Skip to main content

"Health and Environmental Risks: Organisation, Integration and Cross-fertilisation of Scientific Knowledge"

Final Report Summary - HEROIC (Health and Environmental Risks: Organisation, Integration and Cross-fertilisation of Scientific Knowledge)

Executive Summary:


The vision of a sustainable and safe use of chemicals in order to protect human health, preserve the environment and maintain the ecosystem services for the benefit of human welfare requires modern, innovative and more holistic approaches to risk assessment (RA) in order to better inform health- and environmental-based decision making.

Integrated risk assessment (IRA) has been proposed as a solution to these scientific, societal and policy needs, because it better manages the balance between increasing public demand for safety testing and the refusal of additional animal testing, while reducing at the same time the economic burden of the regulatory system on the industry.

In the context of the HEROIC project, IRA is defined as the mutual exploitation of environmental risk assessment (ERA) for human health risk assessment (HHRA) and vice versa in order to coherently and more efficiently characterize overall risks to humans and the environment for better informing the risk analysis process. Extrapolating between species relevant for HHRA and those relevant for ERA is at the core of IRA. This requires a detailed understanding of the pathways of toxicity/modes of action for the various toxicological endpoints.

In recent years significant scientific advances, changes in the chemical legislation and increasing environmental consciousness have created a favourable scientific and regulatory environment to further develop and promote the concept and vision of IRA. Yet the map of each sector-specific RA framework is extremely complex, and there is no one-size-fits-all solution to integrate human and environmental RAs.

In order to foster the incorporation of IRA approaches into the different chemical sectorial regulations and demonstrate the reliability of IRA approaches for regulatory purposes, an initial proof of concept is needed. It is anticipated that once confidence has been gained with IRA this will ultimately contribute to an overall reduction in in vivo toxicity testing requirements. However, significant progress will only be made if long-term support for mode of action-related research is secured.

In the short term, further exchange and harmonization of RA terminology, models and methodologies across chemical categories and regulatory agencies will support these efforts. This will facilitate better mutual understanding between risk assessment experts across disciplines, and lower the existing institutional barriers for implementing IRA.

Since societal values, public perceptions and cultural factors are of increasing importance for the acceptance of risk analysis and a successful implementation of risk mitigation measures, the integration of socio-economic analysis and socio-behavioural considerations into the risk analysis process could help to produce a more effective risk evaluation and consideration of the risks and benefits associated with the use of chemicals.

Project Context and Objectives:
1 Project Summary

1.1 Project Context and Main Objectives

1.1.1 The conceptual environment: science, policy and socio-economic drivers for integration in the EU chemical sectorial legislations

Since the late 1970s, risk assessment (RA) of chemicals has served the needs of health and environment protection policies worldwide (NRC, 1983; Renn, 1998). For historical and practical reasons, the separation of human health RA (HHRA) and ecological (environmental) RA (ERA) of chemicals in general is deeply rooted in the culture and practices of many RA or risk management institutions and organisations at the EU level and beyond. This is mainly considered a consequence of the allocation of the RA of the different chemicals categories to distinct regulatory authorities and scientific disciplines. Confined in the regulatory practice by sector-specific chemical legislations (e.g. REACH, Cosmetic Products Regulation, PPP Regulation), HHRA and ERA have developed independently, using largely separate databases, models and assumptions to evaluate the risk, often sharing the same type of shortcomings and limitations, e.g. in term of availability, accessibility and interpretation of (eco)toxicological information, or lack of exchange and mutual understanding of best RA practices within and across the two scientific disciplines (Bridges, 2003).

While extensive cumulative experience has been gained over time in chemical RA and new regulations have been established for a wide range of chemical stressors, current regulatory RA practice for chemicals faces substantial challenges to meet present and future scientific, ethical and policy needs. On the one hand, there is the desire from politics, non-governmental organisations and the general public to better evaluate the risk to human health and the environment of a large number of chemical products and processes. On the other hand, there is increasing political and public pressure to reduce animal testing used in the safety assessment of those chemicals (viz. the 3R concept). However, in a context of budget restrictions and limited capacity, the demand for RA will continue to increase to address present and future testing needs of a vast number of chemical products, processes and toxicological endpoints.

This situation calls for a paradigm shift to expand from the traditional, high dose extensive standard testing battery to a more cost-effective and knowledge-driven approach (Hartung, 2009, 2010) to address the emerging issues in modern toxicology, such as the potential toxicity of nanomaterials, low dose extrapolation to assess the toxicity of micro-pollutants and endocrine disrupting chemicals (EDCs) and the toxicity of mixtures to assess real life exposure scenarios. Besides, testing every chemical for every possible health and environmental effect is impractical and therefore prioritization is essential. The intensity and depth of toxicity testing for regulatory purposes should primarily be driven by the needs of risk management and therefore prioritization should be given to those chemicals where human and environmental hazard and/or exposure are likely to be high.

1.1.2 The need for more integrated and sustainable RA approaches

The vision of a sustainable and safe use of chemicals in order to protect human health, preserve the environment and maintain the ecosystem services for the benefit of human welfare requires modern, innovative and more holistic approaches to RA in order to better inform health- and environmental-based decision making. There is a need for more cost effective, predictive and rapid tests for high quality sustainable RAs, including a better exploitation of existing data to optimise resources use in HHRA and ERA. Consequently, many regulatory authorities and organisations worldwide have called for and are to an extent already practising modern, innovative and more integrated approaches to RA to address those scientific, societal and policy needs and better inform health- and environmental-based decision making (EC, 2003; NRC, 2007; Suter, 1997).

In response to these challenges, integrated risk assessment (IRA) has been proposed as a potential solution because it better exploits and combines existing (eco)toxicological information, and therefore better manages the balance between increasing public demand for safety and the refusal of additional animal testing, while reducing at the same time the economic burden of the regulatory system on the industry (Suter et al., 2005; Vermeire et al., 2007; WHO, 2001).

1.1.3 Historical perspective, concept and perceived value of integrated risk assessment

Anticipating the changing needs of RA processes, an international expert group involving the EC, US EPA and the OECD was set up in 1998 under the umbrella of WHO-IPCS to advance the integration of approaches for HHRA and ERA. Their report (WHO, 2001) outlined a generic framework for IRA that could be used a guidance to be applied to all chemical categories and that addresses real life multi-chemical, multimedia, multi-route and multispecies exposures.

IRA brings together into one single assessment independent sources of (non)-testing toxicological and/or ecotoxicological data that are usually kept separate in the HHRA and/or ERA process. Integration can be applied in various contexts and at different levels of complexity in chemical RA. IRA builds on the identification of commonalities and interconnections between human health and the environment. One can integrate components such as exposure and effects, in vivo, in vitro and in silico data, multiple chemicals, multiple exposure routes, toxicological endpoints, modes of action, spatial and temporal scales, product life cycle or socio-economic aspects. Integration can be confined to either hazard or exposure assessment (referred to as integrated hazard assessment and integrated exposure assessment, respectively) or to both, for the individual needs of HHRA or ERA, or across the two scientific disciplines. It follows that, by integrating all the existing knowledge of a chemical into the RA process, IRA offers the opportunity for more comprehensive, efficient and cost effective RA (Munns et al., 2003; Suter et al., 2005; Vermeire et al., 2007).

