Integrated Risk Assessment and Risk Governance As Socio-Political Phenomena: a Synthetic

Front page for deliverables

Project no. 003956 Project acronym NoMiracle Project title Novel Methods for Integrated Risk Assessment of Cumulative Stressors in Europe Instrument IP Thematic Priority 1.1.6.3, ‘Global Change and Ecosystems’ Topic VII.1.1.a, ‘Development of risk assessment methodologies’

Deliverable reference number and title:

D.4.3.16b Integrated risk assessment and risk governance as socio-political phenomena: A synthetic view of the challenges

Due date of deliverable: Extra deliverable Actual submission date: Dec 2nd 2009 (revised form) Start date of project: 1 November 2004 Duration: 5 years Organisation name of lead contractor for this deliverable: Finnish Environment Institute, SYKE

Revision [draft, 1]: draft reviewed by 2 anonymous reviewers for Sci Total Environ

Project co-funded by the European Commission within the Sixth Framework Programme (2002-2006) Dissemination Level PU Public PU PP Restricted to other programme participants (including the Commission Services) RE Restricted to a group specified by the consortium (including the Commission Services) 2

CO Confidential, only for members of the consortium (including the Commission Services)

Authors and their organisation: Timo Assmuth, Mikael Hildén, Christina Benighaus Partner 19, SYKE, Helsinki, Finland Partner X, DIALOGIK, Stuttgart, Germany Deliverable no: Nature: Dissemination Date of delivery: 4.3.Xa R level: 2 December 2009 (revised Pu form)

Status: Date of publishing: In press Tbc Reviewed by (name and period): 2 anonymous reviewers, July-Nov 2009

Integrated risk assessment and risk governance as socio-political phenomena: A synthetic view of the challenges

Timo Assmuth, Mikael Hildén, Christina Benighaus 3 Integrated risk assessment and risk governance as socio-political phenomena: A synthetic view of the challenges

Timo Assmuth*

Mikael Hildén

Finnish Environment Institute (SYKE)

PO Box 140, FI-00251 Helsinki, Finland

Christina Benighaus

Dialogik

Lerchenstraße 22, 70176 Stuttgart, Germany

* Tel. +358 540 5907712, e-mail timo.assmuth @y mparisto.fi Correspondence address: timo.assmuth @y mparisto.fi (PO Box 140, FI-00251 Helsinki, Finland)

Abstract

To call for integration in risk assessment and governance as a self-evident goal is deceptively easy. For more insight, we ask: what level and kind of integration and for what purposes is needed and sufficient? What opportunities and obstacles can be identified for integrative treatment of risks? What causes and impacts are there of developments in risk integration? To answer these questions we investigate the socio-political processes and factors surrounding integrated risk assessment and risk governance through a combination of literature reviews and original research. We emphasize regulatory assessment and governance of risks associated with chemicals in the EU, but we link them with other areas to better grasp options and problems in integration. We relate the problems to political factors and barriers in sector and vertical integration, including deviating interests, and further to conflicting information, concepts and mindsets. Risk assessment and risk governance involve varying notions of risks and knowledge, with tensions between stressor- or impact-oriented, exclusive or inclusive, positivist or relativist, and fixed or reflexive notions and approaches. These tensions influence the trajectories of integration between sectors, actors and regions, constraining the fulfillment of ideals of integrated governance. We conclude that risk assessment and governance can be integrated, harmonized and innovated to a limit only, but this limit is variable and flexible, and provides opportunities especially if attention is paid to the socio-political contexts, value choices and decision structures in each case. Generally, the results underline a reflexive approach whereby the meanings, framings and implications of risk integration are probed in open processes of deliberation and negotiation, as a learning process to transcend the formal and prescriptive modes of regulation and knowledge generation.

Keywords: integrated risk assessment; governance; knowledge; political study; convergence, EU 4 INTRODUCTION

The concept of "risk society" (Beck 1986) significantly increased interest in risk as a socio-political issue, prompting criticisms and clarifications (Alexander 1996; Abbinnett, 2000; Elliot 2002, Beck,

2000, 2006; Mythen, 2007; Renn and Keil 2008). Equally important, Luhmann (1993) provided a sociological theory of risk within a general theory of society. Regardless of the differences in perspectives, it is clear that the management of risks is not just about technical assessment and optimization of the risks as quantified entities. Risks have social and psychological dimensions, and are shaped by values, beliefs, political systems and cultural factors (Kasperson et al., 1988; Felt et al., 2007). Risks are multi-dimensional also because risks and responses to them – socio-political and technological - interact. All of these general risk characteristics also apply to chemical risks and to environmental and health risks from multiple stressors, which are in focus in this paper.

The many categories, dimensions and interactions of risks have led to concerns for multiple and cumulative risks. A key concern is that even if single risks could be near-perfectly assessed and managed, cumulative risks from multiple stressors may be unforeseen and unacceptable. One risk even when acceptable per se may exceed limits in combination with others; and the possibility of such surprises increases in complex systems. Assessing these cumulative risks is a major scientific and methodological challenge, but it is also a societal issue, not least because particularly such risks are likely to be distributed unevenly (National Environmental Justice Advisory Council 2004).

Importantly for integrated treatment of risks, the selection of some risks over others is a key point of socio-political contestation, and is also culturally contingent (Douglas and Wildavsky, 1983).

It has consequently become important to develop risk assessment and management processes that facilitate many-sided consideration of the multiplicity of risks and of approaches to them. Risk integration moves from a straightforward task under a rather uniform paradigm of risk (such as biophysical) and of control (such as engineering or normative) to accounting for multiple risks in all 5 their complexity. This task is influenced by the increased importance of dialogues, negotiation and participation among citizens and actor groups including 'epistemic communities' (Haas, 1992;

Hajer, 2003; Meijerink, 2005). These interactions between actors can thus be identified as a focal area in integrated risk assessment and governance (Craye et al., 2009).

Needs for risk integration arise from problems with fragmented knowledge, policy and actions.

Broad overviews of multiple risks and of their contexts are typically needed when priorities are to be set for risk management. Yet, risks have been traditionally analyzed and managed by distinct sectors and groups who can be seen to represent separate Bourdieuan (1975) "fields" with specific rules and characteristics. There are differences in risk governance between regions too (Kaiser and

Prange, 2004; Jasanoff, 2005; Aarden et al., 2008). It is not self-evident how and if such differences can be harmonized or even accounted for. In addition to these relationships between groups, actors and regions, also the functional relations between risk science, risk assessment and risk governance need to be understood in order to manage risks efficiently.

To call for integration in risk assessment and governance as a self-evident goal is deceptively easy.

For more insight, we must ask: what level and kind of integration and for what purposes is needed and sufficient? What opportunities can be identified for integrative treatment of risks? On the other hand, what obstacles are there and what problems are caused by over-ambitious integration? What causes and implications are there of developments in risk integration? To answer these questions we investigate the socio-political processes and factors surrounding integrated risk assessment and risk governance. We discuss specifically how novel methods for the assessment of cumulative risks may affect these processes.

APPROACH AND METHODS 6 We review and synthesize earlier research on socio-political aspects of risks and risk integration, with particular reference to novel methods and approaches in the assessment and management of environmental and health risks from chemicals and other stressors. We complement this by theoretical and conceptual analyses of actor networks in the sense of transient material-semiotic phenomena (Callon et al., 1986; see also Latour, 1997; Kriesi et al., 2006; Schout and Jordan,

2008). We also apply and complement theories of cultural prototypes in risk perception as formulated by Hampton (1982) and co-workers (Douglas, 1989, 1996).

We approach risk assessment and governance as linked, multi-faceted, dynamic processes.

Emphasis is put on the horizontal and vertical relations between actors, and the role of knowledge and its relationships with policy. For coherent discussion, the following frames are selected:

 Geographically, we focus on the EU, but address its interactions with lower (national) and

higher (global) levels of governance, as appropriate for our theme

 We emphasize risks from chemicals to human health and the environment, but apply a dual

perspective on integrated treatment: stressor-oriented and impact-oriented (Figure 1), to clarify

and avoid the inherent limitations in treating it based on only one of these angles (such as

chemicals or other stressors). It should be noted that the impact-oriented approach may

encompass management consequences.

[Fig. 1 approximately here]

DEFINITIONS AND CONCEPTS

A standard definition of risk is "the probability of an adverse effect in an organism, system or population caused … by exposure to an agent" (IPCS and OECD 2003), focusing on quantifiable 7 biophysical entities. As pointed out above, risk is, however, a broader concept. Importantly, it includes social constructs (Macgill and Sie, 2003; Assmuth et al., in press). 'Scientific' risk is also a policy issue, as facts and values cannot be wholly separated (Rudén, 2006; Felt et al., 2007;

Putnam, 2003). The common dimensions in science-driven integration in risk assessment and risk governance thus include physical characteristics of risks (Figure 2, left) that may have also some explicit socio-political aspects (IPCS, 2001). However, additional dimensions are distinguished more directly related to societal contexts (Figure 2, right).

Cumulativeness of risks can mean their accumulation in time or along the stages in risk formation.

Accumulation also takes place in space (hotspots) and in food-chains and other materials.

According to USEPA (2003), cumulative risks mean “the combined effects from aggregate exposures to multiple agents or stressors”. This definition combines the accumulation of responses or effects and that of agents or stressors.

Previously, a strict division was often presupposed between risk assessment and risk management, following the risk paradigm of NRC (1983). This has been often upheld in the EU, partly to assert the independence of assessment from politics such as with the scientific advisory bodies (ElAmin,

2006). An ideal of objectivity in assessment contributes to this division, and reflects a traditional view of expert knowledge (Eduljee 2000). A more nuanced view has, however, developed of assessment-policy relationships (Rayner and Cantor, 1987; Jasanoff, 1995; Horlick-Jones, 1998;

CEC, 2002d; Felt et al., 2006; Assmuth et al., in press). Risk assessment can be seen as reflecting socio-political systems and cultures, and as an activity that mixes facts with valuations of the frames and interpretations surrounding these facts (Wenzel, 1997). Also official definitions of risk assessment have become more inclusive (NRC, 1994, 1996, 2008; OECD, 2003a; Anon., 2009).

Stakeholder involvement in risk assessment and governance has become more common, for 8 example in the model procedures of OECD (2003a, b), Health Canada (2007) and the EU (CEC,

2002a, 2003b). IRGC (2005) went further and placed communication at the center of the risk governance cycle (cf. Renn, 2007).

In definitions of governance, Treib et al. (2007) identified as common denominator "depart[ure] from the traditional model where collectively binding decisions are taken by elected representatives

… and implemented … in public administrations". In general, governance emphasizes a change in the balance between state intervention and social autonomy. Risk governance likewise extends from regulations to multi-actor participation and negotiation and from technical management to legal, institutional, social and economic contexts (Renn 2008: 8). Integration in governance has usually meant vertical and regional integration (Egeberg, 2001) but is multi-dimensional. Horizontal integration between sectors (Chiti, 2004) thus involves also the alignment of risk governance in the sectors. Related concepts include (sector) policy convergence and coordination.

THE EU GOVERNANCE AND RISK INTEGRATION

Actors and activities in multi-level and multi-sector governance

The EU has generated new institutions and processes of regulation and governance (Lyall and Tait,

2004). The vertical relations from supra-national to sub-national levels and associated interregional relations are complemented by the horizontal relations between sectors and policies (Knill, 2001), by the relations between actors groups (extending the horizontal dimension), and by the relations between elected and delegated or knowledge-based powers (Figure 3; Bauer et al., 2007; Radaelli,

2008). Vertical and horizontal integration may compete (Jacob and Volkery, 2004), but have synergies also (von Beyme, 2005).

[Fig. 3 approximately here] 9 Risk governance, and especially integrated risk governance, brings together a wide range of actors.

There is complexity both at a general structural level and within each box in the structure at global,

EU and national levels (Figure 3). In many issues regional and local actors also play an important role in risk related activities. The groups of actors have partly different foci and roles in risk governance, also 'cognitive' (Laffan, 2001). Parliaments for instance contain political groups which have different views and preferences concerning risks (Benedetto, 2005) both specifically and generally, e.g. regarding precaution. Also the national views differ as shown by the debate over the

EU's REACH (Registration, Evaluation and Authorisation of Chemicals) legislation of 2006

(Assmuth et al., this issue).

The involvement of private and other civil society actors is based on accepted policies (CEC, 2001) and in general implies broad, loose and informal integration. As a counterweight, new mechanisms of state influence are needed. For instance, public-private partnerships provide new opportunities through the inclusion of more actors but also new challenges in ascertaining the public interest in risk governance. New procedures and systems of negotiation and mediation emerge to resolve concerns, controversies and interest conflicts. These procedures and systems also include and affect the contents and conduct of risk assessments and what integration takes place in them.

Environmental health is a case in point.

Specifically in chemicals control and risk governance, many actors, with their variable interests in and concepts of risks, influence whether and how integrated treatment of risks is achieved. Formal and legally based multi-actor governance is developing, for instance, under REACH (Assmuth et al., this issue), but is insufficient to cover cumulative risks from multiple stressors. Integration in this connection is often taken to involve mainly integration of substances, environmental compartments or biological effects, such as in the procedures for "Integrated Testing Strategy" and in other procedures e.g. for combining analytical chemistry measures and biomarkers (Blasco and

Pico, 2009). This technical level of integration is important, but from the point of view of integrated 10 governance leaves out many crucial issues. These include the use of testing information in governance and, conversely, the feed-back to testing and other forms of information generation.

Introducing novel methods for the assessment of cumulative risks from multiple stressors is thus not just about providing more sophisticated tools for existing fields and actors. The methods can also innovatively simplify and extend assessments, by bringing in new actors or modifying the roles of existing actors.

Cultures of actor groups in risk governance

A number of basic ways to relate to risks have been collated to "cultural prototypes" of risk perception, based on simple models of adherence to a social group and 'grid' (lacking autonomy).

The resulting prototypes have been given suggestive designations such as 'egalitarians',

'entrepreneurs', 'bureaucrats' and 'hermits' (Hampton, 1982; Thompson and Wildavsky, 1986;

Douglas, 1989, 1992; Wildavsky and Dake, 1993). Such prototypes are relevant in risk assessment and risk governance, but not in simple and uniform ways (Wenzel, 1997; Sjöberg, 1999; Wilkinson,

2001). They co-exist with institutional structures and political processes (Swedlow, 2002) and with social processes and settings that influence identity formation and role enactment (Trondal, 2001).

The approaches of groups to risks also depend on whether the context is dominated by consent, liability or trust (Rayner, 1990). Thus, the cultural roles (in the sense of human customs, institutions and achievements) of these actor groups and networks are an inseparable and interactive aspect of risk assessment and risk governance.

We account for such factors in applying and developing the concept of cultural prototypes within integrated risk governance (Assmuth et al., 2009). For example, entrepreneurs in industry may suspect that the broadening of risk governance leads to stakeholder involvement that may invite criticisms and complicate management. They may nevertheless seek opportunities for avoiding 11 conflicts through a collaborative strategy, as do entrepreneurs in political arenas (Benedetto, 2005).

Egalitarians may also be critical of broad governance, believing that society should strive at zero risk and that multi-actor governance will legitimate risks through stakeholder influence (cf.

Bomberg, 2007). Some NGOs oppose but others support broad participation, and may also be opportunist entrepreneurs. Bureaucrats generally see broad governance as a challenge as it requires changes in routines. However, since the 'Kinnock' reforms (CEC, 2001) the EU apparatus is infused with the idea of such governance, largely to defend its legitimacy and position (Levy, 2006; Ellinas and Suleiman, 2008). 'Eurocrats' therefore endorse broad-based risk governance, if only as lip service.

The consideration of cumulative risks from multiple stressors adds complexity and new dimensions to these roles of actors. All prototypes can perceive needs and opportunities for integrated risk assessment and management but in different ways. Egalitarians will stress the balanced and democratic treatment of risks to all groups; the question will be how far the framings are stretched, e.g. as to sub-groups, and how the necessary weightings and tradeoffs between risks are defined.

Entrepreneurs may see integration needs and opportunities mainly for their positions, and are likely to attack the multi-risk problem innovatively but incrementally. Bureaucrats prioritize integration needs and options within the confines and procedures of institutional structures that are rather formal, hierarchical and sectorized. Non-hierarchical and independent hermits, in an engaged version, may have a special role in the unconditioned flexible assessment and disinterested negotiations needed with multiple and multi-dimensional risks and uncertainties.

Horizontal integration

The types of risks addressed and integrated in the EU governance vary by sectors (Table 1). They are defined variously by economic activities or by entities or properties to be protected (health, 12 environment, safety). The sectors have a range of policy instruments at their disposal, at different levels of intervention (Assmuth and Hildén, 2008). Coordination of the instruments is a key to integrated risk governance. Many attempts have been made, from general policy integration such as the Cardiff process to regulations in specific areas of integration such as IPPC (EP and EC, 2008).

Problems still persist: "across … human health and the environment. … lack of an integrated approach resulted in: … compartmentalization …; duplication of effort; failure of … initiatives to make … impact." (Bridges, 2003: 15). There is also disagreement about the inherent idiosyncrasy and legitimate individuality of fields. The health field may, for instance, claim competence in comprehensively addressing health impacts and benefits of chemicals as well as other health factors and thus to inherently integrate risks; the environmental field, while usually focused on toxic effects of chemicals, may be integrative in other dimensions such as human ecology and environmental risk management (Assmuth and Jalonen, 2005).

[Table 1 approximately here]

There are overlaps between sectors of risk governance. For instance, chemicals control encompasses classification and labeling, registration, evaluation and authorization of industrial and consumer chemicals, pesticides and biocides (Assmuth et al., this issue). However, also many other areas of governance focus on chemicals, such as regulation of pharmaceuticals, cosmetics, food additives, radionuclides and nanomaterials. Other fields close to and even overlapping chemicals control include GMOs; water, soil, and air pollution; waste and product policy; public and occupational health; consumer and citizen safety; and energy (Table 1). These have variable and changing approaches to integrated risk governance. Normative and technological drivers such as standards and aggregated information systems may engender sector integration, but also less formal forces and procedures are important for integrative risk assessment, such as the coupling of health impact assessment with strategic environmental impact assessment (Wright et al., 2005). However, 13 the many levels of governance and the variety of actors complicate integrative treatment (Knill and

Lenschow, 2005; Saurugger, 2005; Kriesi et al., 2006).

Vertical integration

The vertical dimension involves contrasting directions: globalization and localization (Lazer, 2001).

Despite 'europeanization' and transposition of regulations, there are differences between Member

States (Eberlein and Grande, 2005), especially in Eastern Europe (Bauer, 2007; Getimis et al., 2007;

World Bank, 2009). Integration and convergence thus concur with sub-EU variability (Bugdahn,

2005; Bulmer, 2007). As stressed by Knill (2005), the factors influencing this balance are poorly understood. The failures of EU transposition have been explained both by the nature of the specific policy instruments and obligations and by general factors (Etherington, 2006). The latter include endogenous processes such as institutional bargaining and constitutional compromises (Moravsick,

2005) and exogenous such as resource spillover (Farrell and Héritier, 2005). With enlargement and economic and political pressures, also bloc-building plays a role. Benz and Eberlein (1999) took an optimistic view of the possibilities to avoid regulatory overloads in multi-level governance based on

French and German data. The processes elsewhere are more difficult, even in old Member States when the legal role of the state is weak (Giglioli and Swyngedouw, 2008).

The key problem is how to interpret the subsidiarity principle in the context of risk assessment and management: Are Member States or their regions allowed to create their own interpretations of cumulative risks and procedures for assessing these, or only to develop solutions in accordance with harmonized processes and legally specified criteria at the EU level? For chemicals the tendency is towards the latter, justified by the Amsterdam Treaty (1997) which supports community action when “actions by Member States alone … would conflict with the requirements of the Treaty (such as the need to correct distortion of competition or avoid disguised restrictions on trade or strengthen 14 economic and social cohesion) or otherwise … damage Member States' interests”. The appropriate level of Member State influence is a matter of debate and there is great variation between areas of governance. For example land use planning allows specific approaches to integrated treatment of risks, applying EU-level requirements in the geographical context. Here weakly harmonized methods for cumulative risks can account for regional aspects. Vertical coordination is however not simple in any sector (Benson and Jordan, 2008; cf. Blom-Hansen, 2005). It also depends on the circumstances, as found by Krapohl (2003) with the balance between the deliberative supra- nationalism of scientific committees and the inter-governmental function of oversight committees.

The EU is increasingly part of regional and global political and economic systems. This adds a level of complexity that poses a challenge to the EU and specifically to risk assessment and governance

(Luhmann, 1993; Beck, 2003, 2006). There are opportunities and problems for integrated treatment of risks in this expanding scope and connectedness of socio-economic and political systems. The problem is often the lack of political and legal mechanisms at supra-EU levels, as in the governance of global flows of chemicals (Assmuth et al., this issue) and of biotechnology, which is also closely linked with international relations and trade (Millstone et al., 2008).

Change and innovation in integrative risk governance

Many social, political, technological and environmental changes influence integration in risk governance. Societal changes may be intentional or not; they inhibit or promote integration; and they affect both the substance and the conduct of risk assessment and governance. The developments are related to processes and factors in the polity (institutional structures), policy and politics domain, such as constitutional agreements, delegation of powers, enlargement, and the role of state (Kaiser and Prange, 2004; Treib et al., 2007). These developments and factors are reflected 15 in the integration of risk governance in and between the various sectors of the EU, including those that address chemicals, environmental management and health care (Table 1).

In addition to and underlying these political and institutional processes, risk governance is influenced by substantive changes such as increased scarcity of resources and energy; economic and financial turmoil, global and regional crises (including those prompted by risk controversies) and demographic changes. These contribute to increased psychological sensitivity to risks, especially those related to health and security. The EU policies are consequently evolving, albeit unevenly and through 'punctuated equilibria' (Meijerink, 2005). The evolution of policy cycles from conception to follow-up involves vertical and horizontal integration and multi-actor negotiation through organized policy coordination (Bauer, 2006), but also informal policy learning takes place in the EU as a

'system of innovation' (Borrás, 2004).

There is path dependency in the evolution of risk governance, but it is difficult to predict. Complex political, social and cultural processes combine to mold it, and are in turn affected by technological development. On a substructure of historical forces, stochasticity is thus superimposed (Kracauer,

1995 (1969)). This is related to the uncertainty in complex interactions, pronounced in to-day's world (Hajer, 2003). The result is dynamic co-evolution at many levels: between institutions, sectors and polities; between society and technology; between society and nature. New methods play a role in this: for example, the detection of chemicals at smaller concentrations and of new effects has made risks “visible”, leading to demands for regulation and assessment.

While new risks and problems can emerge rapidly, regulatory systems may be slow to react.

However, sometimes changes are made on short notice, even as 'solutions looking for problems'

(Zahariadis, 2008). Multiple time scales are thus imposed on each other, such as the quartile economy on slower legal, organizational and cultural changes, but generally getting faster. The implication is that governance is constrained by inertia, but also doomed to cope with changes. 16 Some punctuation by crises can be discerned as crossroads in the trajectories of EU risk governance in relation to integration and precaution. For example, the accident at Seveso, Italy in 1974 triggered EU-level responses such as the directive 82/501/EEC on industrial accidents and the more comprehensive Seveso II directive 96/82/EC on resultant hazardous wastes and materials; later food dioxin scandals eventually lead to the formulation of the Community dioxin strategy of 2001, and contributed to EU’s activities in the POPs Convention. In some thematic actions, such as the development of a Community chemicals policy, integration of sectors, actors, levels and issues in governance has been stressed. Other actions and instruments such as the REACH legislation have been confined to integration in a limited sense and at detailed level, but have also made attempts at some new areas of integrated risk governance and at the operationalisation of precautionary approaches (Assmuth et al., this issue).

