Journal of Agricultural and Environmental Ethics (2007) 20:53–63 DOI 10.1007/s10806-006-9022-9 Springer 2006 KATE MILLAR, ERIK THORSTENSEN, SANDY TOMKINS, BEN MEPHAM, and MATTHIAS KAISER
DEVELOPING THE ETHICAL DELPHI (Accepted in revised form November 9, 2006)
ABSTRACT. A number of EU institutions and government committees across Europe have expressed interest in developing methods and decision-support tools to facilitate consideration of the ethical dimensions of biotechnology assessment. As part of the work conducted in the EC supported project on ethical tools (Ethical BioTA Tools), a number of ethical frameworks with the potential to support the work of public policy decision-makers has been characterized and evaluated. One of these potential tools is the Delphi method. The Delphi method was originally developed to assess variables that are intangible and/or shrouded in uncertainty by drawing on the knowledge and abilities of a diverse group of experts through a form of anonymous and iterative consultation. The method has hitherto been used by a diversity of practitioners to explore issues such as technology assessment, environmental planning, and public health measures. From the original (classical) Delphi, a family of Delphi-related processes has emerged. As a result of the evaluation of the various Delphi processes, it is proposed that the classical method can be further developed and applied as a form of ethical framework to assist policy-makers. Through a series of exercises and trials, an Ethical Delphi has been developed as a potential approach for characterizing ethical issues raised by the use of novel biotechnologies. Advantages and disadvantages of the method are discussed. Further work is needed to develop the procedural aspects of the Ethical Delphi method and to test its use in different cultural contexts. However, utilizing an ethical framework of this type combines the advantages of a methodical approach to capture ethical aspects with the democratic virtues of transparency and openness to criticism. Ethical frameworks such as the Ethical Delphi should contribute to better understanding of and decisionmaking on issues that involve decisive ethical dimensions. KEY WORDS: ethical frameworks, biotechnology, Delphi method, GM fish, Ethical Delphi
1.
INTRODUCTION
There is increasing interest in the development and application of decisionsupport tools to facilitate bioethical analysis in the field of agri-food production. However, there is a limited number of frameworks available that can support public policy decision-making. In the literature on practical ethics and on ethics in public decision-making (within the agri-food sector and other sectors), there are reports of a number of emergent frameworks.
54
KATE MILLAR ET AL.
Several frameworks have been proposed, but few have been adequately developed and tested to determine their applicability as public policy decision-support tools (Kaiser et al., 2004b). The use of, need for, and expectations surrounding decision-making frameworks are diverse. There is no single generic framework that can be used to assess and manage ethical issues throughout the lifecycle of a biotechnology development, or that would be adequate for all kinds of technologies. As part of the work conducted for the Ethical Bio-TA Tools project,1 a research team consisting of two research groups (one in the UK and one in Norway)2 examined the existing approaches that fall into the category defined as a ‘‘tool’’ for practical ethics. Thirteen decision-making frameworks were identified (Kaiser et al., 2004a). These decision-making frameworks were reviewed on the basis of properties deemed essential in well-functioning decision-making frameworks. Moreover, it is acknowledged that different frameworks are needed at different stages in the research and technology development (RTD) process. A review of these frameworks suggested that a number of the tools are particularly suited, or could be developed further, to act as tools to assist policy-making. These tools are structured frameworks with the following properties: ample substantive ethical content; good opportunities to facilitate transparent decision-making processes; and that include a multiplicity of (stakeholder) viewpoints, ethically relevant information, ethical arguments, and values at stake. One of the methods that the research team considered merited further assessment and development was the classical Delphi method. The Delphi method was designed to combine the knowledge and abilities of a diverse group of experts to the task of assessing variables that are intangible and/or shrouded in uncertainty. The technique has been used to explore a variety of issues such as technology assessment and forecasting (TA and TF), environmental impact assessment (EIA), and public health measures. This widespread use has led to the development of a family of Delphi-related processes. As a result, the classical Delphi method showed potential for further development and application as an ethical framework. Consequently, the research team, through a series of exercises and trials, has developed and assessed the Delphi method as a potential approach for characterizing ethical issues raised by the use of novel biotechnologies. This form of Delphi has been initially designated as an Ethical Delphi. This paper discusses the fabric of the Ethical Delphi method, highlights a 1 www.ethicaltools.info. This paper presents results of the project Ethical Bio-TA Tools as funded by the European Commission, under FP5, Quality of Life Programme. 2 Centre for Applied Bioethics (Nottingham, United Kingdom) and National Committee for Research Ethics in Science and Technology (Oslo, Norway).