The concept of IRA fosters a systematically coordinated exchange of information from the outset between human health and environmental risk assessors. This would translate into an improvement of the quality and scope of assessments, and in turn result in a more coherent, informative, “ready-to-use” RA that is more policy and management relevant. IRA would improve the efficiency of the assessment process because IRA builds on the exploitation of shared data and models, and the transferability of knowledge of mechanisms and modes of actions (MoA) across risk endpoints and stressors. The RA output would become more reliable because a broader scope would reduce uncertainties in the decision-making process, while increasing the likelihood of identifying unexpected and emerging risks, and giving more predictive and accurate risk estimates (Suter et al., 2003; Vermeire et al 2007; Vermeire, 2009). While “integrated thinking” has already reduced the need for further testing in some specific cases (e.g. in REACH assessments), it is anticipated that IRA will reduce overall assessment costs relative to independent ecological and health assessments.

In 2003, the European Commission highlighted IRA as a key element of future action in its European Environment and Health Strategy (EC, 2003), paving the way for the development of the concept as new EU-funded research projects (e.g. FP6 HEIMTSA (http://www.heimtsa.eu/) INTARESE (http://www.intarese.org/) NoMiracle (nomiracle.jrc.ec.europa.eu/) OSIRIS (http://www.osiris-reach.eu/) 2-FUN (http://www.2-fun.org/)) were funded to better characterize the link between environmental risk factors and health-related impacts. Building on this legacy, the FP7 Coordination project HEROIC (Health and Environmental Risks: Organisation, Integration and Cross-fertilisation of Scientific Knowledge) aimed to consolidate the existing knowledge and identify what is necessary to further develop and promote the implementation of IRA.

1.1.4 HEROIC main objectives

HEROIC intended to pave the way for the development of an IRA framework with the aim to identify and explore new ways of improvement, harmonization and cross-fertilization of tools and methods used in environmental and human risk assessments.

HEROIC’s overall objectives were:
• To evaluate the cross-disciplinary use of hazard and exposure data in HHRA and ERA to optimize resources use and meet current needs of risk assessment, risk management and policy-making;
• To give recommendations on how to better harmonize tools and methods used in human and environmental risk assessments;
• To foster cross-disciplinary cooperation and mutual understanding among human and environmental scientists, including risk assessors and risk managers;
• To show the added value of IRA, foster its further development and promote its acceptance to a broad stakeholder audience, including risk managers, policy makers and NGOs;
• To give guidance on how to ultimately establish a framework for IRA across all chemical classes (including industrial chemicals, human and veterinary pharmaceuticals, pesticides, biocides, food additives, cosmetics, and also chemical mixtures).

(Please see attached list for references.)

Project Results:
1.2 Main Science & Technology Results

The HEROIC project structure was built to stimulate efficient bottom-up exchange of information and data between the 6 individual work packages (WP) so as to build a solid process to develop new approaches. Whereas WP2 was completed after the first 12 months (M12) of the project, all other WPs were running until the whole duration (M1-M36) of the project. Figure 1 (attached as separate pdf document) shows the workflow and how the knowledge produced feeds into HEROIC major contribution, the White Paper.

Figure 1: Workflow of the HEROIC Project (2011-2014)
The chart shows how the information flows between the various workpackages (WP). Delivery dates of each WP major outputs are indicated as per project submission month (M).

1.2.1 Overview of the HEROIC project achievements

While WP1 was dedicated to the general coordination and management of HEROIC, the project work itself began with WP2 “Current risk assessment practice” (M1-M12), which served as a starting point for all the other WPs. Based on a comprehensive landscaping exercise to identify common methodological and data needs in current HHRA and ERA, WP2 produced a map of key stakeholders, sector-specific RA processes and regulatory frameworks for six chemical categories (industrial chemicals, pesticides, biocides, human and veterinary pharmaceuticals, food additives, cosmetics) at EU level, and evaluated current approaches in chemical RA and priorities for future harmonisation and improvement. Limitations, barriers and opportunities to the promotion of IRA were further identified through an expert consultation process, including structured targeted interviews of risk assessors and risk managers within regulatory bodies and the private sector at EU level. Our initial findings were validated through the first HEROIC expert workshop held on 12-13 April 2012 at the National Institute for Industrial Environment and Risks (INERIS), Verneuil-en-Halatte, France, which convened over sixty participants from public institutes, academia and the business sector. Based on a gaps and needs analysis and the identification of opportunities and key drivers for success (i.e. political, regulatory, socio-economic, technological), an action list was developed on how to close the gap between current and desired conditions of RA practice, and how IRA could improve the overall quality and efficiency of HHRA and ERA. A set of recommendations for further action was proposed to better promote the development of an IRA framework.

Based on the input of WP2 regarding the scientific and regulatory requirements of an ideal RA practice, WP3 ”Data and models for exposure and effects assessment” (M1-M30) addressed more specifically the issues of current RA practices with a focus on their data gaps and needs and the exposure models used. The initial findings of WP2 and the first HEROIC workshop, together with complementary works inside and outside the Consortium, permitted to propose an inventory of existing endpoints, databases, platforms, models and methods used in hazard and exposure assessment for single compounds and mixtures. This resulted in the recent publication of a paper stating the HEROIC vision of the perspectives for IRA, in conjunction with socio-economic analysis (Péry et al., 2013). Building on this preliminary work, WP3 explored innovative approaches and further opportunities for extrapolating across human and environmental hazard and exposure data, models and methods (with a focus on endpoints, species and mixtures) through two dedicated expert meetings: i) the meeting for hazard extrapolation, held on 7-8 November 2013, at the Helmholtz Centre for Environmental Research (UFZ) in Leipzig, Germany; and ii) the meeting for exposure extrapolation, held on 21-22 January 2014, at Electricité de France (EDF) in Chatou and INERIS in Paris, France. These meetings gathered 28 and 18 participants, respectively, primarily from academia and from regulatory authorities involved in the risk or exposure assessment area. As a result of these activities, guidelines in terms of perspectives and recommendations to better harmonize and extrapolate between HHRA and ERA were drawn up.

Building on the information produced in WP2 and WP3, WP4 “Integrated decision making framework for the future” (M8-M36) comprehensively reviewed and analysed existing Weight of Evidence (WoE) definitions, frameworks and applications in the field of HHRA and ERA as well as approaches for evaluating (eco)toxicological and exposure data quality with a focus on relevance, reliability and uncertainty. Recommendations for improved, more robust, structured and transparent integrated assessment strategies and WoE approach in HHRA and ERA were made, and a general decision-making framework for improving the integration of data in the frame of ERA, HHRA and potentially IRA was proposed. Some of these initial recommendations were tested through dedicated expert consultation activities using three selected case studies to further evaluate how HHRA and ERA can benefit from each other, with a focus on how human and environmental risk assessors cross-evaluate and integrate (eco)toxicological data and to come to a decision based on Best Professional Judgement (BPJ): i) the expert elicitation workshop held on 1-2 October 2013, at EDF in Chatou, France, convened 21 participants from academia, regulatory health agencies and industry, focusing on skin sensitisation and on endocrine disruption of chemicals; and ii) an online consultation exercise in July 2014 on the hazard from veterinary compounds. Through these activities, WP4 was able to analyse and better understand sources of (dis)agreement in expert decision rules and to identify the main drivers for integration in the decision-making process.

