, and they present the synthesis in a format accessible to decision-makers.⁶ Unsurprisingly, systematic reviews are used increasingly to make clinical decisions, write clinical guidelines and set research agendas.

We will call this the ‘classical systematic review’ to distinguish it from later applications. Two features make it classical: the fact that the review answers an empirical question of the specific form we described, and the process used to answer the question.

3. Newer applications of the systematic review

Systematic evaluation and synthesis emerged in the late 1970s in social science to address questions, for example, about the relation between class size and pupil achievement, and later spread to medicine and various fields including qualitative health research.⁷ Recently, one of us argued that the systematic review should be transferred to empirical bioethics, and proposed a model.⁸ Furthermore, Laurence McCullough and colleagues have argued that reason-based bioethics needs systematic reviews.⁹ They outlined a model and illustrated it by conducting a systematic review of a seven-article literature. We will call this outline model the ‘McCullough Model’. To our knowledge, this is the only model for systematic reviews of reason-based bioethics.

There are challenges to each of these newer applications. One of us has shown that a systematic review of empirical bioethics needs to address a different form of research question from the classical systematic review, and to use a modified search strategy.¹⁰ Similarly, Laurence McCullough and colleagues have highlighted challenges to searching for reason-based bioethics literature.

The rationale given for each new application is analogous to that for the classical systematic review. For example, McCullough and colleagues express concern that doctors may lack the skills to retrieve all of the relevant reason-based literature and to assess its quality, and that consequently, if there are no systematic reviews of reason-based literature, doctors' ethically relevant decisions are likely to be biased.¹¹

We will next identify key features of the McCullough Model. Then, we will argue that systematic reviews based on it may comprise an insufficiently informative, and misleading, brief for decision-makers; and they are also of limited use to reason-based bioethicists, and to empirical scientists conducting research relevant to policy.

1. Our objections to the McCullough Model

If a literature omits reasons or inadequately presents reasons that are potentially strong, its all-things-considered ethical conclusion will be uninteresting. Objections arise even if a literature is comprehensive and uniformly excellent. The all-things-considered conclusion may be only one of several conflicting but individually reasonable conclusions that could be drawn from that set of reasons. The same set of reasons might better support a different conclusion given a change in the context or weighting of the reasons. In any case, the reasons underlying the literature's all-things-considered conclusion must be scrutinized prior to making clinical decisions or policy. Thus, we hypothesize, McCullough Model systematic reviews of reason-based bioethics potentially mislead decision-makers, who may fail to appreciate the previous points.

McCullough and colleagues might reply that a systematic review must include disclosures about factors that limit the degree to which we should believe its all-things-considered conclusion. They indeed concede that the burden of proof created by their review's all-things-considered conclusion is ‘modest’ given the reviewed literature's limitations¹⁹ and, in their discussion, they reject a key reason given in the reviewed literature for the conclusion that the majority of the reviewed publications draw. Nevertheless, we fear that policymakers may ignore or discount the disclosures and, in any case, may choose to accept the all-things-considered conclusion for lack of an alternative conclusion. As we will argue below, a full set of published reasons for or against the view in question is more useful than a conclusion that is inadequate – whether due to limitations of the literature reviewed or of the systematic review's methodology.

We also object to the McCullough Model's measure of the quality of reasoning. First, we present a superficial objection, which is that whether a specific instance of reasoning is adequate depends partly on the type of publication in which it appears: (1) greater detail is required in a philosophy article that scrutinizes a specific argument than in a policy report that considers many arguments for a specific view. (2) A good philosophy article can consider just one argument, but a good policy article must consider a range of reasons or else justify why it does not. (3) A good philosophy article need not present evidence and discuss its quality – it may proceed, for example, on the basis of hypothetical scenarios – but a good policy article must do both. (4) For many bioethics questions, the discussion occurs in various fields and genres, such as philosophy articles and policy pieces.

This objection is an unfair objection to the specific systematic review presented by McCullough and colleagues because the only publications it covers are articles in reason-based bioethics. However, the McCullough Model is intended to apply to reason-based bioethics in general. The objection is nonetheless superficial because it is possible to develop criteria for the quality of different types of publication: a publication would then receive its score of 0, or 1 according to the relevant criteria.