Our interpretation is that regulatory changes are often triggered by specific events, which can be random, but the outcomes are affected by general directions and factors in policy and politics, such as converging interests of sectors or countries, and also by overall cultural approaches to risks. On sustained converging interests, responses are revised and broadened (e.g., Seveso and IPPC

Directives). Regulations in turn drive the interest in issues and induce findings. The regulations and the outcomes are related to methodological development. For example, the priority substance list in the Water Framework Directive can be largely traced to 'events' and not only to systematic and ab initio prioritization, but details such as the requirement to measure chemicals in the water phase depend on existing test methods. Some links involve vicious circles: those substances that have been prioritized are mainly measured and can be prioritized. Methods for risk assessment respond to but also influence both the details and overall approach; they mediate and synthesize risk information although do not provide new findings of risks in the way empirical observations do. 17 DISCUSSION

Socio-political aspects of the multiplicity and multi-dimensionality of risks

The socio-political aspects of risk assessment and risk governance are pronounced in the case of cumulative risks from multiple stressors. With such risks, issues arise beyond the usual social amplification (Kasperson et al., 1988) and attenuation of risks. The selection of some risks among many is influenced by the cultural setting and implies downplaying many other risks (Wildavsky,

1995) although there can be some mutual amplification of the risk selected and corollary risks. The consideration of additional risks may reduce the significance of those risks originally in focus, e.g. chemicals as compared to other risks to health and the environment. However, a broadened focus may also turn out to emphasize the risk of initial concern, depending on its perceived characteristics and roles as a part of a whole, such as if accumulated chemicals are seen to break the 'camel's back' or to be the risks that are most amenable to control. In either case, emphasizing or de-emphasizing one risk among many, the multiplicity and multi-dimensionality of risks present challenges particularly for those focused on one risk or interpretation.

Integrated risk assessment and management represent attempts to grasp and solve multiple and multi-dimensional risk problems. This will take place in any case at the integrative level of regulation and resource allocation, but often ad-hoc and a posteriori, on a linear model (Renn and

Keil, 2008). It seems generally better to address such problems up front, to avoid excessive work on some risks and counterproductive conflicts, and instead to enable negotiation and learning between fields. The counter-argument is that also single risks need to be focused on within the limits of existing fields, of which all cannot be involved. A combination of these holistic and specific or incremental approaches seems a natural way out of such argumentation. As the holistic, more ambitiously integrative approach is strengthened, the conflicts between entrenched fields tend to 18 move to framing, the "bottleneck of legitimate institutions" (Kohler-Koch, 2000). Dialogues and negotiations in this case may more easily resolve conflicts and generate innovative joint solutions.

Also the context dependency of risks and of risk governance, specifically of cumulative risks from multiple stressors, suggests that regulation which is flexible and includes elements of reflexivity can provide a basis for integrated assessment and governance of such risks. The challenge is how to balance this 'modernist' approach with the requisite clarity, predictability and normativity (Figure

4). Particularly in the case of integrated risk assessment and risk governance of multiple stressors, broad and loose framings in both substance and conduct have to be balanced with specified and coherent framings. A partial solution may be found in novel combinations of procedures that enhance negotiation of framings, goals and means, as seen in the case of REACH (Assmuth et al., this issue). This is akin to avoiding underestimation of the forms of mobilization affecting both cognitive and normative frames (Surel, 2000), or both the material and semiotic aspects of actor networks (Latour, 1997).

It has been claimed that society has shifted from representative to participative democracy; from centralized government to networked governance; from decisiveness to consensuality (CEC, 2001;

OECD, 2002; Hajer, 2003). This shift is not self-organized but has an element of central steering that influences policy coordination (Schout and Jordan, 2005). As seen above, participation is stressed also in risk governance (Craye et al., 2009); it is particularly important because of the charged nature of risks (Beck, 1986; Johnson, 2005). These 'malign issues' (Nilsson and Persson,

2003) prompt political bargaining that weakens mutual learning and policy integration. Participation can counteract this, enabling actors to frame risks broadly and facilitating many-sided deliberation.

It can however be questioned whether these shifts are real or aspirations, and how strong and lasting they are; a swing of the pendulum to a more bounded interpretation of integration and of expertise 19 is possible (cf. Löfstedt, 2004). There are also caveats to multi-actor governance. It may enable a shift toward private interests, a "crisis in governmentality" instead of governance (Bailey, 2006), as opportunism flourishes in ambiguous coalitions of 'interface actors' (Knill, 2001). As put by

Eberlein and Grande (2005): "the informalization of governance is vulnerable to strong distributive conflict, and … raises … problems of democratic legitimacy". Multi-actor governance is thus no panacea for integrated risk governance. Attention is required to the qualities of the process and the outcomes, to ensure broad and efficient as well as equitable risk management (Figure 4).

Harmonization and integration of risk governance

Harmonization and standardization are often emphasized in international and integrative risk governance (DG-SANCO, 2003a; ISO, 2008; cf. Abbott and Snidal, 2001). In the EU, harmonization is a 'mantra' due to the core idea of common market and the perceived need to eliminate the divergence of goals and means that is judged to unduly limit the market. The idea of harmonization underlies many developments also in EU risk governance (DG-SANCO, 2003a, b).

Integration and harmonization should however not be confused, and neither is to be accepted uncritically. Integrative treatment of risks can involve the comprehensive, e.g. comparative consideration of non-commensurate risks and of individual risk concerns without imposing normative requirements on harmonization.

With diverse risks, harmony is unattainable: "So many circumstances of a small and accidental nature are relevant that no broad and simple uniformities are possible" (Russell, 1948). Limits to harmonization are however often downplayed (Assmuth and Hildén, 2008; Assmuth et al., 2007).

For instance, harmonization of one aspect of risks from a chemical, such as its ecological effects, and the testing and assessment procedures used to define these effects, can jeopardize harmonization of another, such as its health effects or its benefits. As discussed above, there are 20 also justifiable differences between regions and sectors and their respective management goals and approaches. Even when convergence is desired, there are many alternatives to harmonization in achieving it, including coercive imposition, competition, regulatory diffusion and policy learning

(Busch and Jörgens, 2005; Anon., 2006). The tolerance of asymmetries is a key issue in all these measures and in all dimensions of integration (von Beyme, 2005).

Idiosyncrasy and integration are thus not antithetical but complementary (Jones, 2003). Instead of trying to harmonize things at a level unattainable, variation can be embraced by facilitating consideration of specifics and contexts, as seen with vertical integration. This suggests that a context-sensitive pluralistic way of integrating different aspects of risk for governance should be seen as a way forward. Significant barriers are nevertheless likely to be encountered.

Barriers, opportunities and limits of integration of risks

Several types of barriers for integrative risk assessment and management are identifiable (Assmuth and Craye, 2009): in knowledge (epistemological and ontological barriers), in regulation, for instance administrative capacities (Schout and Jordan, 2008), and in socio-political factors (also cultures). These are not separate; for instance, cultural factors influence knowledge and administration, and institutions can be seen as both regulatory and political barriers. Accordingly, institutionalist models can be combined with models of world views and identity formation (Surel,

2000). Further, in a more evaluative view, the above analysis and discussion indicates that the barriers of integration include more and less justified limitations. Overcoming them cannot therefore be a generic objective. The evaluation will depend on the more precise meanings of integration in each case. 21 Basic concepts (ontologies) and methodological approaches can limit possibilities for integrative treatment of risks. Also allegedly integrated risk assessment can have limitations related to the knowledge they engender and deliver. Focusing a priori on one stressor such as chemicals carries a limitation, as noted by NRC (2008) regarding the cumulative risk assessment of the USEPA.

Cumulative risks thus need to be approached also from the perspective of effects. This is a reality check that may, for instance, reduce fixation on synthetic chemicals (Ames and Gold, 2000) or on other stressors believed to cause top risks. Uncertainties arise due to problems in attributing risks to causes. Higher-level questions of the meaning of the cumulative risks involve competing concepts of realism, such as how qualitative differences of risks including social aspects affect risk commensurability and prioritization (e.g. whether the attention on synthetic chemicals or other modern worries is 'really' disproportionate; see Assmuth et al., this issue). Concepts of integration have been extended from physical to social realms also for instance in integrated product policy

(Rubik and Scholl, 2002). However, the socio-psychologically charged nature of risks poses particular challenges that are accentuated by their multiplicity, multi-factoriality and multi- dimensionality. This underlines the need for dialogue and pluralistic deliberation, but also regularly makes those engaged resort to a simplistic notions and claims of knowledge.

Institutional barriers are equally important. Guttier (2004) argued that the coordination between overlapping sector competences in the EU has been too timid. This coherence deficit is common between health and environmental sectors, despite some integration (CEC, 2003a; DG-SANCO

2006, 2008). Cohn (2004) showed that also complex policies can succeed if there is consensus of goals; this lacking, all types of solutions including simplification and restriction hoping for knock- on effects in other fields have drawbacks. For instance, Environmental Policy Integration has been criticized for a lack of concrete 'bottom-up' integration with other sectors (Jordan et al., 2003), but also for diluting environmental policy in other fields (EEA, 2005a, b; Herodes et al. 2007), implicitly assuming a lack of common goals. There are notable institutional barriers also in vertical 22 relations between the EU, its members and regions, and the global community. They can be particularly hard to enforce because of questions of sovereignty and power.

Everything cannot be integrated, and trade-offs are needed (Suter et al., 2003). The capacity of social systems to address risks, old and new, is limited. Simplification of governance (CEC, 2002b) can make some integration easier, for instance by dismantling hierarchies, but it can also prevent some integration, if coordinative functions, regulatory oversight and the ability to take on new cross-cutting issues are diminished. Differences between sectors, regions and actors create barriers, some even permanent, but convergence and integration in some dimensions may be possible. If for instance integration of human health and ecological risks (due to scientific or policy differences), or of stressors (such as chemicals in different contexts) is not practicable in a prescriptive way at detailed level, integration may be achieved flexibly and at general level (CEC-SANCO, 2003b).

The barriers depend also on the risk in question. Stressors that act through the same media, such as food, are for instance in some respects easier to integrate than those that occur and act through different media or compartments, or generally represent different categories.

CONCLUSIONS

The multiplicity, multi-dimensionality and cumulativeness of risks are conspicuous issues in governance, and underlie many controversies about risks. Specifically, along with multiple stressors, their multiple impacts need to be grasped. This complicates integrated risk assessment but also points to some possibilities to develop it for fuller relevance in both scientific and policy terms, i.e. allowing a grasp of the complex reality of risks in ways that better reflect both the scientifically tractable phenomena and the socio-political needs amidst this complexity. The following conclusions emerge from our broad analysis. 23  In developing integrated risk assessment and governance one should recognize the multiple

dimensions of risks and of integration and hence reject the idea that uniformity in concepts,

contents and methods could solve all problems of integration. Communicative learning

processes should be emphasized instead, but these demand development of dialogues across the

different actor fields in risk assessment and governance.

 Integrated risk governance in the EU takes place in and needs to be developed in many sectors

and between different levels of governance from global to local. Both of these involve multi-

actor participation and negotiation. On the other hand, responses on risks influence

developments of overall governance. Risks associated with chemicals are part of a palette of

risks, including risks from loss of benefits and countervailing risks from management. This

requires broader scopes and revised approaches to risk assessment and management.

 There are important knowledge-related, institutional and other socio-political barriers to

developing and applying integrated risk assessments, also of cumulative risks from multiple

stressors. The barriers are of varying strength and permanence. Identifying and

exploringbarriers can create openings for integration also in new and unexpected ways,

especially by bringing in views from actors.

 Methodologically, reflexive, flexible and more transparent approaches to risks are advisable as

part of and support for open processes of risk communication, negotiation and governance.

Different framings, assumptions and valuations of risks can thereby become clearer, and

cumulative risks of multiple stressors can be addressed at the levels and with the types of

integration that is appropriate to the purpose, acknowledging also the need for formal, specific,

constant and robust approaches. An important element and value of such an open reflexive

approach is that is can be revised while being anchored in democratic deliberation and in

existing knowledge. 24

ACKNOWLEDGEMENTS

This work has been part of and has been funded within the project Novel Methods for Integrated Risk Assessment of

Cumulative Stressors in Europe (NoMiracle, contract No 003956) under the EU's 6th Framework Programme for R&D.

Additional support from our employers is acknowledged.

REFERENCES

Aarden E, Van Hoyweghen I, Horstman K, Vos R. Learning from Co-evolution of Policy and Technology. Different

PGDs in the Netherlands, Germany and Britain. J Comp Pol Anal 2008; 10 (2): 191-206. http://www.ingentaconnect.com/content/routledg/jcpa/2008/00000010/00000002/art00006

Abbinett R. Science, Technology and Modernity: Beck and Derrida on the Politics of Risk. Cultural Values 2000; 4:

101-126. http://www.informaworld.com/smpp/content~content=a909628377~db=all

Abbott KW, Snidal D. International 'standards' and international governance. J Eur Public Pol 2001; 8: 345-370. http://www.informaworld.com/smpp/content~content=a713773911~db=all~order=page

Alexander JC. Critical Reflections on Reflexive Modenization. Theory, Culture & Society 1996; 13: 133-138. http://tcs.sagepub.com/cgi/pdf_extract/13/4/133

Ames BN, Gold LS. Paracelsus to parascience: The environmental cancer distraction. Mutat Res 2000; 447: 3-13. http://dx.doi.org/10.1016/S0027-5107(99)00194-3

Anon. Environmental Governance in Europe: The Impact of International Institutions and Trade on Policy Convergence

(ENVIPOLCON). Final Report, Aug. 2006. http://www.uni- konstanz.de/FuF/Verwiss/knill/projekte/envipolcon/downloads/Final_Report_2006.pdf 25 Anon. An Overview of “Science and Decisions: Advancing Risk Assessment”. Risk in Perspective 2009; 17: 1-7. http://www.hcra.harvard.edu/perspective.html

Assmuth T, Benighaus, C, Hildén M, Lyytimäki J, Renn O. Big pictures, close-ups, roadmaps and mind-maps:

Perspectives on integrated treatment of multiple risks. Int J Risk Assess Manage (in press).

Assmuth T, Benighaus C, Craye M, Hildén M, Lyytimäki J. Cultural Types In Integrated Approaches To Cumulative

Risks. Poster, Multiple Stressors Conference, Aarhus, Denmark, Sept 28.-30.2009. http://nomiracle.jrc.ec.europa.eu/Documents/Conference_28-30_September_2009/Proceedings.pdf

Assmuth T, Craye M. Wrestling with a Medusa - Integrated assessment and management of cumulative risks from multiple stressors according to EU chemicals regulators and stakeholder experts. Deliverable Report to NoMiracle

Project, JRC Scientific and Technical Reports (submitted). http://nomiracle.jrc.ec.europa.eu/webapp/ViewPublicDeliverables.aspx

Assmuth, T, Craye M, Hildén M. Beyond REACH: Roadblocks and shortcuts en route to integrated risk assessment and management of chemicals. This issue.

Assmuth T, Hildén M. The significance of information frameworks in integrated risk assessment and management.

Environ Sci Pol 2008; 11: 71-86. http://dx.doi.org/10.1016/j.envsci.2007.07.006 http://authors.elsevier.com/offprints/ENVSCI596/01965ca5d7f3cd89fc86abc7c4ef3f25

Assmuth T, Jalonen P. Risks and management of dioxin-like compounds in Baltic Sea fish: An integrated assessment.

Nordic Council of Ministers, Copenhagen, 2005. http://www.norden.org/pub/sk/showpub.asp?pubnr=2005:568

Assmuth T, Lyytimäki J, Hildén M, Lindholm M, Münier B. What do experts and stakeholders think about chemical risks and uncertainties? - An Internet survey. The Finnish Environment 22/2007. http://www.ymparisto.fi/default.asp? contentid=241573&lan=en

Bailey DJ. Governance or the crisis of governmentality? Applying critical state theory at the European level. J Eur

Public Pol 2006; 13: 16-33. http://www.informaworld.com/smpp/content~content=a727419718~db=all~order=page 26 Bauer MW. Co-managing programme implementation: conceptualizing the European Commission's role in policy execution. J Eur Public Pol 2006; 13: 717-735. http://www.informaworld.com/smpp/content~content=a749282283~db=all~order=page

Bauer MW, Knill C, Pitschel D. Differential Europeanization in Eastern Europe: The Impact of Diverse EU Regulatory

Governance Patterns. J Eur Integr 2007; 29: 405-423. http://www.informaworld.com/smpp/content~content=a781177143~db=all

Beck U. Risikogesellschaft : auf dem Weg in eine andere Moderne. Suhrkamp, Frankfurt am Main 1986.

Beck U. Risk Society Revisited: Theory, Politics and Research Programmes. Adam B, Beck U, van Loon J. (eds.) The risk society and beyond. Critical Issues for Social Theory, London, Thousand Oaks and New Delhi, SAGE Publications,

2000, pp. 211-229. http://www.ingentaconnect.com/content/bpl/cons/2003/00000010/00000004/art00002,. http://books.google.fi/books?id=-XnYEvw_KukC&pg=PA211&lpg=PA211&dq=%22Risk+Society+Revisited:

+Theory,+Politics+and+Research+Programmes

%22&source=bl&ots=3bGPr0x0VP&sig=AxOB4bXSTQ6FaKHdrYh4RU3TrR8&hl=fi&ei=BIdUSoTMOpHumgPng

NUE&sa=X&oi=book_result&ct=result&resnum=2

Beck U. Toward a new Critical Theory with Cosmopolitan Content. Constellations 2003; 10 (4): 453-468. http://www.ingentaconnect.com/content/bpl/cons/2003/00000010/00000004/art00002

Beck U. Living in the World Risk Society. Economy & Society 2006; 35: 329-345. http://www.ingentaconnect.com/content/routledg/reso/2006/00000035/00000003/art00001

Benedetto G. Rapporteurs as legislative entrepreneurs: the dynamics of the codecision procedure in Europe's parliament. J Eur Public Pol 2005; 12: 67-88. http://www.informaworld.com/smpp/content~content=a713948681~db=all~order=page

Benson D, Jordan A. Understanding task allocation in the European Union: exploring the value of federal theory. J Eur

Public Pol 2008; 15: 78-97. http://www.informaworld.com/smpp/content~content=a787034747~db=all~order=page

Benz A, Eberlein B. The Europeanization of regional policies: patterns of multi-level governance. J Eur Public Pol

1999; 6: 329-348. http://www.informaworld.com/smpp/content~content=a713773712~db=all 27 Beyme von K. Asymmetric federalism between globalization and regionalization. J Eur Public Pol 2005; 12: 432-447. http://www.informaworld.com/smpp/content~content=a714024041~db=all~order=page

Blasco C, Picó Y. Prospects for combining chemical and biological methods for integrated environmental assessment

TaAC Trends Anal Chem 2009; 28 (5): 745-757. http://dx.doi.org/10.1016/j.trac.2009.04.010

Blom-Hansen J. Principals, agents, and the implementation of EU cohesion policy. J Eur Public Pol 2005; 12: 624-648. http://www.informaworld.com/smpp/content~content=a714024055~db=all~order=page

Bomberg E. Policy learning in an enlarged European Union: environmental NGOs and new policy instruments. J Eur

Public Pol 2007; 14: 248-268. http://www.informaworld.com/smpp/content~content=a772389395~db=all~order=page

Borrás S. System of innovation theory and the European Union. Sci Public Pol 2004; 31: 425-433. http://www.ingentaconnect.com/content/beech/spp/2004/00000031/00000006/art00002

Borrás S, Jacobsson K. The open method of co-ordination and new governance patterns in the EU. J Eur Public Pol

2004; 11: 185–208. http://www.informaworld.com/smpp/content~content=a713791019~db=all~order=page

Bourdieu P. The specificity of the scientific field and the social conditions of the progress of reason. Soc Sci

Information 1975; 14: 19-47. http://ssi.sagepub.com/cgi/pdf_extract/14/6/19

Bridges J. Report on Priorities in the EU for Risk Assessment in The Non-Food Area. Prepared for the European

Commission’s Directorate General for Health and Consumer Protection (Public Health and Risk Assessment

Directorate) under contract number A0-7050/03/000231, 9 November 2003. http://ec.europa.eu/health/ph_risk/documents/risk_key02_en.pdf

Bugdahn S. Of Europeanization and domestication: the implementation of the environmental information directive in

Ireland, Great Britain and Germany. J Eur Public Pol 2005; 12: 177-199. http://www.informaworld.com/smpp/content~content=a713948676~db=all~order=page

Bulmer S. Germany, Britain and the European Union: Convergence through Policy Transfer? German Politics 2007; 16:

39-57. http://www.informaworld.com/smpp/content~content=a777118258?words=environmental%7chealth%7crisk

%7chorizontal%7cintegration%7ccoordination%7cinstitutional%7ceuropean%7cpolicy%7cgovernance

%7cinstitutions&hash=282779136 28 Busch P-O, Jörgens H. The international sources of policy convergence: explaining the spread of environmental policy innovations. J Eur Public Pol 2005; 12: 860-884. http://www.informaworld.com/smpp/content~content=a714024062~db=all~order=page

Callon M, Law J, Rip A. Mapping the Dynamics of Science and Technology. London, Macmillan, 1986.

CEC. Commission’s White Paper of 25 July 2001 on European Governance, COM (2001) 428. http://eur- lex.europa.eu/LexUriServ/site/en/com/2001/com2001_0428en01.pdf

CEC. Commission Communication. Consultation document: “Towards a reinforced culture of consultation and dialogue

– Proposal for general principles and minimum standards for consultation of interested parties by the Commission”,

COM (2002a) 277 final. http://www.thecre.com/eu-oira/pdf/propose.pdf

CEC. Commission Communication. Action plan: “Simplifying and improving the regulatory environment”, COM

(2002b) 278 final. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2002:0278:FIN:EN:PDF

CEC. Report of the Commission on European Governance, COM (2002c) 705 final. http://eur- lex.europa.eu/LexUriServ/site/en/com/2002/com2002_0705en01.pdf

CEC. Commission Communication on the collection and use of expertise by the Commission: principles and guidelines,

“Improving the knowledge base for better policies”, COM (2002d) 713 final. http://ec.europa.eu/governance/docs/comm_expertise_en.pdf

CEC. A European environment and health strategy. Community strategy for environment and health (COM(2003a)338 final). http://europa.eu.int/comm/environment/health/index_en.htm

CEC. Commission Communication “Dialogue with associations of regional and local authorities on the formulation of

European Union policy”, COM (2003b) 811 final. http://www.cor.europa.eu/migrated_data/com2003_0811en01.pdf

Chiti E. Decentralisation and Integration into the Community Administrations: A New Perspective on European

Agencies. J Eur Law 2004; 10: 402-438. http://www3.interscience.wiley.com/journal/118804920/abstract 29 Cohn D. The Best of Intentions, Potentially Harmful Policies: A Comparative Study of Scholarly Complexity and

Failure. J Comp Pol Anal 2004; 6: 39-56. http://www.ingentaconnect.com/content/routledg/jcpa/2004/00000006/00000001/art00004

Craye M, Funtowicz S, Van Der Sluis JP. A reflexive approach to dealing with uncertainties in environmental health risk science and policy. Int J Risk Assess Manage 2009; 5: 216-36. http://www.inderscience.com/search/index.php? action=record&rec_id=7169&prevQuery=&ps=10&m=or

DG-SANCO. The Future of Risk Assessment in the European Union. The Second Report on the Harmonisation of Risk

Assessment Procedures. Adopted by the Scientific Steering Committee at its Meeting of 10-11 April 2003a. http://ec.europa.eu/food/fs/sc/ssc/out361_en.pdf , http://ec.europa.eu/food/whatsnew/oldnews0304_en.html

DG-SANCO. Opinion of the Scientific Steering Committee on Setting the Scientific Frame for the Inclusion of New

Quality of Life Concerns in the Risk Assessment process Adopted on 10-11 April 2003b as part of its exercise on

Harmonisation of Risk Assessment Procedures. http://ec.europa.eu/food/fs/sc/ssc/out357_en.pdf

DG-SANCO. Future challenges paper: 2009-2014. Brussels 2008. http://ec.europa.eu/dgs/health_consumer/future_challenges/future_challenges_paper.pdf

DG-SANCO and SCHER. Opinion on research priorities for the 7th framework program human health and the environment. Scientific Committee on Health and Environmental Risks (SCHER) 2006. http://ec.europa.eu/health/ph_risk/committees/04_scher/docs/scher_o_044.pdf

Douglas M. The Background of the Grid Dimension: a Comment. Sociol Anal 1989; 50.

Douglas M. Risk and Blame: Essays in Cultural Theory. London, Routledge 1996. 336 pp. [Original 1992.]

Douglas M, Wildavsky A. Risk and Culture – An Essay on the Selection of Technological and Environmental Dangers.

Univ California Press, Berkley, Los Angeles, London 1983. 1st Paperback Printing, [Original 1982.]

Eberlein B, Grande E. Beyond delegation: transnational regulatory regimes and the EU regulatory state. J Eur Public

Pol 2005; 12: 89-112. http://www.informaworld.com/smpp/content~content=a713948682~db=all~order=page 30 Eduljee GH. Trends in risk assessment and risk management. Sci Total Environ 2000; 249: 13-23. http://www.ingentaconnect.com/content/els/00489697/2000/00000249/00000001/art00507

EEA. Environmental Policy Integration in Europe: State of play and an evaluation framework. EEA Technical Report

No. 2/2005a. European Environment Agency, Copenhagen.