DEVELOPING THE ETHICAL DELPHI
55
number of trials that were conducted, and reviews a number of methodological issues. In particular it discusses the role of this method as a substantive and procedural framework for the assessment of technology. One of the key challenges for the assessment of this potential framework, the Ethical Delphi, was to identify to what extent this method does, or could through further methodological development: (i) provide transparency by clarifying the basis of decision-making processes; (ii) allow the inclusion of a multiplicity of (stakeholder) viewpoints; (iii) facilitate systematic inclusion of ethically relevant information; (iv) enable systematic inclusion of ethical arguments; (v) facilitate the explicit inclusion of values at stake. 2.
THE DELPHI METHOD AS AN ETHICAL FRAMEWORK
The classical Delphi method, first developed by the RAND Corporation in the 1950s, was designed to combine the knowledge and abilities of a diverse group of experts to evaluate developments that were deemed to lie outside traditional scientific assessments, because they involve elements of judgment and dealt with uncertainties of various kinds. The method was further structured in such a way as to eliminate, as much as possible, psychological factors (e.g., deferring to dominant individuals, the influence of individualsÕ professional status on group discussions, etc.) that can appear in committee processes. The objective of most Delphi applications is the creative generation of ideas, exploration of future scenarios, improved data collection, and informed decision-making. Application of the Delphi method, in its various forms, can greatly assist policy makers to improve creativity and arrive at decisions when information is incomplete, or the validity of the information is disputed (Adler and Ziglio, 1996). In order to develop a Delphi method that was applicable as an ethical framework, the research group examined the diverse ways in which the classical Delphi method had been used (Hasson et al., 2000). Drawing on the classical Delphi method, an Ethical Delphi is proposed as an iterative process for exchanging views and arguments, including value-based arguments, between a defined group of experts (Rauch, 1979). The method is structured around the notion of a ‘‘virtual’’ committee where the exchange of ideas is conducted anonymously and remotely through a series of opinion exchanges. The committee is convened by the Delphi organizers and represents an ad hoc group or committee that will communicate only through electronic media without ever meeting in person. In each successive round, the panel is asked to comment on the viewpoints expressed by other members in the previous round. From a decision-theoretic perspective, it is
56
KATE MILLAR ET AL.
the combination of iterative rounds and successive embedding of arguments that accounts for the added information value of the end result. The Ethical Delphi is used to map the ethical considerations that experts believe are pertinent and significant. It indicates the extent of agreement as well as drawing out differences of expert opinion on a given topic. Though the Ethical Delphi is based on the classical formulation of the Delphi method, it differs from the latter in that it does not look for consensus on future actions/developments as its target. The presupposition that the end result should be consensus across the panel of experts has been criticized on several grounds. These criticisms concern the following: the social framing of the notion of consensus and whether this is desirable; the lack of precision when defining ‘‘consensus’’ (Crisp et al., 1997); and the level of consensus that might be thought acceptable or whether consensus is in fact always desirable (Williams and Webb, 1994). It has been argued that consensus is not relevant outside of the classical Delphi, because dissent as expressed in the findings of a classical Delphi may be interpreted as a sign of competing scientific paradigms, not just in terms of factual disputes, but also in terms of interpretational dissent and conflicting value judgments. An Ethical Delphi can be used to characterize and map the ethical issues raised by the use of novel technologies. This is most effectively achieved by use of a scoring system (e.g., using a Likert scale). Consequently, one of the benefits of the Ethical Delphi is the combination of scoring and reasoned arguments, where it is possible to see the ‘‘importance’’ or ‘‘significance’’ of an issue and the relevant arguments. Building on the literature surrounding the use of the classical Delphi method, it is suggested that the Ethical Delphi method may be usefully employed by a number of groups to explore ethical issues raised by the development and use of defined technologies. The tool is particularly relevant for use by (i) government advisory or regulatory committees and (ii) non-governmental organizations. The Ethical Delphi method would appear to be particularly valuable when all or the majority of the following conditions are present: • • • • • •
Expert input is required for policies under review or development, Issues are uncertain, controversial, and complex, Judgment and weighing of arguments is essential, Many and diverse, research communities and stakeholders have concerns, Outcomes from the process should have an impact on several issues, including future policy-making, There is a need for cross-sectoral scientific debate.
In terms of its context for use, an Ethical Delphi method may be used to:
DEVELOPING THE ETHICAL DELPHI
• • • • •
57
Identify the diversity of expert value judgments on the use of technology, Identify divergence and convergence in expert opinion, Encourage ethical reflection, Provide a rational basis for ethical decision-making, Clarify the basis of disagreements and highlight relevant values.