The remaining objections are not superficial. The second objection is that independent coders are likely to score differently the quality of reasoning in the same publication. The McCullough Model intends one number (0, or 1) to score both validity and soundness. However, it is unclear, and is left unclear, what score to assign to a valid argument with false premises: how good should we consider such an argument? Similarly, it is unclear, and is left unclear, how to score the quality of reasoning in a publication when it presents some reasons competently but not others. This objection can be avoided to some extent by measuring quality other than by using a summary measure; we note that the Cochrane Handbook, which is the main handbook for writing classical systematic reviews, ‘… recommend[s] against the use of scales yielding a summary score’.²⁰

Nonetheless, a third objection, about the measurement of quality, remains. It may take considerable research to assess whether or not an argument's premises are all true: the truth-value of factual premises may be as yet unknown, and there may be reasonable disagreement about moral premises.

We therefore conclude that the McCullough Model's measure of the quality of reasoning in a publication is an inadequate indicator of the extent to which we should believe the publication's all-things-considered ethical conclusion. In any case, even if one found a satisfactory measure of quality, and used it to replace the Model's measure of quality, our (in our opinion, decisive) objections to the fact that the research question is supposed to be an ethical question would remain.

2. Key features of our alternative model

This section outlines our alternative model, which is a model for writing systematic reviews of reasons.

3. Why systematic reviews of reasons?

One might object that if a literature omits relevant reasons when addressing a question or presents potentially strong reasons inadequately – as do some literatures, in particular, young literatures – a systematic review of reasons will be as uninteresting as one of conclusions. We will next explain, using illustrations from our particular systematic review, why our proposed systematic reviews of reasons are useful even when the literature reviewed is uninteresting, and why they are, in general, less vulnerable to the objections we have advanced against the systematic reviews advocated by McCullough and colleagues. We will also explain why there is a need for systematic reviews of reasons, even though there are already informal reviews of reasons and reports promulgated by official bodies, which also offer decision-makers an intermediate option between reading all the reasons themselves and a summary of select, weighted reasons.

Before we proceed, it is crucial to point out that a systematic review of reasons cannot be the only item in a decision-makers' brief. Such a review does not help decision-makers to select the strong reasons from the published ones or to weight the strong reasons. Depending on the scope of the literature reviewed, it may also fail adequately to represent the arguments and views advanced by industry, regulators, research participants and other stakeholders. The brief should also include, among other items, a distillation of the strong reasons and their implications, and guidance on weighting these reasons that indicates, as appropriate, the possibility of drawing from these reasons, severally and jointly, alternative, individually reasonable conclusions. Yet, as the arguments below imply, distillation should occur only after a systematic review such as this one has captured all the published reasons and how they have been used.²⁶ The systematic review remains an essential ingredient of a decision-makers' brief for the reasons above and because of decision-makers' need to check the distillation and its legitimacy.

By way of introduction, we will need key results from our systematic review: this identified 36 broad and 235 narrow types of reasons, based on a wide variety of moral, legal and practical considerations. None of the 75 publications included in the review reported more than 22 broad and 59 narrow types, even though the publication that reported these was an excellent informal review.²⁷ The mean informal review reported only 24 narrow types of reason, compared to our review's 235.²⁸ Among the publications were major reports by prominent bodies such as the US's National Bioethics Advisory Commission²⁹ and the UK's National Council on Bioethics,³⁰ which both failed to mention major families of reasons identified by our systematic review, as well as frequently published versions of specific reasons. The most-mentioned reason was avoiding exploitation of participants, their countries or communities (96 mentions). For this and many other reasons, publications differed about the reason's interpretation, implications and/or persuasiveness. Publications differed also on the facts about the costs, feasibility and legality of PTA.

These results are from a single – indeed, the only – systematic review of reasons. We hypothesize generally that: (1) current informal reviews of reason-based bioethics and reports produced by official bodies do not identify all the published reasons. (2) For any question in reason-based bioethics, a systematic review of reasons will identify a broader variety of published reasons, alleged implications of reasons and attitudes taken to reasons than searches typically conducted by those writing in reason-based bioethics or by the authors of official reports.