EEA. Environmental Policy Integration in Europe. Administrative culture and practices. EEA Technical Report No.

5/2005b. European Environment Agency, Copenhagen.

Egeberg M. How federal? The organizational dimension of integration in the EU (and elsewhere). J Eur Public Pol

2001; 8: 728-746. http://www.informaworld.com/smpp/content~content=a713774404~db=all~order=page

ElAmin A. EU's food agency battles attempts to hijack science. Food Quality news.com 21-Sep-2006. http://www.foodqualitynews.com/Legislation/EU-s-food-agency-battles-attempts-to-hijack-science

Elliott A. Beck's Sociology of Risk: A Critical Assessment. Sociol 2002; 36: 293-315. http://soc.sagepub.com/cgi/content/abstract/36/2/293

Ellinas A, Suleiman E. Reforming the Commission: between modernization and bureaucratization. J Eur Public Pol

EP, EC. Directive 2008/1/EC of the European Parliament and of the Council of 15 January 2008 concerning integrated pollution prevention and control (Codified version). Offic J EU, L 24/8-29 (EN). http://eur- lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2008:024:0008:0029:EN:PDF

Farrell H, Héritier A. A rationalist-institutionalist explanation of endogenous regional integration. J Eur Public Pol

2005; 12: 273-290. http://www.informaworld.com/smpp/content~content=a713721823~db=all~order=page 31 Felt U, Wynne B, Callon M, Gonçalves ME, Jasanoff S, Jepsen M, Joly P-B, Konopasek Z, May S, Neubauer C, Rip A,

Siune K, Stirling A, Tallacchini M. Taking European Knowledge Society Seriously. Report of the Expert Group on

Science and Governance to the Science, Economy and Society Directorate, Directorate-General for Research, European

Commission. Luxemburg 2007. EUR 22700. http://ec.europa.eu/research/science- society/document_library/pdf_06/european-knowledge-society_en.pdf

Getimis P, Tsoukalis L, Fleischer T, Swierkocki J, Rees N, Cunha Rato MH. EU Enlargement and Multi-level

Governance in European Regional and Environment Policies: Patterns of Learning, Adaptation and Europeanization

Among Cohesion Countries (Greece, Ireland, Portugal) and Lessons for New Members (Hungary, Poland) – ADAPT

Final report, HPSE-CT-2001-00097. www.lse.ac.uk/collections/hellenicObservatory/research/adaptIntro.htm

Giglioli I, Swyngedouw E. Let's Drink to the Great Thirst! Water and the Politics of Fractured Techno-natures in Sicily.

Int J Urban Regional Res 2008; 32: 392-414. http://www.ingentaconnect.com/content/bpl/ijur/2008/00000032/00000002/art00008

Guttier P. Horizontal Coherence and the External Competences of the European Union. J Eur Law 2004; 10: 23-41. http://www3.interscience.wiley.com/journal/118804895/abstract

Haahr JH. Open co-ordination as advanced liberal government. J Eur Public Pol 2004; 11: 209-230. http://www.informaworld.com/smpp/content~content=a713791020~db=all~order=page

Haas PM. Introduction: epistemic communities and international policy coordination. Int Organization 1992; 46: 1-35. http://journals.cambridge.org/action/displayAbstract?aid=3216868

Habermas J. Communicative Rationality and the Theories of Meaning and Action. In: On the Pragmatics of

Communication. MIT Press, Cambridge, MA 1998. Original 1986.

Hajer M. Policy without polity? Policy analysis and the institutional void. Pol Sci 2003; 36: 175-195. http://www.springerlink.com/content/r106762418714510/fulltext.pdf

Hampton J. Making Grid and Group Operational. Douglas M (ed.), Essays in the Sociology of Perception. Routledge &

Kegan Paul, London 1982, pp. 64-82. 32 Health Canada. Health Products and Food Branch Public Involvement Performance Report 2005-2007. Improving stakeholder trust and confidence in the regulatory system. Health Canada, Ottawa 2007. http://www.hc-sc.gc.ca/ahc- asc/alt_formats/hpfb-dgpsa/pdf/pubs/pub-invol-perf-rap_rap-rend-part-pub_2007-eng.pdf

Herodes M, Adelle C, Pallemaerts M. Environmental Policy Integration at the EU Level – A Literature Review.

EPIGOV Paper No. 5, Ecologic, Berlin 2007. http://www.ecologic.de/projekte/epigov/documents/EPIGOV_paper_5_herodes_adelle_pallemaerts.pdf

Horlick-Jones T. Meaning and contextualisation in risk assessment. Rel Eng System Saf 1998; 59: 79-89. http://dx.doi.org/10.1016/S0951-8320(97)00122-1

IPCS. Integrated risk assessment. Report prepared for the WHO/UNEP/ILO International Programme on Chemical

Safety. WHO/IPCS/IRA/01/12, IPCS, Geneva, 2001. http://www.who.int/ipcs/publications/new_issues/ira/en/

IPCS. A framework for the Integration of Health and Ecotoxicological Risks. IPCS, Geneva, 2004. http://www.who.int/pcs/emerg_site/integr_ra/ira_report.htm

IPCS, OECD. Descriptions of selected key generic terms used in chemical hazard/risk assessment. International

Programme on Chemical Safety and Organization for Economic Cooperation and Development, Geneva, Paris, 2003. www.who.int/entity/ipcs/publications/methods/harmonization/definitions_terms/en/

IRGC. White Paper on Risk Governance: Towards an Integrative Approach. International Risk Governance Council,

Geneva, 2005. http://www.irgc.org/IMG/pdf/IRGC_WP_No_1_Risk_Governance__reprinted_version_.pdf

ISO. Draft International Standard ISO/DIS 31000, Risk management -- Principles and guidelines. ISO/PC 992/TMB,

2008. http://www.iso.org/iso/iso_catalogue/catalogue_ics/catalogue_detail_ics.htm? ics1=03&ics2=100&ics3=01&csnumber=43170

Jacob K, Volkery A. Institutions and Instruments for Government Self-Regulation: Environmental Policy Integration in a Cross-Country Perspective. J Comp Pol Anal 2004; 6: 291-309. http://www.ingentaconnect.com/content/routledg/jcpa/2004/00000006/00000003/art00004

Jasanoff S. Procedural choices in regulatory science. Technol Soc 1995; 17: 279-293. http://dx.doi.org/10.1016/0160-

791X(95)00011-F 33 Jasanoff S. Designs on nature. Science and democracy in Europe and the United States. Princeton University Press,

Princeton and Oxford 2005.

Johnson GF. Taking Stock: The Normative Foundations of Positivist and Non-Positivist Policy Analysis and Ethical

Implications of the Emergent Risk Society. J Comp Pol Anal 2005; 7: 137-153. http://www.ingentaconnect.com/content/routledg/jcpa/2005/00000007/00000002/art00002

Jones E. Idiosyncrasy and integration: suggestions from comparative political economy. J Eur Public Pol 2003; 10: 140-

158. http://www.jhubc.it/facultypages/ejones/JEPP-jones.pdf

Jordan A, Wurzel R, Zito A. Has governance eclipsed government? Patterns of environmental instrument selection and use in eight states and the EU. CSERGE Working Paper EDM 03-15, 2003. http://www.uea.ac.uk/env/cserge/pub/wp/edm/edm_2003_15.pdf

Kaiser R, Prange H. Managing diversity in a system of multi-level governance: the open method of co-ordination in innovation policy. J Eur Public Pol 2004; 11: 249-266. http://www.informaworld.com/smpp/content~content=a713791022~db=all~order=page

Kasperson R, Renn O, Slovic P, Brown HS, Emel J, Goble R, Kasperson JX, Ratick S. The Social Amplification of

Risk: A Conceptual Framework. Risk Anal 1988; 8: 177–187. http://www3.interscience.wiley.com/journal/119462808/abstract

Knill C. Private governance across multiple arenas: European interest associations as interface actors. J Eur Public Pol

Knill C. Introduction: Cross-national policy convergence: concepts, approaches and explanatory factors. J Eur Public

Pol 2005; 12: 764-774. http://www.uni-konstanz.de/FuF/Verwiss/knill/projekte/envipolcon/downloads/Knill.pdf

Knill C, Lenschow A. Compliance, Communication and Competition: Patterns of EU Environmental Policy Making and Their Impact on Policy Convergence. Eur Environ 2005; 15: 114–128. http://www.uni- konstanz.de/FuF/Verwiss/knill/projekte/envipolcon/downloads/European_Environment_PDF.pdf 34 Kohler-Koch B. Framing: the bottleneck of constructing legitimate institutions. 2000; 7: 513-531. http://www.informaworld.com/smpp/content~content=a713773838~db=all~order=page

Kracauer S. History. The Last Things Before the Last. Completed by Oskar Kristeller with a new preface. Marcus

Wiener Publishers, Princeton, 1995, 269 pp. Orig. publ. Oxford UP, New York 1969.

Krapohl S. Risk regulation in the EU between interests and expertise: the case of BSE. J Eur Public Pol 2003; 10: 189–

207. http://www.informaworld.com/smpp/content~content=a713773516~db=all~order=page

Kriesi H, Adam S, Jochum M. Comparative analysis of policy networks in Western Europe. J Eur Public Pol 2006; 13:

341-361. http://www.informaworld.com/smpp/content~content=a743896332~db=all~order=page

Laffan B. The European Union polity: a union of regulative, normative and cognitive pillars. J Eur Public Pol 2001; 8:

709-727. http://www.informaworld.com/smpp/content~content=a713774403~db=all~order=page

Latour B. On Actor Network Theory: A few clarifications plus more than a few complications. 14 pp. http://www.cours.fse.ulaval.ca/edc-65804/latour-clarifications.pdf See also: http://www.nettime.org/Lists-

Archives/nettime-l-9801/msg00019.html

Lazer D. Regulatory interdependence and international governance. J Eur Public Pol 2001; 8: 474-492. http://www.informaworld.com/smpp/content~content=a713773912~db=all~order=page

Levy RP. European Commission Overload and the Pathology of Management Reform: Garbage Cans, Rationality and

Risk Aversion. Public Admin 2006; 84; 4123-39. http://www3.interscience.wiley.com/cgi- bin/fulltext/118609443/PDFSTART

Luhmann N. Risk: A Sociological Theory. Aldine De Gruyter, New York 2005. Original 1993. http://books.google.com/books? hl=fi&lr=&id=PzPtFxgv8FwC&oi=fnd&pg=PR7&dq=risk+kirjoittaja:Luhmann+kirjoittaja:N&ots=yKKm_AcSG8&si g=4p2C-gx-yL5eG0bD2ZlbRqWAuNs#PPP1,M1

Lyall C, Tait J. New Modes of Governance. Developing an Integrated Policy Approach to Science, Technology, Risk and the Environment. Ashgate, Aldershot and Burlington, VT 2005. http://books.google.com/books? 35 hl=fi&lr=&id=9uw4US8Xq_UC&oi=fnd&pg=PA3&dq=kirjoittaja:Lyall,

+kirjoittaja:Tait&ots=15Tv_vDGkU&sig=RCdeHkc42fZzwXijth4951sQW5k

Löfstedt R. The swing of the regulatory pendulum in Europe: From precautionary principle to (regulatory) impact analysis. J Risk Uncert 2004; 28: 237-260. http://www.springerlink.com/content/uj172553465k2402/fulltext.pdf

Macgill SM, Sie YL. The nature of risk. J Risk Res 2004; 7: 315-352. http://www.informaworld.com/smpp/content~content=a713704981~db=all

Meijerink S. Understanding policy stability and change. the interplay of advocacy coalitions and epistemic communities, windows of opportunity, and Dutch coastal flooding policy 1945-2003. J Eur Public Pol 2005; 12: 1060-

1077. http://www.informaworld.com/smpp/content~content=a727246136~db=all~order=page

Millstone E, van Zwanenberg P, Levidow L, Spök A, Hirakawa H. Risk Assessment Policy – Differences between jurisdictions. JRC Scientific and Technical Reports 23259 EN. IPTS, Seville, 2008. http://ftp.jrc.es/EURdoc/JRC37719.pdf

Moravcsik A. The European constitutional compromise and the neofunctionalist legacy. J Eur Public Pol 2005; 12: 349-

386. http://www.informaworld.com/10.1080/13501760500044215

Mythen G. Reappraising the Risk Society Thesis. Telescopic Sight or Myopic Vision? Current Sociol 2007; 55: 793-

813. http://csi.sagepub.com/cgi/content/abstract/55/6/793

National Environmental Justice Advisory Council 2004. Ensuring risk reduction in communities with multiple stressors: environmental justice and cumulative risks/impacts. http://www.epa.gov/compliance/resources/publications/ej/nejac/nejac-cum-risk-rpt-122104.pdf [April 17 2009]

Nilsson M, Persson Å. Framework for analysing environmental policy integration. J Environ Pol Planning 2003; 5: 333-

359. http://www.informaworld.com/smpp/content~content=a713624528~db=all

NRC. Risk assessment in the federal government: Managing the process. National Academy Press, Washington DC

NRC. Science and judgment in risk assessment. National Academy Press, Washington DC 1994. 36 NRC. Understanding Risk: Informing Decisions in a Democratic Society. National Academy Press, Washington DC

1996. http://books.nap.edu/openbook.php?record_id=5138&page=2

NRC. Science and Decisions: Advancing Risk Assessment. National Academy Press, Washington DC 2008. http://www.nap.edu/catalog.php?record_id=12209

OECD. Regulatory Policies in OECD Countries. From Interventionism to Regulatory Governance. OECD Publishing,

Paris 2002. http://books.google.fi/books?id=uV99wk_QdfQC&dq=%22Regulatory+policies+in+OECD+countries

%22&printsec=frontcover&source=bl&ots=cp31-xkTME&sig=MK9-GHzlmpGLw54YFwKvNtRE-

PY&hl=fi&sa=X&oi=book_result&resnum=8&ct=result

OECD. Risk Assessment. In: Emerging Systemic Risks in the 21th Century. OECD, Paris 2003a, pp. 63-113. http://www.bvsde.paho.org/bvsacd/cd46/risk-cap2.pdf

OECD. Stakeholder Involvement Tools: Criteria for Choice and Evaluation. Proc Topical Session at 4th meeting of the

NEA Forum on Stakeholder Confidence. May 22, 2003b. http://www.nea.fr/html/rwm/docs/2003/rwm-fsc2003-10.pdf

Putnam H. The collapse of the fact/value dichotomy and other essays. Harvard UP, Cambridge, MA, London 2002.

Radaelli C. Europeanization, Policy Learning, and New Modes of Governance. J Comp Pol Anal 2008, 10: 239-254. http://www.ingentaconnect.com/content/routledg/jcpa/2008/00000010/00000003/art00004

Rayner S. Sociocultural Definitions of Risk. ORNL, Oak Ridge, TN 1990. http://www.osti.gov/energycitations/servlets/purl/6206627-C2T9Or/6206627.pdf

Rayner S, Cantor R. How Fair Is Safe Enough? The Cultural Approach to Societal Technology Choice. Risk Anal 1987;

7: 3-9. http://www3.interscience.wiley.com/journal/119481804/abstract?CRETRY=1&SRETRY=0

Renn O. Risk Governance: Basic elements and requirements for communication and participation. Renn O, Benighaus

C (eds.), Report from the Stuttgart Workshop on Risk Communication, 2007, pp. 14-21. http://nomiracle.jrc.it/Documents/PublicDeliverables/D.4.3.14%20Report%20from%20the%20Stuttgart%20Workshop

%20on%20Risk%20Communication.pdf

Renn O. Risk governance: coping with uncertainty in a complex world. London, Earthscan 2008. 455 pp. 37 Renn O, Keil F. Systemische Risiken: Versuch einer Charakterisierung. GAIA-Ecological perspectives for Science and

Society 2008; 17: 349-354. http://www.ingentaconnect.com/content/oekom/gaia/2008/00000017/00000004/art00009

Rizak S, Hrudey SE. Interdisciplinary comparison of expert risk beliefs. J Environ Eng Sci 2005; 4 (3): 173–185. http://rparticle.web-p.cisti.nrc.ca/rparticle/AbstractTemplateServlet? calyLang=eng&journal=jees&volume=4&year=2005&issue=3&msno=s04-052

Rubik F, Scholl G. Integrated Product Policy (IPP) in Europe—a development model and some impressions. J Cleaner

Prod 2002; 10: 507-515. http://dx.doi.org/10.1016/S0959-6526(02)00016-1

Rudén C. Principles and practices of health risk assessment under current EU regulations. Regulat Toxicol Pharmacol

2006; 44: 14-23. http://dx.doi.org/10.1016/j.yrtph.2005.04.009

Russell B. Human Knowledge, Its Scope and Limits. Part V. Simon and Schuster, New York 1948.

Saurugger S. 2005. Europeanization as a methodological challenge: The case of interest groups. J Comp Pol Anal 2005;

7: 291-312. http://www.ingentaconnect.com/content/routledg/jcpa/2005/00000007/00000004/art00003

Schout A, Jordan AJ. Coordinated European Governance: Self-Organizing or Centrally Steered? Public Admin 2005;

83: 201-220. http://www.ingentaconnect.com/content/bpl/padm/2005/00000083/00000001/art00012

Schout A, Jordan A. The European Union's governance ambitions and its administrative capacities. J Eur Public Pol

Schäfer A. A new form of governance? comparing the open method of co-ordination to multilateral surveillance by the

IMF and the OECD. J Eur Public Pol 2006; 13: 70-88. http://www.informaworld.com/smpp/content~content=a727419721~db=all~order=page

Sjöberg L. Risk Perception in Western Europe. Ambio 1999; 28 (6): 543-549.Surel Y. The role of cognitive and normative frames in policy-making. J Eur Public Pol 2000; 7 (4): 495-512. http://www.informaworld.com/smpp/content~content=a713773846~db=all~order=page 38 Suter GW II, Munns WR, Sekizawa J. Types of Integration in Risk Assessment and Management, And Why They Are

Needed. Hum Ecol Risk Assess 2003; 9: 273-279. http://www.ingentaconnect.com/content/tandf/bher/2003/00000009/00000001/art00018

Swedlow B. Toward Cultural Analysis in Policy Analysis: Picking Up Where Aaron Wildavsky Left Off. J Comp Pol

Anal 2002; 4: 267-285. http://www.ingentaconnect.com/content/routledg/jcpa/2002/00000004/00000003/05097906

Thompson M, Wildavsky A. A Cultural Theory of Information Bias in Organizations. J Mgmt Studies 1986; 23: 273-

286. http://www3.interscience.wiley.com/journal/119489960/abstract

Treib O, Bähr H, Falkner G. Modes of governance: towards a conceptual clarification. J Eur Public Pol 2007; 14: 1-20. http://www.informaworld.com/smpp/content~content=a770441484~db=all~order=page

Trondal J. Is there any social constructivist-institutionalist divide? Unpacking social mechanisms affecting representational roles among EU decision-makers. J Eur Public Pol 2001; 8: 1-23. http://www.informaworld.com/smpp/content~content=a713773878~db=all~order=page

USEPA. Framework for Cumulative Risk Assessment. U.S. Environmental Protection Agency, Washington DC 2003. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=54944

van Asselt MBA, Vos E. The Precautionary Principle and the Uncertainty Paradox. J Risk Res 2006; 9 (4): 313-336. http://dx.doi.org/10.1080/13669870500175063

Van Leeuwen CJ, Vermeire TG. General Introduction. Van Leeuwen CJ, Vermeire TG (eds,) Risk Assessment of

Chemicals. An Introduction (2nd ed.). Springer Publishers, Dordrecht, 2007, pp. 1-36. http://books.google.com/books? hl=fi&lr=&id=-ltZ0K1TcqAC&oi=fnd&pg=PA1&dq=+%22Risk+assessment+of+chemicals

%22+kirjoittaja:CJ+kirjoittaja:Van+kirjoittaja:Leeuwen&ots=vuGlLaOEwP&sig=phhwflGXaIzExcGV9-nFnRc_mvo

Wenzel E. Environment, development and health: ideological metaphors of post-traditional societies? Health Educ Res

1997; 12 (4): 403-418. http://www.ncbi.nlm.nih.gov/pubmed/10176371?dopt=Abstract

Wildawsky AB. But Is It True? A Citizen's Guide to Environmental and Health Safety Issues. Harvard UP, Cambridge,

MA, London 1995. 39 Wildavsky A, Dake K. Theories of Risk Perception: Who Fears What and Why? Daedalus 1990; 119: 41-60. Reprinted in Risk, Burger EJ Jr (ed.), Michigan UP 1993, pp. 41-60. http://books.google.com/books? hl=fi&lr=&id=8AIrQ78VVmQC&oi=fnd&pg=PA41&dq=+%22Theories+of+Risk+Perception:

+Who+Fears+What+and+Why%3F%22&ots=yhDEqB7N0U&sig=EEUYoinTQRw4D7tw0-lljO50A9o#PPP1,M1

Wilkinson I. Social Theories of Risk Perception: At Once Indispensable and Insufficient. Current Sociol 2001; 49: 1-22. http://csi.sagepub.com/cgi/content/abstract/49/1/1

World Bank. Regional Economic Prospects 2009 - Europe and Central Asia. http://go.worldbank.org/G6D92T0L50

Wright J, Parry J, Scully E. Institutionalizing policy-level health impact assessment in Europe: is coupling health impact assessment with strategic environmental assessment the next step forward? Bull World Health Organ 2005; 83 (6): 472-

477. http://www.ncbi.nlm.nih.gov/pubmed/15976899?dopt=Abstract

Zahariadis N. Ambiguity and choice in European public policy. J Eur Public Pol 2008; 15: 514-530. http://www.informaworld.com/smpp/content~content=a792862737~db=all~order=page 40 Table 1. Key traits of EU and state level actors and interventions in regulatory sectors important for risk policy integration (modified and extended from Herodes et al., 2007, based on EEA, 2005a).

Sector Types of State EU Key EU-level Risk issues Risk integration issues principal inter- inter- policy target actors vention ventio instruments n Environme Enterprises, High Low to Legal controls Biodiversity/habita Multistressor nt citizens, CSOs (esp. high Economic and t Multimedia (heterogeneo some information Climate change Adapt/sustainability us) areas) based Invasive sp., GMOS Environ pol integr instruments Chemicals Health Service High Low Some legal Diet, drugs, Multifactoriality providers, controls lifestyle Environ+health clients Information Pathogens Integr food safety based Chronic/ageing Global health instruments Contaminants Citizen Citizens, High Low Crime, disasters Integrated security authorities (nat/tech) preparedness + control Agricultur Farmers, High High Subsidies Chemicals Integrated pest control e (incl. industry, Pests Food-feed quality veterinary consumers GMOs Natural + social factors ) (many, Droughts/meteorol Climate change assess cohesive) ogy Fisheries Fishers, Low to High Legal controls Overfishing Sustainable fisheries industry, mediu (hard- Subsidies Contamination Integr. coastal consumers m enforce Pathogens zonemgmt / marine (few, d) policy cohesive) Industry Producers, Medium Mediu Legal controls Accidents (chem. Integr product policy consumers + m Subsidies etc) Environ+health downstream Nano, bio etc new Total quality tech Product liability/econ Energy Governments, High Low Financial Pollution Fuel cycles producers, support (fuel,/emission) System reliability consumers Legal controls Climate change Energy policy Security integration Transport Governments, High Low Limited Accidents Integral safety industries, financial Air pollution Integral sustainable operators, support Climate change mobility systems users Legal controls Common Producers, Low High Legal controls Trade related Harmonization market consumers Trade-risk links Economics Governments, High Low Fiscal Financial risks Integral security and and banks/financin measures Resource access insurance finance g, taxpayers Legal controls Sustainability Market + regulation Developm Governments, Medium Mediu Financial Water/infra Multi-level integration ent producers, m support Food security Multi-sector integration CSOs Climate adaptation 41 42 43 44 45

PREFACE

In recent years, integrated risk assessment and risk management or governance have attracted considerable interest, both in scientific and policy making communities. In the European Union 46 and in other international settings, various attempts have been made to clarify, develop and implement such integration. In particular, risk assessment of chemicals has been increasingly extended to account for the existence of cumulative risks from multiple stressors.

Attempts at integrated treatment of risks have emerged from straightforward practical grounds in management contexts or following theoretical or research-driven considerations. In both cases the focus has usually been on the natural scientific and technical aspects of risk assessment.