An Ethical Delphi can map expert opinion on the ethical dimensions of the use of a novel technology. The tool highlights issues, as well as divergence and convergence of views and values. However, the method will not provide decision-makers with final judgments or overall opinions. The selection of participants is a key stage of the process. It is, therefore, important that participants: (i) feel directly involved in the problem of concern; (ii) have pertinent information to share; (iii) represent a sufficient diversity of competencies and value commitments, (iv) are motivated to include the Delphi task in their schedule of competing tasks; and (v) feel that the aggregation of judgments of a respondent group will include information which they value and to which they would not otherwise have access. The selection of an adequate representation of competencies and value positions is pre-determined by the Delphi facilitators. The basis for the selection, such as the selection criteria, can be shared with the participants and users of the collected opinions. It is also possible to construct personal profiles for each of the participants so that the depth and breadth of the expert representation on the committee can be verified. An Ethical Delphi does not explicitly include ethical theories in its approach. However, ethical dimensions of the issue may well be drawn out through the initial context of the questionnaire – prepared by the organizers – and the statements or questions included in the initial stage. Values are, therefore, frequently identified by participants, then developed and analyzed in the subsequent rounds of an Ethical Delphi. An Ethical Delphi can be used to identify and map the ethical arguments and value judgments that are used by the expert participants when discussing specific issues. Participants are asked to score the level of importance of these issues, so that the perceived significance of specific issues can be used to highlight ethical judgments (as indicated through semi-quantitative scales). These quantitative outcomes indicate quantitative judgments. However, it should be noted that it is not possible to directly deduce from the data analysis the ethical acceptability of a proposed technology. The Delphi process will identify lines of argument, areas of convergence, and divergence. The outcomes from the process inform public policy decisionmaking in the same way that other lines of expert advice inform the decision-making process. The Delphi process, in a normative form, will aid
58
KATE MILLAR ET AL.
decision-making by, for example, identifying areas of expert consensus and key issues that may require further ethical analysis and reflection.
3. EXPLORING THE DELPHI METHOD THROUGH A CASE STUDY In order to explore a number of methodological issues relating to the method, in particular the substantive and procedural framework dimensions of this framework, a number of trials were conducted. The Ethical Delphi method was tested through two exercises, in each case using genetically modified (GM) salmon in fish farming as a case study. The two pilot studies, one in the UK and one in Norway, were conducted with individuals who had expert knowledge in areas related to GM salmon development and use, e.g., fish farming, GM salmon development, animal welfare, and aquaculture regulation. These experts were geneticists, ecologists, marine scientists, fish veterinarians, social scientists, and philosophers. They were drawn from government institutions, universities, industry, and NGOs.3 Their different backgrounds, institutional affiliations, and training aimed to ensure a diversity of views. The Ethical Delphi method followed the classical Delphi approach by using a series of questionnaires to elicit and develop individual responses to the problems posed and to enable the experts to refine their views as the groupÕs work progressed in accordance with the assigned task. As in the classical Delphi, the process was anonymous and confidential. This facilitated the exchange of information and ideas by enabling each participant to have equal input, preventing bias caused by status or dominant personalities, as the participants did not directly interact with one another. The respondents were, therefore, able to reason individually, which led to the identification of agreements and disagreements. The four key features of the classical Delphi still applied: anonymity; iteration; controlled feedback; and (statistical) aggregation of group response. In terms of the two pilot studies, questionnaires were distributed electronically to the pre-selected group of experts. Within the limitations of the participant numbers, the groups were selected by the research group based on a broad range of expertise and organizational backgrounds. Three rounds were conducted. The essential component of the use of the Ethical Delphi was set out in the structure of the initial questionnaire and the questions used to elicit ethical reflection on the particular topic. An initial open round was 3
These individuals had not been involved with other recent studies on GM salmon conducted by members of the Ethical Bio-TA Tools project.
DEVELOPING THE ETHICAL DELPHI
59
conducted in which participants were asked to list the ethical issues raised by the potential use of GM salmon for a number of potential interest groups. Participants were encouraged to include additional categories and comments. These data formed the basis of the first round questionnaire, and were followed by two further rounds. The participants in both pilot studies not only put forward valuable assessments of the issues raised by GM salmon, both in terms of their factual and value derived aspects, but also commented on the coherence, value, and convenience of using the method (Table 1). As might have been predicted, in both pilot studies there were notable differences in perspectives on the issues. The expert groups were selected at a national level and a number of the key areas discussed reflected different national positions on fishery issues. Thus the degree to which there was disagreement on a particular topic was just as important in this process as discovering points of agreement. The participantsÕ feedback highlighted the importance of clear and welldefined statements. Precise and specific statements draw out detailed discourse that can highlight specific value conflicts as well as the expression of common standpoints. Although, more open statements present risks of
Table 1.