A diverse range of dedicated activities was planned within HEROIC for information exchange, dissemination of scientific knowledge, development of capacity building material and communication outside the project. Those were the primary tasks of WP5 and WP6.

WP5 “Information platform "Tox-Hub" (M1-M36) has developed a web-based tool and open-access platform (web portal "Tox-Hub") to facilitate the access to relevant toxicological and ecotoxicological data and information from various certified sources in an integrative manner, with the aim of improving communication and collaboration in RA as well as supporting capacity building for the professionals involved in HHRA and ERA. The Tox-Hub was released to the public in June 2014 and is accessible freely on the website of the HEROIC project at http://www.heroic-fp7.eu or directly at www.heroic-toxhub-platform.es. A ‘User Manual’ has been produced and is downloadable on line. A special issue of the HEROIC newsletter was released in June 2014 to present the Tox-Hub basic concepts and main functionalities, and is available on the HEROIC website. Currently 20 databases are linked to the Tox-Hub, with a dictionnary containing 12000 words and 14000 references. The inclusion of further external dictionaries and databases is foreseen in the near future. In parallel with the internal work run within the HEROIC Consortium, WP5 has established contacts with another EU FP7 project, diXa (www.dixa-fp7.eu) and the OECD eChemPortal for the purpose of coordination and exchange of best working practice. The first HEROIC-diXa Coordination Meeting was held at JRC, Ispra, Italy, on March 13, 2013, and a follow-up training session was organized on July 3, 2013 with the OECD eChemPortal at the Spanish National Research Council (CSIC) in Barcelona, Spain.

WP6 “Dissemination, capacity building and Communication” (M1-M36) implemented HEROIC communication and dissemination strategy inside and outside the Consortium, raising awareness about the resources of the project by bringing HEROIC major outputs and activities to a broad stakeholder audience. The dissemination action plan was put in place adopting different dissemination tools available, including dedicated quarterly newsletters, flyers, peer-reviewed articles in scientific journals, workshop proceedings, posters at meetings. The HEROIC website (www.heroic-fp7.eu) which was launched publicly in M6, was the main tool for disseminating the project outputs to end users (risk assessors and policy makers, the scientific community) and the general public, and for providing access to the e-learning and data base repository “Tox-Hub” platform as well as providing a connection with other related EU projects. Building on the work of WP2 and as a complement to the approach taken by WP4, WP6 launched a broad open scale stakeholder consultation (online survey) to collect opinions on potential sociobehavioural factors that interplay and may influence the efficiency of the decision making process in the way experts evaluate risk to come to a decision. A follow up roundtable was organized at the OPERA Centre in Brussels on 7 April 2014 to discuss and refine the survey preliminary findings. This work highlighted how the qualitative component, e.g. in term of human values and preferences, may contribute to shape current policy priorities, protection goals and perception among different societal groups, and how knowledge in risk assessment is interpreted by risk managers and translated into risk management options. It also offered some perspectives on how IRA can benefit from the inclusion of socio-economic analysis and social science.

1.2.2 What HEROIC has learned and how it has shaped our vision of IRA

Building on the lessons learnt from previous initiatives, in particular the legacy of the WHO/IPCS framework for IRA (2001), HEROIC has consolidated the existing knowledge related to IRA. As a result of our activities, we have confirmed several major limitations and shortcomings in today’s RA procedures, as well as identified barriers, challenges and opportunities to further develop and promote the implementation of IRA in European regulatory frameworks for chemicals.

It became clear that there are to date no explicit legal mandates for IRA, no consensus about its scope, its potential application to current RA practice and a definitive proof - of concept or usefulness - of its added value. This originates partly from the fact that the map of each sector-specific RA framework (i.e. for industrial chemicals, pesticides, biocides, human and veterinary pharmaceuticals, food additives and cosmetics) is extremely complex, and that there is no one-size-fits-all solution to integrate human and environmental RAs.

Yet, some components of IRA are in place and already used for regulatory purposes. They have benefited from the influence of “integrated thinking“ and more knowledge-based, mechanism-driven approaches to RA that are building on the advances in systems biology and toxicology and new emerging technologies. Those components include the following approaches:
• The integrated testing strategy (ITS) concept for optimizing testing efficiency and minimizing animal use through a combination/integration of in vitro tests and in silico models (e.g. Ahlers et al., 2008, Vermeire et al., 2013) can be regarded as an integral part of IRA.
• The concept of (common) modes of action (MoA) (e.g. Sonich-Mullin et al., 2001; Meek et al., 2014) and adverse outcome pathways (AOP) (e.g. Ankley et al., 2010; Tollefsen et al., 2014; Vinken, 2013) for species to species extrapolations and mixture risk assessments is an integrative process. New technologies (e.g. omics technologies) and collaborative initiatives at international level will enhance our mechanistic understanding of toxicological process and will help to promote integrated approaches. Among those initiatives, one can cite: the US Tox21 programme (Collins et al., 2008; NRC, 2007) which has developed a cost-effective approach to predicting the potential toxicity of chemicals and prioritize the thousands of chemicals that need toxicity testing; the OECD AOP programme (http://www.oecd.org/chemicalsafety/testing/adverse-outcome-pathways-molecular-screening-and-toxicogenomics.htm) which is building a toxicological knowledge framework to support chemical RA based on mechanistic reasoning and adverse outcome pathways (AOP) as a central element (an AOP describes a sequential chain of causally linked events at different levels of biological organisation that lead to an adverse health or ecotoxicological effect); the Human Exposome Project (www.humanexposomeproject.com) which builds on the concept of ‘exposome’ (the exposome encompasses the totality of human life-time environmental exposures from conception onwards, incorporating the body's response to environmental influences and the endogenous metabolic processes that can alter or influence the chemicals to which humans are exposed (Lioy and Rappaport, 2011; Patel and Ioannidis, 2014).
• An integrated framework for the RA of combined exposures to multiple chemicals has been developed under the umbrella of the WHO/IPCS (2009) to support risk assessors in identifying priorities for risk management for a wide range of applications where co-exposures to multiple chemicals are expected (Meek et al., 2011; WHO, 2009).

Integration is often used synonymously with harmonization. IRA is a strong incentive and an important driver for harmonization, but we consider harmonization approaches, including the development of common RA terminology (e.g. for addressing uncertainties in RA) and methodologies as being a necessary prerequisite for IRA rather than an integral part of the IRA concept. This was confirmed in particular through the different activities lead as part of WP3 and WP4. For example, the development and harmonization of methodologies and approaches to RA for more regulatory consistency has been defined by the European Food Safety Authority (EFSA) as one of its key strategic objectives (EFSA, 2011).

What we have learnt during the course of the project enabled us to refine our definition and vision of IRA.