Until now, integrated risk assessments have however not in the first place been based on detailed or explicit consideration of the particular needs they have to address, on their possibilities and limits in this respect, and neither on their (possible) roles and usefulness in the policy context. There has been a lack of structured information regarding the views on these aspects among regulators and stakeholders that play key roles in the development and uptake of more integrated approaches. The various interpretations and underlying concepts of integrated risk assessment, multiple stressors and cumulative effects have not yet been mapped. As a result, knowledge of the barriers to and opportunities for the introduction and uptake of integrated risk assessments has been scarce.

In the present report we will present and discuss the results of a series of interviews that aimed to shed light on the possible role and further uptake of integrated risk assessment in the interface of science and policy making in Europe, in the field of chemicals and of multiple stressors in general. We explore how judgments of integrated risk assessment are influenced by more general underlying views about risks, their assessment and management, and the related uncertainties and complexities. The overall objective of our study is thereby to contribute to future developments that could help to bridge the gaps between production and use of more integrated and richer knowledge on risks, including awareness of the related limits of knowledge. 47 TABLE OF CONTENTS

Executive summary

Foreword

1 Introduction

1.1 The challenge of multiple and complex risks and of narrow risk assessments

1.2 NoMiracle project as a response to the challenge

1.3 Objectives

2 Methods and materials

2.1 General

2.2 Interviews

2.3 Materials obtained

3 Synthesis of results by interview themes

3.1 Framings and concepts of integrated risk assessment

3.2 Addressing cumulative risks and multiple stressors in policy making

3.2 1 The need to develop policies for dealing with cumulative risks

3.2.2 Challenges for addressing cumulative risks of multiple stressors

3.2.3 Participation and communication in addressing cumulative risks

3.3 The roles of information, knowledge, uncertainty and precaution

3.3.1 Status of the knowledge base, opportunities and limitations

3.3.2 Treatment of uncertainties

3.3.3 The role of precaution

3.4 Integration in risks, risk assessment and risk management, and its uptake in the policy sphere

3.4.1 Integration: why, what types and which limits?

3.4.2 Interaction between assessment and management

3.4.3 Obstacles to and possible problems with more integrated assessment and management

3.5 Integrated chemicals assessment and management in relation to REACH legislation

3.5.1 Integrated treatment of risks in REACH 48

3.5.2 Challenges to the implementation of REACH

3.5.3 The generation, collection, processing and transmittal of information in REACH

4 Interpretation and relationships with the survey of expert and stakeholder opinions

4.1 General considerations

4.2 Summary of relevant findings from the survey

4.3 Comparisons of the survey and the interviews

4.4 Variations in views on policy-related dimensions of integration

4.5 Commonalities and key notions in framing integrated risk assessment

4.5.1 The need for integration: addressing the limits of approaches and methodologies to address complex situations

4.5.2 Effective communication between science and policy: assuring management relevance of innovative approaches

5 Barriers and opportunities for integrated risk assessment

5.1 Categories and general characteristics of barriers

5.2 Epistemological and ontological barriers

5.3 Regulatory barriers

5.4 Socio-political barriers

6 Conclusions, directions and opportunities for further development

6.2 Working principles for developing and implementing strategies for integrated treatment of risks

6.2.1 Co-production of policy initiatives and knowledge on cumulative risks

6.2.2 Clarifying and reinforcing links between assessment and management

6.2.3 Addressing complexity in the context of adaptive risk management strategies

6.3 Prioritising areas on which to focus integrated approaches

6.4 Summary of recommendations and suggestions

6.4.1 Research and methods development

6.4.2 Forward-looking expert-stakeholder-policy deliberations

6.4.3 Policy development and application 49

Acknowledgments

References

Annexes

1 Structure of the interviews and specific questions

2 Letter of invitation to participate in interviews 50 EXECUTIVE SUMMARY

As part of the EU's R&D project NoMiracle (Novel Methods for Integrated Risk Assessment of Cumulative Stressors in Europe), views among regulators and stakeholders on integrated risk assessment and management were studied. The study was based on 16 semi-structured interviews that aimed at clarifying underlying concepts of those views. The interviews were carried out in 2007 and complemented an earlier Internet survey among experts and stakeholders. The study took place in the context of research activities that included also other empirical studies such as focus groups and workshops, literature studies and document analyses, and theoretical background studies.

During the interviews, integrated risk assessment of cumulative stressors was first addressed in a general and broad way in order to clarify framings of these concepts. Subsequently, the interviews focused on integrated risk assessment in the context of chemicals control in the EU, specifically in connection with the REACH legislation. Its development and implementation challenges were used as a case to discuss more general issues in (integrated) risk assessment and governance.

A variety of interpretations of key concepts such as integrated risk assessment, cumulative risks and multiple stressors was observed. Some were based on optimistic beliefs that risks, specifically those from chemicals, can be largely known and controlled, even though their multiplicity was commonly seen as a major challenge. Other interpretations were grounded in an increased and more explicit awareness of the myriad uncertainties and ambiguities in risk assessment and management. Respondents also referred to many institutional and broader societal challenges, specifically in the aftermath of controversies that surrounded the development of the REACH legislation.

Typically, cumulative risks from multiple stressors were seen as a challenge that can be addressed in the longer term only, not least because of current administrative pressures in the EU chemicals control. Explicitly addressing uncertainties and ambiguities in assessments was also often seen as a complication rather than an opportunity to direct attention to crucial issues and knowledge options.

Interviewees acknowledged both the need and some opportunities for integrated risk assessment, specifically in the area of chemicals, and also in other areas where more integrative approaches to risks have been adopted, such as environmental and public health and occupational safety. On the other hand, needs to specify, narrow down and decompose risk problems were also expressed. Many felt that the relevance and efficiency of risk assessments could be enhanced through integrating risk-benefit considerations throughout the process, while such an idea was remote to others. 51

Regarding the REACH legislation, some interviewees pointed to the requirements for more integrative assessment that are already present in REACH procedures: a more comprehensive treatment of environmental compartments, target organisms, product life-cycle stages and phases of risk policy cycles, especially assessment and management. All in all, however, despite some provisions for addressing chemical mixtures, the focus of REACH was seen as being still on single substances.

Views about the interaction between the assessment and management phases in risk governance varied and showed some inconsistencies. In general, a need for more interfacing between assessment and management was recognized, although quite some skepticism was expressed that this could endanger the independence of assessors. Statements on the precautionary principle confirmed the existence of its multiple interpretations and the persisting confusion in its implementation.

Notwithstanding the broad variety of views expressed during the interviews, the obtained material allows to discern, at general level, some commonalities in the challenges for integrated assessment: such framing could be seen to converge around two key notions that characterize both the need for integrated assessment and the key challenge in its conduct, namely ‘more realism’ and ‘manageable approaches’ in dealing with complex risk issues. Interviewees repeatedly expressed the concern that integrated approaches should not become too complicated, risking to be irrelevant for policy.

The interview results allowed us to identify a number of conceptual and structural (social and institutional) barriers for the development and deployment of integrated risk assessment. These include:

 The inherent complexity of combined and cumulative risks from multiple stressors, implying conceptual and resource constraints for meaningful knowledge development, fears of lacking relevance and realism of produced knowledge and fears of confusing messages to its use context

 A regulatory environment that favors reasoning in terms of narrowly specified mandates, individual liabilities and separated consideration of risks rather than their integration

 A related difficulty to discern, even when emphasizing integration, between justified and excessive or unfeasible integration of risks, and to combine integrative and differentiating approaches

 A socio-political and cultural environment in which actors tend to be locked in established approaches such as the single substance approach to chemical risks, not least because it gives reasonable certainty of what they can expect in relation to the confines of their field and to their 52

mindsets, liabilities and objectives. The opportunities opened up by considering integration and novel approaches to combined risks were thus often assessed in terms of the possible implications for the interests and particular views of actors.

The interpretation of the obtained material delivers suggestions for furthering the development and deployment of integrated approaches. These include:

 Clear priority setting and concentration of efforts on a number of areas where combined treatment of risks is, in a broad collective deliberation, judged most feasible in terms of research effort, policy developments and information requirements and can lead to tangible and useful results in the not-so-long term. Analysis of the interview material provides clues that can inform the priority setting for integrative approaches:

o Build further on experiences (be it in policy domains or scientific fields) where combined risks and integrated assessment have been practiced

o Concentrate the development of integrated approaches in first instance within regulatory frameworks that are receptive

o Develop integrated approaches in areas where the positions of stakeholders are favourable, i.e. where they think that addressing ‘combined risks’ could lead to policies that fit well within their positions and perspectives

o Build on discussions on the qualities context-dependencies, limits and opportunities of risk integration to define acceptable and workable specific configurations for it.

 Connect integrated approaches firmly with their context of use (social debates and processes, policy, regulation, decision-making) to enhance their relevance. To this end, a number of guiding principles are proposed for accompanying activities: co-production of policy initiatives and knowledge development on cumulative risks; participatory deliberation on the meanings, goals and means of integrated treatment of risks; clarifying and reinforcing links between assessment and management; addressing complexity in the context of adaptive and stepwise approaches. 53 1 INTRODUCTION

1.1 The challenge of multiple and complex risks and of narrow risk assessments

Traditional, well established risk assessment, based on study or prediction of one effect caused by one agent, is often seen by regulators, stakeholders and other actors as inadequate and too narrow. It does not take into account a number of aspects that are increasingly seen as crucially important, such as the possible cumulative, interacting and indirect effects of exposure to multiple agents, the complexity and indeterminacy of risk situations (impacts on different environmental compartments, the possibility of many simultaneous exposure routes, the dependency of risks on hard-predicted factors such as human behavior, and so forth). Furthermore, much of the related uncertainties and conceptual ambiguity of risks – entailing questions about the relevant spatial and time scale, the accumulation and lags in risks and the multiple, also non-technical, dimensions of risk issues – are not explicitly addressed in such traditional risk assessments.

Among other fields, chemicals control shows evidence of an increased awareness to broaden traditional narrow risk assessments in order to address more adequately complex mixtures of potentially harmful (and useful) substances (CEC, 2001). Additionally, the need is recognized to relate risks from chemicals to those from other stressors, e.g. to identify priorities in environmental, health and other policies. On various occasions, in many ways and for different applications, official EU policy documents have stressed the importance of a more integrated treatment of risks (CEC, 2003b; CEC DG SANCO, 2002-2008; EP and EC, 2002, 2008).

The multiplicity and complexity of risks present a challenge both for data gathering, technical information and the development and dissemination of scientific knowledge as well as for management and governance including regulation. While there is increasing awareness in policy making circles of the need to look at risks from a broader and more inclusive perspective, very often narrow framings remain imposed by the regulatory context (specific pieces of legislation and the administrative and institutional structures and functions in place).

Within chemicals control, a key development has been the preparation and enactment of the EU's REACH (Registration, Evaluation and Authorization of Chemicals) legislation. REACH has a certain rationale, framing and structure, and includes some provisions for integrative risk assessment and management (CEC, 2006; see also CEC, 2003a). These are however limited by its fundamental orientation: even REACH, as the most influential new legislation on chemicals, mainly addresses single substances (Assmuth et al., 2009c). Its relationships with other areas of regulation affecting chemicals are also an issue that still needs to be developed. 54

A particular challenge to integrated risk assessment is that it should not only address more complicated risk situations in terms of the simultaneous presence of causative agents and types of effects. Ideally, the efforts to address in a broader and more realistic way the biophysical phenomena in risk situations should go hand in hand with attention to a plurality of meanings and interpretations of these risk experiences (Wynne, 1996). Traditional biophysical interpretations often need to be complemented by considering social aspects of risks, including the concerns of affected citizens and large segments of societies (Kasperson et al., 1988; Horlick-Jones, 1998; Renn and Klinke, 2002; Aven and Christensen, 2005; Renn and Benighaus, 2006; Renn, 2007, 2008).

Interpretations of 'integrated' or 'cumulative' risk assessment have varied widely in regulatory contexts (IPCS, 2001, 2004; CEC, 2003a, CEC DG-SANCO, 2003a, b; Bridges, 2003; Munns et al., 2003; USEPA, 2003; NRC, 2008), but have seldom included societal dimensions (see Renn, 1992; Assmuth and Hildén, 2008; Assmuth et al., 2009a-c). This reflects interpretations of risks in general, often seen mainly in natural scientific or technical terms. For instance, in their discussion of opportunities for integrated risk assessment, Vermeire et al. (2007) consider some socio-political aspects but only in terms of scientific and technical consequences: "The major emerging opportunities for an integrated approach stem from the increasing societal and political pressure to move away from vertebrate testing leading to a demand for scientific integrated approaches to … testing, as well as to computer simulations, in so-called Intelligent Testing Strategies".

R&D projects, also in EU programmes, provide an opportunity to develop and test novel ways to deal with the multiplicity and complexity of risks. However, even here traditional interpretations and approaches to risks often dominate. For instance, among the multitude of EU RTD projects addressing environmental or health risks in the area of chemicals and water, most have had a technical and natural scientific orientation (Assmuth, 2008). A main challenge then is to better match the R&D and policy agendas, enhance multi-disciplinarity, and facilitate a co-production of innovative but practicable approaches to multiple risks.

1.2 NoMiracle project as a response to the challenge

As one response to the challenges of developing alternative approaches to narrow risk assessments, the Integrated Project NoMiracle under the EU's 6th Framework Programme for Research seeks to explore and develop, according to its full name, "Novel methods for integrated risk assessment of cumulative stressors in Europe". As one of several activities that address integrated risk assessment (see www.cordis.europa.eu), NoMiracle focuses on risks from chemical 55 mixtures and on related multi-stressor issues (http://nomiracle.jrc.it/webapp/ViewPublicDeliverables.aspx).

Along with chemical mixtures, the NoMiracle project emphasizes integrated assessment of ecological and human health risks, of exposures and effects, and of different geographical scales. In searching for ways to assess multiple risks more 'realistically' (as specified in the call), NoMiracle has emphasized novelty of methods, often interpreted in terms of new empirical measurements or modeling methods and of improved understanding of the natural processes and phenomena underlying risks from chemicals and other stressors in biological systems.

While most of the work in the NoMiracle project thus has the character of natural scientific R&D, it does include a Work Package (4.3) "Dealing with complex risks and uncertainties in a management context" that addresses risk perception, risk communication and assessment-policy interfaces, and methodological frameworks in this area. Even though other parts of the NoMiracle project also include work on risk communication, this is done in an applied way only, in relation to dissemination.

In Work Package 4.3, an Internet survey on views of integrated risk assessment and management was conducted among ca. 1000 experts, regulators, policy-makers and stakeholders mainly in the EU (Assmuth et al., 2007). The results showed that there is a pronounced variability in views on risks and uncertainties and on the use of risk information. Some respondents regarded information on risks mainly as something that should be confined to expert treatment and analysis in regulatory procedures, whereas others saw most risk information as an essential element of public debate. Overall, the results demonstrated that views on risks and integrated risk assessment cannot be reduced to any simple model without omitting crucial aspects of risks and of approaches to them (see Chapter 4). The findings from the survey prompted us to study in more detail the views and thinking of key persons specifically in the EU administration and among EU-level stakeholders to substantiate, refine and further develop the insights.

In order to develop novel methods for integrated assessment of risks from cumulative stressors that would be relevant for those using the methods and acting on such risks, as addressed in the NoMiracle project, the views of those users need to be clarified. These views include:

 The framings and interpretations of the key concepts involved

 The perceived needs for and aims of a wider deployment of integrated risk assessment

 The expectations with respect to the development and use of integrated risk assessment

 The cautions or qualifications, limitations and possible disadvantages of a focus on integrated risk assessment. 56

1.3 Objectives

The general objective of the present study was to identify and interpret concerns, views and issues among EU-level regulatory risk managers and risk governance actors in relation to integrated treatment of risks. Furthermore, we wanted to relate this information to background concepts of risks. Specifically, we aimed at approaches to risk assessment. Particular attention was paid to assessment and management of chemical risks in a multi-stressor setting.

More specifically, the objectives can be stated as follows:

1. Specify how cumulative risks are framed by key actors, how they are addressed in the policy sphere and how knowledge from integrated assessments fits within those frames. In this respect, the interviews aimed to detail and complement findings from the above mentioned internet survey.

2. Obtain insights on barriers to and opportunities for the uptake and deployment of knowledge on cumulative risks

3. In particular, examine views on integrated risk assessment of cumulative stressors in connection with the REACH legislation of the EU

4. Formulate recommendations or suggestions regarding developments in policy and research contexts to create conditions for fruitful co-production of initiatives on cumulative risks and their integrated assessments.

57 2 METHODS AND MATERIALS

The information for this study was gathered through interviews with key EU-level actors familiar with chemical risk management.

Acknowledging the limits to gaining ascertained and generalizable knowledge, we have chosen a strategy that is at once more modest and more ambitious. Specifically, we have studied views and thoughts of actors in exploratory, semi-structured and informal manner, instead of chasing illusory robustness, reproducibility, objectivity and representativeness.We thus focused on identifying and discussing framings, arguments and preferences in relation to integrated risk assessment in an interactive joint process of exploration.

We have supported the information gained from the interviews with information from other studies including theoretical reasoning and developments. In particular, the interviews have been grounded in the earlier survey of expert and stakeholder opinions.

Particular attention was paid to dealing with uncertainties and to the interface between assessment and management in the case of integrative approaches to risks. Thus, we focused on the roles of knowledge and the relationships between research and methods development, on one hand, and on management and governance, on the other. This included inquiries in how these areas of activity are framed and how their relationships are conceived.

The scope of the interviews was deliberately broad to allow different framings and approaches to risks. However, in the context of the NoMiracle project we focused on risk governance at EU level and on chemical risks (including relevant links with other areas such as parts of food safety).

2.2 Interviews

To solicit views of regulators and stakeholders, we used semi-structured thematic interviews (Grin et al., 1997; Van de Graaf and Hoppe, 2000). This type of interview involves the flexible and creative use of a beforehand developed questionnaire (Annex 1). Initial questions had a very open 58 and exploratory character and the subsequent interview process flexibly allowed interaction between interviewer and respondent. Based on the interviewee’s first answers, the interviewers formulated more detailed questions, whereby the questionnaire was used as a checklist to assure the relevant subjects were discussed during the interview. In this way, the interview could accommodate and acknowledge the particular views and framings of each interviewee without forcing them to interpret cumulative risks and integrated assessment according to a pre-defined set of questions and modes of thinking inherent in these questions. This also allowed the identification of new issues and perspectives.

The following criteria were used to identify interviewees:

 Being representative of key branches in the EU administration (Directorates-General of the European Commission that are involved in chemicals control) and decision making (members of the key committee of the European Parliament) and of relevant EU-level stakeholder groups

 Holding relevant expertise, in particular a mix of sufficient technical background knowledge on chemical and related risks and of knowledge of strategic developments in their particular field

 Having experience in the interfaces between risk assessment and risk management, specifically policy formulation and implementation

 Having preferably replied to the internet questionnaire survey conducted among a larger group of experts and stakeholders (Assmuth et al., 2007)

 Being mentioned by other candidate interviewees as holding relevant insights and positions in relation to integrated treatment of risks.

A request for the interview was sent to candidate interviewees, signed by the coordinators of the Work Package and the NoMiracle project (Annex 2). Confidentiality of the replies and of the identity of the interviewees was assured. With the latter, we hoped to stimulate interviewees to express, as truthfully as possible, their independent and personal views, instead of ‘official’ replies.

The interviewed persons were affiliated with the following organizations at the time of the interviews:

- European Commission Directorate-General Environment (DG ENV)

- European Commission Directorate-General for Health and Consumers (DG SANCO)

- European Commission Directorate-General for Enterprise and Industry (DG ENTR)

- European Commission Directorate-General Research (DG Research)

- European Commission Directorate-General Joint Research Center (DG JRC)

- European Environment Bureau (EEB) 59

- The European Consumers' Association (BEUC)

- European Chemical Industry Council (CEFIC)

A tentative list of the general interview themes was sent to persons willing to participate, to orient them to the interview themes. The themes were developed within the research team in the Work Package of the NoMiracle project during several consecutive comment rounds partly based on the questions and replies that were received in the previous survey.

The questionnaire used by the interviewers was structured around the main themes, including – for each theme – an open ‘headline’ question and a number of more detailed follow up questions to possibly ask for clarifications. However, these follow up questions were used only as an indicative guidance and in most cases large parts of the interviews were conducted as free-form discussions.

The 16 interviews had an average duration of approximately 1.5 hours. They were performed in two rounds in June 2007, in most cases by two interviewers. One interview was conducted with two interviewees (from the same organization unit) as a group interview.

Most interviews were recorded, after permission for this had been obtained. Transcripts of interviews were sent to most interviewees for checking in early 2008.

2.3 Materials obtained

In general, the interviews revealed considerable variation in thinking about risks and uncertainties, (integrated) risk assessment, and its links with management. They allowed deepening insight in the various arguments that underpin views on cumulative risks and integrated assessment. For instance, some interviewees stressed a division of assessment and policy, others expressed a need for more holistic approaches to (chemical) risks, also in the science- policy interface, e.g. with risk-benefit considerations. Different interpretations of precaution and participation also surfaced.

Quantitative analyses could not be made on the basis of the interviews because of their small number. Qualitative variations in the material collected were observed. They were related to:

. The background and experience of the interviewees 60

. The role and organizational status, ranging from experts to strategic management level

. The interviewees’ interest in the theme

. The conception of integration

. The views of the needs and possibilities for integration

. The conception of the role of knowledge and R&D

. The degree of coherence of different statements during the interviews

. The degree of concern with risk management on the one hand, or with knowledge development on the other hand

. The tendency to defend particular interests

. The interviewees’ personal characteristics, e.g. a tendency to scepticism.

The material was treated in two complementary ways: o A bottom-up analysis of the interview material: based on representative statements according to the thematic areas of the interviews. In this way, the variety of views on each of the interview themes is presented; o A top-down analysis of the interview material: according to an interpretative structure distinguishing barriers and opportunities to the deployment of integrated treatment of risks. In this way, the material is processed in view of generating insights for the future development and uptake of integrated risk assessments. 61 3 SYNTHESIS OF RESULTS BY THEMES

3.1 Framings and concepts of integrated risk assessment

A variety of views was expressed on whether cumulative risks are currently an issue of importance.

Some interviewees felt that they are already an issue and as such see a definite need for assessors to address them. However, they often immediately added that it is very difficult to address them. Others thought cumulative risks were hardly an issue of importance now. To underpin this view, reference was made to current legislative frameworks, especially to the requirements for assessment in the frame of the REACH legislation (see Chapter 3.5 below).

The most common view was that there currently is ‘a lack of methods to assess multiple risks’ and, as a consequence, they were more seen as a mid- to long-term issue than a present one, e.g.:

"Integrated risk assessment is an issue in many ways. … Specifically the integrated treatment of cumulative risks from multiple stressors is not now a key area in EU, but probably will become one."

In general, the opinions expressed were influenced by the current status of research and assessment of cumulative risks and by the legislative framework in place. Both are seen as still favouring an approach of ‘decomposition’ of risk situations, pushing towards the future a concern with ‘recombination’ of risk situations to reflect more realistic exposure and impact conditions.

The interviewees framed and conceived integrated risk assessment in different ways and stressed various dimensions of integration (see also paragraph 3.4). Most commonly, it was interpreted as integration of several risk agents. In addition to integration of multiple chemicals e.g. in mixtures, sometimes other stressors were included. Integration was also understood to involve various target organisms, particularly in combined treatment of human health and eco-toxicological risks, various environmental compartments and exposures, various types of effects, and various stages of the life-cycles of chemicals. In relation to the latter, also integration between different phases in the risk governance cycle was mentioned, specifically between assessment and management.

Likewise, cumulative risks were understood in different ways, e.g. ‘accumulation’ was interpreted in terms of accumulation of several agents and as accumulation of effects. Accumulation of risks 62 was conceived both as the simultaneous occurrence of multiple risk factors and as a sequential series of exposures over time.

3.2 Addressing cumulative risks and multiple stressors in policy making

3.2 1 The need to develop policies for dealing with cumulative risks

According to most persons interviewed, there is a clear need to develop policies to deal with cumulative risks and multiple stressors. A typical reply was the following:

"There is a definite need to develop integrative policies on risks … on many levels and in many areas. The cumulative risks of chemicals … are an example. In general, stressors need to be regulated on more holistic grounds. This is already done in some areas like occupational health where it may however be in some respects more readily achievable."

Interviewees referred to factors increasing the awareness of the need for integration, such as strict division of competencies in institutions, leading to lack of synergies or even contradictions in policy initiatives.

"There are needs due to the division of policy areas, e.g. between human health and the environment."