Illustrative results from the Norwegian Ethical Delphi.
Issue 1: One of the future options for European fish producers is the production and marketing of GM salmon. Please list key issues that should be considered as part of any assessment of this option Consensus: SD < 0.50 Topic 5: Safety related to the effect of GM salmon on the environment (0.35) Topic 1: Consumer acceptance (0.45) Topic 11: Buying of GM salmon must be voluntary and separate from conventional salmon (0.45) Relative agreement: 0.50 < SD < 0.75 Topic 8: Secondary GM products such as GM feed and vaccines will be important for farmed fish (0.73) Moderate dissent: 0.75 < SD < 1.00 Topic 12: The GM salmon must be of high quality (0.90) Relative dissent: 1.00 < SD < 1.50 Topic 7: The effect of genetic modifications on the health of the salmon (1.03) Topic 4: Long-term health aspects of eating GM food (1.20) Dissent SD > 1.50 Topic 6: Alleviating food shortage (1.58)
60
KATE MILLAR ET AL.
being misinterpreted, these broader statements can open up the discourse. Through the process of disagreement and cross-examination of other expert positions, the series of iterations (i.e., effectively, ‘‘rounds’’ of the process) can draw out unexpected reflections or factual debates. For example, these can include discussions of intrinsic value that are not necessarily prominent in equivalent expert risk assessment deliberations. In addition, novel interpretations of data sets relating to the technology can stimulate informative discourses. The experience gained for the application of this method parallels similar findings in the literature, where early engagement of a diverse group of stakeholders with a broad spectrum of expertise can highlight issues of conflict and consensus for a given case (COGEM, 2005). It also reflects insights in relation to post-normal science and the need of extended peerreviews (Funtowicz and Ravetz, 1993). A manual to assist users in applying this method has been developed. This manual and details of the pilot studies can be accessed on the website www.ethicaltools.info. The manual sets out a review of the Delphi method and discusses the applicability of the method as an ethical framework. A full description of the method with additional information on application variability is also included. 4.
ADVANTAGES AND DISADVANTAGES OF THIS METHOD
The assessment of the method as ‘‘ethically sound,’’ the evaluation of existing knowledge of the classical Delphi method, and the trials conducted using GM salmon as a case study – all suggest the potential value of the Ethical Delphi as an ethical framework. From the assessment work and the feedback from the pilot studies that have been conducted by the research groups, a number of potential advantages and disadvantages of this method have been identified. 4.1.
Advantages
The method: (i) Allows the group, collectively, to ‘‘set the agenda’’; (ii) Is egalitarian, i.e., it prevents (or at least, inhibits) ‘‘peer domination; because the anonymity of the participants, and the constrained style of the written dialog, impede the emergence of a ‘‘hierarchy of expertise’’; (iii) Encourages (written) dialog through a series of rounds; (iv) Inclusion of a broad range of expertise exposes participants to new ideas and perspectives.
DEVELOPING THE ETHICAL DELPHI
4.2.
61
Disadvantages
The method: (i) Can limit participation to a ‘‘pre-selected group of experts’’; (ii) Can be subject to possible bias in the selection of the ‘‘experts’’; (iii) May omit important ethical concerns, if these are not raised by the participants; (iv) Can entail a somewhat labored dialog, which relies on a high degree of self-motivation; (v) may suffer from ‘‘participant fatigue’’ (or boredom) because of the protracted (and somewhat repetitive) nature of the process; (vi) Can have a poorly defined ‘‘end-point,’’ which may mean the outcome is incomplete or incoherent: this may make it difficult to decide what conclusions have been reached; (vii) For some forms of the method, the role of the moderator may play a prominent role in setting the agenda for the dialog.
5.