If these hypotheses are true, what is their significance? Policymakers considering, for example, whether or not to require that trial participants are ensured access to the trial drug after the trial must identify all the relevant reasons why post-trial access should, or need not be ensured to participants.³¹ Bioethicists seeking to understand when and why post-trial access should be ensured must do the same. As mentioned above, a systematic review of reasons should include a digestible visual presentation of all the published reasons and their alleged implications for the research question. The differences between our systematic review on the one hand, and informal reviews and reports on the other hand, suggest that such a presentation is currently the best tool that decision-makers and bioethicists have to ensure they do not overlook possibly relevant reasons and their alleged implications. In particular, a systematic review of reasons brings to light reasons that have been infrequently published and only inadequately presented, such as – in the case of our systematic review – rights-based reasons. If the publishing world works well, these less prominently published reasons will be a distraction. However, the most prominently published ones may simply be the best publicized, perhaps due to conflicts of interest that induce authors to endorse weak reasons and to ignore or reject strong reasons. In the absence of evidence either way, a systematic review based on a well-constructed search is crucial, in that it gives even such reasons an equal voice prior to selection and appraisal of the relevant reasons.³²

Some reasons given in published discussions of the research question may, in fact, not be strong reasons or perhaps not even relevant to the research question. Although decision-makers and bioethicists need the strong reasons, a systematic review of reasons is still crucial. First, it is a minimally based starting point for identifying the strong reasons. Second, imagination, knowledge of reasons mentioned in other literatures, and analytical skills are not always optimal and, in any case, are sometimes insufficient to identify all the relevant reasons. Thought is stimulated by discovering differences between publications regarding the formulation, implications or persuasiveness of reasons, and differences in how often they are mentioned. Because the readers of a systematic review of reasons do not have to reinvent the wheels or, at any rate, need to invent fewer wheels, we surmise that they are less likely to miss relevant reasons and likelier to formulate relevant reasons plausibly and to understand their implications.³³

There are further ways in which a systematic review of reasons is directly relevant to decision-making. Our systematic finding that the most-endorsed reasons included ones used just for ensuring PTA and others just for the view that PTA need not be ensured, provides a salutary warning to decision-makers against considering only reasons on one side of the case. Other systematic reviews of reasons are likely to provide similar warnings. Also, a systematic review may also imply that a procedural solution will be needed to set guidelines.³⁴ Our systematic review suggested that reason-based bioethicists may fail to reach agreement, in that it showed that the most-endorsed reasons included ones used just for ensuring PTA, and others used just for the view that PTA need not be ensured.

Possible users of systematic reviews of reasons such as ours include the members of a committee, within a pharmaceutical company or government agency, charged with developing guidelines on post-trial access to trial drugs. NS's essential preparation for drafting the guidelines on post-trial access to trial drugs in collaboration with the UK's National Research Ethics Service included writing a systematic review of reasons why such access should, or need not be ensured.³⁵ Such a review would also be useful for designers of research protocols such as researchers, research sponsors, and Contract Research Organizations, particularly given the fact that current guidance on post-trial access is inconsistent, ambiguous and silent on many aspects of post-trial access.³⁶ The public needs these other parties to read systematic reviews of reasons.

We surmise that a systematic review of reasons is also indirectly relevant to decision-making, in that it is a particularly good tool for inspiring and guiding empirical and conceptual research that improves the information base of decisions.³⁷ Various examples follow. First, such a review may identify moral views that differ between publications or indeed trends in attitudes taken to a specific view. Our review identified a trend towards rejecting the view that avoiding exploitation requires post-trial access. This view was common in the early literature and assumed by prominent guidelines written during this early stage (see e.g.³⁸). Thus, our review suggested the need to scrutinize these guidelines. Second, a systematic review may identify, as did ours, reasons that have been presented only inadequately. Third, as mentioned, identifying all the different published formulations of a reason will better enable the selection of the relevant one(s). Fourth, such a review can also show, as did ours, that many authors, who used specific interpretations of a reason, did not express awareness of some or any other interpretations, or relevant literature in the same and other fields, and that some of the reasoning is incomplete or appears invalid. Fifth, the stark differences that our systematic review identified in authors' use of terms for key concepts such as exploitation and reciprocity suggest that bioethicists should investigate how to distinguish such concepts. In some cases, the systematic review will simplify this task into that of deciding which of the entries in its list of very finely individuated reasons should be considered to be reasons of the same type.