In the context of HEROIC, IRA was defined as the mutual exploitation of ERA for HHRA and vice versa in order to coherently and more efficiently characterize an overall risk to humans and the environment for better informing the risk analysis process.

IRA should be considered as an evolutionary process which serves the vision of a sustainable and safe use of chemicals in order to protect human health, preserve the environment and maintain the ecosystem services for the benefit of human welfare (Suter et al., 2005), while at the same time better managing the scientific, societal and economic challenges related to present and future safety testing development and chemical policy making.

We believe that IRA offers the perspective of harmonized RA models and methodologies and a better crosstalk between HHRA and ERA, where the mutual exploitation of all existing data across the two disciplines better informs the risk analysis process on the overall risks to human health and the environment for a more comprehensive and efficient health- and environmental-based decision making.
IRA’s concept and vision extend beyond existing legal toxicity testing requirements set by the various chemical sectorial regulations to encompass more harmonised and transversal legislations aiming at the sustainable use of chemicals and the overall protection of human health or the environment.

Our concept and vision also proposes a more holistic approach to risk analysis, which goes beyond a mere technical analysis based solely on the national environment, spatial framework and economic aspects, to encompass the social and psychological dimensions of risk (Felt et al., 2007; Kasperson et al., 1988). Considering the values, beliefs, political systems and cultural factors as well as the cost of risk reduction and the benefits of risk mitigation measures to society, this reinforces communication, transparency and trust in the risk analysis process.

1.2.3 What are the opportunities for integrated risk assessment in the present European regulatory landscape?

HEROIC has identified some key drivers for the development and implementation of IRA in the EU regulatory landscape for chemicals, and revealed also several challenges, barriers and limitations, in terms of infrastructural and organisational set-ups, scientific and policy issues. However, we believe that many of those can be successfully addressed.

In recent years significant scientific advances, changes in the chemical legislation and increasing environmental consciousness have created a favourable scientific and regulatory environment to further develop and promote the concept and vision of IRA. The HEROIC project has identified several opportunities that address general challenges in current RA practice as well as more specific opportunities in selected European regulatory frameworks for chemicals.

1.2.3.1 Regulatory drivers for IRA

• IRA can improve the cross-talk between the two regulatory science disciplines
IRA can help bridging the gaps and foster stronger interfaces between HHRA and ERA within and across chemical categories and regulatory authorities when different agencies work on the same substance with different intended use/patterns, e.g. triclosan under REACH (EC, 2006), Biocidal Products (EU, 2012) and Cosmetic Products (EC, 2009a) Regulations. IRA builds on a structured and fruitful dialogue at the problem formulation stage between all the stakeholders, fostering mutual understanding and closer collaboration from the outset.
• IRA can foster a better interaction between risk assessors and risk managers
IRA relies on a strong risk assessor-risk manager interaction, helping to better manage expectations, reduce misunderstandings, ensuring a more consistent risk communication from both sides. IRA helps to produce a more holistic and more informative RA that better supports the decision making process, enabling risk managers to define more relevant and explicit management options in a transparent, easy to interpret manner.
• IRA can solve capacity and capability issues in current and future RA practice
High demand for RA, ambitious regulatory timeframes, and limited resources create capacity issues. Those impact on and are amplified by individual capability issues due to the rapid improvement of modern tools and methods for hazard and exposure measuring and modelling, and the recent developments of systems biology/toxicology towards more mechanistic-driven approaches. IRA can solve capacity issues at both the risk assessor and risk manager level by alleviating human resources and time constraints and improving cost-effectiveness of the overall RA process. Creating dedicated training programs on IRA will help risk assessors and risk managers to grasp the growing complexity and multidisciplinary character of regulatory sciences in general.

Where and how HEROIC adds value

Many of these insights were revealed or confirmed through the work conducted as part of WP2 landscaping exercise of chemical RA and survey of risk assessors and risk managers at EU level. The outcome of WP4 case studies also highlighted the need to increase capabilities and awareness of risk assessors regarding the other discipline, because the higher the lack of familiarity and understanding typically orientated human and environmental experts towards more conservatism in the weighting of the risk. A better education and discussion between ecotoxicologists and human toxicologists can thus be of great benefit for forming more rational, consistent and transparent decisions. This supports the HEROIC general recommendation to develop cross-educational programmes between the human and environmental disciplines.

1.2.3.2 Scientific drivers for IRA

The work conducted as part of WP2, WP3 and WP4 has revealed the following opportunities:
• IRA will support efforts to better harmonize, use and share data
The quality of any RA is driven by the quality of the input data. By their nature, data generated by in vivo, in vitro, in silico and/or epidemiological methods is heterogeneous, structured and stored in various formats and independent databases, and are sometimes apparently contradictory. Low quality data used for hazard and/or exposure assessment increase the uncertainty in the risk assessment process. Therefore, harmonization of data, models and methodologies is essential in chemical RA and is a prerequisite for IRA. By integrating all the existing knowledge of a chemical into the RA process, across all levels of biological organisation, incorporating multiple exposure sources, targets and effects as well as life cycle analysis, IRA fosters and optimizes the harmonization, use and sharing of available data and models. Considerable efforts have been taken so far to rank data according to their reliability, relevance and/or uncertainty, before using them in RA. Many frameworks have already been developed to assess robustness and increase transparency in reliability and relevance evaluation of individual tests. These frameworks provide the basis for a common framework allowing the evaluation of the reliability and relevance of individual data across HHRA and ERA. Innovative integrated WoE strategies have been proposed such as the IRA+ scheme proposed by Vermeire (2009) to better manage and reduce the uncertainty if based on more formal, consistent and harmonized WoE procedures, preferably statistically based (Verdonck et al., 2007; Vermeire et al., 2013).

Where and how HEROIC adds value

Integration frameworks were tested as part of WP4 based on two concrete case studies (skin sensitization and endocrine disruption) in order to better understand how human and environmental risk assessors cross-evaluate and (dis)agree to rank and weight the information in the decision making process to conclude on a risk. Activities related to expert consultation and elicitation revealed the limitations of simple BPJ process and highlighted the need for better defined, more robust and transparent WoE approaches for integrated assessment strategies as well as rationale designs, where expertise from different scientific areas contribute to information collection and decision in the context of IRA. WP4 also identified that designing ad hoc expert elicitation protocols is an essential pillar for increasing transparency in decision making, limiting the risk of misunderstanding and promoting consensus-based decisions, which is of great relevance for IRA where data from different sources and scientific areas have to be integrated. WP4 proposed a generic decision-making framework for improving the integration of data in the frame of ERA, HHRA and potentially IRA. The proposed framework can be subdivided in different successive steps that need to be adapted according to the assessment context and are based on the evaluation of ‘Reliability’, ‘Relevance’ and ‘Uncertainty’ criteria.