CONCLUSIONS AND FURTHER DEVELOPMENT OPPORTUNITIES
An Ethical Delphi has been shown to have the potential to map expert positions, in terms of their factual and value derived aspects, for a specific case study. The classical Delphi method has been modified in the way in which the initial statements and first round questions encourage participants to discuss the normative dimensions of the topic. This allows participants not only to explore the scientific basis of the arguments that they are putting forward, but also allows individuals (and the group) to explore value disputes and areas of consensus. The expert opinions derived from this process can be analyzed and presented both to decision-makers and to the public in general. In the EU, the context for leading ethical debates, the degree to which such debates can be conducted and directed by bodies with an ethical mandate, and the expected degree of active involvement of participants, can be affected by the cultural and political traditions of different countries. Thus, how one chooses to conduct an ethical analysis using an Ethical Delphi depends on expectations anchored in socio-political culture. In order to seek feedback on the method, a usersÕ workshop was conducted with decision-makers in Vilnius, Lithuania (August 2005) to discuss the potential role of these tools in national public policy decision-making. The method was also presented during a workshop with users from various
62
KATE MILLAR ET AL.
European countries in Brussels in November 2005. These events provided useful insights into the use of the ethical frameworks under various sociopolitical and cultural conditions. These events highlighted that these frameworks can be used, in some cases with minor modifications, in different national and multi-stakeholder settings without diminishing the inherent usefulness of the tool. Through the analysis of the methods and the feedback from potential users, the research team believes that frameworks of this nature can be flexible enough to be adapted to varying circumstances and national settings. It is important to note that, although these frameworks are valuable decision-support tools, ethical tools are not decision-making machines for ethics. Decision-makers are still required to determine the ethical significance of the positions proposed by these tools. These are decision-support aids, the responsibilities of decision-makers are not diminished by the use of these tools. Detailed, objective reflection is needed with regard to the outcome of these processes. The advantage of the use of ethical frameworks is that they allow a decision-maker to actually enter the substantive debate and state why they favor or put forward a particular conclusion or weighing. The frameworks can clarify value steps in a decision-making process, making decisions more transparent and therefore open to review and possible revision. Utilizing ethical tools, and in particular ethical consultations of various groups, combines the advantage of using a methodical approach to capture ethical aspects with the democratic virtues of transparency and openness to criticism. Given this understanding of the appropriate context, we believe that the Ethical Delphi framework described may contribute in a significant way to better ethical understanding and better decision-making on issues that involve decisive ethical dimensions. REFERENCES Adler, M. and E. Ziglio (eds.). Gazing into the Oracle: The Delphi Method and its application to social policy and public health (Jessica Kingsley Publishers, London, 1996). COGEM. Farm scale evaluations evaluated. What can policy expect from science with respect to publicly-controversial technological innovations? COGEM report CGM/ 050408-04 (Bilthoven, The Netherlands: Netherlands Commission on Genetic Modification (COGEM), 2005) http://www.cogem.net/pdfdb/advies/FSE_evaluated.pdf. Crisp, J., D. Pelletier, C. Duffield, A. Adams, and S. Nagy (1997), ‘‘The Delphi Method?’’ Nursing Research, 46, pp. 116–118.
DEVELOPING THE ETHICAL DELPHI
63
Funtowicz, S., and J. Ravetz (1993), ‘‘Science for the Post-Normal Age.’’ Futures, 25 (7), pp. 739–755. Hasson, F., S. Keeney, and H. McKenna (2000), ‘‘Research Guidelines for the Delphi Survey Technique.’’ Journal of Advanced Nursing, 32, pp. 1008–1015. Kaiser, M., K. Millar, E. -M. Forsberg, O. Baune, B. Mepham, E. Thorstensen, and S. Tomkins. ‘‘Decision-Making Frameworks,’’ in V. Beekman (ed.), Description of Ethical Bio-technology Assessment Tools for Agriculture and Food Production. Interim Report Ethical Bio-TA Tools (QLG6-CT-2002-02594) February, 2004 (Agricultural Economics Research Institute (LEI), The Hague, 2004a) pp. 9–30. Kaiser, M., K. Millar, E. -M. Forsberg, O. Baune, B. Mepham, E. Thorstensen, and S. Tomkins ‘‘Decision-Making Frameworks,’’ in V. Beekman (ed.), Evaluation of Ethical Bio-technology Assessment Tools for Agriculture and Food Production. Interim Report Ethical Bio-TA Tools (QLG6-CT-2002-02594) October, 2004 (Agricultural Economics Research Institute (LEI),The Hague, 2004b) pp. 8–30. Rauch, W. (1979), ‘‘The Decision Delphi.’’ Technological Forecasting and Social Change, 15, pp. 159–169. Williams, P. L. and C. Webb (1994), ‘‘The Delphi Technique: A Methodological Discussion.’’ Journal of Advanced Nursing, 19, pp. 180–186.
Centre for Applied Bioethics, School of Biosciences University of Nottingham Sutton Bonington Campus Loughborough, Leicestershire LE12 5RD UK E-mail:
[email protected]