Another factor mentioned was the confrontation with risk management situations in which the need for more integrative policies to risk was felt (for instance, the combined effects of indoor pollution).

On the other hand, strict divisions based on fragmented institutional structures and legislation and the strong separation of risk management and assessment can work as barriers to develop more integrated approaches:

"There are some artificial divides due to legislation."

Some interviewees argued that developments towards more integrated risk assessment will mainly be policy driven:

"Integration depends on how far regulation and policies are driven. At present, mainly regulatory policies dictate what is needed."

While some general tendencies in policy and society’s dealings with risks were commented, responses stressed the context dependence of the needs for integrated risk assessment:

"The needs and policy drivers for integration depend on the setting. … Society as a whole has moved toward a more hazard-based approach to risks, that is, risk management is done more readily on the basis of exposure information instead of awaiting proof of harmful effects."

Areas in which there is a natural tendency to look at risks in an integrated way were cited as examples, e.g. health care.

3.2.2 Challenges for addressing cumulative risks of multiple stressors

Various kinds of challenges for addressing cumulative risks of multiple stressors, in terms of obstacles, barriers, limitations or problems, were mentioned, in addition to development needs.

Specifically, the interviewees pointed to the following problems for and possible disadvantages of integrative approaches:

1. Lack of data and of formalized, validated and accepted approaches to assess cumulative risks

2. The slow pace of regulatory development and the need to rethink quite fundamentally the established ways of society’s dealings with risks

3. The costs linked with additional testing and information gathering

4. Increasingly global markets pose a challenge for integrated treatment of risks: what should be the geographical scope for an integrated assessment?

5. Integrated assessments could be used as argument to downplay the particular risks posed by a specific product and thus weaken the case for tighter regulations of certain products

"… Inclusion of additional risks in comparative risk assessment may also be used as an argument to downplay a particular risk from a product that is at issue."

3.2.3 Participation and communication in addressing cumulative risks 64

Many interviewees saw possibilities for wider application of participatory processes in addressing cumulative risks and multiple stressors. It was felt that this could lead to more attention to stakeholder views (NGOs, a broad group of industries) and public opinion on e.g. what would be the ‘right’ questions that assessments should try to answer.

On the other hand, some scepticism was expressed regarding the possibilities offered by rendering information more easily and readily available (e.g., in draft documents of assessments): experience shows that availability of information and of possibilities to react on it is no guarantee for engaged participation.

"… It seems the availability of information in principle does not guarantee participation, but it also depends on whether one knows how and is able, e.g. in terms of resources, to participate."

There is a danger of saturation with risk information and alarms, leading to reactions of indifference. As the policy sphere, the sphere of public debate is not served with an overload of confusing risk related information, in which the main messages and relevant insights get lost.

3.3 The roles of information, knowledge, uncertainty and precaution

3.3.1 Status of the knowledge base, opportunities and limitations

The knowledge base available on cumulative risks was generally seen as limited and displaying important deficiencies, although improving. Some of the limitations and deficiencies were however not so much seen as temporary and provisional, but as having a fundamental nature:

" It needs to be defined what is meant by ‘integration’ with regard to multiple risks and uncertainties" … "An important question is the limits of quantification….."

The indeterminacy of cumulative exposures was acknowledged as a key limitation:

"It is virtually impossible to foresee in what mixtures a substance can end up." 65

‘Total integration’ being impossible, it has to be defined which multiple stressors and cumulative impacts will be assessed. It is inevitable and inherent that integrated assessments, meant to overcome the limits of ‘simple, narrow’ risk assessments, have their own limitations. In addition, these ‘limits’ set to the scope of integration were seen as necessary to avoid overload in compiling and processing information:

"Already now there are huge dossiers being compiled and evaluated for instance on food chemicals. So the inclusion of additional information presents definite challenges."

This entails questions of prioritization of risks to be approached by integrated assessments. While such prioritisation is much needed, it will be strongly conditioned by practical constraints as available information and possible measurements.

3.3.2 Treatment of uncertainties

Interviewees held the opinion that, when risks are assessed in an integrated way, the nature of uncertainties change as well as their importance. It seems that the awareness of complexity, through approaching situations marked by multiple exposures and combined effects, entails the need for a changed concept of uncertainty: one that can accommodate inherent uncertainty that remains even after elaborate assessments.

Some interviewees highlighted the limits of modelling approaches and criticized modellers of seeking too high levels of sophistication. Both quantitative and qualitative approaches to risk were thought necessary when developing integrated assessments. Alternatives and substitutes assessment was forwarded as an area in which integrated assessments would be particularly relevant.

Fundamental uncertainties and limitations were pointed out:

"Prior to use of chemicals, inevitably only predictions of risks based on theoretical models are available. Thus there is always considerable uncertainty."

In relation to complexity and uncertainty, the more reflective views of deeper uncertainties and their links with deliberation surfaced when asked directly and in a general way about these concepts, while remarks related to current shortcomings and future possibilities and challenges of 66 integrated assessment implicitly were framed by the more traditional view of science as a truth provider.

The need to better assess and communicate uncertainties was not only mentioned in relation to integrated assessments and multiple stressors, but also to currently well-established risk assessment (e.g., related to the use of safety factors, to extrapolation of toxicological data etc.).

There was however pessimism about the uptake of information on uncertainties in the policy domain ("policy makers want simple and exact figures"); the possibility was mentioned that uncertainty information would further complicate policy debates on dealing with risks. Nevertheless, also benefits of uncertainty analyses and communication were identified:

"Assessment of uncertainties can cause additional complications. But uncertainty analyses can have a big role in reducing unrealistic risk estimates as a result of safety factors piled on each other."

In addition to such specific uses of uncertainty analysis, a general benefit was indicated:

"We have to be able to move from one layer of information to another, and one framing of risks to another. Improved uncertainty analysis and identifying the key uncertainties can be helpful in this".

3.3.3 The role of precaution

Some interviewees expressed a concern that assessments are marked by a bias towards precaution, implicitly recognizing that precaution is not something that is confined to risk management:

"The general evaluation of risks is too biased towards precaution. … In some cases this emphasis has been justifiable. The worst substances have already been phased out. So the picture is not so black-and- white anymore. But some … just want to ban chemicals."

Others stressed however that precaution should be a risk management principle and that an assessment should be as ‘scientific’ as possible. Precaution was conceived as linked with lack of information and in relation to the use of safety factors. Critical comments on the implementation of precaution referred to situations in which it was considered unjustified to invoke and apply it: 67

"The precautionary principle is often misused. Originally it was developed and specified in EU also to combat exaggerated, poorly founded precaution, not as a basis for indiscriminate precaution."

3.4 Interaction and integration of risk assessment and risk management

3.4.1 Integration: why, what types and which limits?

In general, integration was seen as important. But different degrees of this importance and different reasons for it were expressed, as well as different areas in which integration is considered important. Integration was considered important to get a better picture of risks (more ‘realistic’), to regulate more efficiently, to better understand uncertainties about risks. Some interviewees raised the question of the precise meaning or definition of integration.

Limits to the move towards integration were thought to be imposed by:

 The level of complexity that can be addressed through integration;

 The regulatory basis for the integrated treatment and management of risks.

Besides needs for integration, specific focuses were also still seen as necessary in relation to single stressors’ evaluation and management.

As most important types of integrated assessments, the following were mentioned:

 Assessment of chemical mixtures and multiple exposures

 Integrated assessments of risks and benefits of chemicals

 Inter-species integration: limitations for integration in this dimension were pointed out, but its need was acknowledged, based on the relationships between human and non-human animals e.g. in food-chains, and the treatment of both types of risks e.g. in the chemicals area.

 Functional integration of information systems

 Advanced uncertainty analysis

3.4.2 Interaction between assessment and management 68

The interviewees generally recognized an increasing need for more interaction, and even some integration, between assessment and management of risk issues. It was also felt that risk assessors should take into account risk management related questions that are crucial for their assessment:

"This is an increasingly important area. Better interaction is needed, as there has traditionally been a too strict separation. Assessors still decline to ask risk management related questions even when they would be crucial to their assessment…. Generally, the interaction between assessment and management needs to be improved."

In the policy sphere, it was argued that the EU needs to rediscover the linkages between assessment and management, referring explicitly to the developments of risk assessment in the US:

"The rigid risk assessment-risk management separation needs to be overcome, or at least be subjected to discussions and be more clearly delineated. EU needs to rediscover these linkages."

Risk-benefit analyses and comparisons were mentioned as means to improve interaction between risk assessment and risk assessment. It was pointed out that such analyses already are common in some areas, such as medicinal products.

Some interviewees stated that integration, also between assessment and management, is also to some extent already addressed in REACH, as it requires that potential interactions of chemicals are looked at and that all subsequent stages in the uses of the chemical are accounted for (cf. 3.5).

Possible objections to integration of risk assessment and risk management were also formulated. Some of these objections seemed to be based on semantics; for instance, analyses of risks and benefits or of consequences of management options were not questioned as such, but rather how such analyses should be named (in relation to 'pure' risk assessment) and where they should be placed in regulatory schemes. However, there were also objections based on fundamental policy reasons, as a principled and clear definition of responsibilities:

"The Commission has already been criticized for mixing risk assessment and risk management."

This separation of risk assessment and risk-benefit analysis and other management considerations was felt necessary to guarantee objectivity: in the absence of a strict separation, the risk assessment process was explicitly feared to become less reliable, politicized and ‘unscientific’. 69

Opposition to the idea of more interaction between assessment and management came also from industry and from NGOs. For some of the latter, the motivation seemed to be that consideration of risk-benefit relationships and other management aspects in assessments might weaken regulatory measures to decrease or eliminate the risks they perceive. A tension was perceived between risk- benefit analyses and precautionary approaches, even that they would be mutually exclusive:

"The question arises, why should one be so insistent on risk-benefit analysis, and could the precautionary principle be an alternative."

Incommensurability of risks and benefits was another reason for scepticism about the joint evaluation of risks and benefits in risk-benefit analysis and comparison:

"There is also the problem that benefits are of other kinds than risks or harmful effects."

According to another view, risk communication that takes into account the perceptions of actors and publics and is explicit about value judgments and assumptions, is needed to balance science- based approaches with precaution:

"There often is a tendency to blame on some risks and those causing risks." … "Risk communication guidance is needed for more balanced risk assessment and precaution, including e.g. risk-benefit considerations. Value judgments and assumptions need to be explicated."

3.4.3 Obstacles to and possible problems with more integrated assessment and management

Both scientific and regulatory obstacles for the introduction of more integrated treatment of risks were mentioned. Among the latter in particular, it was thought crucial to increase efforts for improved coordination between different actors in regulation, e.g. horizontally. This was felt to be needed to overcome sector divisions.

Potential problems mentioned included a general resistance to introducing new methods and a complication of management through considering complex risks more extensively. The need for ‘simple’ assessments was underlined: 70

It is essential to strike a balance between holistic and narrow assessment. All factors cannot be accounted for due e.g. to legislative divides. Simple assessments are also needed, e.g. decisions to be defensible in courts."

Some interviewees expressed the concern that integration would make it increasingly difficult to standardize assessment and management in formal approaches.

In the policy sphere, an ‘overload’ of evidence, through a multiplication of scientific insights, stemming from several partial integrations, could be a hindrance to the development of policy measures.

Finally, some interviewees stated that a more integrated treatment of risks would need a reinforcement of quality assurance in risk assessment.

The idea of strict separation of assessment and management was defended when asked explicitly about someone’s view on this. However, when commenting precaution or participation, the answers obtained revealed that the interviewee often had a more elaborate view of assessment- management relations, and implicitly admitted that this strict separation has to be seen with a lot of nuance. Often the balancing between the linear and alternative conceptions of the relationship between assessment and management could be brought in relation to whether an interviewee judged the question in the context of its own field or sector of activity: the perspective chosen often served to justify the own interest, position or view.

3.5 Integrated chemicals assessment and management in relation to REACH legislation

3.5.1 Integrated treatment of risks in REACH

In general, in the frame of the EU chemicals legislation, and of REACH in particular, the integrated treatment of risks was seen as a longer-term challenge.

Some actors regretted the lack of explicit consideration of combined effects of multiple substances or stressors in the legislative framework. This lack of attention to multiple stressors was generally considered to be due to limited resources and capabilities to deal with cumulative risks. In 71 addition, some pointed to political factors such as lobbying against broadening the assessment towards chemical mixtures, reflecting the debates on REACH (cf. Pesendorfer, 2006, Assmuth et al., 2009b):

"Some amendments to REACH have been proposed to facilitate better consideration of multiple risks from mixtures. However, some improvements are very slow. The reasons for deficiencies include the lack of information and lobbying against broader assessment to bear on new and more complex risks."

Others argued that industry had already a lot of requirements to fulfill to implement REACH in its current form and that priority should be given to improve the basis for regulation on a ‘single substance’ basis over initiating already now more integrative approaches.

Nevertheless, first steps towards the introduction of some elements of integrated treatment of risks were acknowledged, e.g. through the assessment of product life-cycles, through integrating risk reduction measures in the assessment phase, and through the combined treatment of corresponding groups of chemicals:

"REACH is not an immediate answer, as it is largely based on single substances. However, there are some procedures for handling formulations and products containing many chemicals. Some integrative systems were already in place already before REACH and could be utilized in its implementation and development."

However, there were widely varying views of how more substantial integration of multiple chemicals’ effects could be achievable. Some voices were rather critical in this respect.

Finally, an interviewee saw a possible positive contribution of the concern with multiple stressors to the REACH process, namely that studying combined effects can aid a refocusing of attention to single-substance assessment and management by putting risks in context.

3.5.2 Challenges to the implementation of REACH

Resources were seen as one of the most important challenges for a good start up of the REACH implementation. The information requirements were considered a heavy burden especially for smaller companies who do not have a lot of resources for these tasks. These problems would even increase if more integrated treatment of stressors would be requested. 72

"Under REACH great challenges are posed already by the changes required in product safety assessment; e.g. to develop it toward extended product life-cycle analysis. There is no great additional need for integrated risk assessment of multiple risks from cumulative stressors."

"There is a definite need to keep assessments simple. … Authorities are however getting involved in integrated assessments."

Other challenges to implementation that were mentioned included the following:

 Confidentiality of information

 Sufficient stimulation of innovation

 Specification of the burden of proof

 Application of product life-cycle approaches to chemicals

 Development of risk benefit and socio-economic analyses

 Assessment of alternatives and substitutes

 Communication about risks

 Precautionary and preventive aspects of the legislation

 Co-operation between member states, EU-level and global actors.

3.5.3 The generation, collection, processing and transmittal of information in REACH

A possible problem for the use of information in the context of REACH is related to labeling of chemicals that is thought to contain already too much and too detailed information.

The generation of knowledge in the REACH system could be jeopardized by unwillingness to share information. Information on hazards would perhaps more easily be shared among industries, but exposure related information as well as information on substitutes has a clear economic value and there probably would be attempts to keep such information confidential:

"Hazard information, that is, on adverse effects of substances, can be easily shared by companies, whereas exposure information in particular is linked to the uses of chemicals and may thus have to be kept confidential. … Also in cases where several companies produce same chemicals, it is hard to coordinate information generation and use." 73

The implementation of REACH was feared to lead to a focus on meeting regulatory requirements instead of more attention to innovation in chemicals:

"REACH will create activities around regulation, not innovations. It will ‘lose chemistry’, traditional ways of using chemical information."

However, also positive impacts of the regulatory requirements can be expected, such as promoting novel approaches for integrated assessment of chemical risks and innovation in methods for risk evaluation and socio-economic analyses. Possible positive effects could also be expected of the increased availability of plenty of information:

"REACH requires but also provides plenty of information. Risk information and communication from chemical producers to users including down-stream enterprises are essential."

This increased availability of information, and its broader scope (including socio-economic aspects) will call for due attention to its quality, evaluation, prioritization and utilization in integration. 74 4 INTERPRETATION AND RELATIONSHIPS WITH THE SURVEY OF EXPERT AND STAKEHOLDER OPINIONS

In this section, we first briefly compare the results of the interviews with those from the earlier Internet survey on integrated risk assessment carried out in the frame of the NoMiracle project (Assmuth et al., 2007). The objective of the survey was to give a general picture of perceptions and views among experts and stakeholders concerning risks, risk assessment and risk management. As a preparation to the regulator and stakeholder interviews, the survey also served to identify issues and actors, and to formulate themes and specific questions to be taken up with a restricted group of interviewees. The comparison here is limited to qualitative aspects and to the themes covered during the interviews, and does not include analyses of the co-variation of survey results and of background variables or analysis of views regarding the use of risk maps. Instead, we have striven to use information from the survey to contextualize and support our characterizations and explanations of the interviews especially regarding the epistemic (knowledge-related) and evaluative positions.

The information collected during the interviews on the positions of experts and stakeholders provides further evidence for the existence of contrasting views on a number of themes. Such contrasts in views often suggest fundamental divergences and lead us to conclude that the multiplicity of positions, values and preferences, concepts and background theories as well as of overall world-views is to a certain extent irreducible. However the fundamental divergences in views are not sharp, one-dimensional and fixed, but rather present ambiguous and variable continua.

We therefore attempted to discern, on the basis of the interview material and the survey, the broad contours and common elements of framings regarding the development and use of integrated approaches to risk assessment and management. Such commonalities can be discerned at a general level only, due to limitations in our data and to the great variation (as attested to by the data) in framings and views. This framing of problems and solutions converges around two key notions that characterize both the need for integrated assessment and its main challenge, namely “‘more realism’ and ‘manageable approaches’ in dealing with complex risk issues”.

As such, the basic premise of the work package in the NoMiracle project in which the interviews took place is confirmed: as a rather typical ‘science-for-policy’ activity (Funtowicz and Ravetz, 1993), effective engagement of user groups in the further development and application of integrated risk assessment is desirable. These groups can deliver useful insights on crucial aspects 75 of ‘real risk situations’ and of ‘management needs in the policy context’, aspects that can be of great importance for the design of useful integrated assessment.

The challenge revealed by this interpretation will be to find ways to accommodate, under the umbrella of such general framing, a diverse set of innovative practices towards integrated approaches that does justice to the various positions taken by the ensemble of relevant actors. To meet this challenge, simplistic notions of 'science-for-policy' are insufficient. The relationships between 'science' and, in general, knowledge with policy or application need to be conceived as multi-dimensional and multi-directional processes where also policy influences science, in a recursive instead of linear model; where the positions and roles of actors shape but are also shaped by science and policy; where innovation is not restricted to science or assessment; where innovation and integration are not simple and unproblematic goals; where instead of an idealized progress, controversies and conflicts are redefined and negotiated; and where science, policy and assessment and their relationships thus present topical areas of research in their own right.

4.2 Summary of relevant findings from the survey

The survey showed a pronounced variability of the ways in which risks are defined and approached, of perspectives on the level and characteristics of related uncertainties and of views regarding the use of information related to these. The context of use and the use purpose are important factors that influenced the views on the relevance and usefulness of different types of information on risk.

Both the need for and the difficulties to arrive at integrated risk assessment were commonly acknowledged. The evaluations of the difficulties to arrive at meaningful integrated risk assessment differed and were expressed by judgments that ranged from pessimistic to optimistic regarding:

 The potential of quantitative approaches to arrive at objective and reliable knowledge on multiple and cumulative risks (pending on notions of objectivity and reliability)

 The inherent possibilities to develop sound knowledge on and effective management of complex risks (pending on notions of soundness and effectiveness).

Survey results indicated the following average assessments of the need for and importance of different types of integration and integrated information:

 Integrated information on both risks and benefits of chemicals was considered as the most important item.

 Integrated information on risk reduction was also highly valued 76

 Closer integration of assessment and management was supported by some, but contested by others

 Divergences were also found in the attitudes towards broad public engagement, and integration of broader social concerns, in risk management.

Much of the observed variability in views, for instance on the role of uncertainty, could be traced back to divergent theoretical and principle-level positions on the development and status of knowledge on complex issues and on the relations between assessment and management.

Instead of establishing 'right' views, the key message of the survey was the ambiguity in positions, perceptions and underlying theories and valuations. The complexity of risks is reflected in complexity of multiple views. Comparative evaluations of risks and of the importance of various kinds of information reveal different overall concepts of risks as well as of science, expertise, and society. These have direct bearing on what is meant by, and what can be meaningfully aimed at by, 'integrated' risks assessment and 'novel' methods.

The relevant policy conclusion was that the multiplicity of perceptions and views of risks and uncertainties and of the need for asymmetry in dealing with them are genuine and to a great extent fundamental and cannot be reduced to any simple model, nor wholly dispelled by mere information. Instead, respecting this variability and multi-dimensionality of risk perceptions, experiences and assessments can lead to more meaningful and inclusive concepts and communication processes on risk issues.

This non reducible multiplicity should be taken into account, in both the application of existing risk assessment and management procedures and in associated communication processes, as well as in the development of novel methods for risk assessment. It should lead to an increased transparency of methods through communicating their underlying framings, assumptions and interpretations.

The challenge remains however how to fruitfully inject such pluralism and transparency specifically regarding integrative approaches to risks into the regulatory (EU) machinery that can be, even when plurality and transparency are in principle encouraged, dominated by particular views on risks, knowledge and management and also particular interpretations of and views on integration. This will require an understanding of the workings of the machinery and its capacity for change, including an appreciation of the need for some uniformity in regulation. 77 4.3 Comparisons of the survey and the interviews

In a general evaluation, the interviews seemed to confirm, or offer support for, some of the key observations and interpretations of the survey. However, both the confirming and the contrasting interview results provided additional nuances of the reasoning on integrated assessment among EU-level regulators and stakeholders. The interviews gave insight in the underlying concepts, theories and preferences that back and justify their views. The pronounced variability observed in the survey results was also substantiated during the interviews, for instance regarding the definitions and interpretations of core concepts such as 'risk' (e.g., to whether loss of benefits is included), 'cumulative' (in time and other dimensions), 'multiple stressors' (e.g., whether non- chemical stressors are considered), 'risk assessment' (e.g., whether it includes consequences of management options), and 'integrated' risk assessment in its various dimensions (Assmuth and Hildén, 2008). Regarding the latter, variable views on the desirability of integration as opposed to specification of risks were noted.

The interviews made also clear that the views held by a stakeholder or expert on the possibilities to know and manage complex risks in an integrated way are not uniformly pessimistic or optimistic. Vacillation of views between pessimism and optimism could be observed and seemed often related to the framing and the time scale the respondents adopted: when discussing near-term developments, the interviewees’ evaluations, being more focused on present institutional and other boundaries found in their own professional sphere, were usually more constrained and pessimistic.

In both the survey and the interviews, the influence of positivist scientific and policy ideals was clearly present. However, especially in the interviews the belief in the intellectual tractability and controllability of risks was tempered by scepticism and by awareness of the limits of scientific assessment. Views could vary also within the same organization and even responses given by one and the same person contained inconsistencies and conflicting elements for instance regarding the desirability and feasibility of (some kind of) integrated treatment of risks.

As can be seen from the synthesis of results, especially regarding views on uncertainties and limits of knowledge (see for instance 3.3), both a linear model of science-policy relations (‘science speaks truth to power’) and alternative conceptions on complexity and uncertainty and on the links between assessment and management seem to co-exist (on these, see Jasanoff, 1995). Depending on the situation, the question, the context to which is referred, one seems to dominate over the other.

The interview results suggest that the tensions that were observed in the survey between established, routine methodologies for dealing with risks, on one hand, and novel approaches including more integrated treatment, on the other hand, were in some cases related to 78 epistemological factors such as new concepts of risks and risk assessment as opposed to traditional concepts, while in other cases they were related to political and institutional factors such as normative requirements. In the latter case, the need for permanence in mandates, framings and procedures of institutions or in political principles regarding risks was contrasted with the need the open up new perspectives crossing the institutional divides and overcoming inertia.

We obtained confirmation in the interviews of the observations made earlier in the survey that some actors feared that integrated risk assessment or novel methods could be confusing through diverting the attention from tangible basic-level assessment and straightforward precautionary management actions. Some interviewees were even suspicious that the move towards novel and more integrated assessment methods could be part of a misleading strategy aimed at relaxing requirements to be put specifically on chemical risks. The interviews substantiated the interpretation that this suspicion is partly based on definitions of integration and thus artificial; some concepts and types of integrated assessment are compatible with both broad and some with narrower framings. There seemed also to be a related fear that more integrated assessment would strengthen 'paralysis by analysis', as was seen to happen in other areas of integrated assessment (Blanco, 2008).