Exploring new opportunities to foster efficient data mining and sharing was one of the key objectives of WP5, through the web based, open access interactive and integrative tool ‘Tox-Hub’, to facilitate data mining and accessibility to available toxicological and ecotoxicological data from vetted information sources in order to manage this information altogether. This will contribute to create a better mutual understanding and management of data, methods and procedures generated in both scientific disciplines. Also WP5 contributes to better use and dissemination of scientific knowledge through the development of distance e-learning and capacity building and will be maintained 5 years after the completion of the project.
• IRA is building on a better mechanistic understanding
Mutual exploitation of data in HHRA and ERA is only possible if we have a detailed mechanistic understanding of the toxicological processes in different species. Predictive HHRA is essentially based on rodent models (including surrogates for rodent toxicity data) and the same rodent data are routinely also used for ecotoxicological assessments of risks for mammalian wildlife. Within ERA, species to species extrapolations are standard procedures, but only within certain taxonomic groups (e.g. algae to algae, fish to fish), similar to the standard extrapolations from rats to humans. MoA/AOP approaches are a prerequisite to extrapolate from toxicity findings in one species to the other. Based on common MoAs and AOPs across species, the vast amount of toxicokinetics and toxicodynamics data available used for the categorisation of e.g. human pharmaceuticals could serve as a substitute for ecotoxicological data in situations where insufficient ecotoxicological data are available to run a comprehensive ERA. Indeed, Vermeire (2009) anticipated that the IRA+ framework would reduce the uncertainty in the effect assessment if based on MoA and WoE of multiple lines of evidence (LoE) as part of the decision-making process.

Where and how HEROIC adds value

In order to get closer to IRA, WP3 aimed to get closer to a better individual understanding of integrated hazard assessment, and to assess how human and environmental hazard assessment can benefit from each other. The same approach was used for integrated exposure assessment. The aim was to better understand the drivers for integration for hazard and exposure extrapolations of both human and environmental data and models. Those are presented below:

Opportunities for an integrated hazard assessment

For species to species and endpoint-to-endpoint extrapolation, opportunities exist in the regulatory context but a lot of data is needed. Extrapolation may work for the same under¬lying mechanism (Benigni and Bossa, 2011; Blaschke et al., 2012; Böhme et al., 2009, 2010; Mekenyan et al., 2010, 2012; Péry et al., 2013; Scholz et al., 2013; Schramm et al., 2011; von der Ohe et al., 2005) and the same target of the molecular initiating event (MIE) (Leist and Hartung, 2013). Generally, more than one endpoint should be involved (1:n extrapolation). Pitfalls are time dependent mechanistic differences and different metabolic capacities (Coecke et al., 2006). For species-to-species extrapola¬tion, life style and life stage have to be taken into account. Clarifying mechanistic common¬alities between human and non-human species may form a promising route and lead to an integrated assessment.

To extrapolate from non-animal to animal endpoints, mechanistic understanding of the cause of a toxicological effect is required. Questions of concern are: i) What is the MIE and can we model it?; ii) Do these approaches address the right activity (Ji and Schüürmann 2012)?; iii) Is the relevant biology re¬flected rightly?; iv) Do the test compounds belong to the applicability domain (Dimitrov et al., 2005; Kühne et al., 2009) of the non-animal approach?; v) Are there any other additional issues?; and vi) The concept of AOP (Ankley et al., 2010) appears to be prom¬ising for a step-wise description of adverse outcomes based on mechanistic knowledge. Due to loss of information using non-animal tests, ITS combining in silico, in chemico, in vitro and in vivo methods instead of a 1:1 replacement are recommended (Buist et al., 2013; Lombardo et al., 2014a, 2014b; Nendza et al., 2013; Rorije et al., 2013; Tluczkiewicz et al., 2013).

Extrapolation from single compounds to mixtures is discussed controversially (Hartung, 2013; Belden, 2007), but considerable progress towards harmonized approaches for human and environmental mixture hazard and risk assessments has been made during recent years (EC, 2011a). For the purpose of regulatory tier zero screening level assessments, the assumption of con¬centration addition is generally considered to provide reasonable worst case estimates, irrespective of the modes of toxicant action. For higher tier assessments, differential mechanism-driven modelling approaches may be used: concentration addition for chemicals with a common MoA, inde¬pendent action for dissimilarly acting chemicals, and so-called mixed models for multi-component mixtures with partly similarly and partly dissimilarly acting components. Extrapolation has been demonstrated to provide fairly good estimates for situations of simultaneous exposure, if all mixture con¬stituents vary only regarding their potency and if no physicochemical and toxicokinetic inter¬actions cause a non-additive outcome. More research is needed for the assessments of sequential exposure and for mutually activating or deactivating components. A current bottleneck is the lack of single-compound toxicity data for mixture constituents. All efforts for closing this gap (Schüürmann et al., 2011; Kühne et al., 2013) will directly improve the capability of predictive toxicity assessments for complex mixtures (Altenburger et al., 2003).

Opportunities for an integrated exposure assessment

Traditionally recognized as the weakest part of risk assessment, expert consultation conducted by HEROIC WP2 and WP3 confirmed the large need for data generation, development and harmonization of exposure metrics, endpoints and models (i.e. area, concentration, species, routes, multiple exposure to single chemicals or mixtures, pattern of uses, life cycle analysis). Based on the commonalities in transport and fate models for human health and environmental RA, opportunities for integration have been previously identified within the WHO/IPCS IRA framework (2001) in the modelling of chemical transport, fate, and exposure, in particular for environmental exposure models, where concentrations in water, soil, air and different food items must be estimated (Vermeire et al., 2007).

New opportunities to better integrate human health and environmental exposure assessments were evaluated in WP3. The key point for the integration is the move from environmental fate and exposure estimations to the internal dose in the exposure assessment, which will be facilitated by the sharing of emission and exposure data (i.e. distribution, fate and exposure models, monitoring data on concentrations in environmental media and food, toxicokinetic and physiologically - based pharmacokinetic models, dose -response models and assessment of the variability for the critical effect across and within species).

Another promising approach for screening and prioritisation purposes is the application in human health and environmental exposure assessments of the Threshold of Toxicological Concern (TTC) concept to define common exposure-based waiving rules for chemicals and their metabolites and impurities (EFSA, 2012); the ‘waiving’ concept advocated by REACH assumes that (eco)toxicological tests may be waived if it can be shown that humans or organisms in the environment are either not or only minimally exposed to the investigated substance).

Expanding from the TTC concept for human food safety (Kroes et al., 2004), De Wolf et al. (2005) proposed a transposition of the TTC approach into an ‘Environmental Thresholds of No toxicological Concern’ concept (ETNC), which was subsequently used by Hollander et al. (2011) to define an Exposure-Based Waiving system in ERA. Application of such a scheme would facilitate the setting of common exposure threshold values for all chemicals below which no significant risk is expected. However, the concept of ETNC is limited to freshwater organisms as no sufficient data on industrial organic chemicals for the sediment, marine or soil environments are available.

1.2.3.3 The role of socioeconomics and social science
• IRA can benefit from the inclusion of socio-economic analysis and social science
The majority of regulatory recommendations or guidelines indicate that: i) RA outputs should be expressed in terms of value-relevant impacts on humans and ecosystems rather than in terms of the somewhat technical surrogates often used in the routine risk characterizations (EC, 2013); and ii) be more policy and management relevant to facilitate the dialogue and the acceptance of the risk amongst all stakeholders. Integration with socio-economic analysis and inclusion of socio-behavioural issues in IRA at the problem formulation stage may initially increase the complexity of integrations between disciplines, but in turn will provide a better and more useful estimation of the risk. This will also ensure a common language and facilitate the translation of risk evaluations into socio-economic impacts. A more holistic approach to risk analysis, which also considers the cost of risk reduction and the benefits of risk mitigation measures to society, would reinforce transparency of the process, reduce risk aversion among politicians and the public, and drive a more efficient use of risk management resources.