Sixth, a systematic review of reasons may identify, as did ours, empirical assumptions that differ between publications, for example, assumptions about the cost, feasibility or legality of ensuring post-trial access to trial drugs. The empirical testing of these assumptions – which will require prior operationalization of appropriate concepts, e.g. of cost – will improve the information base of decisions, whether in practice or policy. For example, prominent international research ethics guidelines state that payments to research subjects should not be so large that they induce individuals to participate in research against their better judgment.³⁹ Many ethics commentators argue that payments to research participants that do more than reimburse their costs should not be permitted because they may lead individuals to underestimate or undervalue the risks of participating.⁴⁰ Limited evidence now shows that increasing payments to research participants is not associated with reducing the ability to assess risk.⁴¹ It may be that the case for the current guidelines on payments to research subjects remains compelling, even if the reason (apparently incorrectly) based on the effect of payments on risk assessment is removed from the case. Nonetheless, the legitimacy of a guideline based on a falsehood is questionable. A systematic review seems best placed to identify empirical assumptions that need testing.

Also, our systematic review aids bioethicists' identification of reasons relevant to determining whether there are other obligations, such as ensuring PTA to trial results,⁴² providing ART to participants who seroconvert in HIV vaccine trials,⁴³ or providing participants with care not necessary to conduct the research, to prevent or address research-related injury, or to fulfil morally optional promises.⁴⁴

For all the above reasons, a systematic review of reasons remains helpful even if some of the published reasons turn out to be irrelevant, or the implications of the reasons for the research questions have been misunderstood. We surmise that a systematic review of a young literature may hasten its maturation by identifying shortfalls and areas for further research. We surmise also that the main value of a systematic review of a mature literature is to identify the major trends and reversals in what will typically be a literature too vast, fragmented and complex for most decision-makers to collect and appraise. Possibly, the direct relevance to decision-makers of a systematic review of reasons increases as the field matures. As with any systematic review, decision-makers may lack the time to wait for one to be written. While ours was extremely time-consuming, the process should be speedier now that the methodology has been developed. That methodology, which we report elsewhere as a step-wise process,⁴⁵ could be further automated, increasing its value to decision-makers. If the methodology is also applied to write reviews within large fields such as law or economics, the incentive to automate the process will increase and (amortized) cost decrease.⁴⁶

The limitations of the literature revealed by our systematic review suggest that a systematic review that answers the question ‘should PTA be ensured to former participants in a clinical trial?’, as the only alternative model of SRs of reason-based literature proposes,⁴⁷ would not be useful to decision-makers, because the answers given by many publications are unreliable. For the same reason, such a review would not place a burden of proof on those who disagree with the most common answer given in the literature to this question. Furthermore, all-things-considered conclusions may distract decision-makers, given that different weightings of values and different contexts could yield different conclusions. Because the aim of systematic reviews of reasons is not to present the literature's all-things-considered conclusion, they avoid giving decision-makers the impression that this conclusion is the correct one. Furthermore, a systematic review of conclusions that does not extract data on reasons will be uninteresting to philosophers, whatever the completeness or quality of the literature reviewed, as they will still need to ‘go back to first principles’: to consider all the relevant reasons, their implications for the specific issue and their relative weight.

A merely pragmatic consideration in favour of systematic reviews of reasons is that there is no need to measure the quality of reasoning: a complete answer does not require this. However, a key feature of the classical systematic review is ‘an assessment of the validity of the findings of the included studies’.⁴⁸ If the relevant finding is that a specific reason has been published, the publication of the reason suffices to establish the validity of the finding. Thus, our model is less vulnerable to objections to the measurement of the quality of reasoning.