On the other hand, some commonly held opinions on risks and uncertainties observed in the survey differed from those of the interviewees. For instance, relatively many respondents in the survey agreed that economic considerations should be completely excluded from risk assessments, while the views of the interviewees on this area of analysis were more favourable. Some differences were discernible also in views on uncertainties. Many respondents agreed that the key to dealing with uncertainties is more exact measurements and better validated models; while these were emphasized by some interviewees, there was also awareness of deeper uncertainties that resist such approaches. In keeping with this, survey respondents commonly held that most risk controversies are caused by a lack of expert information, while the interviewees often acknowledged that it alone is not the key.

Part of these differences can be explained by the working environment and experience of the interviewees. Most of them had inside and detailed knowledge of regulatory issues in the chemicals control area, both regarding policy aspects and scientific-methodological aspects.

4.4 Variations in views on policy-related dimensions of integration 79

In earlier research (Assmuth et al. 2007, Craye 2006a,b) a diversity of positions had already been documented regarding the relations between risk assessment and management, regarding the place and role of ‘participation’ in risk policies, regarding the salient uncertainties in scientific risk assessment and in risk management, regarding the implementation and role of precaution. This diversity had been interpreted as having roots in diverging models of science and policy (Funtowicz and Strand, 2007) and concepts of uncertainty, that ultimately even are culturally influenced.

As both the survey and the interview material confirmed this diversity of views on issues as risk- benefit analysis, the conception of the different phases in the risk governance cycle and the relations between them and on the interpretation of precaution, we have commented the interview material in this paragraph to display this diversity even more clearly. However, not all causes of divergent positions have to be sought in different underlying heuristics: positions were also influenced by calculations how they could further or impede one’s interests and preferences.

In particular, risk-benefit analyses seemed an important area both in terms of their role and of their methodological approaches. While the need for valid ways to comparing risks and benefits was acknowledged, wide support for such methods was seen to be still lacking. Some responses suggested that this lack of support could be partly explained by the unwillingness to open up still more inclusive analyses. Such unwillingness might be based on the suspicion that risk-benefit considerations would not provide for the straightforward restrictions conceived as the crucial ones on some actors’ agenda’s (environmentalist or other), but instead could offer counter-arguments.

Other critical thoughts feared that risk-benefits could contribute to blurring the boundaries between assessment and management. Some of these objections though seemed to be based on semantics; for instance, analyses of risks and benefits or of consequences of management options was not contested as such, but rather with respect to the labelling of such analyses (in relation to traditional 'pure' risk assessment) and under what phases in the risk policy cycle they should be placed.

Some views on risk communication still showed a strong belief that the ‘right’ definition of a risk issue exist, supported by a reliable assessment. Following such logic, perceptions are to be seen as deviations of this correct assessment, for instance expressed through the following statement:

"Risk communication can also lead to ‘wrong’ (or otherwise problematic) perceptions."

Here risk communication is seen as a support to convey the correct information on risks and a problematic situation in which risk communication could not achieve this objective is signalled. Such view is however in stark contrast with findings in social risk research showing that 80 definitions of risk issues, also those by experts carrying out assessments, are, at least partly, constructed (Krimsky et al., 1992). Following this reasoning, a key objective of communication would be to debate what is “wrong” risk perception and on what grounds it can be called wrong, instead of tacitly imposing normative statements, on what is to be considered ‘right’ or ‘wrong’, from the outset.

In trying to clarify their positions on the precautionary principle, some interviewees’ responses clearly showed that they had difficulties in interpreting precaution as a policy principle strictly limited to risk management.

Some statements on how the invocation of the precautionary principle should be triggered were based on a perceived need for more close interaction between knowledge production and management of risk situations:

"When there’s not enough information, the precautionary principle is invoked. This does represent another approach to knowledge and uncertainty than checking hypotheses generated."

In other responses, precaution was approached as being closely linked to the use of safety factors, an area that bridges assessment and management:

"Safety factors for chemical mixtures are an important area of practical application of uncertainty …. For some classes of substances it is impossible to define safe levels due to their particular properties."

The applicability to mixtures of safety factors, used in single substance assessments, was seen as critically dependent on how much certainty in this extrapolation is required, and what the criteria are for similarity of the single substance vs. multiple chemicals situations.

Some respondents said that no strict rules can be developed for the invocation of the precautionary principle on the basis of the status of knowledge: a judgement and interpretation will always be necessary as “there is no clear division between mixtures that are completely or not at all tractable”.

Nevertheless, when asked directly about their definition of precaution in relation to the phases of assessment and management, some expressed a very principled and categorical view on the role of the precautionary principle: 81

"Precautionary principle is a risk management, not a risk assessment consideration. Risk assessment needs to be scientific and quantitative to the extent possible."

Such position did not leave any room for the question whether concepts of uncertainty and precaution and their interpretation (need to) change when multiple complex risks are addressed. Although phrased as a defence of rational and reliable approaches to risk, the statement is marked by cultural and normative beliefs in the characteristics of knowledge production and its use in society. If risk assessment would instead be defined more inclusively as an activity that overlaps with management, it could accommodate important aspects of precaution. This would be particularly the case if assessment is taken to include not only quantifiable risks and natural scientific facts, but also framings of risks and consequences, and contextualized evaluations of uncertainties that utilize various forms of knowledge and deliberation.

Discussing the precautionary principle in the context of integrated approaches also revealed that some interviewees still tended to see a clear dichotomy between science-based approaches and precautionary based approaches, even judging a priori that 'risk-based' assessment would always be more lenient than a precautionary approach:

"There are definite needs to address also multiple risks by more fully utilizing the scientific evidence instead of denying it a priori on the basis of allegedly precautionary policies, such as with PBT substances for which the European Commission has declared that risk assessment (as defined in regulation) is 'not needed'. This amounts to self-imposed ignorance and stagnation, on purely political grounds. It is even more illogical and surprising, as for some similar compounds such as short-chain chlorinated paraffins (SCCPs), now suggested to be formally included in persistent and bio-accumulative toxic compounds (PBTs), extensive risk assessments have been previously required by the CEC, and have been performed by competent authorities themselves."

While such a self-imposed ignorance certainly can occur towards broader assessments, it is not implicit in a precautionary approach not to pursue scientific research on issues for which currently a lack of knowledge is felt and precautionary measures taken. In this respect, it might be argued that, depending on further research, the thinking of authorities regarding the seriousness of the risks e.g. of PBTs could change.

Finally, the difficulties with the aim to have strict decision rules on when precaution is to be applied, is illustrated by following quote on the role of a precautionary approach in different areas: 82

"In general, the precautionary principle is not used in assessing chemicals in food as it refers to cases where there is no information, whereas in these cases there always is a considerable amount of information before assessment even gets to be done."

Such statement seems to suggest that the availability of a considerable amount of information always implies a ‘certainty about what to do’ and avoids the questions on the quality of information and knowledge required for policy action and its possible limits.

4.5 Commonalities and key notions in framing integrated risk assessment

Despite the sometimes fundamental differences in approaching and defining concepts as ‘risk’, ‘uncertainty’, ‘precaution’, ‘integration’, the interviews allow to sketch, in a very general way, contours of a possible consensual framing of the problem and challenges with integrated assessment.

This framing should recognize the need to make risk assessments more ‘realistic’, while keeping/improving their usefulness and relevance for contributing to decision making on measures to deal with risks: complexity should be addressed, but this should not endanger meaningful action.

4.5.1 The need for integration: addressing the limits of approaches and methodologies to address complex situations

Overall, the interviews showed a broad consensus about the limits of current risk assessment and about the need to explore how these can be overcome to address risk situations in a more ‘realistic’ way.

As an example of the problems with current methodological approaches to environmental health risks, the limitations of epidemiological studies were mentioned by some interviewees:

“Epidemiological evidence will always be debated, and especially the resolution of chronic effects is difficult." 83

These limitations are all the stronger in multi-stressor risk assessment, as the causality and the attributable share of effects and risks are hard to ascertain or to estimate. Although the latter is especially clear in epidemiology, it is a problem for any evidence-based or ‘science-based’ assessment. As a response to these very challenges, highly advanced and efficient methods have been developed in epidemiology for resolving attributable risks among a multitude of risks, co- factors and confounders, also in ways that can be utilized in other areas of risk assessment. This makes clear that a deliberate effort to arrive at results that are meaningful for real complex risk situations can indeed lead to improved methods that are an important part of the inquiry process.

Nevertheless, lack of methodological progress for integrated approaches inspired many interviewees to take a more sceptical stance regarding a rapid uptake of the explicit consideration of accumulation and complexity in risk situations:

"As there is a lack of methods to assess multiple risks, they seem as a mid- to long-term issue. In general, we are now in a decomposition stage, and can only then move to recombination."

This reflects an awareness of the complexity of the task to develop suitable methods to address cumulative risks.

However, the current lack of methods might lead to exactly the opposite assessment on the status of multiple risks as an issue of importance: possible they present an issue because of, and in the form of, the lack of methods. Furthermore, recognizing it as an issue now would create fruitful ground to involve the possible users of information on multiple risks upstream in the development of novel methods.

All in all, the interviews indicate a will to overcome current limitations, combined with a strong awareness of the difficulty to address complex situations ‘scientifically’ in a meaningful way. As any more integrated approach will unavoidably entail reductions of a complex reality, the answers obtained suggest it is crucially important to explicitly address the setting of problem boundaries, to follow a stepwise approach in developing more integrated approaches as well as to give due attention to priority setting in this area.

4.5.2 Effective communication between science and policy: assuring management relevance of innovative approaches 84

Interviewees repeatedly expressed a fear that newer, more elaborate and broader approaches to combined risks, would not be able to provide clear messages and could lead to confusion as they would not clarify how to manage those complex risks, i.e. they would neither deliver clear conclusions nor indications for management.

Many of the replies seemed to reflect an ambiguity with respect to balancing the need to expand assessments and to keep them focused and manageable. This was often tied explicitly to handling of new information:

"If you can measure the existence of a potential risky agent, it gets more easily perceived as a problem. Thus, we need tools to sort out the meaning of these multiple measurements."

This comment is linked to vicious circles in being constrained by available information in assessments.

Some pessimistic evaluations and interpretations of integrative risk models also were based on an assumption that they would necessarily complicate management:

"Modelling tends to increase sophistication in representation and explanation of risks, adding to the complexity. Reality can be easier to grasp without getting entangled in the particularities and in highly sophisticated models that are not appropriate to the task."

Models may indeed increase complexity in comparison with earlier representations of risks, but not necessarily; sometimes improved models are simpler than earlier ones. In any case, models simplify and distort reality, but this applies to specific models as well. The key question is whether more inclusive models could reduce some of the distortions, e.g. by getting rid of unnecessary detail and addressing the lack of linkages in narrow models (Assmuth and Hildén, 2008).

Some evaluations of the policy relevance of explicit uncertainty analyses were rather pessimistic:

"Probabilistic risk assessment to account for uncertainty would be too complicated for policy-makers. In addition to the unfamiliarity with concepts and the difficulty to comprehend such analyses, policy- makers don't want to make decisions if given information on alternatives. They want to get exact figures." 85

Policy-makers will however make decisions on uncertain information and between different options anyway, and it can be argued that it would be better if they would be given these up front. The single exact figure provides illusory remedy, as no consensus on a magic number can develop, with scientific and expert opinions also spanning a range of values. It seems rather a matter of developing capability to discern what kinds of treatment of uncertainties are appropriate in given cases.

The statement reflects scepticism about the will and/or the ability of policy-makers to learn complex risk issues. However, the signalled or expected problems can also be seen as an argument to plea for necessary changes in the communication processes between risk science and policy making.

To sum up: while the need for developing broader approaches is recognized, as well as the scientific difficulties to address complex risks in a ‘sound’ way, there is also a widespread awareness that ‘communication’, ‘usefulness’ and ‘relevance’ of the resulting information can pose a problem.

Those possible problems include: the creation of an information overload, inadequate communication patterns for complex information, a possible lack of receptivity in policy and society for non-simple or non-straightforward scientific results, a lack of social and policy culture to deal with ‘normal’ inherent scientific uncertainty.

Some of the interviewees’ responses are to be understood as framed by mainstream simple models of how scientific knowledge can be linked to policy measures in a direct and straightforward way, or by experiences of attempts to use science in an instrumental way. A fear can also be discerned in these responses that such direct influence would not be possible anymore if science loses itself in complicated assessments.

More reflexive analysis have however rejected as too linear such narrative in which science first finds evidence, and that further discoveries and inventions inevitably lead to informed social responses via avenues of prediction, communication, rational choice and control. Rather, in the case of successful science-for-policy, a lot of boundary work is performed between the worlds of science and policy (and society), an interaction in which natural knowledge and political decisions are co-produced, shoring up each other’s legitimacy (Jasanoff, 1990). The successful production of factual, or ‘certain’, knowledge on policy issues thus goes hand in hand with a reduced account of these issues, itself a product of (sometimes implicit) negotiations, inside the scientific communities as well as with social and policy actors. 86

The fear of possible confusion (and lack of management relevance) that would be created by multiple partially integrated assessments or by more inclusive approaches should lead us to conclude that the above mentioned processes of negotiation, in the case of cumulative risks, should be reinforced, and perhaps carried out more openly and explicitly, in order to lead to a successful closure.

Thus, within a more reflective or reflexive interpretation of how science and policy relate to each other (Jasanoff and Wynne, 1998; Felt, 2007), processes of dialogue and interaction between the scientific and policy communities seem advisable to mutually influence agenda setting, to reduce the complex risk situations to still meaningful scientific risk problems and to comment and assess with nuance the obtained results. As work on cumulative risks lends its justification from its possible policy relevance, it seems all the more necessary, for advances in this area, to invest time and resources in such dialogues. 87 5 BARRIERS AND OPPORTUNITIES FOR INTEGRATED RISK ASSESSMENT

5.1 Categories and general characteristics of barriers

The aim of this section is to characterize and interpret, based on the interview materials, on theoretical developments and on insights from literature, barriers for the introduction of more integrative approaches to risks, specifically to cumulative risks from multiple stressors. In particular, we address the question why such integration has not yet moved up the policy agenda in a decisive way. Corollary questions pertain to the interpretations of 'integration' and the reasons for the various barriers, and to what degree forcing them is desirable, or feasible. That is, we do not take the needs for and the benefits of all kinds of risk integration as granted, and do not regard the barriers for its development and uptake as necessarily and only problematic.

For our purposes, we divide the barriers hindering the development and uptake of more integrative approaches to risks according to the spheres of activity in our societies that deal – each in their own ways – with complex risk issues: the scientific-technical sphere, the administrative- legal sphere, and the socio-cultural and political sphere. This results in the following main categories:

 Epistemological and ontological, knowledge-related barriers (e.g. concepts and interpretations of risks and risk integration; workable epistemological approaches towards complex issues)

 Regulatory and institutional barriers (e.g., legal classifications and provisions, regulatory approaches, institutional structures, including such structures of multi-actor governance)

 Socio-political barriers (e.g. particular interests, political priorities, cultural or 'attitude' and other psychological barriers).

These categories are not to be conceived as strictly separated from each other or as wholly independent: rather they overlap and interact. For instance, institutional systems of classification and categorization for dealing with risks are influenced by conceptual thinking and by the contents and modes of operation in scientific support, such as its organization as a discipline-based activity. The psychological aspects underlie barriers in all categories, even knowledge-based and regulatory, as these are not solely based on facts and choices devoid of psychological content. Resource constraints, often a consequence of political priorities, can also create what is perceived as a knowledge barrier, namely in scientific-technical thinking and information that is judged as too narrow and insufficiently advanced to inform policy making and policy implementation.

The barriers for integrated risk assessment can be characterized in general terms as follows: 88 o Some barriers are strong, visible or persistent, while others are weak, non-evident or temporary o Some barriers are structural, some functional o In some subject areas or spheres of governance controversies may wane (such as perhaps with REACH), in others they may wax o There are differences between different regions and even between different sectors regarding the obstacles for integrated risk assessment and management o Some obstacles are technical and some are related to incompatible mindsets and conceptions, even world-views. o Many barriers are detrimental for efficient governance, but some are legitimate due to the need to differentiate between risks and fields, also to account for their qualitative differences.

5.2 Epistemological and ontological barriers

Epistemological (pertaining to ways of knowing) and ontological (entities or phenomena to be known) barriers, or generally knowledge-related barriers in integrated risk assessment center around the question of how to define and account for multiplicity, multi-dimensionality and complexity in risks and in responses to risks.

Nearly all interviewees mentioned a lack of knowledge as a major reason why cumulative risks from multiple stressors, and integrated risk assessment in general, have not yet moved higher on the list of policy priorities. An example of this is given by the statement of one interviewee: “We can’t shift the regulatory paradigm if the knowledge is not available”, implying that knowledge has to be developed first before policies can follow.

This lack of knowledge is characterized by expressed doubt about the soundness of hypotheses of synergistic as opposed to additive effects and by a perceived lack of tools to address complex risk situations. However, some interviewees did already see serious indications of cumulative effects, mentioning in this respect the adverse effects of indoor air quality and work done by the EU’s Scientific Committee on Emerging and Newly Identified Health Risks (SCENIHR of Directorate- General for Health and Consumer Protection). Nevertheless, the limits of ascertaining and attributing cumulative effects of multiple stressors even in well-defined populations and settings kept coming up, and the interviewees had widely varying views on the implications of this uncertainty for assessment and policy approaches, specifically with regard to what level and kind of integrated treatment of risks would be possible in the various contexts. 89

In general, the interview materials suggest that awareness of complex risk situations is increasing (‘we know that single substance – single exposure assessment is not realistic’; ‘it is desirable to address combined risks’). On the other hand, the responses also suggest that the organizations and experts involved, in short and medium term, remain locked up in what they know and do best. They are reluctant to give up approaches that have shown to lead to 'clear' results, and that have become well established in terms of assigning responsibilities: actors know to a fairly good degree what to expect when they continue to deal with risks on an ‘individual’ basis. There are thus knowledge barriers also due to organizational and institutional factors such as sector divides (see below).

The lack of knowledge was seen in relation to the pronounced complexity of environmental and health risks when they are approached in a combined way, accounting for cumulative effects of multiple stressors. Deficiencies in the knowledge base were thought to be caused by a number of factors:

1. Effectively addressing complex risk situations through generation of the needed broad knowledge would need a lot of resources, which nowadays are not directed to this issue

2. When addressing this complexity, there is still a high probability that possibly important combined risks will not be looked at, i.e. there is a high risk that it will never be entirely clear when a combination of stressors is still safe: "If there are too many unknowns it is difficult to address the issue." The influence of co-factors along with actual stressors adds to the difficulty. This expresses a kind of de-motivation to address issues that are thought to be too complex, as there is skepticism whether one would ever arrive at robust results. The meaning of robustness, the requirements put on it, and their context-dependency emerge as key corollary questions.

3. If on the other hand risks are approached from the perspective of adverse effects that inherently integrate many stressors and co-factors, typically on epidemiological (or eco- epidemiological) and etiological inference of causality, the key knowledge barrier is the attribution of the effects to specific causes also in quantitative terms, to better sort out main causes and negligible causes. Even if this could be done, awaiting effects to show may risk irreparable damage especially in the case of lagged nonspecific effects accumulating in time, for which the cause-effect relationships are particularly hard to prove, justifying precautionary approaches focused on stressors. The problem however remains of harmful unintended consequences of actions thought to attack the right stressors in right ways, i.e. overdoing risk reduction. The interviewees on many occasions acknowledged these problems and some obvious solutions such as integration of improved effect and exposure measures and markers, 90

but seldom expressed the option of coupling causal inference with analyses of risk management options and thereby avoiding the dilemma between stressors-based or effect- based assessment (cf. Finkel, 1996).

4. The experienced probable lack of robustness is not only de-motivating in scientific terms but possibly also disappointing from an organizational and socio-political point of view. Interviewees expressed the fear that if no convincing evidence can be produced there will never be robust approaches to deal with cumulative risks as policy issues, as the produced knowledge will have difficulty to stand the test of policy debates in which high stakes are involved. This can be part of a vicious circle, as without the policy recognition the evidence may be unlikely to be produced, partly because of lacking resources (cf. the first point above)

5. Addressing risks in terms of their combined effects, both from multiple stressors as well as on multiple species, systems and properties, multiplies the requirements of information gathering. Respondents referred to the huge needs and efforts of information gathering and communication under the REACH legislation (largely based on single substance assessment and management), and thought it improbable or even unfeasible that, in the short term, new and even more extensive information collection and analysis procedures would be initiated.

6. Not only the magnitude of the information collection and treatment effort to address combined risks is thought to be a barrier for knowledge development, but also the changed nature of that collection and treatment under a ‘cumulative’ regime. Under current legislation, manufacturers are for instance often requested to provide some information on the risk-related properties of their products. It is however unclear whether an individual manufacturer may be expected to give information on risks of its products in combination with products of other manufacturers. Disclosure of information of one manufacturer to others could be problematic as this information is often considered confidential, e.g. regarding exposures to chemicals in products. Thus, integrative information is not only about technicalities but also about organization and interaction principles (see the sections below).

While such analysis of reasons for deficient knowledge can prove helpful in developing measures and actions to improve information, it also suggests that, until now, the confrontation with and awareness of the complexity and multiplicity of risks has only initiated a limited reflection on knowledge characteristics and its relation with policy making, More specifically, the mentioned reasons for a lack of knowledge reveal that interviewees interpreted complexity mainly as complicatedness (a lot of interdependencies in systems with lots of causal factors and output variables), leading to problems for information gathering and adequate modeling of risk situations, rather than as ‘thick’ complexity (Strand, 2002), which would mean that they acknowledge there is no unique way to describe and analyze such risk issues. The latter approach would include much of the complexity that is due to the value-laden character of risks and to the 91 uncertainties of risk management goals, means and consequences; this level of complexity was more seldom referred to, although crucial both as an obstacle and as a pointer to opportunities for integrated consideration of risks.

A common first reaction to the confrontation with the lack of knowledge and with the barriers to the development of knowledge was to look for areas where feasible and manageable approaches seem possible, such as the combined effects by stressors who have similar modes of action (e.g., effect of combination of pesticides on crops). Additionally, one may look at the combined effect of one stressor on different receptors (workers, consumers, and other human groups). Until now, looking into cumulative risks has most often been limited to this. However, at some level the integration of human and non-human risk assessment has been attempted in various connections (IPCS, 2001), even under previous EU chemicals control procedures (CEC, 1993). Under REACH, integrative approaches additionally include further attempts to take into account metabolites and substances in articles (solid products), life-cycle stages of chemicals and environmental compartments, and also include downstream uses of chemicals (Assmuth et al., 2009c). Even these procedures are so complex that additional integration presents strong barriers, underlining the need for trade-offs between dimensions of integration and for attention to the context and level of use of information.

In general, it became clear from the interviews that the role of knowledge in policies on cumulative risks is considered very important but that there nevertheless is a lack of explicit consideration of how policy making, implementation and evaluation can and have to develop in the case of not only inconclusive evidence but also persistent and serious lack of knowledge or plurality of knowledge (which can be considered a ‘natural’ condition in situations of complexity).

There was also optimism regarding the knowledge barriers, often based on a view of science and expertise that emphasizes quantification, specification and measurements. The possibilities to use specific integrative biomarkers of exposure and effects were mentioned both in connection with chemicals testing and with real-population studies combining exposure and effects measures. Some interviewees also expressed optimism in the developing methods for broad screening and predicting effects and for extrapolating across chemicals and biological systems, although this was tempered by a realization of the limits of knowledge because of the complexity of the systems.

A small minority held the general view that science will be able to deliver hard and convincing evidence on complex risks in a more or less unproblematic way, but then contradicted themselves by saying that "validation of new methods will be exponentially more difficult". Most interviewees expressed simultaneously hope and skepticism that science, along with related other areas of knowledge generation such as testing and monitoring, will provide major breakthroughs in addressing cumulative risks. However, awareness of the limitations of knowledge did not lead to more fundamental questioning of the relationships between policy and scientific knowledge. The 92

‘ideal’ strived at, and considered necessary for good policy making, is still a linear model of ‘first the scientific facts, then the decision on policy measures’.

In our interpretation, this lack of reflection on the implications of complexity for knowledge development and policy making, and of the multi-directional and co-evolving relationships between knowledge and policy, can be regarded as another important barrier to the uptake of cumulative risks as a policy issue. It is related to the barrier of interpreting risks in multi- dimensional terms, not only as physical and quantifiable entities but also as socio-psychological constructs (Assmuth et al., in press); this adds yet another layer to integrated treatment of risks but at the same time opens up new insights and possibilities for dealing with their complexity.