Where and how HEROIC adds value

The purpose of WP6 activity was to contribute to the future developments of policy frameworks and guidelines on the factors influencing the effectiveness of risk assessment process. These results are reflected in the HEROIC White paper. One of our key objectives was to gain new insight on the socio-behavioural factors that need to be taken into account in the processes of risk analysis and policy making, referring to all its phases of risk assessment, risk management and risk communication, to be able to produce a better estimate and more effective response to the risks associated with the use of chemicals. An analysis of the state of the art on SEA (Socio Economics Analysis) frameworks was conducted and a stakeholder consultation process was developed, as judged as valuable opportunity to improve the system and information source.

After the overall debate experts agreed that some changes are needed to strengthen the credibility of the RA process and improve the effectiveness of policies. It has emerged that it is important to integrate social sciences in the whole process to overcome cognitive barriers, without attributing to the risk assessor skills and responsibility that are not part of his cultural background. The strong influence of socio-psychological factors and the cultural ones on the overall quality and effectiveness of ex ante risk evaluations were confirmed during our process as well as that, due to the growing complexity and multiculturalism of the real world, a cross-cultural and interdisciplinary research is essential. The stakeholder consultation revealed that the society is presenting the RA community with the challenge to ensure more transparency, on input data quality, assessment procedures and on resulting uncertainty. This should allow better risk communication with the aim to regain consumer/public trust and to give unambiguous guidance for improved risk management.

Some issues and suggestions for further research have been therefore identified, and are listed below.

SEA integration in IRA

While IRA would improve the efficiency of the assessment process it is important to factor in the risk assessment the SEA needs in terms of outputs so as to ensure a common language and to facilitate the translation of risk evaluations into socio-economic impacts. SEA could be a powerful tool to check and improve the level of integration of HHRA and ERA, from the point of view of their outputs (Péry et al., 2013). Indeed, outcomes of the assessment may differ substantially depending both on the nature of the analysis and the stakeholders or users concerned.

RA relies on worst case assumptions – economic analysis generally is based on more realistic assumptions. Risk Characterisation Ratios (RCR) approach is excellent at meeting RA purposes but is of limited value in providing quantifiable estimates of impact to inform socio-economic analyst and there is the need for fuller information on relationship between exposures and impacts to characterise effects in a way which is meaningful to economists and others. There is the need to ensure that all effects are properly assessed and that uncertainties as well as safety/assessment factors are explicit to avoid the comparison of a ‘worst case‘ with an ‘expected’ case.

Commonly proposed or used metrics, such as total mortality or morbidity are not always appropriate. For these reasons, measures such as years of life lost (YLL), disability-adjusted life years (DALYs) or quality-adjusted life years (QALYs) have been proposed, while economic measures may also be applied, as in cost-benefit analysis (Briggs, 2008). However these measures can reflect only impacts on health and longevity and do not take into account the nature of the risk itself and societal perceptions, and values related to it. Particular care should be taken to present not only quantified results, but also non-quantified outcomes (often in the social and environmental areas) that may be highly significant to raise a qualitative RA especially in the case of substances persistent, or very persistent, and bioaccumulative or very bioaccumulative properties.

There is a need to develop collaborative frameworks with other knowledge areas or disciplines like social sciences or economics and to consider different approaches in order to conduct a qualitative and more useful RA at the problem formulation stage of the RA process (EC, 2013; EFSA, 2010; Péry et al., 2013).

Socio-behavioural aspects to be considered in IRA

Early consideration of the social-behavioural factors influencing the risks associated to the use of certain chemicals can improve the efficiency of the risk analysis process. Factors linked to psychological and sociological realty, such as responsibility, trust, reliance and beliefs, can strongly influence and limit the decision quality and the choice of the approach to be taken, which was suggested by the overall HEROIC stakeholder consultation. However the decision on the level and details of considering such factors should be made context specific and in relation to the properties and context of use of the chemical concerned.

In the risk assessment phase experts have to make decisions, without all the necessary information or enough time that would enable them to choose the best option. Experts make decisions but are subject to biases, heuristics, and a number of other influencing factors. Despite the evidence-based approach of the RA process, results could be sometime controversial due to different experts beliefs and views, biases in processing scientific evidence. Accordingly, approaches for characterising scientific evidence, transparency and definition of uncertainty are needed. For example, so called ‘Evidence maps’ (Wiedemann et al., 2011) have been proposed to meet such needs.

The improvement of measuring risk perception is one of the most important issues we have to take into account. Risk perception has its own role in the political debate regarding the level of acceptance of certain risks. The HEROIC Expert Roundtable, held on 7 April 2014 in Brussels, Belgium, highlighted the need of an improvement of measuring risk perception to increase trust in the risk assessment process. The CORA (COmmunicating health Risk Assessments to the general public) framework was suggested as pragmatic tool (Wiedemann et al., 2013). It was designed to help people without specialised scientific risk knowledge to judge the trustworthiness of a risk assessment report and to give suggestions on what kind of information are needed to be communicated. Using such instruments will support non-experts to judge the credibility of the overall risk assessment activity.

The overall HEROIC stakeholder survey put into perspective that a potential involvement of social sciences would allow to better target the risk analysis process to produce more accurate predictions of the risk and more efficient suggestions for its management and its communication, therefore allowing to overcome the psychological and cognitive barriers without attributing to the risk assessor skills and responsibility that are not part of his cultural background; and to early predict behavioural aspects related to the use patterns of certain chemicals or related to risk perception.

1.2.4 Opportunities for IRA in specific regulatory frameworks for chemicals

There are a number of opportunities to further promote the development of IRA in the various regulations for authorization of chemical products, yet the map of each sector-specific RA framework is extremely complex, and there is no one-size-fits-all solution to integrate human and environmental RAs. The regulatory RA frameworks for chemicals (REACH) and plant protection products (PPP Regulation) were chosen as examples to demonstrate the benefits of developing IRA, initially for the assessment of single substances but in the long-term with a focus on mixture RA (MRA).

1.2.4.1 REACH

The sectorial chemical legislation REACH (EC/1907/2006) (EC, 2006) has the potential to substantially increase demands in toxicity testing, which motivates the search for a shift from the extensive standard testing paradigm to a more cost-effective and knowledge-driven approach. In this way, REACH is expected to play a major role in driving the scientific validation and regulatory acceptance of alternative testing methods (Hartung, 2009, 2010).

Many components of the WHO/IPCS IRA Framework (2001) can already be found in the REACH RA approach (Vermeire, 2009). Indeed not only the REACH Regulation but also institutional RA practice and organizational set up at the European Chemicals Agency (ECHA) create a favourable environment to further promote the development and implementation of IRA.