This consideration is superficial because, as noted, decision-makers will also need information on quality. Worse, one might object that decision-makers may confuse the most commonly-presented reasons with the strongest reasons, just as they may confuse the literature's all-things-considered conclusion (presented by a McCullough Model systematic review) with the truth. The explanation of why the most commonly-presented reasons may fail to be the strongest ones presumably varies with context, as mentioned above. Whatever the explanation, this objection threatens our view that, with regards to reason-based bioethics, systematic reviews of reasons are superior.

In reply, particularly because our systematic review showed that publications presenting the same common reason differed regarding its implications and persuasiveness, we consider it unlikely that readers would assume that the more commonly-presented reasons are the stronger ones. However, we concede that a common reason for a specific conclusion might be commonly presented, always endorsed, yet invalid. We therefore propose that systematic review methodology should be improved to enable it to identify possible conflicts of interest,⁴⁹ and that – in the absence of a measure of quality – systematic reviews should warn readers against assuming that the more commonly presented reasons are the stronger reasons.⁵⁰ Furthermore, research should also be conducted to understand whether or not such a warning suffices to prevent readers from making this assumption. The results should be used to assess whether the risk that invalid reasons will mislead policymakers is more, or less, serious than the risk that policymakers will fail to take into account potentially strong reasons that were excluded from the review because the literature presented them only as invalid reasons.

If it turns out that such a warning does not suffice, we recommend writing different systematic reviews for bioethicists versus policymakers. Bioethicists should be given all the published reasons, because this furthers their interest of identifying all the published reasons and because they are trained to assess reasons. Policymakers should, instead, be given a subset of the published reasons, that is, the strong reasons; if necessary, the data on how often the (strong) reasons were presented should be withheld. It may be necessary to construct a measure of quality; and this poses various challenges. For example, it will be necessary to consider whether the review should include only valid reasons, or the narrower set of reasons that are valid and sound. It will also be necessary to decide whether to include a reason when there is legitimate disagreement about its validity and soundness. It is important to note that, even in the absence of a measure of quality, the systematic reviews we advocate can be coupled with publications that discuss the quality and implications of the reasons identified by the systematic review.

4. Why systematic reviews?

As mentioned above, the rationale given by McCullough and colleagues for writing systematic reviews of reason-based bioethics is the same as in the case of classical systematic reviews, namely, that systematic reviews make possible maximally informed, minimally biased decisions. We add here that society has an interest in clinical decisions and policies that are maximally informed and minimally biased. Furthermore, various clinical decisions, policies and policy tools that appear to be based on solely factual considerations are in fact influenced by ethical considerations – in some cases ethical judgments that are implicit and unjustified.⁵¹

The rationale for writing systematic reviews of reason-based bioethics may, in some respects, be stronger than in the case of clinical epidemiology: reason-based bioethics is particularly contentious, and the reason-based bioethics literature is especially difficult to retrieve, not only for medical professionals, who are not trained to retrieve it, but also for reference librarians. This is because the literature is fragmented across different fields, a greater proportion of reason-based bioethics literature appears in books, and the literature has been inadequately indexed. A systematic review's complete reference list is an important aid for both decision-makers and scholars, obviating the need for hundreds of hours of hard-to-find, skilled labour to retrieve relevant publications. Furthermore, medical professionals may be better appraisers of clinical epidemiology than of reason-based bioethics. Also, a systematic review's summary of the positions taken by individual publications (reasons endorsed and any final conclusion reached) enables decision-makers to grasp key publications immediately without trawling through typically large amounts of text. Both these additional items – the complete reference list and summary of positions – save decision-makers time and greatly reduce the scope for error. Last, in the case of our particular systematic review at least, guidance is inconsistent and incomplete, and so an additional aid to decision-making is necessary. Admittedly, a systematic review makes more demands on decision-makers than an informal review, but its advantages justify these demands. Unlike the informal reviews and reports covered by our systematic review,⁵² our systematic review identifies all the empirical or ethical points of dissent and so sets the research agenda. Our systematic review's identification of stark differences between publications, for example, on matters of fact, shows that multi-disciplinary research is urgently needed to understand the costs,⁵³ feasibility, legality and effects of requiring PTA, and any tendency of PTA offers to potential participants to boost recruitment or to reduce the quality of informed consent. Its identification of inter-publication differences in the formulation and implications of moral reasons suggests that further reason-based research is needed.