It seems indeed necessary – seeing the situations of ‘thick’ complexity that are typical of cumulative risks of multiple stressors – that, in order to assure that such risks would get a firmer place on the policy agenda, approaches to co-produce policy processes, information gathering and scientific knowledge would be developed and get implemented.

5.3 Regulatory barriers

The regulatory barriers of integrated risk assessment and governance include questions about how legal, institutional and other (such as weaker procedural) structures of regulation such as sector divides allow or do not allow such assessment and governance. In addition, these barriers involve tensions and balances between individual liability and collective co-responsibility.

Next to the epistemological barrier – the awareness of situations in which sufficiently robust knowledge cannot be developed in the short term – a second reason why cumulative risks do not move higher on the policy agenda is its incongruence with current dominant reasoning underpinning regulation and legal frameworks (“The way we regulate is linked with the way we assess”). When approaching cumulative risks, continued reasoning within a framework in which individual (also individual sector) liability is the central concept poses the problem of evaluating and deciding on the importance of each stressor in a combined effect. In addition, problems of assigning liability can be foreseen in case norms in relation to combined effects would not be respected.

Some interviewees expressed that one always needs to assess how regulations can be defended in court, hence stating that the best approaches to regulation may be the simple ones (“Separating 93 things is an easier way to follow than trying to integrate.”) To the contrary, a small minority expressed optimism that ‘it will not be difficult to develop a novel regulatory framework for combined risks’.

Others thought it was questionable whether the ‘successes’ that now often are accomplished, for instance some substances being phased out because of their individual risks, would be achieved if one would stress too much the combined nature of risks to humans and the environment. Hence there was a fear that risk regulation on other grounds than based on a single substance or clear responsibility would be less strict and lead to less benefits for the environment and human health. Risk integration is however inherently neither alarmist or risk-averse; the relative weight of the aggravating and attenuating aspects among the multitude of risks depends on policy (and life) choices.

All in all, a majority of interviewees saw more possibilities for integrating the concern for cumulative effects in legislation that deals with a posteriori evaluation rather than in product legislation which is based on individual liability. An important consideration in relation to regulation in Europe is that the EC policy framework entails forceful drivers for product legislation and food legislation, as food and products are traded in the internal market. As the functioning of the European internal market is seen as a cornerstone of European economic (and political) integration, it is an important driver for harmonization and integration of protective measures and also of risk assessments (CEC, 2000; CEC DG SANCO, 2000a, b, 2002, 2003a). However, overly rigid harmonization of assessment may hinder the development and application of flexible procedures for integrated assessment which would allow variable framings to suit the task at hand (Assmuth and Hildén, 2008, see also Millstone et al., 2008). EU legislation in the environment area leaves much more room for accommodating different national approaches (Jones, 2007).

Integrative approaches might thus be much more adequate and welcomed to contribute to the collective management of environmental compartments, based on deliberated management plans that include voluntary schemes for actions and commitments to comply with certain agreed upon objectives. Explicitly mentioned was the Water Framework Directive, which has, from the ‘nature’ of the problems it addresses, to manage cumulative effects ("River systems are multi-stressor environments.").

In relation to the opportunities for considering combined effects, offered by current chemicals legislation, it was thought that, through the generation of information that it will promote, REACH can provide incentives to consider integrated effects in the frame of other legislation. Some expressed the view that within REACH integrated assessment of mixtures will only make sense for some classes of substances, such as mineral oil products, a highly complex, variable and partly non-definable mixture. However, critical is that REACH made actors aware of a possible barrier to 94 approaches to regulating combined risks, which runs parallel to a perceived barrier for knowledge development, namely the immense requirements of information collection.

The experience with REACH, generally based on simple ‘one substance, one use’ approaches, has evidently made regulators and stakeholders (in particular industry that has a radically increased responsibility for assessments themselves) aware that capacity to assess risks is critical. REACH is already considered such a big and demanding project; requirements for assessing risks in an integrated way are then easily experienced as burdens with no real justification (Assmuth et al., 2009b).

If all industrial players – also the many small and medium size enterprises (SMEs) – have to use the chemical registration system, such a system has to be simple enough that it is usable by non- experts. The need for simple guidance becomes all the more a problem when regulators would want to address combined risks. When integrated approaches would be asked in the frame of regulation, this would probably need an involvement of authorities, who have the expert capacity to deal with this (one can not ask this e.g. of SMEs). 5.4 Socio-political barriers

Socio-political barriers to more integrated risk assessment and management or governance are related to questions such as, what is the cultural attitude and capabilities to address and to cope with the multiplicity of risks in a broader way. In the political domain, the key question is who’s winning with addressing cumulative risks.

Finally, another barrier is a lack of societal concern about cumulative effects. Representatives of civil society have, until now, not taken up the issue of combined risks in a forceful way. While there have been instances of increased attention to combined effects, for instance a media campaign on ‘chemical cocktails’, this finally did not lead to a decisive breakthrough to the list of immediate policy priorities. An interviewee mentioned that in the case of pesticides, there is some pressure from green NGOs and consumer groups to consider cumulative effects, although they also point to the lack of robust insights from integrated risk assessment until now.

Industry, on the other hand, might tend to be interested in cumulative effects of multiple stressors, expecting that an overall assessment will show that exposure to, for instance, chemicals is responsible for only a minor fraction of the actually incurred overall risk (and that generally underestimated factors, e.g. physical stressors, will be shown to be much more important). Projects to evaluate this possibility are sponsored by industry. 95

Beyond the interests, more or less overt, and the power play of actors, many other societal and political factors and processes constitute or influence barriers, as well as opportunities, for integrated risk assessment. These factors and processes include technological and cultural strives toward specialization and, as a counterforce and opportunity, strives toward greater integration, based on a realization of the need to reduce fragmentation. The political integration of the EU is another example of such undercurrents or meta-narratives influencing the risk field in Europe (Assmuth et al., 2009a).

Specifically with REACH, political obstacles may not impede its implementation directly, partly as stakeholders and players presently premium a status quo and ‘Arbeitsruhe’ after exhausting debates, controversy and lobbying. Therefore, additional development toward integrated risk assessment easily is dropped from the agenda. However, needs for such development, and associated disagreements on principle-level issues such as the interpretations of precaution, transparency, participation and subsidiarity, can surface even rapidly after a ‘honeymoon’. 96 6 CONCLUSIONS, DIRECTIONS AND OPPORTUNITIES FOR FURTHEIR DEVELOPMENT

In this section, we discuss and outline strategies for developing and deploying integrated risk assessments for multiple stressors. These strategies and evaluations are based on needs and opportunities mentioned and assessed by individual interviewees as well as on our analysis and interpretation of the interview material as a whole, in particular the sketched contours of a common problem framing in chapter 4 and the analysis of barriers in chapter 5.

The contours of commonalities in problem framing revealed an increased awareness of complex risk situations, which entailed the insight that current ‘non-realistic’ approaches to risks have limitations that invite reflection. There was on the other hand also a concern that, while more ‘realistic’ approaches will necessarily address more numerous causal and influencing factors in risk situations, including their interdependencies, attention needs to be paid to keeping the novel approaches manageable. This relates to the resources that the approaches and methods would need, to the feasibility of information gathering, to the abilities to communicate the approaches and the results to target audiences, and to their usefulness and acceptance in the policy context (‘how can they inform and contribute to developing policy measures and decisions’).

The barrier analysis explained that a current lack of public awareness of combined effects and cumulative risks is leading to a circular problem. There is lack of awareness because of lack of information; the lack of information stems from very limited (and isolated) data gathering and from the absence of accepted methodologies; but due to the lack of awareness there are limited incentives (financial and political) for a more intense research and development effort.

As resources thus remain limited, there is a clear need to concentrate research efforts to areas where integrated approaches would seem to be able to make a difference. This concentration is all the more important because of the risk of a kind of de-motivation to address an area that is in principle boundlessly complex.

Optimal concentration of R&D efforts is however not attainable partly because the needed and possible results and methods are not known beforehand. R&D always needs a considerable freedom to focus on the issues and areas which it considers interesting. This is particularly important with R&D in integrated risk assessment, inherently broad, amorphous and sensitive to findings of risks. In this field the dimensions of and approaches to integration to be chosen are 97 affected by both science-internal and contextual factors and cannot be prescribed, even if precisely prescriptive steering is easily required by stakeholders and regulators because of the high stakes involved.

Therefore, only general, partial and flexible concentration and steering is feasible, guided by the findings and by deliberation and negotiation of their significance for further work. To overcome such dilemmas between exaggerated optimism and scepticism regarding focus, and between politicization and disregard of policy contexts, a stepwise approach seems important, starting with issues and approaches that could deliver useful tangible results and new insights in cumulative risks.

The interview material in general shows there are openings towards more integrated approaches and that there is certainly no outright rejection of them. Rather, there are possibilities for growing support for them, provided scepticism among stakeholders, policy makers and regulators is addressed, not only about the feasibility of such assessments, but also whether addressing cumulative risks will actually lead to better protection and risk management. However, answers to the interview questions showed there is a tendency of ‘lock in’ in current assessment and regulatory approaches.

Quite a few actors are comfortable with a single substance approach based on the assumption of reasonable worst case assessment, because of its familiarity and what they can expect in terms of their own objectives, interests and liabilities.

Thus, it is advisable to clearly communicate initiatives towards integrated treatment of cumulative risks as complementary, and in many cases reinforcing, the familiar single substance-single exposure approaches.

Besides a need for focus in R&D efforts, our analysis also suggests that the relevance of novel methods will be enhanced by accompanying initiatives that connect the development of integrated approaches firmly with their context of use. As such, this analysis is in line with insights, documented in the scientific literature, highlighting the specific characteristics of such emerging science-for-policy and policy-for-science relationships.

In the next paragraph we propose elements that can underpin possible strategies for integrated assessment of cumulative risks from multiple stressors: a number of working principles (co- production of knowledge and policy; interface activities between management and assessment; adaptive and stepwise approach), complemented by guidance for prioritizing and choosing areas and issues of cumulative risks, and by concrete initiatives to start to implement the principles and inform the priority setting and choice process. 98

6.2 Prioritising areas on which to focus integrated approaches

As noted above, the confrontation with a boundless complexity of risks and limited resources for their management calls for priority setting. Despite the limited ability to define specific research objectives, due in part to unpredictable needs, outcomes and opportunities, concentration of efforts is needed on general areas where integrated treatment of risks is judged feasible in terms of research effort and information requirements and can lead to tangible and useful results in the not- so-long term.

Analysis of the interview material delivers clues that can inform the priority setting for integrative approaches:

1. Build further on experiences (be it in policy domains or scientific fields) where cumulative risks and integrated assessment have been practiced to some extent, e.g. assessment of combined effects of pharmaceuticals; practices within industrial safety and occupational health; food safety and nutrition area. There are also examples in which addressing complex risks has led to the development of cooperative frameworks instead of strictly separated assessment-management activities and ‘command and control’ type regulation (e.g., some of the EU technology platforms; sustainable chemistry; nanotechnologies). Further activities can focus on the continued development of integrated approaches in these areas and on developing approaches in other areas that are thought to have similar characteristics and able to utilize integrated approaches in established areas as examples. The exemplary or forerunner function relates both to the scientific aspects of integrated assessment as well as on the ideas in the policy making communities related to the use of the integrated information, i.e. to scientific or technical and policy innovation.

2. Concentrate the development of integrated approaches in first instance on regulatory frameworks that are receptive, i.e. in which the concern with and the assessment of cumulative risks will be (or is already) explicitly and ‘naturally’ an issue, e.g. Water Framework Directive (WFD) and integrated river basin management, Integrated Pollution and Prevention Control (IPPC) Directive (EP and EC, 2008), environmental health (CEC, 2003b), integrated product policy, and general environmental policy integration with other sectors e.g. on the Cardiff process (EP and EC, 2002). However, some elements of integrated treatment of risks can 99

naturally arise also in other fields even when it is not explicitly emphasized, such as with pharmaceuticals (e.g., in integrating considerations of risks and benefits of chemicals).

3. Prioritize research and methods development that explicitly addresses the interaction of scientific knowledge base of assessment and its societal contexts in risk governance (Henry et al., 1997; NRC, 2008; cf. CEC DG SANCO and SCHER, 2006)

4. Develop integrated approaches in areas where the positions of stakeholders, despite limited ‘action’ until now, are favourable or where they can become favourable, despite or sometimes because of controversies, i.e. where they think that addressing ‘cumulative risks’ could lead to policies and practices that fit well within their positions and perspectives. Examples of such areas include the following:

o Industry could be interested that integrated assessments of combined risks will prove that the impact of chemicals (or other products) on human health is much less than that of other stressors; industry could on the other hand also be interested in exploring in a broader outlook how even the lesser risks can be more efficiently managed e.g. by considering cradle-to-grave life-cycles of products, including avoidance of 'collateral' damage and liabilities, and provision of safer alternatives

o NGOs could be interested in that this type of assessment may show that current standards for individual substances will have to be lowered because of simultaneous exposure to multiple stressors (e.g., a consumer who uses several chemicals); they could hope there will be a (restrictive) effect on product legislation. They will also be interested in the prospects of justifying risk reduction by non-health such as ecological risks; they may even buy in to still more integrative assessment to choose between or at least discuss alternatives involving different risks and benefits, instead of automatically opposing all of them.

o Policy makers and stakeholders could see possibilities for saving money (in social security, in health care) if environmental causes for health problems are assessed, i.e. when it would be known that improving environmental conditions would lead to improvements in health, there will be more willingness to address this. On the other hand, they may also identify, especially by inter-sector communication, possibilities to save resources by reducing concerns with some minor, phasing-out or otherwise less significant risks, and by reducing inefficient measures.

In any case, it is worthwhile to reflect on how, in different areas, the frames and positions of actors and stakeholders would be affected by a focus on integrated treatment of risks (of different kinds and in various dimensions), and to take these reflections into account in the development of integrative approaches (Craye et al., 2005). It is already an important step to get actors to deliberate on framings and preferences, exposing disagreements and reasons and assumptions underlying these (Benighaus and Renn, 2007). 100

A concrete suggestion heard during the interviews that can support the process of priority setting, was to develop an information network in which available pieces of scientific information on cumulative risks would be integrated in a coherent way. This would help to generate, for some issues, a critical amount of credible knowledge, and also facilitate multi-actor deliberation on the risks and on what is critical knowledge, instead of just spending more money for research which may not result automatically into very relevant knowledge. There needs to be an upfront effort for organizing but also for processing, discussing and disseminating the broad-based, structured and issue-oriented risk information gathered. Such systems should also interact with and feed back to the generation of information (in research, testing, monitoring and associated methods development). 6.3 Working principles for developing and implementing strategies for integrative treatment of risks

6.3.1 Development of working principles

The interview material strengthens the case to take into account the specific characteristics of science-for-policy, as widely described in science (policy) studies. Our results also highlight the need to broaden and to specify 'sciences' (not only natural science, and including testing and monitoring) and scientists (including multiple, variable and hybrid actors), and to consider the reverse mechanisms of policy-to-science, as a combined 'pull and push' factor in R&D funding and in other areas of interaction between research and application where scientific and policy understandings are co-produced, translated and negotiated. The typical contextual elements for such science-policy assessments include that scientific input is sought on complex issues, involving important uncertainties, while being under close scrutiny of a range of stakeholders, holding various perspectives and having different interests (Funtowicz and Ravetz, 1990).

Studies of successful scientific input into policy making have revealed that rather than being generated as basic scientific insights, removed from considerations of application, and then becoming applied as knowledge specified to solve specific problems, the scientific claims that prove influential in policy making are co-produced with political decisions and social order (Jasanoff and Wynne, 1998; Jasanoff, 1990). The science domain, although more strongly than other domains aspiring and better able to grasp risks objectively, is never free from subjectivity, values and bias, and the stakeholders, besides seeking inputs and providing support to gain these, provide information on their part on both valuations and 'facts' in the applied domain. Assessment and particularly integrated assessment is the middle ground where the domains and aspirations come together. 101

Complexity both in terms of natural scientific aspects and in terms of social and political elements inspires to accompany the development of novel methods by working principles that aim at pro- actively co-producing the 'sciences' underpinning integrative approaches to cumulative risks, and an innovative regulatory environment, and more broadly an innovative system of multi-actor governance, to deal with these combined risks, at developing effective interfacing between assessment and management and at an approach to deployment and implementation that is adaptive and stepwise.

6.3.2 Co-production of policy initiatives and knowledge development on cumulative risks

The interview material highlights a need for a pro-active strategy for the co-production of new, more integrative risk assessment concepts and methods on one hand and regulatory procedures on the other hand, by both researchers and experts and others engaged in the generation of scientific knowledge and assessment methods as well as by appliers of such knowledge and methods, including groups outside administrations.

Simultaneously with a concentrated research effort to develop integrative approaches in some risk areas and in some specific and over-arching dimensions of integration (such as risks and benefits), research into appropriate and innovative regulatory instruments could be carried out, including adequate coupling and feedback between the two lines of research. Concrete policy-attentive, yet multi-disciplinary science-based assessment approaches could then be co-developed with related management initiatives.

When rethinking policy and regulatory frameworks and instruments for dealing with cumulative risks in which integrated assessments can prove relevant, one would naturally think of collaborative and collective approaches, implying the need for more (profound) deliberation (van Asselt and Rijkens-Klomp, 2002). This would cater to the needs and aspirations, also in the frame of EU policies, to improve the participation of groups in European civil society in the reflection, deliberation, negotiation and responses on risks.

Intense dialogue is therefore needed between scientific and regulatory experts (including regulators), stakeholders and policy makers. Scientific information resulting from integrated assessment of risks, notably from social, psychological, communication, decision and management sciences, could serve as the ‘platform’ through which policy measures are discussed and developed, rather than being an input in a linear form of regulation (first assessment, then management). 102

6.3.3 Clarifying and reinforcing links between assessment and management

Stakeholder involvement and participation in the value debates and negotiations on risk assessment principles as well as in its practical conduct (e.g. by industry in the chemicals area) are particularly important. This dimension of participation was often referred to by our interviewees, along with the challenges for communication it entails. Some barriers for the development and adoption of more participatory integrated assessment were pointed out, for example in sharing delicate product information between enterprises, and in reconciling the interests and policy principles of stakeholders in defining the precise meaning of phrases like precaution. On the other hand, it was often mentioned especially by regulators that also in the administrative regime a sensible and reasonable work-sharing and division of mandates and powers need to be assured between different regulatory sectors and levels of governance (Renn, 2998; Assmuth et al., 2009).

It seems advisable that the evolution towards integration across substances, environmental compartments, exposure routes or biological effects would be accompanied by an evolution in thinking about the interface between assessment and management of multiple cumulative risks. These interface activities should counter fears of confusing policy-oriented with politicized assessment and the inherent resistance to new ways of thinking. However, integration at the assessment-management interface is not only about basic concepts or policy principles (NRC, 2008). At a concrete and also technical level, this dimension of integration includes research and methods development in exposures and their causes, as this is often a natural key route to risk reduction (cf. CEC, 1998), specifically when precautionary solutions and actions by those causing the risk are at a premium.

The interviews revealed there is a rather widespread realization of the need for risk assessments that are attentive, in new ways, to the management context and that include elements explicitly addressing management questions, such as in risk evaluation and consequence analysis. In particular, comparative assessment of risks and benefits of alternatives (in the case of chemicals, typically substituting substances) was identified as an emerging need. It was also expected that REACH will bring some improvements in this dimension of integration, e.g. through its emphasis on extended life-cycles of chemicals, on direct linkages between exposure assessment and risk reduction (based on proactive modes), and on risk-benefit analyses.

In relation to the policy and governance domain, the implications of integrated assessment for the degree of precaution in dealing with single risks emerged as a key concern. There was a palpable fear among some interviewees that more integrated assessment might jeopardize the attention on their 'pet' risk. This is likely to become an ever more important issue as integrated assessment and governance evolve and as the more precise interpretations of precaution are sought. However, co- 103 production and negotiation of integrated knowledge about risks can clarify the priority of the multitude of risks and the grounds for prioritization. Such evaluations and negotiations will in any case be essentially supported when the levels of complexity and knowledge are given attention in a multi-directional deliberation process.

6.3.4 Addressing complexity in the context of adaptive risk management strategies

Integrated treatment of cumulative risks from multiple stressors is a complex and in principle boundless area. However, integrated approaches should not become too complicated and irrelevant for policy and application. This has been discussed in the area of integrated assessment and management of chemicals by Hart and Jensen (1992) and later by Suter et al. (2003) and Assmuth and Hildén (2008), all of these authors stressing the need for attention to the context and purpose of assessment, and for dialogue between those involved, in order to define the degree and kind of integration that is appropriate and to make informed, informative and workable trade-offs between the many possible dimensions of integration.

Confronted with this complexity and with the related uncertainties, it may be preferable in some instances to start with ‘bounded’ integration in areas for which:

 There is an existing regulatory, legal or policy framework that ‘naturally’ calls for integrated treatment of risks (cf. above)

 Stepwise development towards more integration could take place within (or starting from) regulatory frameworks that are in place (e.g., integrative strategies for sustainable use of pesticides).

Following this logic, integration would be approached through an adaptive strategy, based on intense and sustained dialogue and on timely and continuous feedback between actors involved both in research and policy making and application (see Benighaus and Renn, 2007). Otherwise there is a risk that the knowledge generated and processed will be irrelevant for practical purposes and that policy and regulatory processes get blocked through confusion and disagreement about the meaning of a multiplicity of risks and pieces of information that result from approaches that are integrated only partially.

The interviews indicated areas in which the development of novel methodologies can connect with evolutions in policy frameworks. The opposite influence comes from science on policy; science therein functions also as a source of ‘early warnings’ on risks, including warnings against non- reflected and non-deliberated management actions that may cause countervailing risks. 104

6.4 Summarizing recommendations and suggestions

6.4 1 Research and methods development

 R&D in the panoply of risks supporting priority setting: where do we have to look at combined effects and combined solutions and where not

 Research and methods development in the policy conditions and uptake of integrated risk assessment

 Utilization of results from EU R&D on integrated risk assessment before embarking on new ventures, and avoiding repetition of work (e.g., on risk communication only as an 'add-on')

 Closer international cooperation, particularly between the EU and the US, in R&D on integrated risk assessment, including integration of natural scientific and technical assessment (of chemicals and other stressors) with socio-economic analyses and assessments

 Overcoming artificial divides in R&D stemming e.g. from programme structures (such as the arbitrary separation of R&D in food and non-food 'integrated' risk assessment)

6.4.2 Forward-looking expert-stakeholder-policy deliberation

As a follow up to the interview analysis and as input to the process of focusing the research effort, a concrete initiative could be proposed, building on the principles described above. The core of the initiative would be to develop and test ideas for a stepwise research agenda for integrated treatment of combined risks. Experts, stakeholders and policy makers would, in a series of discussions, deliberate on:

 Research priorities and concentration of efforts

 Provisions for exploratory 'blue sky' research and methods development

 Communication strategies in the frame of integrated assessment

 Research priorities in the regulatory sphere

 Structures for continued dialogue between assessors, managers and stakeholders

 Approaches to strengthen the interface between assessment and management (risk-benefit, alternatives assessment, precautionary strategies)

 Accompanying activities dedicated to science-policy interfaces on complex issues (e.g. on dealing with assumptions and system boundaries in integrated assessment). 105

6.4.3 Policy development and application

 Adaptive risk management practices allowing flexible development and co-production of 'bold' and limited integration in dealing with risks

 Inter-agency coordination

 Development of the links between statutes and their implementation

 Utilization of the Cardiff process (integration of environmental concerns in other sectors) and other integrative policy processes in the EU

 Global collaboration on integrated risk assessment and management, e.g. in WHO and other UN bodies

 Improved interplay between levels of governance in addressing cumulative risks

 Utilization of uncertainty analyses as a means to focus assessments

 Clearer but also extended roles for advisory bodies, including integrative socio-economic considerations of risks (Note: does not imply 'politicization' of the bodies)

 Multi-actor dialogues and 'multilogues' on integrated risk assessment. 106 ACKNOWLEDGEMENTS

This report has been produced as a deliverable within the integrated project NoMiracle ("Novel methods for integrated risk assessment of cumulative stressors in Europe"), in Work Package 4.3 "Dealing with complex risks and uncertainties in a management context".

The NoMiracle project, nr. 003956, is funded under the European Union's 6th Framework Programme for research, within the Thematic Priority 'Global Change and Ecosystems'. Funding from this program and from the participating organizations is gratefully acknowledged.

The anonymous interviewees are warmly thanked for their interest and time.