In common with the principles of IRA, REACH specifies the optimal use of all existing data (in vitro, in vivo, in silico) to avoid duplication of studies, enabling a flexible, tiered approach to RA, considering in vivo testing only as a last resort. REACH promotes the development and use of alternative testing methods such as ITS based on consideration of MoA and WoE schemes. Notably, REACH will contribute to increase the availability of data through enhanced data sharing (ECHA, 2012) and increased accessibility to proprietary toxicity data, which will help to address the problem of data-poor chemicals.

In addition, the existing close interactions at ECHA between human and environmental risk assessors, and between the Scientific Committees for Risk Assessment (RAC) and Socio-Economic Analysis (SEAC) offer further opportunities for better integration into the RA process of cost-benefits and value-relevant impacts on human health and ecosystems services to more efficiently inform about the pros and cons of possible risk management options (Péry et al., 2013). Furthermore, ECHA liaises with other EU regulatory bodies, which enables the assessment of environmental safety in a cross-sectorial manner (REACH, EC/1907/2006, Art.122). This may create new IRA opportunities to populate ecotoxicological datasets for data-poor chemicals. Regarding exposure assessment, it is noted that REACH promotes close interactions with industry across the supply chain in order to integrate all information available on a chemical, especially on exposure sources and intended uses.

1.2.4.2 PPP Regulation

Regulation (EC) No. 1107/2009 (EC, 2009b) sets out the provisions for the placing of plant protection products (PPP) on the market of the European Community. The data requirements for PPPs and their active substances are the most stringent among all the classes of chemicals and in general, pesticides are recognized as the most extensively tested chemicals.

In the future, the application of the concept of the IRA within the frame of the PPP Regulation could be used in order to facilitate the better exploitation of the available (eco)toxicological data and reduce the number of required studies, as this is also a requirement of the PPP Regulation especially for studies conducted in vertebrate species. Beyond the mere harmonisation of principles and methodologies, the vision of an actual integration of human and environmental MRAs has been suggested for pesticides with common modes of action in both humans and wildlife species, using organophosphates as an example. However, convincing examples for pesticide groups other than acetylcholine esterase inhibitors have not been brought forward.

The IRA concept could be applied in the case of active substances that either share the same MoA/AOP or some commonalities in the exposure assessment such as common sources and emissions, distribution routes and exposure scenarios for humans and non-target species. Even if commonalities in MoA cannot be established, IRA supports a more coherent grouping and ranking for MRA, which is at present not possible.

1.2.5 Opportunities for IRA in transversal legislations

An additional benefit and opportunity for IRA in the regulatory context could be to underpin multisectorial harmonised legislations dealing with sustainability and overall protection of human health or the environment, such as the Directive on sustainable use of pesticides (EC, 2009c), on nature protection in the Birds Directive (EC, 2009d) and the Habitats Directive (EEC, 1992) or the protection of waters in the Water Framework Directive (EC, 2000). IRA would allow the overall inclusion of inputs from currently fragmented risk characterizations in relation to these various pieces of legislations. This would provide additional support for sound decision making by fostering a better management and setting of protection goals and establishment of sustainability principles considering overall cost-benefits and risks.

1.2.5.1 Water Framework Directive (WFD)

The WFD (2000/60/EC) does not regulate the placing of the market of individual substances (such as REACH) or chemical products (such as the PPP Regulation), but establishes a framework for the protection of waters. For surface waters, the aim is to achieve ‘good ecological status’ as well as ‘good chemical status’, i.e. low levels of individual chemical contaminants. This entails a focus not only on individual chemicals but also on their combined hazards and risks. Under the WFD, environmental quality standards (EQS) are established for substances or groups of substances that are prioritized by means of a risk-based ranking.

In this context, several contributions from IRA to the implementation of the WFD are readily conceivable. One can assess candidate substances for RA prioritisation and for the setting of EQS on the basis of compound properties such as Persistence, Bioaccumulation and Toxicity (PBT). One can run a specific assessment of pharmaceuticals and personal care products to assess their effects in aquatic systems, as required by Directive 2013/39/EU (EU, 2013). Since a substantial amount of human-related toxicology data is generated for the development and registration of a pharmaceutical product, it seems obvious that using such body of scientific information on an integrated perspective can be of great value in order to extrapolate from human to environmental effects. Knowledge about MoA and toxicity of metabolites can provide highly valuable information. In this respect, pharmaceutical products are particularly suitable for IRA, although the issue of data confidentiality needs careful consideration.

While EQS have been currently established for 45 priority substances (primarily individual chemicals), the need to consider mixture toxicity has been recognized, and a chapter on ‘substances occurring in mixtures’ has been recently included in the revised ‘Technical Guidance for Deriving Environmental Quality Standards’ (EC, 2011b). Further opportunities to integrate HHRA and ERA of mixtures and the overall risks for humans and aquatic organisms exist in the WFD and it is anticipated that it will benefit from initiatives such as the WHO/IPCS integrated framework for mixture risk assessments (Meek et al., 2011; WHO, 2009) and the EU FP7 SOLUTIONS project (www.solutions-project.eu/) which aims to produce consistent solutions for the large number of legacy, present and future chemicals posing a risk to European water resources with respect to ecosystems and human health.

(Please see attached list for references.)
Potential Impact:
1.3 Expected Impact of the Project

HEROIC explored how data generated in ecotoxicology could be applied to human toxicology and vice versa for integrated RA, aiming to develop a framework for integrated RA methodologies and approaches. It is expected that this will result in:
• Better coordination and exploitation of resources by establishment of a common framework for human and environmental risk assessment, resulting in optimal use of resources;
• Improved harmonised tools and methods in RA;
• Improved quality of human and environmental risk assessments based on novel integrated testing strategies in data-poor situations;
• Improved capacity building for risk assessors across EU member states, through the dedicated data and training platform ‘Tox-Hub’;
• Improved stakeholder and public understanding and acceptance of integrated risk assessment by open-access communication of the knowledge produced within the project;
• Increased transparency in RA and better risk communication to maintain public trust and to give unambiguous guidance for improved risk management decisions.

HEROIC key findings, opportunities and challenges for IRA have been summarized in the HEROIC ‘White Paper on the promotion of an integrated risk assessment concept in European regulatory frameworks for chemicals’. The White paper is regarded as the major contribution of the project. It brings forward to the decision-makers and the IRA community the HEROIC vision of IRA and proposes policy recommendations on how to develop further and promote the implementation of IRA. It also gives pointers on how to advance the dialogue within and across the human and environmental scientific disciplines.

To facilitate the implementation of IRA, the HEROIC project evaluated short-term and medium-term opportunities for the further development, promotion and stepwise implementation of IRA as a feature of European regulatory frameworks in the following areas:
1. Development of RA science (hazard assessment, exposure assessment, mixtures assessment; harmonization of RA terminology, models and methodologies)
2. Regulatory development (including harmonization of legal requirements for HHRA and ERA; institutional dialogue; development of capabilities; handling of data confidentiality issues)
3. Integration with other disciplines (socio-economic analysis, socio-behavioural considerations)

In order to disseminate the vision and concept of IRA, the HEROIC project has developed recommendations and a roadmap for a tiered implementation of IRA.