One might object that systematic reviews of reason-based bioethics are unnecessary to enable maximally informed, minimally biased decisions: informal reviews suffice. According to this view, it suffices to sample qualifying publications until no new reasons emerge from additional publications, a point that qualitative researchers call ‘theoretical saturation’. Arguably, this is standard philosophical practice, although philosophers neither use standard techniques to maximize the chance of retrieving all relevant material nor report reproducible searches.

We have already argued why a review of reasons must be based on an exhaustive search and explained the direct and indirect relevance to decision-making of a list of all the published reasons. We add here that appropriate aggregation of exhaustive data on reasons, and appropriate data on publications, can be combined to identify bias due to the fact that some arguments are published more frequently than others, in higher impact publications, or across a broader variety of journals. Our systematic review suggested that all these sources of bias were present. A systematic review of reasons has, due to its methodology, a much lower risk of bias than the individual publications it covers, which decision-makers would otherwise use to identify relevant reasons (or should use, in the absence of a systematic review of reasons). Bias minimization is crucial: most bioethics issues are highly contentious and stakeholders often have unequal power.

Furthermore, while much of the value of a systematic review of reasons lies in its list of all the published reasons, two other components are of direct use to decision-makers. One is its complete reference list, as mentioned previously. Particularly when the literature is mature and fragmented, decision-makers are likely to lack the time and skills to retrieve all the publications. The other is the table that summarizes the position taken by each included publication: the reasons it endorsed, and the all-things-considered conclusion reached. When a publication is long and difficult, such a table can greatly help decision-makers who wish to understand the positions taken by a specific publication. A systematic search seeks to be not only exhaustive but reproducible. Describing a reproducible search enables readers to assess whether the search is likely to have retrieved all relevant literature, and thus helps them reach the appropriate degree of confidence in the systematic review's findings. The analysis provided by a systematic review written according to our model should interest policymakers more than the limited analysis given in informal reviews. Only the former is designed to identify differences in the alleged implications of a reason or differences in how persuasive different publications find the reason. Furthermore, only the former provides rankings of reasons in terms of how frequently they were mentioned, endorsed, or rejected.

Thus, paradoxically, a systematic review of reasons is better placed than a systematic review of conclusions (as well as better placed than an informal review) to show both when we should take the field's all-things-considered conclusion seriously and when that conclusion places a burden of proof on those who hold different views.

For very helpful comments, the authors would like to thank Bioethics' anonymous reviewers, Penney Lewis, Marcel Mertz, Reuben Thomas, Leif Wenar and the audience of the LABTEC (London & Brighton Translational Ethics Centre) colloquium at King's College London. Neema Sofaer was supported by a Research Fellowship in Biomedical Ethics from the Wellcome Trust, grant number 088360. Daniel Strech was partly supported by a grant from the German Research Society (DFG), grant number STR 1070/2-1.

Neema Sofaer is a Wellcome Trust Research Fellow at the Centre of Medical Law and Ethics at King's College London. She has a three-year grant from the Wellcome Trust to conduct research on the ethical and legal aspects of post-trial access to trial drugs, healthcare and information. She obtained a double-first BA Honours in English and Philosophy, and a M.Phil in Philosophy, both from Trinity College, Cambridge. After studying philosophy and classics at Harvard as a Kennedy Memorial Scholar, she received a PhD in Linguistics and Philosophy from Massachusetts Institute of Technology, and then returned to Harvard as a Research Fellow in Ethics and Health to train in social scientific methods. Her research focuses on the ethics of research and of resource allocation.

Daniel Strech is Assistant Professor of Medical Ethics at Hannover Medical School and a member of the advisory board for the German Network of Evidence-based Medicine (DNEbM). He holds a PhD in Philosophy and a MD degree, both from the University of Düsseldorf. His research focuses mainly on the ethics of public health and of clinical research; he also works on how to improve transparency in medical decision-making and how to make such decision-making more systematic.