The report has been reviewed by partners and coordinators in the NoMiracle project. The authors are responsible for the materials presented, and they do not necessarily reflect those of the funding, coordinating or employing organizations. 107 REFERENCES

Assmuth, T. 2008. Integrated risk assessment and management of water resources: R&D-application links. Abstract of presentation at Water4Health Workshop, Lyon, Dec. 2.-3. 2008.

Assmuth, T., Benighaus, C., Hildén, M., Lyytimäki, J. and Renn, O. 2009a. Big pictures, close-ups, roadmaps and mind-maps: Perspectives on integrated treatment of multiple risks. Int. J. Risk Assess. Manage. (in press).

Assmuth, T., Craye, M., and Hildén, M. 2009c. Beyond REACH: Roadblocks and shortcuts en route to integrated risk assessment and management of chemicals. Sci. Total Environ. (submitted).

Assmuth, T. and Hildén, M. 2008. The significance of information frameworks in integrated risk assessment and management. Environ. Sci. & Pol. 11; 71-86. http://dx.doi.org/10.1016/j.envsci.2007.07.006

Assmuth, T., Hildén, M. and Benighaus, C. 2009b. Integrated risk assessment and risk governance as socio- political phenomena: A synthetic view of the challenges. Sci. Total Environ. (submitted).

Assmuth, T., Lyytimäki, J., Hildén, M., Lindholm, M. and Münier, B. 2007. What do experts and stakeholders think about chemical risks and uncertainties? - An Internet survey. Helsinki: Finnish Environment Institute. The Finnish Environment 22/2007. 117 pp. http://www.ymparisto.fi/default.asp?contentid=241573&lan=en , http://nomiracle.jrc.ec.europa.eu/Documents/PublicDeliverables/D.4.3.5%20What%20do%20experts%20and %20stakeholders%20think%20about%20risks%20and%20uncertainties.pdf

Aven, T. and Kristensen, V. 2005. Perspectives on risk: review and discussion of the basis for establishing a unified and holistic approach. Rel. Eng. System Saf. 90, 1–14. http://dx.doi.org/10.1016/j.ress.2004.10.008

Benighaus, C. and Renn, O. (eds.). 2007. Report from the Stuttgart Workshop on Risk Communication: Communicating chemical risks - The role of risk perception and communication for characterizing and managing cumulative stressors, Stuttgart: University of Stuttgart. http://nomiracle.jrc.it/Documents/PublicDeliverables/D.4.3.14%20Report%20from%20the%20Stuttgart %20Workshop%20on%20Risk%20Communication.pdf

Benighaus, C. and Renn, O. 2009. Communication Needs of Chemical Risks – a Consumer/Public Perspective (previous title: Report of the three focus groups: Good practice of the information and communication of multiple risks from the NoMiracle context.) Stuttgart: Dialogik. Deliverable D. 4.3.6. of the NoMiracle project. 31 pp. http://nomiracle.jrc.ec.europa.eu/Documents/PublicDeliverables/Del_4_3_6_final.doc.pdf

Blanco, J. 2008. Integrated Water Resource Management in Colombia: Paralysis by Analysis? Int. J. Water Res. Devel. 24; 1, 91–101. http://www.informaworld.com/smpp/content~db=all~content=a790182365

Bridges, J. 2003. Report on Priorities in the EU for Risk Assessment in The Non-Food Area. Prepared for the European Commission’s Directorate General for Health and Consumer Protection (Public Health and Risk Assessment Directorate) under contract number A0-7050/03/000231, 9 November 2003. Brussels: Commission of the European Communities, DG SANCO. http://ec.europa.eu/health/ph_risk/documents/risk_key02_en.pdf

CEC. 1993. Commission Regulation (EEC). No. 793/93 of 23 March 1993 on the evaluation and control of the risks of existing substances. Brussels: Commission of the European Communities. Official J. No. L 84, 1-15. http://ecb.jrc.ec.europa.eu/legislation/1993R0793EC.pdf

CEC. 1998. Technical Guidance Document on development of risk reduction strategies. Brussels: Commission of the European Communities, and Luxembourg: Office for Official Publications of the European Communities. 101 pp. http://ec.europa.eu/environment/chemicals/exist_subst/pdf/tg_risk_reduction.pdf 108

CEC. 2000. First report on the harmonisation of risk assessment procedures, Part 1. 26-27 Oct 2000. Brussels: Commission of European Communities, Health & Consumer Protection DG, Directorate C - Scientific Opinions. (Published on the Internet 20.12.2000.) 261 pp. http://ec.europa.eu/food/fs/sc/ssc/out84_en.pdf

CEC. 2001. White paper - Strategy for a future Chemicals Policy. COM(2001) 88 final. Brussels: Commission of European Communities 27.2.2001. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do? uri=COM:2001:0088:FIN:EN:PDF

CEC. 2006. Regulation (EC) No 1907/2006 of the European Parliament and of the Council of 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH), establishing a European Chemicals Agency, amending Directive 1999/45/EC and repealing Council Regulation (EEC) No 793/93 and Commission Regulation (EC) No 1488/94 as well as Council Directive 76/769/EEC and Commission Directives 91/155/EEC, 93/67/EEC, 93/105/EC and 2000/21/EC. Official J. EU; L 396, 49, 1-849. http://eur-lex.europa.eu/JOHtml.do?uri=OJ:L:2006:396:SOM:en:HTML

CEC. 2003a. Technical Guidance Document in support of Commission Directive 93/67/EEC on Risk Assessment for new notified substances, Commission Regulation (EC) No 1488/94 on Risk Assessment for existing substances and Directive 98/8/EC of the European Parliament and of the Council concerning the placing of biocidal products on the market. Brussels: Commission of the European Communities. http://ecb.jrc.it/Technical-Guidance-Document/

CEC. 2003b. A European environment and health strategy. Community strategy for environment and health (COM(2003)338 final). Brussels: Commission of the European Communities Jun 2003. http://europa.eu.int/comm/environment/health/index_en.htm

CEC DG-SANCO. 2002. The Future of Risk Assessment in the European Union. The Second Report on the Harmonisation of Risk Assessment Procedures. Adopted by the Scientific Steering Committee at its Meeting of 10-11 April 2003. Brussels: Commission of the European Communities, Health and Consumer Protection Directorate-General, Scientific Steering Committee. 95 pp. + 4 + 8 + 89 + 14 + 57 + 25 + 37 pp. appendices. http://ec.europa.eu/food/fs/sc/ssc/out361_en.pdf , http://ec.europa.eu/food/whatsnew/oldnews0304_en.html

CEC DG-SANCO. 2003a. Updated Opinion of the Scientific Steering Committee on Harmonisation of Risk Assessment Procedures Adopted on 10-11 April 2003. Brussels: Commission of the European Communities, Health and Consumer Protection Directorate-General, Scientific Steering Committee. 5 pp. http://ec.europa.eu/food/fs/sc/ssc/out355_en.pdf

CEC DG-SANCO. 2003b. Opinion of the Scientific Steering Committee on Setting the Scientific Frame for the Inclusion of New Quality of Life Concerns in the Risk Assessment process Adopted on 10-11 April 2003 as part of its exercise on Harmonisation of Risk Assessment Procedures. Brussels: Commission of the European Communities, Health and Consumer Protection Directorate-General, Scientific Steering Committee. 6 pp. http://ec.europa.eu/food/fs/sc/ssc/out357_en.pdf

CEC DG-SANCO. 2006. Report of the first meeting of the chairs of Scientific Committees of Community bodies involved in risk assessment (7 AND 8 December 2005). Brussels: Commission of the European Communities, Health and Consumer Protection Directorate-General. 28 February 2006 C7/MM D(2006) 370004. http://ec.europa.eu/health/ph_risk/committees/documents/ev_20051207_co07_en.pdf

CEC DG-SANCO. 2008. Future challenges paper: 2009-2014. Brussels: Commission of the European Communities, Directorate-General for Health and Consumer Protection. http://ec.europa.eu/dgs/health_consumer/future_challenges/future_challenges_paper.pdf

CEC DG-SANCO and SCHER. Opinion on research priorities for the 7th framework program human health and the environment. Scientific Committee on Health and Environmental Risks (SCHER) 2006. Brussels: Commission of the European Communities, Directorate-General for Health and Consumer Protection and Scientific 109

Committee on Health and Environmental Risks. 10 pp. http://ec.europa.eu/health/ph_risk/committees/04_scher/docs/scher_o_044.pdf

Craye, M., Funtowicz, S. and Van Der Sluis, J.P. 2005. A reflexive approach to dealing with uncertainties in environmental health risk science and policy. Int. J. Risk Assess. Manage. 5; 2-4, 216-36. http://www.inderscience.com/search/index.php?action=record&rec_id=7169&prevQuery=&ps=10&m=or

Craye, M. 2006a. Reflexively dealing with uncertainty and complexity in policy related knowledge : What can it mean ? In: Pereira, A. and Guedes Vaz, S. (eds.,) Interfaces between science and society. Sheffield, UK: Greenleaf.

Craye, M. 2006b. Stakeholder involvement and precautionary science-policy relations. In: Pistocchi, A. (ed.), Proceedings of the 1st Open International NoMiracle Workshop. Technical Report Series (EUR 22625), European Commission DG Joint Research Centre, Ispra. http://nomiracle.jrc.it/Lists/Events/Attachments/2/proceedings.pdf

EP and EC. 2002. Decision No 1600/2002/EC of the European Parliament and of the Council of 22 July 2002 laying down the Sixth Community Environment Action Programme. Official J., L 242, 1–15, http://eur- lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32002D1600:EN:NOT

EP and EC. Directive 2008/1/EC of the European Parliament and of the Council of 15 January 2008 concerning integrated pollution prevention and control (Codified version). Official J. EU, L 24/8-29 (EN). http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2008:024:0008:0029:EN:PDF

Felt, U., Wynne, B., Callon, M., Gonçalves M.E., Jasanoff, S., Jepsen, M., Joly, P.-B., Konopasek, Z., May, S., Neubauer, C., Rip, A., Siune, K., Stirling, A. and Tallacchini, M. 2007. Taking European Knowledge Society Seriously. Report of the Expert Group on Science and Governance to the Science, Economy and Society Directorate, Directorate-General RTD. Luxembourg: Commission of the European Communities. EUR 22700. 96 pp. http://ec.europa.eu/research/science-society/document_library/pdf_06/european-knowledge- society_en.pdf

Finkel, A.M. 1996. Comparing risks thoughtfully. Risk - Health Saf. Environ. 7, 325-46.

Funtowicz, S.O. and Ravetz, J.R. 1990. Uncertainty and Quality in Science for Policy. Dordrecht: Kluwer.

Funtowicz, S. and Ravetz, J. 1993. Science for the post-normal age. Futures 25; 7, 739-755. http://dx.doi.org/10.1016/0016-3287(93)90022-L

Funtowicz, S. and Strand, R. 2007. Models of science and policy. In: Ching, L.L. and Traavik, T. (eds.), Biosafety First: Holistic Approaches to Risk and Uncertainty in Genetic Engineering and Genetically Modified Organisms. Trondheim: Tapir Academic Press.

Grin, J., Graaf, van der H. and Hoppe, R. 1997. Technology Assessment through Interaction – a guide. Den Haag: Rathenau Instituut. Hart, J.W. and Jensen, N.J. 1992. Integrated risk assessment or integrated risk management? Regul. Toxicol. Pharmacol. 15; 1, 32-40. http://dx.doi.org/10.1016/0273-2300(92)90081-J Henry CJ, Moses MS, Henderson R. 1997. The need for policy and risk analysis--the Department of Energy experience. Toxicol Pathol. 25; 1, 27-31. http://tpx.sagepub.com/cgi/reprint/25/1/27 Horlick-Jones, T. 1998. Meaning and contextualisation in risk assessment. Rel. Eng. System Saf. 59, 79-89. DOI: 10.1016/S0951-8320(97)00122-1. http://dx.doi.org/10.1016/S0951-8320(97)00122-1 IPCS. 2001. Integrated risk assessment. Report prepared for the WHO/UNEP/ILO International Programme on Chemical Safety. WHO/IPCS/IRA/01/12. Geneva: IPCS. http://www.who.int/ipcs/publications/new_issues/ira/en/index.html 110 IPCS. 2004. A framework for the Integration of Health and Ecotoxicological Risks. Geneva: IPCS. http://www.who.int/pcs/emerg_site/integr_ra/ira_report.htm Jasanoff, S. 1990. The fifth branch : science advisors as policymakers. Cambridge, Massachusetts: Harvard University Press. Jasanoff, S. 1995. Procedural choices in regulatory science. Technol. Soc. 17, 279-293. http://dx.doi.org/10.1016/0160- 791X(95)00011-F

Jasanoff, S. and Wynne, B. 1998. Science and decisionmaking, In: Rayner, S. and Malone, E. (eds.), Human choice and climate change, vol.1. The societal framework. Columbus, Ohio: Battelle Press.

Jones, J. 2007. Regulatory Design for Scientific Uncertainty: Acknowledging the Diversity of Approaches in Environmental Regulation and Public Administration. J. Environ. Law 19; 3, 347-365. http://jel.oxfordjournals.org/cgi/content/abstract/19/3/347

Kasperson, R., Renn, O., Slovic, P., Brown, H. S., Emel, J., Goble, R., Kasperson, J. X., and Ratick, S. 1988. The Social Amplification of Risk: A Conceptual Framework. Risk Anal. 8; 2, 177–187. http://www3.interscience.wiley.com/journal/119462808/abstract

Krimsky, S. and Golding, D. (eds.). 1992. Social theories of risk. Westport, Connecticut: Praeger.

Millstone, E., van Zwanenberg, P., Levidow, L., Spök, A. and Hirakawa, H. 2008. Risk Assessment Policy – Differences between jurisdictions. JRC Scientific and Technical Reports 23259 EN. Seville, Spain: IPTS. http://ftp.jrc.es/EURdoc/JRC37719.pdf

Munns, W.R., Suter, G.W. II, Damstra, T., Kroes, R., Reiter. L.W. and Marafante, E. 2003. Integrated Risk Assessment -- Results from an International Workshop. Human & Ecol. Risk Assess. 9; 1, 379-386. doi: 10.1080/1080703031877177. http://www.ingentaconnect.com/content/tandf/bher/2003/00000009/00000001/art00017

NRC. 2008. Science and Decisions: Advancing Risk Assessment. Washington, D.C.: U.S. National Research Council. National Academy Press. http://www.nap.edu/catalog.php?record_id=12209

Renn, O. 1992. Concepts of risk: a classification. In: Krimsky, S. and Golding, D. (Eds.), Social theories of risk. Praeger, Westport, CT, pp. 53–79.

Renn, O. 2007. Risk Governance: Basic elements and requirements for communication and participation, In: Renn, O. and Benighaus, C. (Eds.), 2007. Report from the Stuttgart Workshop on Risk Communication: Communicating chemical risks - The role of risk perception and communication for characterizing and managing cumulative stressors. pp. 14-1. http://nomiracle.jrc.it/Documents/PublicDeliverables/D.4.3.14%20Report %20from%20the%20Stuttgart%20Workshop%20on%20Risk%20Communication.pdf

Renn O. 2008. Risk governance: coping with uncertainty in a complex world. London: Earthscan . 455 pp.

Renn, O. and Benighaus, C. (eds.) 2006. Report on approaches to the characterization of knowledge of risks, uncertainties and ambiguity and their use and quality assurance in the IP domain. Deliverable 4.3.2., part DIA of NoMiracle project. Stuttgart: Dialogik. 55 pp. http://nomiracle.jrc.ec.europa.eu/webapp/ViewPublicDeliverables.aspx 111

Renn, O. and Klinke, A. 2002. A new approach to risk evaluation and management: risk-based precaution- based and discourse-based strategies. Risk Anal. 22, 1071–94. http://www3.interscience.wiley.com/cgi- bin/fulltext/118950700/PDFSTART

Strand, R. 2002. Complexity, Ideology, and Governance. Emergence 4; 1, 164–183. http://www.informaworld.com/smpp/content~db=all~content=a787464101 Suter, G.W. II, Munns, W.R. and Sekizawa, J. 2003. Types of Integration in Risk Assessment and Management, And Why They Are Needed. Hum. & Ecol Risk Assess. 9; 1, 273-279. doi: 10.1080/1080703031877339. http://www.ingentaconnect.com/content/tandf/bher/2003/00000009/00000001/art00018

USEPA. 2003. Framework for Cumulative Risk Assessment. Risk Assessment Forum, U.S. Environmental Protection Agency, Washington, D.C. EPA/630/P-02/001F, May 2003. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=54944 van Asselt, M.B.A. and Rijkens-Klomp, N. 2002. A Look in the Mirror: Reflection on Participation in Integrated Assessment from a Methodological Perspective. Global Environ. Change 12, 167–184. http://dx.doi.org/10.1016/S0959-3780(02)00012-2

Van de Graaf H. & Hoppe R. 2000. Beleid en politiek. Bussum, The Netherlands: Coutinho. Vermeire, T., Munns, W., Sekizawa, J., Suter, G. and Van der Kraak, G. 2007. An Assessment of Integrated Risk Assessment. Human & Ecol. Risk Assess. 13; 2, 339-354. doi: 10.1080/10807030701226848. http://www.ingentaconnect.com/content/tandf/bher/2007/00000013/00000002/art00009 Wynne, B. 1996. May the sheep safely graze? A reflexive view on the expert-lay knowledge divide. In: Lash S. et al. (eds.), Risk, environment and modernity: towards a new ecology. Pp. 44-83. London: Sage. 112 Annex 1: Structure of the interviews and specific questions

Short intro by interviewer

- The interview addresses cumulative risks of multiple stressors. For the contributions of the NoMiracle project to dealing with them in novel integrative ways, various concepts of cumulative risks and integrated assessment are being examined. Cumulative risks in this connection typically mean risks caused by many chemicals and stressors jointly. Integrated risk assessment means analytical treatment of risks in ways that integrate many aspects of them, such as risks from all such agents, and risks to humans and ecosystems.

- The main objective of the interview is to build further on earlier survey of expert and stakeholder views on risks, risk assessment and risk management by soliciting views from key EU actors on the needs, contexts and conditions for use (and even development) of integrated assessment methodologies; especially on whether and how they could contribute to policy-making on cumulative risks.

Theme 1) Framing cumulative and comparative risks and impacts of multiple stressors

Purpose of this theme: To understand how the concepts of cumulative risks and multiple stressors are framed and dealt with by actors (organizations or institutions)

1.1 Are cumulative risks or risks from multiple stressors an issue in your work, or in the work of your organization? In what ways?

(Assistance: what are the key problems in relation to cumulative risks seen by your organization, and what are the goals of your organization in relation to dealing with cumulative risks)

- Follow-up: If they are not and issue in the work of your organization, should they be?

Optional:

1.2 What do you think are the most important cumulative risks to humans, and why?

1.3 What do you think are the most important cumulative risks to ecosystems, and why? 113

Theme 2) Addressing cumulative risks and multiple stressors in policy making

The purpose of this theme is to identify issues in how the policy formulation and implementation processes address cumulative risks from multiple stressors, and in development of these processes.

2.1 Is there a need to develop policies and their implementation to deal in an appropriate way with cumulative risks and multiple stressors ? If so, which are the most important elements to develop?

- Follow-ups (if needed):

- What are the main organizational and institutional changes needed?

- Would responsibilities for assessing and managing risks have to be changed, and how?

2.2 Do you see needs and possibilities for wider application of participatory processes? Which needs or possibilities?

Theme 3) The role of knowledge, uncertainty and precaution in policy approaches

The purpose of this theme is to explore to what extent the governance of cumulative risks is framed as something fundamentally different from "standard" risk governance, in particular when it comes to information, uncertainty and precaution

3.1 Do the concepts of uncertainty and precaution change when the focus is on cumulative risks and multiple stressor? How?

3.2 What would these changes imply for the interpretation of the precautionary principle?

- Follow-up: Do you see problems and opportunities in combining the precautionary principle with evidence- based assessment? Which ones? 114

3.3 How would you characterize the knowledge base available and the knowledge base needed for addressing cumulative risks

Theme 4) Integration in risks, risk assessment and risk management and its uptake in the policy sphere

The purpose of this theme: To understand how integration is framed: what does it include and what does it not? Which are the institutional and organizational barriers to integration that are considered particularly relevant (note barriers are obviously conditional on the framing – if a particular kind of integration is excluded by the framing then any corresponding barrier is obviously irrelevant)?

4.1 Is more integrated treatment of risks important, and why? Are there policy-driven needs for it, and which?

4.2 In which sense should risk assessment become more integrated, and why? What kind of integrated treatment of risks do you consider the most important, and why?

(Assist: Do you think the most important thing is generally speaking integration of agents, target organisms, outcomes, stages of risk formation, risks and benefits, or what?)

4.3 What problems do you identify or foresee with more integrated risk assessments? Why?

Assist1: Specifically, do you think attempts to address a multitude of complex risks can complicate assessment and management?

Assist2: Are there institutional or organizational barriers, and which are the most important?

4.4 Specifically, should risk assessment and risk management be better integrated, and why or why not?

4.5 (Optional) Specifically, can assessment of risks to human health and ecosystems be integrated, and how?

Theme 5) Integrated chemicals assessment and management in relation to REACH 115

Purpose of this theme: To document and understand the direction in which REACH (and closely related regulatory mechanisms) is moving and will move the whole chemical risk governance, and to explore if it is indeed raising demands or opportunities for addressing cumulative risks and multiple stressors

5.1 What are the most important challenges for implementation of REACH?

(- Assist: Are they in information, policy, structures, resources, communication, trust, or where?)

- Follow-up: How could these challenges be met?

5.2 How do you see the importance of integrated treatment of risks in connection with REACH? Is it for instance an important challenge immediately, or more in the longer term?

- Specifically, what integration is needed with other areas of legislation and sectors of governance?

5.3 What are the roles of generation and use of information in this implementation? Do they involve problems, and what kinds?

- Assist1: Can the information flows become too extensive and detailed, and hard to manage? If so, is the solution mainly in technical management or in more fundamental development of the system?

- Assist2: Is the combination of a confidential and ordered process with transparency and participation a problem, and how?

- Specifically, what are the roles of experts in relation to chemicals regulation? Should researchers and experts mainly supply the information deemed necessary in the procedures, or do you see a role for them in producing unanticipated information and in questioning and developing the procedures?

4.4 Is there a (rising) demand for novel methods particularly for integrated assessment of chemical risks, and for what kinds in particular?

- Follow-up: what kind of R&D will be initiated/continued through the REACH agenda?

Round-off

Are there any important elements in relation to cumulative risks and the role of integrated assessments in dealing with these risks that you feel we have not treated yet during the interview? 116 117 Annex 2: Letter of invitation to participate in interviews

Dear recipient,

As part of the NoMiracle (Novel Methods for Integrated Risk Assessment of cumulative Stressors in Europe) project of the EU's Framework Programme 6 (nomiracle.jrc.it), interviews will be conducted with key experts and actors.

The NoMiracle project involves 38 institutions from 17 European countries, and deals mainly with complex risk assessment of chemical mixtures, integrating environment and human health.

The objective of the short semi-structured interviews is to gain further insights for the development of risk assessment methodologies, as a follow-up of a recent Internet survey (see http://www.finenvi.org/questionnaire/questionnaire.html). One of the particular aims of NoMiracle is to support the development of integrated approaches that are sensitive to their contexts of use and to their users’ needs. To this end, expert opinions are solicited from people who hold key positions at EU level in the field of chemicals and, more generally, environment and health: either in the formulation of policies or in steering the conception and application of risk assessment. The general idea is to utilize their experiences, take into account their concerns, and to give them the opportunity to participate in the development of assessment.

The interviews will be conducted during weeks 25-28 (June 19th - July 12th) by Timo Assmuth (Finnish Environment Institute) and Matthieu Craye (EC DG Joint Research Center). Both are active in NoMiracle Work Package 4.3 on Risk perception, communication and policy.

As coordinators, we kindly ask your cooperation in this communication endeavour. You may indicate additional person(s) in your organization who could be available for the interviews. All material gained through the interviews will be treated confidentially and the results will be reported anonymously.

The interviewers will subsequently be in touch with you or, alternatively, with the contact persons indicated by you, to find suitable times for the interviews. In case you have questions or would need further information, you may also contact the interview team at [email protected] and [email protected], tel. +358-40-5907712.

Thanking you in advance for your important contribution and cooperation, yours sincerely

Hans Løkke, National Environmental Research Institute of Denmark Coordinator, NoMiracle Integrated Project

Mikael Hildén

Professor, programme director, Finnish Environment Institute, Coordinator of Work Package 4.3