1.3.1 Development of RA science

• Expand harmonization of principles and methodologies: The harmonization of methodologies and approaches to risk assessment for increased regulatory consistency are not only critical for addressing uncertainties in traditional RA, but are a prerequisite for IRA of both single substances and mixtures. Thus we recommend expanding these harmonization efforts across chemical categories through better coordination between regulatory agencies (as outlined in EFSA’s science strategy (EFSA, 2011) and other international initiatives, e.g. the WHO/IPCS Harmonization Project (http://www.who.int/ipcs/methods/harmonization/en/) which aims to globally harmonize approaches to RA by developing basic principles and guidance on specific chemical RA issues). It is recommended in particular to use transparent WoE approaches and clear decision strategies, preferably statistically based; to use advanced and innovative toxicological concepts, like AOP, to improve causal inference in WoE, in particular on the ‘biological plausibility’ criteria; and to use non-conventional approaches to incorporate uncertainty in WoE like NUSAP (Funtowicz and Ravetz, 1993), MCDA (Figueira et al., 2005; Köksalan et al., 2011), fuzzy logic and Bayesian approaches such as ITS (Jaworska et al., 2010). We recommend to build on innovative approaches and expert elicitation protocols and apply to human and environmental RAs those methodologies pertaining to decision making modeling such as those used to analyze industrial or economic risks.
• Improve scientific knowledge: Advancing our mechanistic understanding of toxicological pathways is the key driver for mutual exploitation of data between HHRA and ERA, but the establishment of MoA/AOP is a relatively new initiative and a full understanding of the major AOPs for the most relevant toxicological endpoints will take time to achieve. Existing support for current initiatives such as the OECD AOP process needs to be continued in the long term and with this, a less fragmented, more structured and scientifically robust approach should be encouraged, to improve regulatory confidence. Research should be focused on commonalities in MoA/AOP across species, in particular across mammalian and non-mammalian animals (vertebrates in the first place) and their usability for hazard and risk extrapolations to humans. Ideally, regulatory testing requirements should be revised so that resulting data provide some evidence on MoA/AOP.
• Demonstrate validity and advantages of IRA: A proof-of-concept case study with “real life” data under the auspices of a regulatory authority is advocated as a pragmatic and convincing way forward to demonstrate the added value of IRA. This case study should clearly illustrate real cost-benefit advantages and demonstrate where, when and to what extent IRA would add greatest value to current RA processes:
➢ Choose a lead authority where both HHRA and ERA are run within the same organization (e.g. ECHA) to illustrate how institutional cooperation between human and environmental risk assessors (e.g. exchange of data and insights; use of similar models) affect efficiency and cost-effectiveness;
➢ Choose a data-rich chemical with shared human and environmental concern, and known and common MoA (effects observed both in humans and wildlife, e.g. PPP, biocide).
OR
➢ Choose a chemical for which safety assessment is cross-sectorial based on product type (e.g. cosmetics, HHRA/DG SANCO, ERA/ECHA) or use (e.g. triclosan or resorcinol under REACH, Pharmaceuticals, Biocide and Cosmetics Regulations).
• ‘Start small’ with IRA of single substances: Cumulative (mixture) risks may result from exposure of humans and wildlife to many different chemicals from various sources and via multiple routes. However, more complex feasibility studies for integrated MRAs may be considered as a second step, after the advantages and the validity of integrated uses of hazard and exposure data have been established for single substance RAs.

1.3.2 Regulatory development

• Disseminate the concept and vision of IRA to a broad audience to promote stakeholder and institutional dialogue: Several activities should be organized to disseminate the concept and vision of IRA to the scientific community, the industry and the regulators, as part of targeted meetings with e.g. ECHA and/or EFSA; ILSI-HESI and ECETOC activities; or joint meetings, e.g. SETAC-EUROTOX, to bring academics and scientists from different areas together and explore the IRA concept as well as opportunities for new educational programmes. These activities should aim to increase familiarity with the IRA concept, promote its acceptance and create an IRA community.
• Create dedicated IRA education and training opportunities: The complexity and multidisciplinary character of IRA requires dedicated, cross-functional education and training programs for risk assessors and risk managers. These need to address both the scientific basis for IRA as well as its implementation across institutional boundaries.
• Produce new guidance which would define standards, expectations and goals for integrated approaches to RA, and where, when and how to apply IRA.
• Promote the inclusion of harmonized legal requirements for HHRA and ERA in the different sectorial chemical regulations and seek governmental acceptance to drive policy-making and enactment of new legal mandates for IRA.

1.3.3 Integration with other disciplines

• Augment technical RA by including socio-economic analysis and socio-behavioural considerations into the risk analysis process: Early consideration of the social-behavioural factors influencing the risks can improve the efficiency of the risk analysis process. It would address issues related to transparency and trust; improve risk perceptions and risk attitudes as well as addressing the challenges in risk communication, while positively impacting on the measures taken through risk management decisions. Integration of the social sciences in the whole process could support communication of risk assessment results to increase RA credibility, to reach agreement, to demonstrate elimination of biases by appropriately weighting different kinds of evidence, and to communicate uncertainty, contributing to drive a more efficient and sustainable use of risk management resources. Stakeholder participation can help to define scenarios and population exposed, yet a tiered approach is suggested so that the process does not become too complex to operate.

1.3.4 Roadmap

Building consensus to develop clear, easily understood, transparent and unambiguous integrated RA procedures will require time and continuous support. The promotion of the development and implementation of an IRA framework in regulatory practice will require a concerted effort and sustained dialogue between all stakeholders, and will necessitate policy and regulatory changes.
Implementation of a full IRA at larger scale will require international agreement, and major coordination and harmonisation efforts.

The following roadmap is proposed:

Short-term
• Establish a multi-stakeholder expert working group / Task Force for IRA at EU level in collaboration with leading EU regulatory authorities to coordinate these approaches based on a common understanding and joint definition of IRA and to provide a framework for a tiered approach on how IRA can best be implemented in regulatory frameworks;
• The Task Force should define the most promising setting for a real-life case study on a current open issue to establish proof of concept and usefulness with a leading EU regulatory authority (e.g. ECHA, EFSA); this pilot case study should be based on the most promising identified scientific research areas with reliable and robust data quality (e.g. systems toxicology, MoA, AOPs, Tox21, Exposome).

Mid-term
• Harmonization efforts should be encouraged at every level and considered as a prerequisite for IRA.
• The Task Force, in liaison with the leading EU regulatory authorities, should provide guidance on where, when and how to apply IRA.

Long-term
• Policy-makers should enact new legal mandates for IRA in the most promising identified EU regulations identified by the Task Force (e.g. REACH, PPP, WFD) to support its implementation into regulatory practice.

(Please see attached list for references.)

List of Websites:
1.4 Contact details for the HEROIC Project

Prof. Dr. Martin Wilks
HEROIC Coordinator
University of Basel
Swiss Centre for Applied Human Toxicology (SCAHT)
Klingelbergstrasse 61
4056 Basel
Switzerland
Tel: +41 61 267 19 55
Fax: +41 61 267 19 61
E-mail: martin.wilks@unibas.ch

HEROIC website: http://www.heroic-fp7.eu